National Emission Standards for Hazardous Air Pollutants from the

Portland Cement Manufacturing Industry

Response to Comments Received on Proposed Rule

Published on May 06, 2009, 74FR 21135

U.S. Environmental Protection Agency

Office of Air Quality Planning and Standards

Sector Policies and Programs Division

Metals and Minerals Group

Research Triangle Park, NC 27711

August 6, 2010

Table of Contents

  TOC \o "2-5" \h \z \t "Heading 1,1"    HYPERLINK \l "_Toc268873222" 
1.0	Introduction	  PAGEREF _Toc268873222 \h  1  

  HYPERLINK \l "_Toc268873223"  2.0	Proposed Amendments to the Current
NESHAP from Portland Cement Manufacturing Industry	  PAGEREF
_Toc268873223 \h  34  

  HYPERLINK \l "_Toc268873224"  2.2	General	  PAGEREF _Toc268873224 \h 
34  

  HYPERLINK \l "_Toc268873225"  2.2.1	Support	  PAGEREF _Toc268873225 \h
 34  

  HYPERLINK \l "_Toc268873226"  2.2.2	Oppose	  PAGEREF _Toc268873226 \h 
35  

  HYPERLINK \l "_Toc268873227"  2.3	Overall Approach to Setting MACT
Floor Limits	  PAGEREF _Toc268873227 \h  36  

  HYPERLINK \l "_Toc268873228"  2.3.1	Variability and Adequacy of Data	 
PAGEREF _Toc268873228 \h  36  

  HYPERLINK \l "_Toc268873233"  2.3.2	Pollutant-by-Pollutant MACT Floors
  PAGEREF _Toc268873233 \h  42  

  HYPERLINK \l "_Toc268873237"  2.3.3	Lowest Emitters Versus Best
Performers	  PAGEREF _Toc268873237 \h  45  

  HYPERLINK \l "_Toc268873238"  2.3.4	Top 5 versus 12 Percent	  PAGEREF
_Toc268873238 \h  48  

  HYPERLINK \l "_Toc268873239"  2.3.5	Area Sources in HCL Floor
Development	  PAGEREF _Toc268873239 \h  49  

  HYPERLINK \l "_Toc268873242"  2.3.6	Legal	  PAGEREF _Toc268873242 \h 
50  

  HYPERLINK \l "_Toc268873250"  2.3.7	Format of Emission Limits	 
PAGEREF _Toc268873250 \h  58  

  HYPERLINK \l "_Toc268873255"  2.3.8	Other	  PAGEREF _Toc268873255 \h 
60  

  HYPERLINK \l "_Toc268873257"  2.4	General Legal Issues	  PAGEREF
_Toc268873257 \h  61  

  HYPERLINK \l "_Toc268873258"  2.4.1	Subcategorization	  PAGEREF
_Toc268873258 \h  61  

  HYPERLINK \l "_Toc268873259"  2.4.2	Fly Ash Ban	  PAGEREF
_Toc268873259 \h  62  

  HYPERLINK \l "_Toc268873260"  2.4.4	Other	  PAGEREF _Toc268873260 \h 
63  

  HYPERLINK \l "_Toc268873262"  2.5	Mercury Standards	  PAGEREF
_Toc268873262 \h  64  

  HYPERLINK \l "_Toc268873263"  2.5.1	Legal Issues	  PAGEREF
_Toc268873263 \h  64  

  HYPERLINK \l "_Toc268873266"  2.5.2	Adequacy of Floor Determinations
(Existing and New)	  PAGEREF _Toc268873266 \h  69  

  HYPERLINK \l "_Toc268873273"  2.5.3	Beyond the Floor Determinations	 
PAGEREF _Toc268873273 \h  77  

  HYPERLINK \l "_Toc268873274"  2.5.4	Compliance Assurance	  PAGEREF
_Toc268873274 \h  77  

  HYPERLINK \l "_Toc268873287"  2.5.4.1	Performance Specification 12A,
12B, and Procedure 5	  PAGEREF _Toc268873287 \h  86  

  HYPERLINK \l "_Toc268873291"  2.5.5	Impacts	  PAGEREF _Toc268873291 \h
 105  

  HYPERLINK \l "_Toc268873304"  2.6	THC Standards	  PAGEREF
_Toc268873304 \h  115  

  HYPERLINK \l "_Toc268873305"  2.6.1	Legal Issues	  PAGEREF
_Toc268873305 \h  115  

  HYPERLINK \l "_Toc268873308"  2.6.2	Adequacy of Floor Determinations
(Existing and New)	  PAGEREF _Toc268873308 \h  116  

  HYPERLINK \l "_Toc268873319"  2.6.4	Compliance Assurance	  PAGEREF
_Toc268873319 \h  123  

  HYPERLINK \l "_Toc268873330"  2.6.5	Impacts	  PAGEREF _Toc268873330 \h
 132  

  HYPERLINK \l "_Toc268873333"  2.6.6	Other	  PAGEREF _Toc268873333 \h 
136  

  HYPERLINK \l "_Toc268873335"  2.7	HCl Standards	  PAGEREF
_Toc268873335 \h  136  

  HYPERLINK \l "_Toc268873336"  2.7.1	Legal Issues	  PAGEREF
_Toc268873336 \h  136  

  HYPERLINK \l "_Toc268873340"  2.7.2	Adequacy of Floor Determinations
(Existing and New)	  PAGEREF _Toc268873340 \h  140  

  HYPERLINK \l "_Toc268873344"  2.7.4	Compliance Assurance	  PAGEREF
_Toc268873344 \h  143  

  HYPERLINK \l "_Toc268873353"  2.7.5	Impacts	  PAGEREF _Toc268873353 \h
 152  

  HYPERLINK \l "_Toc268873362"  2.8	PM Standards	  PAGEREF _Toc268873362
\h  156  

  HYPERLINK \l "_Toc268873363"  2.8.1	Legal Issues	  PAGEREF
_Toc268873363 \h  156  

  HYPERLINK \l "_Toc268873371"  2.8.4	Compliance Assurance	  PAGEREF
_Toc268873371 \h  164  

  HYPERLINK \l "_Toc268873384"  2.8.5	Impacts	  PAGEREF _Toc268873384 \h
 179  

  HYPERLINK \l "_Toc268873387"  2.8.6	Other	  PAGEREF _Toc268873387 \h 
182  

  HYPERLINK \l "_Toc268873392"  2.9	Compliance Dates	  PAGEREF
_Toc268873392 \h  184  

  HYPERLINK \l "_Toc268873395"  2.10	Startup, Shutdown, Malfunction
(SSM)	  PAGEREF _Toc268873395 \h  186  

  HYPERLINK \l "_Toc268873404"  2.11	Economic Impacts	  PAGEREF
_Toc268873404 \h  193  

  HYPERLINK \l "_Toc268873411"  2.12	Environmental Justice	  PAGEREF
_Toc268873411 \h  215  

  HYPERLINK \l "_Toc268873412"  2.13	Open Clinker Piles	  PAGEREF
_Toc268873412 \h  217  

  HYPERLINK \l "_Toc268873413"  2.14	Crushing Operations	  PAGEREF
_Toc268873413 \h  219  

  HYPERLINK \l "_Toc268873414"  2.15	Coal Mills	  PAGEREF _Toc268873414
\h  219  

  HYPERLINK \l "_Toc268873415"  2.16	Compliance Assurance	  PAGEREF
_Toc268873415 \h  221  

  HYPERLINK \l "_Toc268873417"  2.16.1	Flow Rate Monitoring	  PAGEREF
_Toc268873417 \h  221  

  HYPERLINK \l "_Toc268873418"  2.16.2	Kiln Feed/Clinker Production	 
PAGEREF _Toc268873418 \h  221  

  HYPERLINK \l "_Toc268873419"  2.16.3	Electronic Reporting	  PAGEREF
_Toc268873419 \h  223  

  HYPERLINK \l "_Toc268873420"  2.16.4	Other	  PAGEREF _Toc268873420 \h 
223  

  HYPERLINK \l "_Toc268873428"  2.17	Other – Miscellaneous Comments	 
PAGEREF _Toc268873428 \h  226  

 

List of Tables

  TOC \h \z \t "Title_Table,1"    HYPERLINK \l "_Toc268873466"  1-1.
List of Commenters on Proposed Amendments to the Current NESHAP from the
Portland Cement Manufacturing Industry, Proposed May 06, 2009 (74 FR
21135)	  PAGEREF _Toc268873466 \h  2  

 Acronyms and Abbreviations

µg

micrograms

ACFM

actual cubic feet per minute

ACI

activated carbon injection

ASTM

American Society for Testing and Materials

AWMA

Air and Waste Management Association

BACT

best available control technology

BART

best available retrofit technology

BAT

best available technology

BLD

bag leak detector

BLDS

bag leak detection system

BOD

biochemical oxygen demand

BTEX

benzene, toluene, ethylbenzene, xylene

CAA

Clean Air Act

CAM

compliance assurance monitoring

CAMR

Clean Air Markets Rule

CARB

California Air Resources Board

CASAC

Clean Air Scientific Advisory Committee

CD

calibration drift

CEMS

continuous emission monitoring system

CIBO

Council of Industrial Boiler Owners

CKD

cement kiln dust

CMM

continuous mercury monitor

CO

carbon monoxide

CO2

carbon dioxide

COMS

continuous opacity monitoring system

CWA

Clean Water Act

D/F

dioxin and furan 

DAS

data acquisition system

dscf/t

dry standard cubic feet per ton

dscfm

dry standard cubic feet per minute

EGU

coal- and oil-fired electric utility steam generating unit

EO

executive order

EPA 

U.S. Environmental Protection Agency 

EPRI

Electric Power Research Institute

ESP

electrostatic precipitators

FID

flame ionization detectors

FTIR

fourier transform infrared spectroscopy

GACT

generally available control technology

GC-MS-MS

gas chromatography-mass spectrometry

gr/dscf

grains per dry standard cubic feet

HAP

hazardous air pollutants

HCl

hydrogen chloride

ISIS

Industrial Sectors Integrated Solutions

lb/MWh

pound per megawatt hour

lb/t

pound per ton

lb/TBtu

pound per trillion British thermal unit

LWS

limestone wet scrubbers

MACT

maximum achievable control technology

MMbtu

million British thermal units

NAAQS

National Ambient Air Quality Standards

NaOH

sodium hydroxide

NESHAP

National Emission Standards for Hazardous Air Pollutants

NIST

National Institute of Standards and Technology

NOEL

no observable effects level

NOx

nitrogen oxides

NSPS

New Source Performance Standards

NSR

New Source Review

OAQPS

Office of Air Quality Planning and Standards

OSW

Office of Solid Waste

PCA

Portland Cement Association

PCDF

polychlorinated dibenzofuran

PH/PC

preheater/precalciner

PM

particulate matter

PM2.5

PM with an aerodynamic diameter less than or equal to a nominal 2.5 μ

PMD

particulate matter detectors

PMDS

particulate matter detection systems

POLVITEC

Polysius Environmental Technology

POM

polycyclic organic matter

ppmv

parts per million by volume

ppmvd

parts per million by volume dry 

PQL

practical quantitation limit

PS

Performance Specification

PSD

Prevention of Significant Deterioration

QA

quality assurance

RATA

relative accuracy test audits

RCA

response correlation audit

RCRA

Resource Conservation and Recovery Act

RIA

Regulatory Impact Analyses

RRA

relative response audit

RTO

regenerative thermal oxidizer

SBREFA

Small Business Regulatory Enforcement Fairness Act

scfm

standard cubic feet per minute

SCR

selective catalytic reduction

SO2

sulfur dioxide

SO3

sulfur trioxide

SRE

system removal efficiency

SSM

startup, shutdown and malfunction

STMS

sorbent trap monitoring system

syngyp

synthetic gypsum

TDS

total dissolved solids

THC

total hydrocarbons

Tl

thallium

TNMHC

total non-methane hydrocarbons

tpd

tons per day

tpy

tons per year

TSS

total suspended solids

UBA

The Federal Environment Agency (Umweltbundesamt) Mandate and
Organization

UPL

upper prediction limit

U.S. 

United States

UTL

upper tolerance limit

VOC

volatile organic compound

μm

micrometers

1.0	Introduction

The U.S. Environmental Protection Agency (EPA or the Agency) requested
comment on May 06, 2009 (74 FR 21135), on a proposed amendments to the
current National Emission Standards for Hazardous Air Pollutants
(NESHAP) from the Portland Cement Manufacturing Industry. These proposed
amendments would add or revise, as applicable, emission limits for
mercury, total hydrocarbons (THC), and particulate matter (PM) from
kilns and in-line kiln/raw mills located at a major or an area source,
and hydrochloric acid (HCl) from kilns and in-line kiln/raw mills
located at major sources. These proposed amendments also would remove
the following four provisions in the current regulation: the operating
limit for the average hourly recycle rate for cement kiln dust; the
requirement that cement kilns only use certain type of utility boiler
fly ash; the opacity limits for kilns and clinker coolers; and the 50
parts per million by volume dry (ppmvd) THC emission limit for new
greenfield sources. EPA is also proposing standards which would apply
during startup, shutdown, and operating modes for all of the current
§112 standards applicable to cement kilns.

Finally, EPA is proposing performance specifications for use of mercury
CEMS, which specifications would be generally applicable and so could
apply to sources from categories other than, and in addition to,
portland cement, and updating recordkeeping and testing requirements. A
60-day comment period (ending July 06, 2009) was initially provided for
the public to submit comments to EPA regarding the proposed amendments.
Following a request for an extension to the comment period, the comment
period was extended to September 4, 2009.

A total of 16,055 sets of comments were received by EPA regarding the
proposed information collection request. Some of the comment sets were
signed or submitted on behalf of multiple commenters. Also, in some
cases duplicate comment sets from the same commenter were submitted to
EPA’s Air and Radiation Docket. Table 1-1 lists the names of the
commenters, the commenters’ affiliations, and the Air and Radiation
Docket entry number for each of the comment sets submitted to EPA. Some
commenters also submitted supplemental information and data to support
their comments.

This document presents EPA’s response to each substantive comment
received by EPA on the proposed amendments. The comment summaries are
grouped by topic in sections.

Table 1-1. List of Commenters on Proposed Amendments to the Current
NESHAP from the Portland Cement Manufacturing Industry,
Proposed May 06, 2009 (74 FR 21135)

Docket No. EPA-HQ-OAR-2002-0051	Date Received	Commenter Name and
Affiliation

2137	May 20, 2009	EarthJustice Mass Mailing Campaign

2138	May 21, 2009	Craig S. Campbell

Vice President, Environment and Government Affairs

Larfarge North America – Cement Division

2142	June 09, 2009	Nancy Wittenberg

Assistant Commissioner

State of New Jersey Dept of Environmental Protection

2144	June 16, 2009	A.A. Linero, Program Administrator

Special Projects Section

Florida Dept of Environmental Protection

2145	June 18, 2009	George Morris

Organizer, North Texas Environmental Meetup

2146	June 19, 2009	J. Manion, Private Citizen

2147	June 19, 2009	Multiple Comment Letters to EPA – Submission of 19
citizen letters

2148	June 23, 2009	E. Eggleston, Private Citizen

2149	June 26, 2009	Anonymous public comment

2150	June 26 2009	K. D. Haglund, Private Citizen

2151	June 30, 2009	Anonymous public comment

2152	June 30, 2009	Anonymous public comment

2153	June 30, 2009	Anonymous public comment

2154	June 30, 2009	Anonymous public comment

2155	June 30, 2009	Karen Nicholson, President

League of Women Voters of Texas

2156	July 09, 2009	M. Wassenich

2157	July 09, 2009	Fort Worth Mayor, Mike Moncrief

2158	July 21, 2009	Ed & Carol Soph, Private Citizens

2159	July 21, 2009	Joan Eddy, Private Citizen

2160	July 22, 2009	Ellis Gilleland, Private Citizen

2161	July 22, 2009	Liz Walley, Private Citizen

2162	July 22, 2009	Fran Gloor, Private Citizen

2163	July 22, 2009	Mary Faler, Private Citizen

2164	July 22, 2009	Anne Randolph, Private Citizen

2165	July 22, 2009	Sr. Elizabeth Riebschlaeger, Private Citizen

2166	July 22, 2009	Emily Willoughby, Private Citizen

2167	July 22, 2009	Edward and Elisabeth Bala, Private Citizens

2168	July 23, 2009	Tara West, Private Citizen

2169	July 27, 2009	Emily Bergeron, Private Citizen

2170	July 27, 2009	Christopher Ebey, Private Citizen

2171	July 27, 2009	Evelyn Shepard, Private Citizen



Table 1-1. (continued)

Docket No. EPA-HQ-OAR-2002-0051	Date Received	Commenter Name and
Affiliation

2172	July 27, 2009	Estela Moreno, Private Citizen

2173	June 28, 2009	Fabian Mendez, Private Citizen

2174	July 28, 2009	Leyland DelRe, Private Citizen

2175	July 28, 2009	Linda Sears, Private Citizen

2176	July 28, 2009	Elaine Fassel, Private Citizen

2177	July 28, 2009	Barbara Besold, Private Citizen

2178	July 28, 2009	Mark Warford, Private Citizen

2179	July 28, 2009	Brian Dodge, Private Citizen

2180	July 28, 2009	Kym Fraher, Private Citizen

2181	July 28, 2009	Betty J. Walters, Private Citizen

2182	July 28, 2009	Barbara Fitzpatrick, Private Citizen

2183	July 29, 2009	Robert Brennan, Private Citizen

2184	July 29, 2009	Virginia Velez, Private Citizen

2185	July 29, 2009	Brenon Duff, Private Citizen

2186	July 29, 2009	Patricia Mullen, Private Citizen

2187	July 29, 2009	Virginia Catanese, Private Citizen

2188	July 29, 2009	Sabine Axt, Private Citizen

2189	July 29, 2009	V. Catanese

2190	July 29, 2009	Maureen Nealon, Private Citizen

2191	July 29, 2009	Richard McGonagle, Private Citizen

2192	July 29, 2009	James Davies, Private Citizen

2193	July 29, 2009	Althea Day, Private Citizen

2194	July 29, 2009	Tim Mullins, Private Citizen

2195	July 29, 2009	Sandy Schmidt, Private Citizen

2196	July 29, 2009	Terrence and Elizabeth McCloskey, Private Citizens

2197	July 29, 2009	Careen Jennings, Private Citizen

2198	July 29, 2009	Carolyn Bell, Private Citizen

2199	July 29, 2009	Cindy Cohen, Private Citizen

2200	July 29, 2009	Charles R Shelly, Private Citizen

2201	July 29, 2009	Tria Shaffer, Private Citizen

2202	July 29, 2009	Charles Trost, Private Citizen

2203	July 29, 2009	Mary Crider, Private Citizen

2204	July 29, 2009	Christopher Walter, Private Citizen

2205	July 29, 2009	Cynthia Crutcher, Private Citizen

2206	July 29, 2009	Paula Katz, Private Citizen

2207	July 29, 2009	Arnold Simmel, Private Citizen

2208	July 29, 2009	Carlene Meeker, Private Citizen

2209	July 29, 2009	Robert Elliott, Private Citizen

2210	July 29, 2009	Patricia Gorospe, Private Citizen

2211	July 29, 2009	James Carrell, Private Citizen

2212	July 29, 2009	Tim Nistler, Private Citizen

2213	July 29, 2009	Marcia Brelsford, Private Citizen

2214	July 29, 2009	Alexis Raney, Private Citizen

2215	July 29, 2009	Myra Bailes, Private Citizen

2216	July 29, 2009	Judith Sanders, Private Citizen

2217	July 29, 2009	Maija Dravnieks, Private Citizen

2218	July 29, 2009	William Reeves, Private Citizen

2219	July 29, 2009	David McIntosh, Earthjustice

2220	July 29, 2009	Pat Westwater-Jong, Private Citizen

2221	July 29, 2009	Wayne Horlacher, Private Citizen

2222	July 29, 2009	Wendy Snyder, Private Citizen

2223	July 29, 2009	Mercedes Michalski, Private Citizen

2224	July 29, 2009	Kristen Holtvedt, Private Citizen

2225	July 29, 2009	Kathryn Rose, Private Citizen

2226	July 29, 2009	Michael Nola, Private Citizen

2227	July 29, 2009	Carol Collier, Private Citizen

2228	July 29, 2009	David Bridgeman, Private Citizen

2229	July 29, 2009	Burnell Vincent, Private Citizen

2230	July 30, 2009	Winfield Solin Coleman, Private Citizen

2231	July 30, 2009	M. Fitzsimmons, Private Citizen

2232	July 30, 2009	Sergio Monteiro, Private Citizen

2233	July 30, 2009	Liza Elliott, Private Citizen

2234	July 30, 2009	Margaret Lange, Private Citizen

2235	July 30, 2009	Marcia Maroon, Private Citizen

2236	July 30, 2009	Carla Rowton, Private Citizen

2237	July 30, 2009	Magdalena Hoersch, Private Citizen

2238	July 30, 2009	Nancy Thompson, Private Citizen

2239	July 30, 2009	Michael Ochs, Private Citizen

2240	July 30, 2009	Nasus Aransu, Private Citizen

2241	July 30, 2009	Nathan Drozd, Private Citizen

2242	July 30, 2009	Nathan Huerkamp, Private Citizen

2243	July 30, 2009	Anita Yandle, Private Citizen

2244	July 30, 2009	Lara Boeck, Private Citizen

2245	July 30, 2009	Lynne Hardesty, Private Citizen

2246	July 30, 2009	Terry Jensen, Private Citizen

2247	July 30, 2009	Maureen Porter, Private Citizen

2248	July 30, 2009	Padmaja Iyer, Private Citizen

2249	July 30, 2009	Paula Thornton, Private Citizen

2250	July 30, 2009	Nancy Minges, Private Citizen

2251	July 30, 2009	Martha P. Nochimson, Private Citizen

2252	July 30, 2009	Nick Heath, Private Citizen

2253	July 30, 2009	Pat Okerlund, Private Citizen

2254	July 30, 2009	David Harris, Private Citizen

2255	July 30, 2009	Keith Wells, Private Citizen

2256	July 31, 2009	Sherman Murray, Private Citizen

2257	August 03, 2009	Rick Breckenridge, Private Citizen

2258	August 03, 2009	Christy M. Woo, Private Citizen

2259	August 03, 2009	Donna Carr, MD, Private Citizen

2260	August 03, 2009	Dot Snowden, Private Citizen

2261	August 03, 2009	Ed Dickie, Private Citizen

2262	August 03, 2009	Virginia Kilgore, Private Citizen

2263	August 03, 2009	Jane Voisard, Private Citizen

2264	August 03, 2009	Bradley Esdale, Private Citizen

2265	August 03, 2009	Joen Mein, Private Citizen

2266	August 04, 2009	Zia Shields, Private Citizen

2267	August 04, 2009	Amanda Moore, Private Citizen

2268	August 04, 2009	Laura Arntz, Private Citizen

2269	August 04, 2009	Polly Bunce, Private Citizen

2270	August 04, 2009	Linda Mulka, Private Citizen

2271	August 04, 2009	Penelope Summers, Private Citizen

2272	August 05, 2009	Angel Lugo Santiago, APCAS, Inc.

2273	August 05, 2009	Cynthia Erville, Private Citizen

2274	August 05, 2009	Arvin Eyre, Private Citizen

2275	August 05, 2009	Paul Haseman, Private Citizen

2276	August 05, 2009	Pamela Ruediger, Private Citizen

2277	August 05, 2009	Carol Edwards, Private Citizen

2278	August 05, 2009	Quilla Miralia, Private Citizen

2279	August 05, 2009	Robert Burns, Private Citizen

2280	August 05, 2009	Linda McCracken, Private Citizen

2281	August 05, 2009	Rainbow Di Benedetto, Private Citizen

2282	August 05, 2009	Randolph Worsham, Private Citizen

2283	August 05, 2009	Jerry Rosenkoetter, Private Citizen

2284	August 05, 2009	M. Kissinger, Private Citizen

2285	August 05, 2009	Pete Richardson, Private Citizen

2286	August 05, 2009	Robert Posch, Private Citizen

2287	August 05, 2009	Robin Chandler, Private Citizen

2288	August 05, 2009	Rose Bertrand, Private Citizen

2289	August 05, 2009	Robert Ouradnik, Private Citizen

2290	August 05, 2009	Ann Weiner, Private Citizen

2291	August 05, 2009	Rob Sadler, Private Citizen

2292	August 05, 2009	Marjorie Rhodes, Private Citizen

2293	August 05, 2009	Salma Mushir, Private Citizen

2294	August 05, 2009	Sandy Justis, Private Citizen

2295	August 05, 2009	Sandra Carrubba, Private Citizen

2296	August 05, 2009	M.A. McDonald, Private Citizen

2297	August 05, 2009	Scott Brooks, Private Citizen

2298	August 05, 2009	Eugene L. Stackle, Baker County Economic
Development

2299	August 05, 2009	Robert Marquis, Private Citizen

2300	August 05, 2009	Ray Brammer, Private Citizen

2301	August 05, 2009	Mark Lea, Private Citizen

2302	August 06, 2009	Seth Edelman, Private Citizen

2303	August 06, 2009	Steven Gaber, Private Citizen

2304	August 06, 2009	Martin Pall, Private Citizen

2305	August 06, 2009	Daniel Salomon, Private Citizen

2306	August 06, 2009	Randy Juras, Private Citizen

2307	August 06, 2009	Kristin Reed, Private Citizen

2308	August 07, 2009	Fred Osterholtz, Private Citizen

2309	August 07, 2009	Dave Kisor, Private Citizen

2310	August 07, 2009	Gretchen Goodman, Private Citizen

2311	August 07, 2009	Anne Rehm, Private Citizen

2312	August 07, 2009	Gina Giacona, Private Citizen

2313	August 07, 2009	Ginny Lentz, Private Citizen

2314	August 07, 2009	Virginia Juliano, Private Citizen

2315	August 07, 2009	Ellie McCaffrey, Private Citizen

2316	August 07, 2009	Patricia Mann, Private Citizen

2317	August 07, 2009	Georgia Yelton, Private Citizen

2318	August 07, 2009	Jan Scott, Private Citizen

2319	August 07, 2009	Patricia Heithaus, Private Citizen

2320	August 07, 2009	Irvin Uphoff, Private Citizen

2321	August 07, 2009	Linda Burlingame, Private Citizen

2322	August 07, 2009	Annie Steele, Private Citizen

2323	August 07, 2009	Iman Latifpour, Private Citizen

2324	August 07, 2009	Ian Smith, Private Citizen

2325	August 07, 2009	John Ferriday, Private Citizen

2326	August 07, 2009	Jerry Comer, Private Citizen

2327	August 07, 2009	Joan Eliasoph, Private Citizen

2328	August 07, 2009	Jeremy Halinen, Private Citizen

2329	August 07, 2009	Jack Lockhart, Private Citizen

2330	August 07, 2009	Joanne Beeman, Private Citizen

2331	August 07, 2009	John Abrams, Private Citizen

2332	August 07, 2009	Susan Virostek, Private Citizen

2333	August 07, 2009	Janet Rees, Private Citizen

2334	August 07, 2009	Katherine Blum Russell, Private Citizen

2335	August 07, 2009	Judith Emerson, Private Citizen

2336	August 07, 2009	Kimberly Jarvis, Private Citizen

2337	August 07, 2009	Thomasin Kellermann, Private Citizen

2338	August 07, 2009	Mary Warren, Private Citizen

2339	August 07, 2009	Laurie Harris, Private Citizen

2340	August 07, 2009	Leah Boyd, Private Citizen

2341	August 07, 2009	Leif (Captain) Knutsen, Private Citizen

2342	August 07, 2009	Leon Tomlinson, Private Citizen

2343	August 07, 2009	Suzanne Yaeger, Private Citizen

2344	August 07, 2009	Susan Martin, Private Citizen

2345	August 07, 2009	Sandra White, Private Citizen

2346	August 07, 2009	Margo Sprenger, Private Citizen

2347	August 07, 2009	Robin Evans, Private Citizen

2348	August 07, 2009	Stephanie Hammond, Private Citizen

2349	August 07, 2009	Michael Thomas, Private Citizen

2350	August 07, 2009	Patricia Finder-Stone, Private Citizen

2351	August 07, 2009	Fran Sullivan-Fahs, Private Citizen

2352	August 07, 2009	Suzette Druzik, Private Citizen

2353	August 07, 2009	Susan Ebershoff-Coles, Private Citizen

2354	August 07, 2009	Sara Ross, Private Citizen

2355	August 07, 2009	Susan Cooper, Private Citizen

2356	August 11, 2009	Charles Schmalz, Private Citizen

2357	August 11, 2009	Hugh Curtler, III, Private Citizen

2358	August 11, 2009	Cyndi Norwitz, Private Citizen

2359	August 11, 2009	Cas Waugh, Private Citizen

2360	August 11, 2009	Doug Wood, Private Citizen

2361	August 11, 2009	Christopher Viavant, Private Citizen

2362	August 11, 2009	David Kleiman, Private Citizen

2363	August 11, 2009	David Payne, Private Citizen

2364	August 11, 2009	Rahim Shaikh, Private Citizen

2365	August 11, 2009	Barbara Dotson-Brooner, Private Citizen

2366	August 11, 2009	Donald Dougall, Private Citizen

2367	August 11, 2009	David Edwards, Private Citizen

2368	August 11, 2009	Deane Lavender, Private Citizen

2369	August 11, 2009	Diane Balter, Private Citizen

2370	August 11, 2009	Catherine Carpenter, Private Citizen

2371	August 11, 2009	Donald Walker, Private Citizen

2372	August 11, 2009	Doris Cellarius, Private Citizen

2373	August 11, 2009	Dorothy Patent, Private Citizen

2374	August 12, 2009	Louise Pillai, Private Citizen

2375	August 12, 2009	Janice Jacobson, Private Citizen

2376	August 13, 2009	Dee Rossi, Private Citizen

2377	August 13, 2009	Ann Cockrell, Private Citizen

2378	August 13, 2009	D. Shoemaker, Private Citizen

2379	August 13, 2009	Diane Dubendorff, Private Citizen

2380	August 14, 2009	Barbara Fukumoto, Private Citizen

2381	August 14, 2009	Keith Bussen, Vice President/General Manager,
Independent Concrete Pipe-St. Louis (Missouri Division)

2382	August 17, 2009	Kenneth R. Schmick, Controller, Art Wilson Company

2383	August 17, 2009	Darcy Cronin, Private Citizen

2384	August 18, 2009	Bruce Kosanovic, Private Citizen

2385	August 18, 2009	David Low, Private Citizen

2386	August 18, 2009	Deborah K. Dorsch, Private Citizen

2387	August 18, 2009	Deb Brown, Private Citizen

2388	August 18, 2009	Harold Denenberg, Private Citizen

2389	August 18, 2009	Dennis Feichtinger, Private Citizen

2390	August 18, 2009	Alex Oshiro, Private Citizen

2391	August 18, 2009	D. Johnson, Private Citizen

2392	August 18, 2009	Donna Bonetti, Private Citizen

2393	August 18, 2009	Amy Lowery, Private Citizen

2394	August 18, 2009	Ashley Holtorf, Private Citizen

2395	August 18, 2009	Anthony Drew, Private Citizen

2396	August 18, 2009	Dwight Johnson, Private Citizen

2397	August 18, 2009	Edward Janus, Private Citizen

2398	August 18, 2009	E. Clark, Private Citizen

2399	August 18, 2009	Elsa Bondar, Private Citizen

2400	August 18, 2009	Frank Spadazzi, Private Citizen

2401	August 18, 2009	Meredith Garmon, Private Citizen

2402	August 18, 2009	Gail Denemark, Private Citizen

2403	August 18, 2009	Julie Schaar, Private Citizen

2404	August 18, 2009	George Hassinger, Private Citizen

2405	August 18, 2009	Gerald Daley, Private Citizen

2406	August 18, 2009	Madeline von Foerster, Private Citizen

2407	August 18, 2009	Helen Corneli, Private Citizen

2408	August 18, 2009	Jerry Best, Private Citizen

2409	August 18, 2009	Paula D. Brennecke, Private Citizen

2410	August 18, 2009	Paula Olch, Private Citizen

2411	August 18, 2009	Patrick Hall, Private Citizen

2412	August 18, 2009	James A. Pierson, Private Citizen

2413	August 18, 2009	William Bell, Private Citizen

2414	August 18, 2009	Rachael Garcia, Private Citizen

2415	August 18, 2009	William Grant, Private Citizen

2416	August 18, 2009	Roger Holmen, Private Citizen

2417	August 18, 2009	Ted Terrasas, Private Citizen

2418	August 18, 2009	Sara Bhakti, Private Citizen

2419	August 18, 2009	Sherry Lewis, Private Citizen

2420	August 18, 2009	Evan Corrales, Private Citizen

2421	August 18, 2009	Jamie Specht, Private Citizen

2422	August 18, 2009	Karen Lizars, Private Citizen

2423	August 18, 2009	Susan Curry, Private Citizen

2424	August 18, 2009	Pat Turk, Private Citizen

2425	August 18, 2009	Caryl Wolstenholme, Private Citizen

2426	August 18, 2009	John Feins, Private Citizen

2427	August 18, 2009	Trevor Sheldon, Private Citizen

2428	August 18, 2009	Colin Forwood, Private Citizen

2429	August 18, 2009	Peter Wilsnack, Private Citizen

2430	August 18, 2009	Norman F. Ramsey, Private Citizen

2431	August 18, 2009	Susan Hathaway, Private Citizen

2432	August 18, 2009	Cathy Church, Private Citizen

2433	August 18, 2009	Linda Swanson, Private Citizen

2434	August 18, 2009	George Manning, Private Citizen

2435	August 18, 2009	Vivian Layton, Private Citizen

2436	August 19, 2009	Anson Chi, Private Citizen

2437	August 19, 2009	James Conroy, Private Citizen

2438	August 19, 2009	Jim Tower, Private Citizen

2439	August 19, 2009	John Neville, Private Citizen

2440	August 19, 2009	Joan Fuhry, Private Citizen

2441	August 19, 2009	John Stifter, Private Citizen

2442	August 19, 2009	James Snively, Private Citizen

2443	August 19, 2009	James Sorrells, Private Citizen

2444	August 19, 2009	Robert Summers, Private Citizen

2445	August 19, 2009	Katherine Nelson, Private Citizen

2446	August 19, 2009	Ken Bugler, Private Citizen

2447	August 19, 2009	Karen Schoemer, Private Citizen

2448	August 19, 2009	Rose Lernberg, Private Citizen

2449	August 19, 2009	Jolene Rutherford, Private Citizen

2450	August 19, 2009	Lorie Vorraro, Private Citizen

2451	August 19, 2009	Richard McBane, Private Citizen

2452	August 19, 2009	Deborah Jamini, Private Citizen

2453	August 19, 2009	Michael Spink, Private Citizen

2454	August 19, 2009	Mike Ewall, Private Citizen

2455	August 19, 2009	Lori Martell, Private Citizen

2456	August 19, 2009	M'Lou Christ, Private Citizen

2457	August 19, 2009	Michael Vinciquerra, Private Citizen

2458	August 19, 2009	Nikhil Lucas Kamat, Private Citizen

2459	August 19, 2009	Joy Schochet, Private Citizen

2460	August 19, 2009	Hudson Ansley, Private Citizen

2461	August 19, 2009	Connie Mariottini, Private Citizen

2462	August 19, 2009	Doug Gawoski, Private Citizen

2463	August 19, 2009	Don Brennecke, Private Citizen

2464	August 19, 2009	Steven Calender, Private Citizen

2465	August 19, 2009	Evelyn Callaghan, Private Citizen

2466	August 19, 2009	Gerrit Crouse, Private Citizen

2467	August 19, 2009	Gregory and Renate Syme, Private Citizens

2468	August 19, 2009	Wanda Nichols, Private Citizen

2469	August 19, 2009	Jeffrey Peck, Private Citizen

2470	August 19, 2009	Alan Rosenthal, Private Citizen

2471	August 19, 2009	Anneta Buster, Private Citizen

2472	August 19, 2009	Janet Foster, Private Citizen

2473	August 19, 2009	Kevin Scott, Private Citizen

2474	August 20, 2009	Jon Halstein, Private Citizen

2475	August 20, 2009	Julia Sendor, Private Citizen

2476	August 20, 2009	David Price, Member of US Congress

2477	August 20, 2009	Jim (Surname illegible), Private Citizen

2478	August 20, 2009	Kenneth H. Baker, Private Citizen

2479	August 20, 2009	Martha B. Anderson, Private Citizen

2480	August 20, 2009	Ted Ferrioli, Oregon State Senator District 30

2481	August 20, 2009	Winnie Chang, Private Citizen

2482	August 20, 2009	Marjorie Rogalski, Private Citizen

2483	August 20, 2009	MaryJane Barnes, Vice President, Lar-Ken Concrete
Products, Inc

2484	August 20, 2009	Howard Marica Miller, Private Citizen

2485	August 20, 2009	Peter Sullivan, Private Citizen

2486	August 20, 2009	Patricia Mena, Private Citizen

2487	August 20, 2009	Steven Skal, Private Citizen

2488	August 20, 2009	Amy Mall, Private Citizen

2489	August 20, 2009	Susan Innes, Private Citizen

2490	August 20, 2009	Tom Matarrese, Private Citizen

2491	August 20, 2009	Wilma Bradbeer, Private Citizen

2492	August 20, 2009	Jo Greenwald, Private Citizen

2493	August 20, 2009	Beth Stewart, Private Citizen

2494	August 20, 2009	Robert Mazairz, Private Citizen

2495	August 20, 2009	Sharon Keys, Private Citizen

2496	August 20, 2009	Amanda Gerwig, Private Citizen

2497	August 20, 2009	Barb Cole, Private Citizen

2498	August 21, 2009	Anonymous public comment

2499	August 21, 2009	Jeff C. Calkins, Private Citizen

2500	August 21, 2009	Richard F. Burly, II, Private Citizen

2501	August 21, 2009	Mary Perry, Private Citizen

2502	August 21, 2009	Chris Barrett, Private Citizen

2503	August 21, 2009	Mel McQuarrie, Binggeli Rock Products, Inc.

2504	August 24, 2009	Wayne A. Mayer, Private Citizen

2505	August 24, 2009	J.F. Griffin, Private Citizen

2506	August 24, 2009	Sandra McDonough, President & CEO, Portland
Business Alliance

2507	August 24, 2009	John Ledger, VP, Associated Oregon Industries (AOI)

2508	August 24, 2009	Alan Cuperus, Private Citizen

2509	August 24, 2009	Paulette M. Switzer-Tatum, Private Citizen

2510	August 24, 2009	Barbara Smith, Private Citizen

2511	August 24, 2009	Dana Allen, Private Citizen

2512	August 24, 2009	Candy Halliburton, Private Citizen

2513	August 24, 2009	Bill Himelhoch, Private Citizen

2514	August 17, 2009	Ruth Fischer, Private Citizen

2515	August 17, 2009	Reihana Robinson, Private Citizen

2516	August 17, 2009	Robert Grady, Private Citizen

2517	August 17, 2009	Robert Coleman, Private Citizen

2518	August 17, 2009	Robert Bliss, Private Citizen

2519	August 17, 2009	Robert Bruner, Private Citizen

2520	August 17, 2009	Robert Warren, Private Citizen

2521	August 17, 2009	Richard Winkler, Private Citizen

2522	August 17, 2009	Scott Callaway, Private Citizen

2523	August 17, 2009	Scott Johnsen, Private Citizen

2524	August 17, 2009	Greta Bliss, Private Citizen

2525	August 18, 2009	Peg Diaz, Private Citizen

2526	August 18, 2009	Stephen Gustafson, Private Citizen

2527	August 18, 2009	Charles Shackelford, Private Citizen

2528	August 17, 2009	Shelly Thomas, Private Citizen

2529	August 18, 2009	Jeffrey Green, Private Citizen

2530	August 17, 2009	Peter Stickney, Private Citizen

2531	August 18, 2009	Rhiannon Lincoln, Private Citizen

2532	August 17, 2009	Phyllis Hasbrouck, Private Citizen

2533	August 17, 2009	Susan Chandler, Private Citizen

2534	August 17, 2009	Suzie Kidder, Private Citizen

2535	August 17, 2009	Rosemary Hicks, Private Citizen

2536	August 17, 2009	Ted Robinson, Private Citizen

2537	August 17, 2009	Tim McMullen, Private Citizen

2538	August 17, 2009	Elizabeth Kenyon, Private Citizen

2539	August 17, 2009	Thamar Wherrit, Private Citizen

2540	August 17, 2009	Don Segraves, Private Citizen

2541	August 17, 2009	David Bellos, Private Citizen

2542	August 18, 2009	Edra Bogle, Private Citizen

2543	August 18, 2009	Alan McCauley, Private Citizen

2544	August 18, 2009	Evelyn Kountoupes, Private Citizen

2545	August 17, 2009	Richard Dwyer, Private Citizen

2546	August 18, 2009	Christopher Eliot, Private Citizen

2547	August 17, 2009	Rebecca Pontikes, Private Citizen

2548	August 17, 2009	Jeanne Ellgar, Private Citizen

2549	August 18, 2009	Eric Cartier, Private Citizen

2550	August 17, 2009	Elaine Ransom, Private Citizen

2551	August 17, 2009	Maxim Fetissenko, Private Citizen

2552	August 17, 2009	David Beebe, Private Citizen

2553	August 17, 2009	Gay Kramer-Dodd, Private Citizen

2554	August 18, 2009	Mike Desmond, Private Citizen

2555	August 18, 2009	Susan Lippman, Private Citizen

2556	August 17, 2009	Richard Kipp, Private Citizen

2557	August 18, 2009	Alicia Mein, Private Citizen

2558	August 17, 2009	Robert Toon, Private Citizen

2559	August 17, 2009	Linda Mattox, Private Citizen

2560	August 18, 2009	Lara Valigorsky, Private Citizen

2561	August 17, 2009	Lucille Bertuccio, Private Citizen

2562	August 17, 2009	Lisa Butterfield, Private Citizen

2563	August 18, 2009	Susan Hawes, Private Citizen

2564	August 17, 2009	William Smith, Private Citizen

2565	August 17, 2009	William Audette, Private Citizen

2566	August 18, 2009	Alice Littlefield, Private Citizen

2567	August 17, 2009	George Loveday, Private Citizen

2568	August 21, 2009	Ione Woods, Private Citizen

2569	August 17, 2009	Zorine Rinaldi, Private Citizen

2570	August 17, 2009	Marilyn Britton, Private Citizen

2571	August 18, 2009	Keith Allen, Private Citizen

2572	August 17, 2009	Lois Reborne, Private Citizen

2573	August 18, 2009	Barbara Toshalis, Private Citizen

2574	August 17, 2009	Bonnie Pilcher, Private Citizen

2575	August 18, 2009	Karen Bunkowski, Private Citizen

2576	August 17, 2009	Cindy Cox, Private Citizen

2577	July 20, 2009	Bill Cagle, Private Citizen

2578	August 17, 2009	Alyn Goldsmith, Private Citizen

2579	August 17, 2009	Barry Rein, Private Citizen

2580	August 17, 2009	Caroline Herzenberg, Private Citizen

2581	August 18, 2009	Carroll English, Private Citizen

2582	August 17, 2009	Cat Koehn, Private Citizen

2583	August 18, 2009	Charles Trebes, Private Citizen

2584	August 18, 2009	Christina Muniz, Private Citizen

2585	August 18, 2009	Greg Ciapponi, Private Citizen

2586	August 17, 2009	Doug Richardson, Private Citizen

2587	August 17, 2009	Charles Wonka, Private Citizen

2588	August 17, 2009	Anne Humphrey, Private Citizen

2589	August 17, 2009	R.H. Kroell, Private Citizen

2590	August 17, 2009	David Kleiman, Private Citizen

2591	August 17, 2009	Dawn Nelson, Private Citizen

2592	August 18, 2009	Deirdre Wilson, Private Citizen

2593	August 17, 2009	David Hucknall, Private Citizen

2594	August 17, 2009	Delcie Light, Private Citizen

2595	August 18, 2009	Donna Bonetti, Private Citizen

2596	August 17, 2009	David Lewton, Private Citizen

2597	August 17, 2009	Lenore Bentz, Private Citizen

2598	August 17, 2009	Laurel Herendeen, Private Citizen

2599	August 17, 2009	Mark Robertson, Private Citizen

2600	August 17, 2009	Marlena Machol, Private Citizen

2601	August 17, 2009	Martha Burton, Private Citizen

2602	August 17, 2009	Melanie Tang, Private Citizen

2603	August 17, 2009	H. Michael Laybourn, Private Citizen

2604	August 17, 2009	Margaret Fontenot, Private Citizen

2605	August 17, 2009	Linda Griggs, Private Citizen

2606	August 17, 2009	Matthew Comber, Private Citizen

2607	August 17, 2009	Anita Trenner, Private Citizen

2608	August 17, 2009	Marie Lloyd, Private Citizen

2609	August 18, 2009	Katherine Narveson, Private Citizen

2610	August 17, 2009	George Thomas, Private Citizen

2611	August 17, 2009	Megan Stephenson, Private Citizen

2612	August 17, 2009	Gary Neu, Private Citizen

2613	August 17, 2009	Hal Nixon, Private Citizen

2614	August 17, 2009	Helen Putera, Private Citizen

2615	August 17, 2009	Jerry Best, Private Citizen

2616	August 17, 2009	Joanie Alexander, Private Citizen

2617	August 18, 2009	Judith Arcana, Private Citizen

2618	August 17, 2009	John Pierson, Private Citizen

2619	August 18, 2009	Jennifer Jackson, Private Citizen

2620	August 18, 2009	Nora Gallaher, Private Citizen

2621	August 17, 2009	Anesa Miller, Private Citizen

2622	August 20, 2009	G. Vinson Hellwig, Michigan, Co-Chair, NACAA Air
Toxics Committee

Robert H. Colby, Chattanooga, TN, Co-Chair, NACAA Air Toxics Committee

2623	August 17, 2009	James Sanders, Private Citizen

2624	August 17, 2009	Janet Jones, Private Citizen

2625	August 18, 2009	Joan Wilce, Private Citizen

2626	August 17, 2009	John Thomas, Private Citizen

2627	August 17, 2009	Joseph Belisle, Private Citizen

2628	August 18, 2009	Judith Eiseman, Private Citizen

2629	August 18, 2009	Judith Gallagher, Private Citizen

2630	August 17, 2009	Kathryn Bankston, Private Citizen

2631	August 17, 2009	Keith Powell, Private Citizen

2632	August 17, 2009	Kevin Shilling, Private Citizen

2633	August 18, 2009	Angele Ellis, Private Citizen

2634	August 17, 2009	Kevin Riley, Private Citizen

2635	August 17, 2009	Catherine Harned, Private Citizen

2636	August 17, 2009	Jon Krampner, Private Citizen

2637	August 18, 2009	Nicolas Pekari, Private Citizen

2638	August 22, 2009	John A. Cooper, Cooper Environmental Services LLC
(CES)

2639	August 17, 2009	Ted Taylor, Private Citizen

2640	August 17, 2009	Barbara Roberts, Private Citizen

2641	August 17, 2009	Vanessa Weeks, Private Citizen

2642	August 18, 2009	Sue Bond, Private Citizen

2643	August 17, 2009	Harriet Irby, Private Citizen

2644	August 18, 2009	Thomas Martinez, Private Citizen

2645	August 17, 2009	Tom Ragouzis, Private Citizen

2646	August 17, 2009	Victoria Baker, Private Citizen

2647	August 17, 2009	Scott Ohlman, Private Citizen

2648	August 17, 2009	Paulette Bliss, Private Citizen

2649	August 20, 2009	Dave Bennett, Kaman Industrial Technologies -
Ontario Oregon

2650	August 20, 2009	Neil Sulivan, Granite Concrete Co, Inc

2651	August 20, 2009	Mike Jeffries, Goat Rock Mining Corp.

2652	July 14, 2009	Mass Comment Campaign, Sponsoring Organization
Unknown (8,715)

2653	August 23, 2009	Kenneth Carlson, C&A Equipment Company

2654	August 20, 2009	Cass R. Vanderwiele, President/Owner, Triple C.
Redi-Mix

2655	August 20, 2009	Gary and Kathy Bloomer, Private Citizens

2656	August 14, 2009	Rhonda Fox, Private Citizen

2657	August 24, 2009	Wesley B. Prouty, Private Citizen

2658	July 20, 2009	Mass Comment Campaign, Sponsoring Organization
Unknown (80)

2659	August 21, 2009	Ronald Verini, Private Citizen 

2660	August 25, 2009	John C. Colbert, Vice President and General Manager
Automatic Wilbert Vault Co., Inc.

2661	August 25, 2009	Bob Vandecar, Business Representative and
Organizer, International Brotherhood of Electrical Workers

2662	September 09, 2009	J. Robenolt, Private Citizen

2663	September 09, 2009	M. Minow, Private Citizen

2664	August 08, 2009	Carmelita Holland, Private Citizen

2665	August 16, 2009	Thomas Bunch, Private Citizen

2666	August 18, 2009	Heidi Martin, Private Citizen

2667	August 20, 2009	Jeff Mendes, Regional Sales Manager, Ash Grove
Cement Company

2668	August 21, 2009	Michael A. Henigan, Private Citizen

2669	August 20, 2009	Diana L. Henigan, Private Citizen

2670	August 17, 2009	Robert J. Schreiber, Jr., President, Schreiber,
Yonley & Associates

2671	August 10, 2009	Mark Holliday, Business Manager and Financial
Secretary, Local 701, International Union of Operating Engineers (IUOE)

2672	July 30, 2009	Dennis Dorrah, Mayor, City of Baker City, Oregon (OR)

2673	August 17, 2009	Sarita Raney, Private Citizen

2674	August 19, 2009	Herbert L. Keyser, Private Citizen

2675	August 26, 2009	Thomas Mumley, Assistant Executive Officer, San
Francisco Bay Regional Water Quality Control Board, California EPA

2676	August 31, 2009	Robert Minister, Ash Grove Cement Company

2677	August 25, 2009	David Hayes, Private Citizen

2678	August 24, 2009	Deb Santa, Private Citizen

2679	August 24, 2009	Dennis Swaer, Private Citizen

2680	August 19, 2009	Devon Downeysmith, Private Citizen

2681	August 28, 2009	Diane Dillard, Private Citizen

2682	September 01, 2009	Jon Trygve Grey, Process Engineer, Seattle
Plant, Ash Grove Cement Company

2683	August 28, 2009	Dr. Dick Schoech, Private Citizen

2684	August 21, 2009	J. Ding, Private Citizen

2685	August 28, 2009	Edgar and Dora Stahl, Private Citizens

2686	August 26, 2009	Edwina Allen, Private Citizen

2687	August 29, 2009	Barbara Lawrence, Private Citizen

2688	August 28, 2009	Fred Kennedy, Private Citizen

2689	August 28, 2009	Carol Fulton, Private Citizen

2690	August 28, 2009	Rollie Blackwell, Private Citizen

2691	August 28, 2009	Ruth Cook, Private Citizen

2692	August 22, 2009	Sara Freed, Private Citizen

2693	August 19, 2009	Susan Lewis, Private Citizen

2694	August 19, 2009	Sean Edmison, Private Citizen

2695	August 29, 2009	Gail Krueger, Private Citizen

2696	August 28, 2009	Gary and Christy Richards, Private Citizens

2697	August 28, 2009	Gayle E. Marshall, Private Citizen

2698	August 24, 2009	Grace Koehler, Private Citizen

2699	August 19, 2009	Holly Berkowitz, Private Citizen

2700	August 28, 2009	Howard Tubre, Private Citizen

2701	August 28, 2009	John Rath, Private Citizen

2702	August 28, 2009	John White, Private Citizen

2703	August 28, 2009	Shari Buhan, Private Citizen

2704	August 28, 2009	Ilana Shumsky, Private Citizen

2705	August 25, 2009	Shayne Vinagre, Private Citizen

2706	August 30, 2009	James Presley, Private Citizen

2707	August 26, 2009	Beth Duke, Private Citizen

2708	August 24, 2009	Sheila Boester, Private Citizen

2709	August 20, 2009	Sheriel Hubble, Private Citizen

2710	August 19, 2009	Sandra Sawyer, Private Citizen

2711	August 29, 2009	Carolina Vornberg, Private Citizen

2712	August 25, 2009	Jane Janzen, Private Citizen

2713	August 24, 2009	Bernard Feiss, Private Citizen

2714	August 24, 2009	Jay Krajic, Private Citizen

2715	August 27, 2009	Jean Benedict, Private Citizen

2716	August 24, 2009	Jean Gonzales, Private Citizen

2717	August 28, 2009	Jonathan Grimm, Private Citizen

2718	August 28, 2009	Judith Godinez, Private Citizen

2719	August 25, 2009	Betsy Goodman, Private Citizen

2720	August 26, 2009	Mike Chandler, Private Citizen

2721	August 24, 2009	Michael Lanza, Private Citizen

2722	August 25, 2009	Jean Maryborn, Private Citizen

2723	August 26, 2009	Michael Rutledge, Private Citizen

2724	August 28, 2009	Mona Kandeler, Private Citizen

2725	August 30, 2009	Mona Mehdy, Private Citizen

2726	August 24, 2009	Muriel Roberts, Private Citizen

2727	August 20, 2009	Michele Mason, Private Citizen

2728	August 29, 2009	Nancy Holcomb, Private Citizen

2729	August 28, 2009	Elizabeth Carman, Private Citizen

2730	August 28, 2009	Stacy Alaimo, Private Citizen

2731	August 19, 2009	Steven Libby, Private Citizen

2732	August 28, 2009	Stewart Snider, Private Citizen

2733	August 25, 2009	Susan McKean, Private Citizen

2734	August 26, 2009	Susan Norton, Private Citizen

2735	August 30, 2009	Tamara Teague, Private Citizen

2736	August 22, 2009	Anthony Wilson, Private Citizen

2737	August 25, 2009	Thomas Heinrich, Private Citizen

2738	August 25, 2009	Todd Davis, Private Citizen

2739	August 28, 2009	Charles Hobbs, Private Citizen

2740	August 24, 2009	Tom Godard, Private Citizen

2741	August 19, 2009	Patricia M. Williams, Private Citizen

2742	August 28, 2009	Virginia Gewax, Private Citizen

2743	August 19, 2009	Kathleen Wagner, Private Citizen

2744	August 29, 2009	Wayne Greene, Private Citizen

2745	August 28, 2009	Jeffrey Jacoby, Private Citizen

2746	August 24, 2009	Jeffrey Davis, Private Citizen

2747	August 29, 2009	Jo Adams, Private Citizen

2748	August 24, 2009	John Laz, Private Citizen

2749	August 29, 2009	Olive Hershey, Private Citizen

2750	August 22, 2009	Candia McNeal, Private Citizen

2751	August 25, 2009	Pam Colesworthy, Private Citizen

2752	August 24, 2009	Paul Martin, Private Citizen

2753	August 19, 2009	Philip K. Kreitner, Private Citizen

2754	August 24, 2009	Peggy Jordan, Private Citizen

2755	August 28, 2009	Peggy Sakach, Private Citizen

2756	August 24, 2009	Perry Brown, Private Citizen

2757	August 19, 2009	Arvin Eyre, Private Citizen

2758	August 28, 2009	Raymond E. Bowden, Private Citizen

2759	August 25, 2009	Regi Teasley, Private Citizen

2760	August 28, 2009	William Blackley, Private Citizen

2761	August 22, 2009	Caryl Wolstenholme, Private Citizen

2762	August 28, 2009	Dr. Yvonne Hansen, Private Citizen

2763	August 24, 2009	Charles Trost, Private Citizen

2764	August 24, 2009	Claire Casey, Private Citizen

2765	August 25, 2009	Craig Willey, Private Citizen

2766	August 21, 2009	Paige Zhang, Private Citizen

2767	August 29, 2009	Dan Smith, Private Citizen

2768	August 28, 2009	Vic Gilliam, House District 18, Oregon House of
Representatives

2769	August 28, 2009	Carol Coulston, Private Citizen

2770	August 20, 2009	Dan Stabel, Private Citizen

2771	August 24, 2009	David Anderson, Private Citizen

2772	August 19, 2009	Alan Mahrenholz, Private Citizen

2773	August 28, 2009	George L. Gilman, State Representative, District
55, Oregon House of Representatives

2774	August 24, 2009	Andrea Sparrow, Private Citizen

2775	August 28, 2009	David Davidson, Private Citizen

2776	August 28, 2009	Jenoa Ingram, Oregon Independent Aggregate
Association (OIAA)

2777	August 30, 2009	John E. Huffman, State Representative, House
District 59

2778	August 31, 2009	Norma Job, Private Citizen

2779	undated	Patrick Flanagan, Private Citizen

2780	July 20, 2009	Brittni Philips, Private Citizen

2781	August 31, 2009	Donna Piercy, Private Citizen

2782	July 20, 2009	William Edgar, Private Citizen

2783	July 22, 2009	Tim Burrows, Private Citizen

2784	July 28, 2009	Hilary Cox, Private Citizen

2785	August 31, 2009	Wendy Hosman, Private Citizen

2786	September 1, 2009	Deborah Behles, Staff Attorney

Lucas Williams, Graduate Fellow

Environmental Law and Justice Clinic at Golden Gate University on behalf
of  West Valley Citizens Air Watch and Bayview Hunters Point Community
Advocates

2787	August 31, 2009	Daniel Kocurek, Private Citizen

2788	August 31, 2009	Jerry R. Boggs, Private Citizen

2789	August 31, 2009	Charles Morgan, Private Citizen

2790	August 31, 2009	Jensie Madden, Private Citizen

2791	August 28, 2009	Jules Dylan Stuer, Private Citizen

2792	August 24, 2009	Julie Grove, Private Citizen

2793	August 24, 2009	Karen Pfleger, Private Citizen

2794	August 30, 2009	Kathleen Tanaka, Private Citizen

2795	August 24, 2009	Kathy Patlovich, Private Citizen

2796	August 27, 2009	Ken Dunbar, Private Citizen

2797	August 25, 2009	Laura Hartman, Private Citizen

2798	August 19, 2009	Kim Young, Private Citizen

2799	August 24, 2009	Betsy Bridge, Private Citizen

2800	August 26, 2009	Kirsten Severud, Private Citizen

2801	August 24, 2009	Laura Todd, Private Citizen

2802	August 19, 2009	Leonard and Nancy Deadman, Private Citizens

2803	August 25, 2009	Leonora Midgley, Private Citizen

2804	August 24, 2009	Leslie Bradshaw, Private Citizen

2805	August 28, 2009	Linda Knowles, Private Citizen

2806	August 24, 2009	Lorna Emdy, Private Citizen

2807	August 20, 2009	Mary Ann Lambert, Private Citizen

2808	August 25, 2009	Marlene Strong, Private Citizen

2809	August 28, 2009	Martha Clinkscale, Private Citizen

2810	August 31, 2009	Mayor Joe Dominick and City of Ontario City
Council, City of Ontario, Oregon

2811	August 27, 2009	Cliff Bentz, State Representative, District 60,
Oregon State House of Representatives

2812	August 27, 2009	Randall B. Hampton, President, Hampton's Inc. DBA
Rogers Asphalt Paving Company

2813	August 28, 2009	Art Wilson, Cal-CM Plus

2814	August 28, 2009	Wendt, Private Citizen

2815	August 25, 2009	Flack, Private Citizen

2816	September 18, 2009	Martha H. Keating, Director of Research
Translation, Nicholas School of Environment, Duke University on behalf
of the Children's Environmental Health Initiative (CEHI)

Attachment – Letter

2817	August 21, 2009	Gould, Private Citizen

2818	August 28, 2009	Faidley, Private Citizen

2819	August 24, 2009	Richard Heimsch, Private Citizen

2820	August 28, 2009	Richard Wilder, Private Citizen

2821	August 24, 2009	Rob Santa, Private Citizen

2822	August 28, 2009	Robert Burns, Private Citizen

2823	April 2009	Supporting Statement for NESHAP for Portland Cement
Plants (40 CFR part 63, subpart LLL) – Not a public comment

2824	September 18, 2009	E. Rafacz, Private Citizen

2825	August 25, 2009	Robsert Vestal, MD, Private Citizen

2826	August 24, 2009	Rocci Johnson, Private Citizen

2827	August 24, 2009	Roger Harrer, Private Citizen

2828	September 18, 2009	Al Armendariz, Ph.D., Private Citizen

2829	September 18, 2009	Al Armendariz, Private Citizen

2830	September 18, 2009	Gary A. Molchan, Vice President, Environmental
Affairs, Essroc Cement Corporation

2831	September 18, 2009	Jan Alexander, Minerals Policy Coordinator,
Eastern Oregon Mining Association (EOMA)

2832	September 3, 2009	Howard L. Gilberg, Guida, Slavich & Flores on
behalf of the Tile Council of North America, Inc. (TCNA)

2833	September 18, 2009	Laura Skaer, Executive Director, Northwest
Mining Association (NWMA)

2834	September 18, 2009	Geoffrey R. Gisler, Staff Attorney, Southern
Environmental Law Center (SELC)

2835	September 18, 2009	James R. Roewer, Executive Director, Utility
Solid Waste Activities Group (USWAG)

2836	September 18, 2009	James C. Morris III, Thompson & Knight LLP on
behalf of Eagle Materials, Inc.

2837	September 4, 2009	John A. Cooper and Drag A. Petterson, Cooper
Environmental Services, LLC (CES)

2838	September 4, 2009	John A. Cooper and Krag A. Petterson, Cooper
Environmental Services, LLC (CES)

2839	September 18, 2009	Alva R. Mascall, et al., Ash Grove Cement
Company

2840	September 18, 2009	T. Kerry, Private Citizen

2841	September 18, 2009	David Perkins, Corporate Director,
Communications and Government Affairs, TXI Operations, LP

2842	September 18, 2009	Andrew M. Cuomo, Attorney General, State of New
Yhork et al.

2843	September 21, 2009	Susie Devergranne, Chairperson, Davenport
Environmental Health Coalition

2844	September 21, 2009	Patrick M. Raher, Hogan and Hartson, LLP

On behalf of National Cement Company (NCC)

2845	September 21, 2009	Andrew T. O'Hare, Vice President, Regulatory
Affairs, Portland Cement Association (PCA)

2846	September 21, 2009	Robert D. Bessette, President, Council of
Industrial Boiler Owners (CIBO)

2847	September 2, 2009	C. Puljan, Private Citizen

2848	September 21, 2009	Andrew T. O'Hare, Vice President, Regulatory
Affairs, Portland Cement Association (PCA)

2849	September 4, 2009	Jim Griffin, Senior Director, Regulatory and
Technical Affairs, American Chemistry Council (ACC)

2850	September 21, 2009	Robert J. Kaleel, Manager, Air Quality Planning
Section, Bureau of Air, Illinois Environmental Protection Agency
(Illinois EPA)

2851	September 21, 2009	Andrew T. O'Hare, Vice President, Regulatory
Affairs, Portland Cement Association (PCA) [Part 3 Daily Chart]

2852	September 01, 2009	Chuck Buchanan, Private Citizen

2853	September 01, 2009	David Baker, Private Citizen

2854	September 01, 2009	John Unger, Private Citizen

2855	September 21, 2009	Andrew T. O'Hare, Vice President, Regulatory
Affairs, Portland Cement Association (PCA)

Part 4: Appendix 2 to Appendix 12

2856	September 01, 2009	Louis T. Scharbrough, Private Citizen

2857	September 01, 2009	Swanna Tanger, Private Citizen

2858	September 21, 2009	Edward X. Junia, Esquire, Attorney-at-Law

2859	September 21, 2009	Craig S. Cambell, Vice President, Environmental
& Government Affairs, Lafarge North America - Cement Division

2860	September 21, 2009	Anne Hedges, Program Director, Montana
Environmental Information Center (MEIC)

2861	August 27, 2009	Bruce Schacht, Environmental Engineer, Columbia
Steel Casting Co., Inc

2862	September 01, 2009	Basil D. Jacobson, Vice President, Concrete
Planet Manager, United Materials of Great Falls, Inc.

2863	September 04, 2009	Curtis Lesslie, P.E., Director of Environmental
Affairs, Ash Grove Cement Company

2864	September 08, 2009	Hector Ybanez, Holcim (US), Inc.

2865	September 01, 2009	W.C. Jacobson, Private Citizen

2866	September 01, 2009	T. Kerby, Private Citizen

2867	September 04, 2009	Alvaro A. Linero, P.E., Program Administrator,
Special Projects Section, Bureau of Air Regulation, Division of Air
Resources Management, Florida Department of Environmental Protection

2868	September 21, 2009	Christopher Lish, Private Citizen

2869	September 02, 2009	Scott Burns, Chief Executive Officer (CEO),
BURNCO Rock Products Ltd (BURNCO)

2870	September 03, 2009	Bryan Brendle, Director, Energy and Resources
Policy, National Association of Manufacturers (NAM)

2871	September 03, 2009	B. Johnson, Private Citizen

2872	September 03, 2009	Geoff Brosseau, Executive Director, Bay Area
Stormwater Management Agencies Association (BASMAA)

2873	September 22, 2009	Andy O'Hare, Portland Cement Association (Part
5: Appendix 13 and Appendix 14)

2874	September 04, 2009	Daniel B. Nugent, Vice President, Environmental
Affairs, Buzzi Unicem USA, Inc.

2875	September 04, 2009	E.X. Junia, Private Citizen

2876	September 04, 2009	F. Pinto, Private Citizen

2877	September 01, 2009	Raymond L. Calkins, DVM

Lynn M. Calkins

Private Citizens

2878	September 05, 2009	Charles Hofmann, MD

2879	September 04, 2009	John F. Chadbourne, Ph.D., President, Chadbourne
Environment & Safety Programs, L.L.C. (CESP)

2880	September 01, 2009	Richard Angstrom, Oregon Concrete & Aggregate
Producers Association (OCAPA)

2881	September 04, 2009	Mariah S. Chen, Private Citizen

2882	September 04, 2009	Virgil L. Thomas, Private Citizen

2883	September 01, 2009	Lisa Kerby, Private Citizen

2884	September 02, 2009	Judge Dan P. Joyce, Commissioner Louis Wettstein

Commissioner Jim Nakano, Malheur County Court

2885	September 01, 2009	Norma Job, Private Citizen

2886	September 01, 2009	Anonymous Public Comment

2887	September 04, 2009	Peter T. Kimmel, Vice President (VP),
Operations, Armstrong Cement & Supply Co.

2888	September 04, 2009	Doug Roark, Manager, Environmental and Process,
GCC of America

2889	September 09, 2009	John Richards, Ph.D., P.E., Air Control
Techniques, P.C.

2890	September 04, 2009	Susan Miller, Vice President, Environmental,
Health & Safety, Brick Industry Association (BIA)

Terry Schimmel, BIA MACT Task Force Lead, Vice President, Manufacturing,
Boral Bricks

2891	September 04, 2009	Joseph Otis Minott, Executive Director, Clean
Air Council

2892	September 04, 2009	Scott Trainor, General Manager, Thiessen Team
USA Inc.

2893	September 04, 2009	Joyce M. Eden et al., West Valley Citizens Air
Watch (WVCAW) ad Monica Wilson, Global Alliance for Incinerator
Alternatives (GAIA)

2894	September 03, 2009	Frank Schaedlich, Vice President Research &
Development, Tekran Instruments Corporation

2895	September 03, 2009	Jennifer Mesaros, Private Citizen

2896	September 04, 2009	Lee Jamison, Private Citizen

2897	September 05, 2009	Lindsey L. Hrizuk, Private Citizen

2898	September 09, 2009	Avinash Kar, Natural Resources Defense Council
(NRDC) et al.

2899	September 04, 2009	Kevin Blankenship, Private Citizen

2900	November 12, 2009	Timothy L. Matz, Director, Environmental Affairs,
Lehigh Hanson, Inc. (Lehigh Cement Company)

2901	September 09, 2009	Melvin E. Keener, Ph.D., Executive Director,
Coalition for Responsible Waste Incineration (CRWI)

2902	September 04, 2009	Jacqueline K. Clark, Private Citizen

2903	September 04, 2009	Charles Simpson, Mayor, City of Huntington,
Oregon (OR)

2904	September 01, 2009	Alan Finch, Private Citizen

2905	September 01, 2009	James Walton and Rebecca L. Wood, Private
Citizens

2906	September 04, 2009	Kelly Campbell BA, Executive Director, Joseph
Miller Ph.D., and Susan Katz M.D., Members, Board of Directors and Maye
Thompson RN, Ph.D., Environmental Health Program Director, Oregon
Chapter of Physicians for Social Responsibility

2907	September 04, 2009	Allie Sheffield, Chair, Committee on Titan
Cement, Pender Watch and Conservancy

2908	September 04, 2009	Bishop Titus B. Deas, Jr., National Education
and Outreach Spokesperson, EmPower Consumers

2909	September 04, 2009	Susan Falzon, Friends of Hudson (FoH)

2910	September 04, 2009	Verie C. Martz, Vice President Engineering and
Environment, Salt Riber Materials Group (SRMG), on behalf of Phoenix
Cement Company

2911	September 04, 2009	Sierra Club et al.

2912	September 04, 2009	Adam W. Olivieri, Dr. PH, P.E., Program Manager,
Santa Clara Valley Urban Runoff Pollution Prevention Program (SCVURPPP)

2913	September 04, 2009	Natural Resources Defense Council (NRDC) et al.

2914	September 04, 2009	Michael R. Benoit, Executive Director, Cement
Kiln Recycling Coalition (CKRC)

2915	September 09, 2009	Lee B. Zeugin and Lauren E. Freeman, Hunton &
Williams LLP,

on behalf of 

Utility Air Regulatory Group (UARG)

2916	September 10,2009	Jim Griffin, Senior Director, Regulaotyr and
Technical Affairs, American Chemistry Council (ACC)

2917	September 10, 2009	Michael Stewart, President & CEO (Chief
Executive Officer), Texas Aggregates & Concrete Association (TACA)

2918	September 04, 2009	Lane H. Smith, Manager, Environmental Projects,
Giant Cement Holding, Inc. (GCHI)

2919	September 10, 2009

	Dick Pedersen, Director, Oregon Department of Environmental Quality
(ODEQ)

2920	September 01, 2009	Hurchel Young, Instrument & Electrical
Supervisor, Seattle Plant, Ash Grove Cement

2921	August 31, 2009	Aaron Mott, Private Citizen

2922	September 04, 2009	Andrew T. O'Hare, Vice President, Regulatory
Affairs, Portland Cement Association (PCA)

2923	August 24, 2009	Brenton Kidder, Private Citizen

2924	September 26, 2009	Brian Good, Private Citizen

2925	August 12, 2009	Brian G. Finch, Private Citizen

2926	August 30, 2009	Bryan Pintok, Private Citizen

2927	August 05, 2009	Gail Cheeseman, Private Citizen

2928	September 10, 2009	Jim Schermbeck, Director, Downwinders At Risk

2929	September 04, 2009	Stephen Minick, Vice President, Governmental
Affairs, Texas Association of Business

2930	August 26, 2009	Chris Barrett, Private Citizen

2931	August 26, 2009	Christ Hasser, Private Citizen

2932	August 25, 2009	Chris Hughes, Private Citizen

2933	September 01, 2009	Anonymous Public Comment

2934	August 24, 2009	Corey Younger, Private Citizen

2935	August 14, 2009	Alan Finch, Private Citizen

2936	August 25, 2009	Courtney Washburn, Private Citizen

2937	September 04, 2009	Neil J. Carman, Ph.D., Director, Clean Air
Porgram, Lone Star Chapter of Sierra Club et al.

2938	August 27, 2009	Anonymous Public Comment

2939	August 28, 2009	Don and Dawn O'Grady, Private Citizen

2940	August 26, 2009	Denise Kemp, Private Citizen

2941	August 27, 2009	Donald R. and Ann Clark, Private Citizens

2942	August 25, 2009	Daniel Peters, Private Citizen

2943	August 17, 2009	Lisa De Golier, Private Citizen

2944	August 25, 2009	Logan A. Loening, Private Citizen

2945	August 25, 2009	Lynelle Lewis-Raymond, Private Citizen

2946	August 25, 2009	Margaret Schiff, Private Citizen

2947	August 12, 2009	Dan Shaffer, Dykman Electrical, Inc.

2948	August 31, 2009	Vicky Ho, Private Citizen

2949	August 24, 2009	William and Wanda Jennings, Private Citizen

2950	August 28, 2009	Ken Ware, Private Citizen

2951	August 27, 2009	Leland H. Page II, P.E., Private Citizen

2952	September 01, 2009	Dustin Scott, General Manager, Umatilla
Ready-Mix, Inc.

2953	September 01, 2009	Kevin Bell, Clarke & Clarke Insurance

2954	September 03, 2009	Tim and Jan Kerns, Kerns Rainbow Ranch, Inc.

2955	September 04, 2009	Ananda Lee Tan, U.S. Coordinator, Global
Alliance for Incinerator Alternatives (GAIA)

2956	August 26, 2009	Maxine J. Murray, Private Citizen

2957	August 24, 2009	Paul J. Lyons, Private Citizen

2958	August 26, 2009	Rick L. Wallace, Private Citizen

2959	August 31, 2009	Elmer M. and Sheila M. Hill, Private Citizens

2960	August 26, 2009	Sherman W. Murray, Private Citizen

2961	September 02, 2009	Bob Groppo, Private Citizen

2962	September 02, 2009	Boris Degtyar, Private Citizen

2963	September 03, 2009	Bradley Walker, Private Citizen

2964	September 03, 2009	Kathleen Smith, Private Citizen

2965	September 03, 2009	A. Burr Tyler, Private Citizen

2966	September 04, 2009	Sandra R. Dickey, Private Citizen

2967	September 03, 2009	Cathy Trimble, Private Citizen

2968	September 02, 2009	David H. Jones, Private Citizen

2969	September 04, 2009	Mike Schaufler, State Representative, District
48, Oregon House of Representatives

2970	September 03, 2009	Chad Williams, Private Citizen

2971	September 03, 2009	Steve Bochenek, Private Citizen

2972	September 04, 2009	Denny Larson, Exec. Director, Global Community
Monitor

2973	August 31, 2009	James Wells, District Sales Manager, Ash Grove
Cement Company

2974	September 03, 2009	Kathryn Ballard, Private Citizen

2975	September 04, 2009	Diana Matley, Private Citizen

2976	September 03, 2009	Donnas Smith, Private Citizen

2977	September 01, 2009	D.T. Iddins, MD, Private Citizen

2978	September 01, 2009	Elwood Wirth and Family, Private Citizens

2979	September 02, 2009	Wade Christensen, Private Citizen

2980	September 03, 2009	Heather Ward and Family

2981	September 04, 2009	Linda K. Hughes, Private Citizen

2982	August 29, 2009	John Martin, Laborer/Physical Tester, Ash Grove
Cement Plant

2983	September 03, 2009	Anonymous Public Comment

2984	August 28, 2009 	Justin Hayes, Program Director, Idaho
Conservation League

2985	September 03, 2009	Janet Hedley, Private Citizen

2986	September 02, 2009	Jim Wheeler, Private Citizen

2987	September 01, 2009	Joan Stempniak, Private Citizen

2988	September 03, 2009	Kyle Hancock, Private Citizen

2989	September 03, 2009	Laina Raveendran Greene, Private Citizen

2990	September 02, 2009	Lareen Jacobs, Private Citizen

2991	September 04, 2009	Leslie Fowler, Private Citizen

2992	September 04, 2009	Lethika Richter-Addo, Private Citizen

2993	September 04, 2009	Christopher Lish, Private Citizen

2994	September 03, 2009	Gregor Becker, Private Citizen

2995	September 02, 2009	Gregory Becker, Private Citizen

2996	August 24, 2009	Ryan Deckert, President, The Oregon Business
Association (OBA)

2997	September 03, 2009	T.C. Johnston, Sr., Private Citizen

2998	August 14, 2009 	Gary D. McCall, McCall Industrial Supply

2999	August 20, 2009	Walter L. Johnson, Private Citizen

3000	August 13, 2009	Warren Weatherspoon, Private Citizen

3001	August 25, 2009	Yvonne Shevnin, COO, Enfinittee, LLC

3002	September 04, 2009	Dr. J. O'Donnell, Private Citizen

3003	September 01, 2009	Julia Olson, Private Citizen

3004	September 03, 2009	Julie Gutierrez, Private Citizen

3005	September 04, 2009	Katherine D. Allen, Private Citizen

3006	September 01, 2009	Charlotte (no surname provided), Private Citizen

3007	September 03, 2009	Louise Pillai, Private Citizen

3008	September 03, 2009	Louise Pillai, Private Citizen

3009	August 24, 2009 	Ryan Deckert, President, The Oregon Business
Association (OBA)

3010	September 02, 2009	Lynn Metrulas, Private Citizen

3011	September 03, 2009	M. and T. Manickam, Private Citizen

3012	August 21, 2009 	Bill Hubbell, General Manager, Wilco-Winfield LLC

3013	August 28, 2009	Evelyn M. Christensen, Private Citizen

3014	August 27, 2009	Anonymous Public Comment

3015	September 04, 2009	Margie Gong, , Private Citizen

3016	September 03, 2009	Diane Soffe, Private Citizen

3017	September 04, 2009	Rhoda Fry, Private Citizen

3018	September 01, 2009	Rhonda Givson, Retired Teacher, Private Citizen

3019	September 03, 2009	Mark Soffe, Private Citizen

3020	September 03, 2009	Marta Tingdale, RRT, TN, Private Citizen

3021	September 03, 2009	Mary Ann Mayer Redmond, Private Citizen

3022	September 04, 2009	Mary Harrison, Private Citizen

3023	September 04, 2009	Mary Peveto, Private Citizen

3024	August 26, 2009	Thomas Mumley, Assistant Executive Officer, San
Francisco Bay Water Board

3025	September 03, 2009	Richard Lines, Private Citizen

3026	September 04, 2009	M. G. Bloom, Private Citizen

3027	September 03, 2009	Roxanne Christal, Private Citizen

3028	September 03, 2009	Sam Pratt, Private Citizen

3029	September 02, 2009	Samantha Bornhorst, Private Citizen

3030	September 01, 2009	Sandra Breakfield, Private Citizen

3031	September 02, 2009	Martin Schiltz, Private Citizen

3032	September 03, 2009	Maureen Porter, Private Citizen

3033	September 03, 2009	Michele Westlaken, Private Citizen

3034	September 04, 2009	Michelle Craig, Private Citizen

3035	September 03, 2009	Nancy Johnstone, Private Citizen

3036	September 01, 2009	Sarah Castor, Private Citizen

3037	September 03, 2009	Sarah Laurie, Private Citizen

3038	September 02, 2009	Seth Cohen, Private Citizen

3039	September 04, 2009	Shannon Huges, Private Citizen

3040	September 04, 2009	Amy Parenti,, Private Citizen

3041	September 04, 2009	Jennifer Swearingen, Private Citizen

3042	September 03, 2009	Peeyush Jain, Private Citizen

3043	September 03, 2009	Philip Johnson, Private Citizen

3044	September 03, 2009	Allison Sunshine, Private Citizen

3045	September 04, 2009	Sheila Jones, Private Citizen

3046	September 03, 2009	Sue Blair, Private Citizen

3047	September 03, 2009	Susan Waskey, Private Citizen

3048	September 02, 2009	Susanne Endisch, Private Citizen

3049	September 03, 2009	Suzanne Baker, Private Citizen

3050	September 03, 2009	Terri Rimmer, Private Citizen

3051	September 04, 2009	Jeanne Schumacher, Private Citizen

3052	September 04, 2009	Arran E. Thomson, Private Citizen

3053	September 03, 2009	Vicki Oppenheim, Private Citizen

3054	September 02, 2009	Vinita Soin, Private Citizen

3055	September 04, 2009	Gerald Yeoumans, Private Citizen

3056	August 31, 2009	Gail Bower, Private Citizen

3057	August 24, 2009	George and Frances Alderson, Private Citizen

3058	August 27, 2009	Ginger Younger, Private Citizen

3059	August 25, 2009	Ignatius and Josephine Ding, Private Citizen

3060	August 26, 2009	James Robenolt and Mary Minow, Private Citizens

3061	August 20, 2009	Jenny Pompilio, M.D., MPH, Private Citizen

3062	August 24, 2009	Jerry Randolph, Private Citizen

3063	August 7, 2009	Jessica Shlager, Private Citizen

3064	August 28, 2009	James E. Finch, Private Citizen

3065	August 21, 2009	John Condon, Private Citizen

3066	August 19, 2009	Rodd and Rhea Bunch, Private Citizen

3067	September 03, 2009	Robert R. Geddis, QEP, Private Citizen

3068	September 01, 2009	Glenn Foote, Eastern Oregon Service Company
(EOSC)

3069	September 03, 2009	J. Jared Snyder, Assistant Commissioner, Office
of Air Resources, Climate Change and Energy, New York State Department
of Environmental Conservation

3070	September 02, 2009	Dawn Dawson, Secretary, Dawson Trucking, Inc.

3071	September 03, 2009	Vivian Zikmund, Private Citizen

3072	August 27, 2009 	Jerri Wickert, Private Citizen

3073	August 05, 2009	J. Amberg, Private Citizen

3074	August 07, 2009	Jane S. Philipson, Private Citizen

3075	August 28, 2009	Kari Van Artsdalen, Private Citizen

3076	August 25, 2009	Kay Hummel, Private Citizen

3077	August 08, 2009	Kim Jelfs, Private Citizen

3078	September 01, 2009	Kathryn C. White, Private Citizen

3079	August 31, 2009	Larry D. and Maxine J. Cole, Private Citizens

3080	September 09, 2009	Congresswoman Eddie Bernice Johnson, Member of
Congress 

3081	September 01, 2009	Patrick Flanagan, Private Citizen

3082	September 15, 2009	Anonymous Public Comment

3083	September 03, 2009	Bill Zikmund, Private Citizen

3084	August 29, 2009 	Anonymous Public Comment

3085	August 24, 2009	Dick and Laura Parsons, Private Citizen

3086	August 25, 2009	Linda Olson, Private Citizen

3087	September 04, 2009	Anonymous Mass Comment Campaign sponsoring
organization unknown (23)

3088	August 24, 2009	Anonymous Mass Comment Campaign sponsoring
organization unknown (129)

3089	August 25, 2009	Anonymous Mass Comment Campaign sponsoring
organization unknown (46)

3090	September 03, 2009	B. Sap, Private Citizen

3091	September 03, 2009	Gary Gable, Private Citizen

3092	September 03, 2009	William B. McNabb, President, McNabb Trucking

3093	September 03, 2009	Anita McKinney, Private Citizen

3094	September 03, 2009	Mr. and Mrs. W. Holoboff, Private Citizen

3095	September 03, 2009	Dale Smith, Private Citizen

3096	September 03, 2009	D. and B. Shumway, Private Citizen

3097	July 22, 2009	Maxine J. Murray, Private Citizen

3098	August 29, 2009	Anonymous Public Comment

3099	August 13, 2009	Michael Anderson, Private Citizen

3100	August 28, 2009	Michael Van Artsdalen, Private Citizen

3101	August 23, 2009	Michael Condon, Private Citizen

3102	August 12, 2009	Mike G. Eiselein, Private Citizen

3103	September 01, 2009	Nancy Sanchez, Private Citizen

3104	September 01, 2009	Nicholas O'Neil, Private Citizen

3105	August 20, 2009	Nick Littlejohn, Private Citizen

3106	August 25, 2009	Anonymous Public Comment

3107	August 8, 2009	Phil Jelfs, Private Citizen

3108	August 25, 2009	Breanna Teintze, Private Citizen

3109	August 21, 2009	Barbara G. Sidway, Private Citizen

3110	August 12, 2009	R. Kaminsky, Private Citizen

3111	September 01, 2009	Reece M. Hrizuk, Private Citizen

3112	August 21, 2009	Anonymous Public Comment

3113	August 24, 2009	Richard Kuhnet, Private Citizen

3114	August 25, 2009	Anonymous Public Comment

3115	August 21, 2009	Rod F. Cook, Private Citizen

3116	August 20, 2009	Roger L. Weber, Private Citizen

3117	August 25, 2009	Rosemary Marks, Private Citizen

3118	August 24, 2009	Roy Heberger, Private Citizen

3119	September 03, 2009	J. McKinney, Private Citizen

3120	September 03, 2009	Ron and Janice Parsons, Private Citizen

3121	September 03, 2009	Janine R. Little, Private Citizen

3122	September 03, 2009	Wilma J. Rogers, Private Citizen

3123	September 03, 2009	State Representative Ron Maurer, Ed., D., House
Republican Whip, District 3, Oregon House of Representatives

3124	September 03, 2009	William C. Parrish, Jr., Chief Executive Officer
(CEO), Santa Clara County Medical Association (SCCMA)

3125	August 21, 2009 	Ken Smyth, Private Citizen

3126	August 28, 2009	Kevin Harris, Private Citizen

3127	July 29, 2009	Kie and Cindy Hale, Private Citizen

3128	August 10, 2009	Amy L. Lansky, Private Citizen

3129	August 26, 2009	Sandra R. Dickey, Private Citizen

3130	August 5, 2009	Barbara Fukumoto, Private Citizen

3131	August 26, 2009	Cathy Helgerson, Private Citizen

3132	August 10, 2009	Teri Wiss, Private Citizen

3133	August 28, 2009	Tom Beverly, Private Citizen

3134	August 10, 2009	Marianne F. Tranter, Private Citizen

3135	August 11, 2009	Richard R. Tranter, Private Citizen

3136	August 31, 2009	Fred D. Blackham, Private Citizen

3137	September 14, 2009	Alvaro Linero, P.E., Program Administrator,
State of Florida Department of Environmental Protection

3138	September 04, 2009	Anne Hedges, Program Director, Montana
Environmental Information Center (MEIC)

3139	September 03, 2009	Cecil Mark Benintendi, Private Citizen

3140	September 03, 2009	Diane Schrack, Private Citizen

3141	September 01, 2009	Baker County Board of Commissioners, Baker City,
Baker County, Oregon

3142	September 08, 2009	Randall G. Gross, Jr., Director of Environmental
and Regulatory Policy, Michigan Manufacturers Association (MMA)

3143	September 08, 2009	Jim Folsom, Jr., Lieutenant Governor, Alabama
State House, State of Alabama

3144	September 08, 2009	Mark Murphy, VP, CPM Development Corporation

3145	September 09, 2009	Joan Howe, Private Citizen

3146	September 11, 2009	Donald. E. Morgan, Private Citizen

3147	September 11, 2009	Steve Robinson, President, Spokane Rock
Products, Inc.

3148	September 10, 2009	Candy Howland, Owner, Candy's Corner Market

3149	September 14, 2009	P.J. Lier, Private Citizen

3150	September 15, 2009	B. Vanderwall, Private Citizen

3151d	September 15, 2009	Andrew T. O'Hare, Vice President, Regulatory
Affairs, Portland Cement Association (PCA)

3152	August 24, 2009	John Woolley, President, Olympic Forest Coalition,
(OFCO)

3153	September 15, 2009	Christopher I. West, Private Citizen

3154d	September 16, 2009	R. Latane Montague, Partner, Hogan and Hartson
on behalf of Ted Kniesche, Vice President, Fulcrum Bioenergy, Inc.

3155	September 15, 2009	Representative Greg Walden, Second District,
Oregon, U.S. House of Representatives

3156	September 03, 2009	Sharla Vaughan, Private Citizen

3157	July 28, 2009	Lori Finch, Private Citizen

3158	August 28, 2009	Heather Finch, Private Citizen

3159	August 14, 2009	Herbert L. Keyser, Private Citizen

3160	August 07, 2009	Allison Marks, Private Citizen

3161	September 04, 2009	Hazel Lambert, Private Citizen

3162	August 28, 2009	Mohsen Nazemi, P.E., Deputy Executive Officer,
Engineering and Compliance, South Coast Air Quality Management District
(SCAQMD)

3163	September 04, 2009	K. and J. Dyke, Private Citizen

3164	September 03, 2009	Anonymous public comment

3165	September 04, 2009	Alva R. Mascall, S. C. Mascall, and T. Orr

3166	September 03, 2009	Ed Hardt, President, Eastern Oregon Mining
Association (EOMA)

3167	September 02, 2009	Elyse Kunz, Board of Directors, Community
Advocates for Safe Emissions (CASE)

3168	September 01, 2009	R. L. Moore, Private Citizen

3169	September 15, 2009	Lewis F. Gossett, President and Chief Executive
Officer (CEO), South Carolina Manufacturers Alliance (SCMA)

3170	September 15, 2009	Jon Baggett, Director, Dorchester County
Economic Development, Dorchester County, Maryland

3171	September 03, 2009	Comment submitted by State Senator Jason Allen,
Majority Whip, 37th District, Michigan State Senate

3172	September 15, 2009	State Senator Kevin Engler, District 3, Missouri
State Senate

3173	September 03, 2009	Carl Stiff, M.D., Baker County Commissioner,
Baker County, Oregon

3174	September 08, 2009	D. (surname illegible), Private Citizen

3175	September 08, 2009	Tim Pope, Private Citizen

3176	 September 08, 2009	Wade L. Boxberger, Private Citizen

3177	September 08, 2009	Richard D'Ewart, Private Citizen

3178	September 08, 2009	John R. McLean, Private Citizen

3179	September 08, 2009	Representative Bart Stupak, (MI-1) and
Representative Charles W. Dent, (PA-15) Co-Chairs, Congressional Cement
Caucus, Congress of the United States

3180	August 31, 2009	Kenneth W. Carlson, President, C and A Equipment
Company

3181	September 08, 2009	Chris Combs, Private Citizen

3182	September 08, 2009	J.T. "JABO" Waggoner, Senate Minority Leader,
District 16, Alabama State Senate

3183	September 08, 2009	Douglas L. Whitley, President and Chief
Executive Officer (CEO), Illinois Chamber of Commerce

3184	September 08, 2009	Seth Hammett, Speaker, House of Representatives

3185	September 08, 2009	Representative Mike Hill, House of
Representatives, Alabama State House

3186	September 08, 2009	Craig A. Dally, State Representative, 138th
Legislative District, House of Representatives, Commonwealth of
Pennsylvania

3187	September 08, 2009	Julie Harhart, State Representative, 183rd
Legislative District, House of Representatives, Commonwealth of
Pennsylvania

3188	September 08, 2009	Jari Askins, Lieutenant Governor, State of
Oklahoma

3189	September 08, 2009	Ray Salva, State Representative, District 51,
Missouri House of Representatives

3190	September 08, 2009	Cam Ward, House of Representatives, Alabama
State House

3191	September 08, 2009	Joseph F. Brennan, State Representative, 133rd
Legislative District, House of Representatives, Commonwealth of
Pennsylvania

3192	September 08, 2009	Hank Erwin, State Senator 14th District, Alabama
State Senate, Alabama State House

3193	September 08, 2009	Congressman John Shimkus, 19th District of
Illinois, Congress of the United States, House of Representatives

3194	August 06, 2009 	Eugene Stackle, Manager, Baker Economic
Development

3195	September 04, 2009	Howard L. Giberg, Counsel, Guida, Slavich and
Flores on behalf of the Tile Council of North America, Inc. (TCNA)

3196	September 08, 2009	George H. Winner, Jr., Member of Senate, The
Senate State of New York

3197	September 08, 2009	Tony Stamas, State Senator, 36th District, The
Senate State of Michigan

3198	September 04, 2009	Neil D. Breslin, Member of Senate, New York
State Senate

3199	September 08, 2009	Representative Jimmy Martin, House of
Representatives, Alabama State House

3200	September 08, 2009	Patrick M. Browne, Senator 16th District,
Commonwealth of Pennsylvania, Senate of Pennsylvania

3201	September 08, 2009	Doug Roberts, Jr., Director of Environmental and
Energy Policy, Michigan Chamber of Commerce

3202	September 04, 2009	George W. Roy, Mayor, City of Calera

3203	September 04, 2009	Andy Neumann, State Representative for Alcona,
Alpena, Crawford, Montmorency, Oscoda and Presque Isle Counties,
Michigan House of Representatives

3204	September 02, 2009	Ed Hardt, President, Eastern Oregon Mining
Association (EOMA)

3205	September 04, 2009	Roger Smitherman, President Pro Tempore, Alabama
State Senate, Alabama State House

3206	September 04, 2009	Mayor Stanley J. Salva, City of Sugar Creek,
Missouri

3207	September 04, 2009	Glen Zorn, Assistant Commissioner, State of
Alabama Department of Agriculture and Industries

3208	August 24, 2009	Wyatt Pegg, President , Western Interlock, Inc.

3209	August 31, 2009	John C. Colbert, Vice President and General
Manager, Automatic Wiber Vault Co., Inc.

3210	September 08, 2009	M. D'Ewart Logsdon, Private Citizen

3211	September 04, 2009	Mark Mullins, Private Citizen

3212	September 08, 2009	Norma Job, Private Citizen

3213	September 08, 2009	Elmer M. and Sheila M. Hill, Private Citizens

3214	September 08, 2009	Robert M. Hardy, Private Citizen

3215	September 08, 2009	Charles D. Carey, Private Citizen

3216	September 08, 2009	Robert Harris, Private Citizen

3217	September 08, 2009	Claude and Jane Treanor, Private Citizens

3218	September 08, 2009	Michael J. Hrizuk, Private Citizen

3219	September 04, 2009	Clifford T. Smith, Private Citizen

3220	September 04, 2009	Edward P. Stocks, Private Citizen

3221	September 04, 2009	Jeff Wiest, Private Citizen

3222	September 08, 2009	Tony DeFalco, Private Citizen

3223	September 05, 2009	Julia Chen, Private Citizen

3224	September 08, 2009	Dennis L. West, Private Citizen

3225	September 09, 2009	Phillip Teintze, Private Citizen

3226	September 13, 2009	Kerrie Utsumi, Private Citizen

3227	September 20, 2009	Panos Kougiouris, Private Citizen

3228	October 05, 2009	Sally and Wick Drummon, Private Citizens

3229	October 05, 2006	Craig Clapsaddle, Mechanical Systems, Inc.

Notes:

Request for extension of public comment period.

Duplicate copies of same comments entered in docket.

Comment is incomplete.

d.	These items were submitted to the wrong docket.

2.0	Proposed Amendments to the Current NESHAP from Portland Cement
Manufacturing Industry (05/06/09)

2.2	General

2.2.1	Support

Comment: More than 700 commenters supported the proposed amendments to
the NESHAP for portland cement manufacturing. The commenters supported
the proposal because it would reduce emissions of hazardous air
pollutants (HAP), especially mercury, and reduce risks to health and
protect the environment. Many of the commenters that supported the
proposed amendments also stated that they did not support the use of
subcategorization especially in the case of the mercury emission
standards.

Response: EPA acknowledges the commenters’ support. 

2.2.2	Oppose

Comment: More than 200 commenters stated that they opposed all or parts
of the proposed amendments to the NESHAP for portland cement
manufacturing. For example, many of the commenters opposed the proposed
standards for mercury because naturally high mercury concentrations in
limestone would mean that a Portland cement plant in their community
would be unable to comply with the emission limit for mercury even after
installing control devices. These commenters pointed out that the plant
had reached an agreement with the State to reduce mercury emissions by
75 percent by installing an activated carbon injection system and to use
other means to reduce mercury emissions by 85 percent by the year 2012.
These commenters stated that closure of the plant would have negative
impacts on the community as a result of job losses and lost income for
individuals and municipalities and would result in the importation of
cement from countries where there are fewer environmental controls. Many
of these commenters requested that, in order to avoid these negative
impacts, EPA create a subcategory for cement plants that must use
limestone that contains naturally high mercury concentrations and that
EPA require that plants in this subcategory apply air pollution control
technologies that represent maximum achievable control technology
(MACT). 

Many commenters also opposed the pollutant-by-pollutant approach EPA
used to set MACT floor standards due to the likely negative impacts on
the cement industry. 

Response: As explained in the preamble to the final rule, EPA has
choosing not to create a subcategory for the two plants with
exceptionally high mercury emissions due to limestone mercury levels.
EPA also explains in the preamble why setting floors HAP-by-HAP is both
permissible and better promotes statutory objectives than setting floors
based on performance of entire facilities.

2.3	Overall Approach to Setting MACT Floor Limits

2.3.1	Variability and Adequacy of Data

Many comments were received on various aspects of the data and the
adequacy of the data used by EPA in setting MACT floor limits and on the
statistical methods EPA used to estimate variability in the data for
purposes of setting MACT floor limits.

Adequacy of data

Comment: On the issue of the adequacy of the data used to develop the
MACT floor limits, many commenters (2832, 2841, 2844, 2845, 2859, 2863,
2864, 2874, 2889, 2900, 2901, 2914, and 2918) stated that EPA should
have collected data from a larger portion of the kiln population.
Commenters 2844 and 2845 offered assistance in improving the data
supporting the standards and the subcategorization. Commenter 2889
suggested that EPA should either initiate requests for additional
industry-specific data or determine if relevant control technology and
emissions data are available in other industrial categories. Commenter
2844 stated that for EPA to devise the variability methodologies
consistent with Brick MACT and other relevant case and statutory law,
EPA can exercise its discretion to seek appropriate data from regulated
entities and other stakeholders and make discretionary methodological
determinations of how best to estimate the necessary information, so
long as EPA’s approach is fact-based and reasonable. 

On the question of the adequacy of the data used to set MACT floors,
commenter 2832 stated that the use of as few as two data points for the
creation of the MACT standard for existing sources does not reflect a
statistically significant number and should not be used for the
establishment of a standard.

Many commenters (2841, 2844, 2846, 2858, 2863, 2874, 2915, and 2929)
stated that EPA’s limits were based on limited testing and/or
short-term operations. Short-term data may not capture variability that
occurs in a facility, and since the industry is expected to be in
compliance all of the time, not simply most of the time, it is critical
to account for normal variability. 

Commenter 2844 argued that short-term testing is insufficient to capture
emissions variability associated with nontechnological operational
factors.

Response: As described in more detail in the preamble, EPA has made
every effort to generate data for this rule. For mercury, EPA compelled
plants in the industry to gather 30 days of sampling data by means of
section 114 requests. EPA has repeatedly requested, indeed implored, the
industry to provide longer-term sampling data to support their
assertions that mercury input variability is greater, but industry
largely chose not to do so. Cf. Association of Pacific Fisheries v. EPA,
617 F. 2d 794, 817 (9th Cir. 1980) (burden on industry to provide data
in its possession). EPA also adjusted its data base where commenters
provided reliable information about mercury content of inputs not
reflected in the EPA-compelled sampling effort. See in particular
comment 2845 at pp. 32-33. For THC, EPA has compiled data from best
performing sources comprising years of operating data, measured
continuously. (Industry participated constructively in supplementing the
database used at proposal, although delays in providing certain
supplemental information and significant methodological issues relating
to that supplemented information led to EPA’s decision not to use
those additional data in developing the THC standard. In any case, EPA
views the late-submitted information as corroborating the promulgated
standard, as explained in the preamble to the final rule.) For HCl, EPA
again used its section 114 authority to gather further data on emission
levels. For PM, EPA also supplemented its data base at proposal
(including adding data supplied by LaFarge North America Inc. which
better characterized the performance of two of its kilns). EPA further
applied well-recognized statistical methodologies to the mercury and PM
data to further project sources’ variability. EPA believes that these
efforts result in floor determinations that reasonably estimate
performance over time of best performing sources. 

Furthermore, in response to commenter 2844 and others, in no instance is
EPA relying solely on stack testing to assess sources’ performance.
The mercury, PM and HCl standards utilize the 99th percentile UPL
formula to further assess performance and variability. The mercury data
comes from total facility input testing over 30-days, not from
individual stack tests. EPA further adjusted these data to account for
best performers’ changes in mercury inputs outside the 30-day sampling
period (where properly documented), and further adjusted the estimated
intra-quarry mercury variability for the pool of best performers based
on the reasonably-estimated long-term intra-quarry variability of the
two best performing sources. This yielded a floor standard of 55 lb/MMT
clinker. At least one industry commenter (commenter 2830, Essroc Cement)
stated that proper use of statistics to estimate variability for mercury
should result in a standard between 43 and 50 lb/MMT clinker (slightly
lower than the standard EPA adopted, am indication to EPA that it is
adequately accounting for variability of inputs and ultimate
performance).  The THC standard relies on hundreds of observations per
facility, obtained by continuous emission monitoring, not by individual
stack tests. All of the standards use a pooled variability developed
from the pool of best performing sources. This comment is thus factually
mistaken.

Comment: Commenters 2844 and 2846 stated that Brick MACT held that EPA
must base floors on achieved emissions control rather than control
technology choices and this may require additional data from
stakeholders. The commenter notes that floors now must address the
impact of testing variability, technological variability and
nontechnological factors if they impact emission levels achieved in
practice. Non-intentional factors must also be taken into consideration,
such as a technology that dramatically lowers its emissions of a
particular HAP but at the same time increases its emissions of other
HAPs or other air pollutants. EPA must take those factors into account
when setting the floor and must devise a reasonable way to address such
factors in its methodology. 

Response: EPA has accounted for testing variability, technological
variability (i.e. variability in control device performance), and
variable HAP content of raw materials and fuels utilized by best
performing sources in developing each of the standards. However, as
explained in the preamble to the final rule, there is no instance where
technology which controls one HAP increases emissions of another. Put
another way, means of control are available to the industry which are
mutually compatible, not mutually exclusive. 

Comment: One commenter (2844) stated that EPA should use the emissions
data for the large number of kilns in the docket to get the most
accurate perspective on emissions and control needed.

Response: See previous response

EPA’s use of statistical methodologies in assessing variability

Comment: On the EPA’s consideration of variability in setting the MACT
floor limits, one commenter (2845) provided a detailed discussion of the
statistical methods used by EPA to derive the 99th percentile values
which EPA developed the proposed MACT floors for mercury, HCl, THC, and
particulate matter (PM). The commenter stated that the EPA should use
the upper tolerance limit (UTL) method to account for variability where
data is limited and use actual data in large data sets. Where there is
very limited data, which is the case for HCl, it is not possible to
perform any statistical analysis or extrapolate a 99th percentile;
therefore, it is not appropriate to use existing data to establish a
MACT floor for HCl using even the upper tolerance limit (UTL)
methodology. This commenter also provided an extensive and detailed
analysis of EPA’s analysis and data as well as an alternative analysis
and MACT floors using different statistical methods and updates to
EPA’s data sets.

Commenter 2832 supported EPA’s acknowledgement and recognition that
data variability is a statistical issue that is required to be taken
into account in setting a MACT floor. EPA should explain its selection
of the Upper Prediction Limit (UPL) statistical approach and how that
approach was applied to the data. 

Commenter 2864 agreed with EPA’s use of the 99th percentile for the
MACT floor sources, stating that its use accounts for emissions
variability. However, the commenter disagreed with the approach used to
estimate the upper 99th percentile. The UPL statistical method does not
actually generate the 99th percentile for the data sets in the MACT
floor in the Proposal. Instead, it generates a probability for a single
value that is possible in the future, which is not a 99th percentile.
The commenter stated that EPA should use an UTL to estimate the 99th
percentile for small data sets and an actual 99th percentile where large
data sets are available.

One commenter (2829) disagreed with EPA’s calculation of the MACT
standard emission limits. The commenter stated that to account for
variability, EPA should first establish standards that are the actual
average emission rates of the 12 percent of existing sources, and then
use longer averaging times (perhaps 60 or 90 day averages) to allow for
short term variability and spikes in emissions.

Response: EPA’s legal obligation in establishing MACT floors is to
reasonably estimate the average performance achieved by the best
performing 12 percent of best performing sources, or for new sources,
the performance of the best controlled similar source. Although
acknowledging that variability in best performing sources’ performance
exists, commenter 2829 nonetheless maintained that EPA is compelled to
select the arithmetic mean or median of the best performers and adjust
for variability by taking expressing those values in longer averaging
periods. EPA disagrees that it must establish floors based on the
numerical average (or median). Caselaw, much of it cited in comments
2844 and 2845, as well as EPA in the preamble to the proposed and final
rule, as well as at 70 FR at 59437, indicates that variability is to be
accounted for by EPA in establishing MACT floors. This is because
sources must comply at all times, so that their overall performance must
be assessed. See, e.g. Mossville v. EPA, 370 F. 3d at 1242 (EPA properly
used a highest data point to estimate best performing source’s
performance, since this was part of its documented variability). Caselaw
likewise indicates that EPA, in determining MACT floors, can make
reasonable estimates of sources’ performance, including estimates of
sources’ variability. Cement Kiln Recycling Coalition v. EPA, 255 F.
3d at 862; Sierra Club v. EPAu, 167 F. 3d at 661. 

The statute does not speak to the issue of permissible averaging times
in constructing standards, so it may be that in factually appropriate
cases, a choice of averaging time to account for best performing
sources’ variability could be legitimate. However, automatic (i.e., a
priori) determinations to adopt long-term averages for any type of data
set regardless of circumstances of sources’ performance and robustness
of the data set would go too far. In addition, if the commenter is
suggesting that EPA simply average short-term test results and then hope
to account for variability by assigning a longer averaging period, EPA
rejects such a methodology as both qualitative and arbitrary – it
could easily result in estimates of performance that are either overly
stringent or lax due to the rough estimate of variability underlying the
methodology. It is better to more accurately assess sources’ initial
performance (for example, as with the mercury standard in this rule
where EPA changed its baseline for best performing plants based on
information about materials these sources process even though they did
not do so in the 30-day sampling regime), and to use statistical
methodologies to account for the variability reflected in the resulting
data set. 

Other factors that contribute to variability

Comment: Many commenters stated that varying levels of HAP in inputs
contribute more to kiln emissions variability than variations in
at-the-stack control performance, and that EPA should take these factors
take into account when setting the MACT floor limits.

Commenters 2830, 2836, 2840, 2844, 2845, 2863, and 2929 stated that the
chemical composition of inputs into the cement manufacturing process is
highly variable and EPA’s consideration of this variability and its
effect on emissions will dictate whether the final regulations will
result in workable emission standards or standards that could adversely
impact the portland cement industry, for example, by forcing the
shutdown of domestic cement production capacity in the US, thereby
inhibiting economic development in the US. The commenters stated that
the limestone quarry is intrinsically linked to the cement manufacturing
facility and is the main reason for the cement plant’s location. A
typical plant will process more than a million tons and a half of
material annually. Limestone represents approximately 70-80 percent of
this volume. A cement plant will not be constructed unless there is a
sufficient quantity and quality of limestone to last 50-100 years. 

Commenters 2844, 2845, and 2915 stated that the trace chemical
composition of the fuels which come in direct contact with the raw feed
materials include coal, petroleum coke and alternative fuels that
include tires and industrial solvents. The variability of trace
compounds in the conventional and alternative fuels may vary widely. The
concentration of trace compounds present in the cement and other plant
emissions may be highly variable, reflecting changes in the chemistry of
conventional fuels and raw materials and their alternatives. EPA’s
consideration of this variability will dictate whether the final
regulations will result in workable emission standards or standards that
could adversely impact the portland cement industry by forcing the
shutdown of domestic cement production capacity in the United States,
inhibiting economic development in the United States.

Commenter 2874 stated that EPA should revise the MACT floor for mercury
by more comprehensively accounting for actual and reasonably foreseeable
variability in the operations and delay promulgation of the THC emission
standard until data has been gathered that will appropriately account
for variability.

Commenters 2844, 2845, 2846, 2915, and 2929 stated that EPA should
consider process variability and its effect on emissions. A single
cement kiln produces various products throughout the year and may use
different raw materials and fuels. Each product, raw material, and fuel
results in different emission profiles from a kiln.

Response: As discussed in the preambles to both proposed and final
rules, as well as other comment responses, EPA is aware that input (raw
material and fuel) variability is the predominant contributor to both
mercury and THC emissions, as well as their variability. EPA explains
there at length how it accounts for these varying levels in inputs. EPA
particularly emphasizes the pooled intra-quarry variability estimates
for mercury as well as adjustments in particular sources’ average
performance, plus use of the UPL 99th percentile statistic as means by
which short-term and long-term variability are accounted for in the
mercury standard. EPA has also carefully assessed the role of input
variability in constructing the HCl and PM standards, although for these
standards different factors (variability in analytic detection and
control device variability) played the predominant role in assessing
sources’ performance over time. 

Other factors that contribute to variability

Comment: One commenter (2844) stated that the terms in §112(d)(3)
permit EPA to address the unique factors that affect emissions control
achieved by sources. Such factors, even when not relevant to determining
the actual emissions control of a particular HAP achieved by a source,
may be relevant to determining whether a source that demonstrates low
achieved emission levels is to be considered a best performing source
for the purpose of setting a floor. The commenter stated that EPA may
define best performers as the lowest emitters as long as EPA’s
approach is reasonable in light of all relevant factual and data
circumstances and so long as EPA reasonably applies variability that
impacts the lowest reported emissions in order to accurately estimate
the level of control actually achieved by the best performing sources.

Commenter 2844 stated that a lawful floor setting methodology must
estimate relevant emissions variability resulting from nontechnological
factors. EPA has the discretion under §112 to devise methodologies for
estimating variability that impacts emissions control. Variability in
emissions results from test-related variability, technology-related
variability, and nontechnological-related variability. The commenter
provided discussion and examples. To be consistent with Brick MACT, EPA
must address all the factors that impact variability of emissions,
especially not ignoring nontechnological factors.

Response: EPA is not ignoring nontechnological factors in assessing
variability. Indeed, for mercury, there are essentially no other factors
to consider since the industry has not installed technological controls
to reduce mercury emissions (hence EPA’s mass balance approach based
on sampling of mercury levels in all kiln inputs as a means of assessing
emission levels and overall performance). EPA agrees that determining
which performers are best requires consideration of sources’
performance over time. See 72 FR at 54877 (Sept. 27, 2007). 

Comment: Commenters 2845 and 2846 stated that EPA should not rely solely
on short-term data from similar sources and ignore the variability and
long term impact of the proposed emission rates. In EPA’s proposed
rule to set a NESHAP for mercury at coal-fired electric utilities, the
agency identified a number of factors that contribute to variability in
emission test data. The commenter suggests that these factors should be
considered in this rulemaking, and for cement kilns, the influence of
raw materials should also be considered.

Response: In no instance is EPA relying exclusively on short-term data
in establishing floors and standards. EPA has carefully considered the
specific comments as to why mercury levels in inputs to the best
performing sources could vary beyond the levels in the data base and
made appropriate adjustments as a result. Use of the UPL statistic
further accounts for this variability (both from non-technological and
technological factors), as explained in the preamble to the final rule.

2.3.2	Pollutant-by-Pollutant MACT Floors

CAA intent

Comment: Many commenters (2830, 2832, 2836, 2841, 2844, 2845, 2858,
2859, 2863, 2864, 2874, 2890, 2908, 2910, 2914, 2915, 2916, and 2917)
stated that setting MACT floors on a pollutant-by-pollutant basis
violates the law and results in MACT floors that bear no relation to
emission limits that are being achieved at the best performing existing
sources. According to commenters, this method violates the plain
language and intent of §112(d) of the Clean Air Act (CAA) and its
effect is a MACT floor that reflects a standard that no one plant in
existence currently achieves. The CAA requires a two-step process to set
a MACT standard: first, EPA identifies a MACT floor for each pollutant
and source category based on emission limitations that have already been
achieved by the best-performing 12 percent of existing sources (Brick
MACT, 479 F.3d at 877; Northeast Maryland Waste Disposal Auth. v. EPA,
358 F.3d 936, 954, D.C. Cir. 2004, and Cement Kiln Recycling Coalition
v. EPA, 255 F.3d 855, 866, D.C. Cir. 2001). Then the CAA requires EPA to
determine whether stricter standards—beyond the floor limits—are
achievable, taking into account cost and feasibility factors.

Commenters 2832, 2841, 2844, 2845, 2846, 2910, 2914, 2915, and 2916
stated that §112(d)’s use of the terms best-performing and existing
clearly means that sources in a category or subcategory that are used to
set the MACT floor are to be real, not theoretical or hypothetical,
sources (42 U.S.C. §7412(d), 2006 and Northeast Maryland Waste Disposal
Authority, 358 F.3d at 954). The phrase achieved in practice can only
mean that Congress intended actual sources, performing under real-life
conditions, to be the benchmark for determining the MACT floors.
Furthermore, the language of the statute does not speak in terms of the
best-performing source or sources for each listed pollutant or group of
pollutants (42 U.S.C. §7412(d)). Rather, the focus is on the best
existing source or sources for all pollutants, and what these sources
truly can achieve on an overall basis. EPA’s pollutant-by-pollutant
methodology is also at odds with the legislative history underlying
§112(d) (S. Rep. No. 228, 101st Cong., 1st Sess. 169, 1989). 

According to the commenters, the focus on overall performance is not
surprising because in the 1990 CAA Amendments Congress abandoned
§112’s previous focus on individual pollutant standards, and adopted
the technology-based multi-pollutant approach to regulating toxics in
use under the Clean Water Act (CWA). See S. Rep. No. 228, 101st Cong.,
1st Sess. 133-34 (1989). Thus, if one source can achieve a firm degree
of control for one pollutant but not for another, there may be no
justification for including it in the set of sources from which the
floor is calculated (Tanners’ Council of America v. Train, 540 F.2d
1188, 1193 (4th Cir. 1976) deeming CWA effluent limitations guidelines
not achievable where plants in EPA’s database were capable of meeting
the limitations for some, but not all, of the pollutant parameters).

Several of these commenters provided a table (the result of an industry
survey) listing EPA floor units and the ability or inability of cement
plants to comply with the proposed existing source MACT limits for
mercury, THC, HCl and PM. 

Several commenters (2845, 2910) stated that EPA’s previous use of a
pollutant-by-pollutant analysis was based on authorities not applicable
to the CAA. EPA previously used a pollutant-by-pollutant methodology to
set MACT floors in the context of the proposed NESHAP for Hazardous
Waste Combustors, 69 FR 21198 (Apr. 20, 2004) [hereinafter Hazardous
Waste NESHAP.] Several parties submitted public comments questioning
EPA’s approach and pointing to the fact that EPA had failed to cite a
single existing source which met the various MACT floor standards. See
EPA’s Response to Comments on April 20, 2004 HWC MACT Proposed Rule,
Volume 1: MACT Issues, 126 (Oct. 2008). In response, EPA attempted to
defend its practice of establishing pollutant-by-pollutant MACT
standards by citing Chemical Mfr. Ass’n. v EPA, 870 F.2d 177, 239
(1989), clarified 885 F.2d 253, 264 (5th Cir. 1989), cert. denied, 495
U.S. 910, (1990), a case where the court held that, under the CWA, best
available technology (BAT) referred to the single best-performing plant
on a pollutant-by-pollutant basis.

According to commenters 2845 and 2910, EPA’s reliance on Chemical Mfr.
Ass’n is misplaced as the CAA’s procedure regarding the selection of
MACT technologies differs on a textual basis from the CWA’s procedure
for identifying best available technology. Under the CWA, BAT standards
are to be set based on the best practicable control technology currently
available. 33 U.S.C. §1311(b)(1)(A)(i)(2006). This has led to
pollutant-by-pollutant determinations. The CAA more narrowly limits the
basis for MACT designation to what has been achieved at existing
sources, not what could be hypothetically achievable on a per-pollutant
basis.

One commenter (2890) stated that EPA appears to be forgetting that the
floor is only the first step in the process. Once EPA has established a
floor based on physical sources, it is directed to go back and look at
options beyond the floor. Those beyond the floor options would include
the best control for each pollutant on every source. By correcting the
floor approach, the EPA would also correct the issue identified by Judge
Williams in his concurring opinion to the Brick vacatur, where a floor
that is designed to represent what has been achieved is more stringent
than what would be deemed achievable under a MACT.

Response: See preamble for response

Impacts of pollutant-by-pollutant approach

Comment: Several commenters (2831, 2844, 2845, and 2874) stated that in
evaluating the economic cost of achieving emission reductions, looking
at one plant’s emission control of only one pollutant to the exclusion
of all other emission controls produces a disjointed view of cost
implications and compliance feasibility. While an individual MACT floor
for one pollutant might not appear cost-prohibitive, when combined with
all of the other MACT floors for other pollutants, the total cost
implications could become especially onerous. While the CAA was authored
with the intent of reducing air pollution, Congress did not intend to
disrupt the productive capacity of the United States through the
promulgation of economically unachievable standards. 42 U.S.C.
§7401(b)(1)(2006). By setting MACT floors individually and ignoring the
collective cost implications of the entire NESHAP, EPA would effectively
disregard the CAA’s requirement that air pollution control be advanced
while promoting the nation’s productive capacity. Id.

Response: See preamble for response

Comment: Several commenters (2845, 2889, and 2890) stated that EPA’s
argument that EPA believes that because all our standards are not
technically interdependent (i.e., implementation of one emission control
technology does not prevent the source from implementing another control
technology), the fact that sources are not achieving all the standards
simultaneously does not indicate a flaw in the methodology is not valid.
Commenter 2845 cites the example of activated carbon injection to reduce
mercury from hazardous waste combustion, which has been observed to
increase the emissions of dioxin/furans, calling into question EPA’s
legal justification that control requirements for one pollutant do not
impact another. Commenter 2889 notes that the interaction between unit
characteristics and control capabilities must be taken into account.
Although there is no one kiln or no small number of kilns that
simultaneously define the MACT floors for mercury, total hydrocarbons,
hydrogen chloride, and particulate matter, all existing kilns will have
to achieve all four requirements despite that fact that no single
existing kiln achieves these requirements. 

Response: Although the commenter cited the example of activated carbon
increasing dioxin emissions they provided no data to substantiate this
claim. A thorough review of one of the references cited indicated there
was no mention of dioxins or activated carbon injection. The other
reference cited, which was not provided with comments, seems to imply
that dioxin emissions increase relative to dioxin inputs if activated
carbon is used. First, there is not sufficient information provided to
determine the accuracy of this claim. The reference authors cited in the
text did not match the information in the footnote, so we cannot
ascertain if the correct reference was cited. Second, the statement said
dioxins increased relative to dioxin inputs, and did not indicate if the
dioxin emissions from the stack increased. So this information does not
substantiate the claim that ACI increases dioxin emissions at the stack.
Even assuming that dioxin emissions at the stack did increase, there are
other factors that could influence dioxin emissions. For example, the
temperature at the inlet of the control device has been determined to be
a critical factor in formation of dioxins in these types of processes.
In fact, proper control of temperature is the preferred method used to
control dioxin emissions for this source category, and is part of the
emissions standard (see 63.1343 (b)(3)(ii) and (c)(3)(ii)). Given the
lack of information provided with this comment, and the errors noted, we
do not believe that commenters have provided information that actually
supports their statement. We also note that activated carbon injection
is currently used to reduce emissions of dioxins from municipal solid
waste incinerators. 

The commenters have likewise provided no information to demonstrate that
a facility cannot simultaneously control PM, HCl, THC, and mercury. For
example, several cement kilns simultaneously control NOx, SO2 (which
also reduced HCl), and PM using a fabric filter followed by a wet
scrubber. Two older wet kilns controlled THC using activated carbon
injection followed by a PM control. A new preheater precalciner kiln had
PM (fabric filter), SO2 (wet scrubber) and THC (regenerative thermal
oxidizer) controls in series. We also note that in the utility industry,
simultaneous control of NOx, SO2, and PM are accomplished using
selective catalytic reduction and wet scrubber controls in series.
Municipal waste combustors simultaneously control dioxins and mercury
using activated carbon injection, acid gases using spray dryers, and PM
using fabric filters or electrostatic precipitator. In the specific case
of cement kilns, we assume that a source needing to control all four
pollutants would use a primary PM control (mainly to capture valuable
producer or raw materials that are emitted from the kiln), activated
carbon injection followed by a polishing fabric filter (which would
capture additional PM, remove the mercury, and remove the THC (or
organic HAP if the alternative standard is used)) and a wet scrubber for
HCl control. We costed this configuration for many kilns in our cost
estimates for the rule. (See Regulatory Impact Analysis: Amendments to
the NESHAP and NSPS for the Portland Cement Manufacturing Industry,
Final Report, August 2010.) Depending on the uncontrolled levels of the
pollutants, it may be very possible for many facilities to combine
activated carbon injection and alkaline sorbent injection downstream of
the primary PM control followed by a polishing fabric filter and control
all four pollutants, PM HCl, THC, and mercury. Based on these facts, we
see no technical impediment to simultaneously meeting all four emission
standards in the NESHAP. 

Subcategory for each HAP

Comment: One commenter (2890) asked EPA to clarify the statement on page
21145 of the proposal preamble that EPA, may create a subcategory
applicable to a single HAP, rather than a source category, if the facts
warrant (so that, for example, a subcategory for kilns emitting mercury,
but a single category for kilns emitting HCl, is legally permissible
with a proper factual basis). The commenter took this to mean that EPA
is creating a subcategory for a single HAP.

Response: It is legally permissible to create HAP-specific
subcategories, for example, where a difference in size, type or class of
facility only affects emissions of one HAP but not others, although EPA
is not doing so in this rule. See fuller discussion where EPA discusses
whether to subcategorize on the basis of quarry limestone mercury
levels. If EPA had chosen to do so, the subcategory would have contained
two sources and would apply to emissions of mercury only. 

2.3.3	Lowest Emitters Versus Best Performers

Comment: One commenter (2834) stated that the Brick MACT ruling of the
D.C. Circuit Court reinforces earlier holdings in National Lime
Association vs. EPA. The court again held that floors are to be based on
the emission level actually achieved by the best performers (those with
lowest emission levels) not the emission level achievable by all
sources.

Response: See preamble for response

Comment:  Many commenters (2841, 2844, 2845, 2846, 2858, and 2914)
stated that EPA established its proposed floors equating best performing
sources with those that have the lowest emissions for particular HAPs
even though there are other ways to measure performance and, in some
cases other methodologies may comply with the statute where the "lowest
emitter" approach does not. Commenter 2845 noted that equating best
performer with lowest emitter contravenes a Congressional directive that
in developing MACT standards, EPA cannot require substitution of raw
materials in mineral processing industries, such as cement
manufacturing. The Joint Explanatory Statement of the Committee of
Conference for the 1990 CAA Amendments stated: For categories and
subcategories of sources of [HAPs] engaged in mining, extraction,
beneficiation, and processing of nonferrous ores, concentrates,
minerals, metals, and related in-process materials, the Administrator
shall not consider the substitution of, or other changes in, metal- or
mineral- bearing raw materials that are used as feedstocks or material
inputs . . . in setting emission standards, work practice standards,
operating standards or other prohibitions or requirements or limitations
under this section for such categories and subcategories. H.R. Rep. No.
101-952, at 339 (1990), quoted from Sierra Club v. EPA, 353 F.3d 976,
988 (D.C. Cir. 2004). According to the commenters, enormous amounts of
limestone are fed into a kiln to manufacture clinker, and it is
cost-prohibitive to import limestone from further away. If the plant’s
quarry contains limestone with high concentrations of mercury or high
concentrations of organics, the kilns will emit more mercury or THC and
potentially more organic HAPs. Because limestone with high mercury or
organic emissions will result in higher HAP emissions, and it is not
cost-effective to import limestone from far away, equating the lowest
emitters with the best performing sources makes no sense in the context
of cement facilities. It also would be squarely in opposition to the
Joint Explanatory Statement.

Response: See response in preamble

Comment: Commenter 2844 stated that EPA could interpret §112(d)(3) as
Brick MACT appears to do, as one unitary concept meaning sources with
the lowest emission levels, or EPA can interpret it as a more complex
concept that the EPA may determine the emission control (using any of
the various definitions in the CAA) that sources have achieved in
practice (as estimated by reasonably predictive variability factors) and
rank them according to their relative emissions levels (i.e., a
quantitative measure of achievement). Having done so, the Agency can
then evaluate each of the lowest emitters in terms of whether they meet
the Agency’s criteria for best controlled similar source. With regard
to best controlled, EPA may evaluate this from a purely quantitative
angle (lowest emissions) or from more qualitative aspects (reduction
efficiency, environmental and health (or cross-media) impacts,
cost-effectiveness of reductions achieved, impacts on other HAP or
pollutant emissions, and so on.

Commenter 2845 provided several examples of court MACT decisions
endorsing a technology approach to setting the standards, in which EPA
selected the best performing sources based on the relative performance
of air pollution control technology.

Commenter 2844 stated that EPA also has the discretion to define best
performers as sources other than those with the lowest achieved emission
levels. In the current proposal, the many difficulties associated with
evaluating the impact of HAP content in the raw material inputs to
mercury emission control and other factors could support a decision by
the agency to establish a standard based on efficiency (i.e., a percent
reduction standard) if not for the source category as a whole, then such
a standard might be established for a particular subcategory as
relevant, or as an alternative compliance strategy. EPA’s discretion
is sufficiently broad to encompass many reasonable decisions identifying
and estimating the emission control of best performing sources on bases
other than lowest emissions data, assuming the floor for the standard is
based on a reasonable methodology estimating the percent reduction
achieved in practice by the best performing sources under the reasonably
foreseeable worst operating conditions.

Commenter 2844 stated that before EPA can determine a floor, EPA must
define the following terms in regard to the selection of a best
performer for new sources: Emission control; Achieved in practice; Best
controlled; and Similar source.

To set the floor for existing sources, the commenter stated that EPA
should define the following terms: Average emission limitation;
Achieved; and Best performing.

Response: See response in preamble to final rule.

Comment: Commenters 2832, 2846, and 2890 stated that rather than
selecting sources with the lowest emissions for particular HAP as best
performing sources, EPA could use the relative performance of air
pollution control technology to select the best performing sources,
applying the best reasonable method for determining best-performers,
which does not necessarily have to equate to lowest emissions. 

Response: See response in preamble to final rule.

Comment: One commenter (2859) stated that EPA is making an unrealistic
conclusion that best performing is equivalent to lowest emitting when
applied to the Portland Cement NESHAP. In cement manufacturing, lowest
emitting may be strictly related to a cement plant located adjacent to a
source of limestone that has the lowest level of contaminants. The goal
of the NESHAP is to reduce emissions of HAPs to the lowest possible
level using the best technology. There are several examples in this
rulemaking where the lowest emitting sources are not the most modern
cement manufacturing technologies or utilizing the most modern pollution
control measures. Conversely, some of the highest emitters are the most
modern cement manufacturing technologies that employ the most modern
pollution control measures. 

Response: The statement (without supporting citation) that “the goal
of NESHAP is to reduce emissions of HAPs to the lowest possible level
using the best technology” again ignores the contribution of raw
material and fuel inputs. 

Comment: One commenter (2832) stated that §112 did not define best
performing major sources as those that emit the least HAP or HAPs.
Unless a source is the lowest emitting for all HAPs of concern to EPA
for a given industry, best performing cannot be defined as lowest
emitting.

Response: EPA believes that selecting as best performers those with the
lowest emissions of the HAP at issue is a reasonable and permissible
approach for this rule.

Comment: One commenter (2844) stated that EPA has not taken the
necessary steps to evaluate or validate whether its methodology in fact
accurately estimates emissions control achieved in real world
circumstances at sources.

Response: EPA disagrees strongly with this comment. Data for mercury and
THC is long-term. EPA has adjusted underlying data further where
commenters provided reasonable quantitated information.

Comment: One commenter (2889) stated that EPA should avoid using
outliers when determining the MACT floor. The sources should have
process and raw material characteristics that are reasonably
representative of the population of sources in the industrial category
being regulated. With respect to the cement industry, this evaluation
must now include an evaluation of the variability of contaminants such
as mercury and total hydrocarbons throughout the industry.

Response: This comment appears to misapprehend the MACT floor process.
In assessing which performers are best, EPA cannot look to
‘representative’ raw materials for the industry as a whole, any more
than EPA could examine ‘representative’ baghouse performance. EPA
must focus on the performance of the best performers. This means
assessing the inputs of these sources, not others (unless performance of
best performers can reasonably be interpreted from aspect of other
sources’ performance).

2.3.4	Top 5 versus 12 Percent

Comment: One commenter (2898) disagreed with an alternative reading of
the MACT floor requirements that allowed EPA to consider the emissions
of the 5 lowest-emitting existing sources instead of the emissions of
the lowest-emitting 12 percent of existing sources in setting the
standard. The CAA unambiguously states that emission standards
established pursuant to the MACT floor mechanism shall not be less
stringent than the average limitation of the best performing 12 percent
of existing sources, for which the EPA has emissions information, in a
category or subcategory with 30 or more sources. Basing the standard on
the 5 best performing sources would violate this mandate by letting the
standard be less stringent than the emissions performances of the
best-performing 12 percent of existing sources and is not allowed by the
language of the CAA.

Commenter 2816 stated that EPA should not adopt an alternative
interpretation of data required to set an existing floor. The commenter
stated that basing the floor on 5 sources rather than the best
performing 12 percent is contrary to the language of CAA §112. This
interpretation also favors industries whose members, for whatever
reason, are unwilling to submit test data, or are unwilling to
participate in emission stack tests for the purposes of data collection
by EPA. The result is that EPA has limited data to work with and is
subsequently criticized for establishing standards based on limited
data. It is incumbent upon industry to submit sufficient data so that
the range of facility types and emission levels are adequately
represented.

Commenter 2824 recommended that the floor limit be the best-performing
12 percent of existing sources with an energy efficiency requirement
such as the energy star award.

Response: EPA believes its interpretation to be compelled by the
statutory text. 

Comment: Several commenters (2845, 2846, 2874, and 2915) stated that EPA
is proposing to calculate MACT floors by averaging the top 12 percent of
sources for which CEMS data are available (even if that amounts to less
than 30 sources), rather than by considering the top 12 percent of
sources for which EPA has emissions information. As a result, EPA is
proposing to establish the MACT floor based on data from only 2 sources.
The commenters stated that CAA §112(d) obligates EPA to set the MACT
floor looking at no fewer than 5 sources, recognizing the value of
relying on the maximum amount of data available.

Commenter 2841 stated that the use of a minimum of five facilities
should be adopted in the establishment of THC standards as well as the
other standards in this proposed regulation. The establishment of
requirements based on a small amount of data would run counter to the
intent of the CAA in utilizing data that is truly representative of the
best-performing facilities throughout an entire industry.

Commenter 2841 stated that in previous MACT rulemakings, EPA used the
five best performing facilities if the number of facilities was less
than 30. Consistent with these prior rulemakings, the commenter stated
that this approach should be used for this proposed Portland Cement
NESHAP rule and that EPA needs additional data points in order to
appropriately set limits for the industry as a whole.

Response: See response in preamble

Comment: Two commenters (2890 and 2916) stated that EPA must base their
emission limits on 12 percent of the total number of sources, and not
exclude facilities’ data. The commenters recommended that EPA not make
emission estimates for only a small subset of sources to determine the
MACT floor. If sufficient information is available for a source, such as
its size, control level, and other source-specific parameters, an
estimate should be developed, as allowed by the Courts, and the source
included in the count of sources for which the Administrator has
emissions information. Limiting this estimation to sources at the same
site or same company is arbitrary and should not be done. Emission
estimates based on the average emission factor for a subcategory should
be allowed.

Response: EPA believes its interpretation to be compelled by the text of
the statute.

2.3.5	Area Sources in HCL Floor Development

CAA intent

Comment: One commenter (2832) opposed the inclusion of actual and/or
synthetic area sources, with or without control devices, for any purpose
in this major source NESHAP standard. The commenter stated that there is
nothing in §112 that allows EPA to deem an area source as a best
performing source for major sources and nothing that allows EPA to
include area sources as part of a MACT pool.

Commenter 2916 stated that it is unlawful for EPA to group major and
area sources for purposes of setting MACT. By blending area and major
sources together for standard setting purposes, EPA has created a hybrid
major/area source category that is not authorized under §112. 

Response: In the final rule, no data from natural or synthetic minor
area sources are being used to establish the standard for HCl.

Establish GACT standards 

Comment: One commenter (2887) stated that EPA should develop generally
available control technology (GACT) standards (or management practices)
for area sources under §112(d)(5) for the pollutants listed in
§112(c)(6). EPA could minimize the economic impact on all area sources
by subjecting them to GACT for the §112(c)(6) pollutants.
Alternatively, the EPA could create a subcategory for area sources that
are small entities and subject this smaller number of sources to GACT
(or management practices). This would satisfy the policy behind Small
Business Regulatory Enforcement Fairness Act (SBREFA) as implemented by
EPA policy (to minimize the economic impact on small entities while
still meeting the mandates of the CAA). 

Response: EPA explains in the preamble to the final rule its various
dispositions for applying standards to area cement kiln sources.

Comment: One commenter (2887) stated that for HAPs that are not listed
under §112(c)(6), EPA should establish GACT (or management practices)
for area sources under §112(d)(5), and not regulate area sources under
MACT for non-section 112(c)(6) pollutants. The commenter stated that,
specifically, EPA proposed limits for PM emissions as a surrogate for
non-volatile metal HAPs (arsenic, cadmium, beryllium and lead) on cement
manufacturing facilities that are area sources. Alternatively, the EPA
could choose to not regulate area sources or a subcategory of area
sources that are small entities for these pollutants. Emissions from
area sources, and even more so from the subcategory of area sources that
are small entities, are very small.

Response: EPA explains in the preamble to the final rule its various
dispositions for applying standards to area cement kiln sources.

2.3.6	Legal

Removal efficiency standards vs emission limits

Comment: One commenter (2898) supported EPA’s decision to not rank
best performers based on their relative mercury removal efficiency.
Relying on mercury removal efficiency in setting the MACT floor for the
portland cement manufacturing industry would downplay the role of HAP
inputs on emissions. EPA characterizes Brick MACT’s statement that
best performers are those emitting the least HAP as appearing arguably
in dicta. However, the Brick MACT Court itself characterizes the
statement as the holding of the Cement Kiln case. Brick MACT, 479 F.3d
at 880 (relying on Cement Kiln’s holding that §7412(d)(3) requires
floors based on the emission level actually achieved by the best
performers or those with the lowest emission levels). The proposed
alternative of setting the MACT floor on the basis of percentage of
emission reduction achieved by sources would minimize, if not eliminate,
the consideration of cleaner inputs in setting MACT floors, as EPA
acknowledges, and is therefore contrary to statutory dictates and case
law.

Response: See response in preamble to final rule.

Comment: Several commenters (2831 and 2901) stated that EPA retains
considerable discretion on how to set MACT floors. The commenters
supported the Agency’s authority to set floor standards based on
control efficiency, or any method as long as their method reasonably
estimates the performance of the relevant best performing plants. There
is nothing in the court’s decisions that requires EPA to use the
straight-emissions approach favored in this rule. The commenter stated
that the court has expressly decided that a straight-emissions or
arithmetical methodology is not required, and EPA’s technology based
approach that estimated performance rather than deriving the standards
through an arithmetic straight emissions approach, as long as they
result in a reasonable estimate of the performance of the best
controlled units, is supported by the courts. According to the commenter
(2901), Brick MACT does not endorse a straight emissions approach; nor
could it. To do so would mean that the Brick MACT court was overturning
the Chevron step one holding in Sierra Club and National Lime II,
something that it cannot do.

Response: See response in preamble to final rule.

Comment: One commenter (2832) stated that EPA has incorrectly focused on
the term emission limitation in proposing a numeric allowable HAP
emissions MACT floor. In fact, EPA is required to set emission
standards, not necessarily numeric emissions limitations in creating a
MACT floor. The Court in the Brick decision did not require otherwise.
In fact, the wording of §112, a reduction in emissions that is achieved
in practice by the best controlled similar source contemplates air
emissions control equipment whose manufacturer’s specified HAP
emissions removal efficiency for all HAPs is the highest percentage
achieved in practice in the industry in question. Inlet and outlet stack
testing that confirms the efficacy of air emissions control equipment is
the legal requirement of the §112(d)(3). The law does not state that
EPA is required to set numeric HAP emissions limits. In identifying the
best performing sources, the Administrator is directed by §l12(d)(3) to
review emissions information, not numeric emissions data or emissions
stack test reports. Sections 112(j)(5) and 1l2(i)(5) also support this
reading of §112(d)(3). Each refers to demonstrations of emissions
reductions of a stated numeric percentage as a means of meeting MACT
requirements. EPA’s discussion of the beyond-the-floor evaluation
confirms this approach as well. 

Response: The commenter does not address the issue of how non-numeric
percent reduction standards address the issue of raw material and fuel
HAP levels and their contribution to emissions. EPA cannot ignore such
contributions in determining which sources are best performing, and
their level of performance. At the least, EPA is adopting a reasonable
approach of establishing numerical standards, and using a ranking
methodology based on which sources emit the least of the HAP in
question, on a normalized basis (for Hg and PM) and considering their
performance over time (variability).

Standards must be achievable

Comment: One commenter (2844) stated that the CAA requires that lawfully
promulgated NESHAP standards must be achievable. Section 112(d)(2) of
the Act, 42 U.S.C. §7412(d)(2), required EPA to establish emission
standards for HAPs that require the maximum degree of reduction in
emissions taking into consideration the cost of the emission reduction
and non-air quality health and environmental impacts and energy
requirements, that the EPA Administrator determines is achievable for
new or existing sources. Further, House Rep. 101-490, Part 1 (328)
stated that The Committee expects MACT to be meaningful, so that MACT
will require substantial reductions in emissions from uncontrolled
levels. However, MACT is not intended to require unsafe control
measures, or to drive sources to the brink of shutdown. The commenter
noted that the proposed portland cement proposed NESHAP standards do not
comply with §112’s achievability requirements.

Response: See preamble to final rule for response.

Comment: Commenter 2844 stated that EPA’s conclusion that §112(d)(3)
and/or Brick MACT requires or even permits the Agency to ignore the
achievability requirements of §112(d)(2) is an unreasonable reading of
the statute and of Brick MACT. The Agency retains more than sufficient
discretion to devise NESHAP standards that successfully bridge the
tension between achieved and achievable in §112’s standard-setting
provisions by appropriately using both subcategorization and variability
methodologies.

Response: See response in preamble.

Comment: One commenter (2844) stated that EPA’s floor setting
methodology does not comply with three of Brick MACT’s requirements: 

Floors must be based on emissions achieved in practice by
best-performing sources; 

EPA’s use of variability factors and methodologies to adjust reported
emissions data must be based on demonstrated relationships, so that the
floor setting methodology serves to reasonably estimate or predict the
performance of the best performing sources; and 

EPA must consider the impact of nontechnology factors, such as raw
material and fuel inputs, on a source’s emission control levels.

Commenter 2844 stated that in the portland cement proposal, EPA set MACT
floor levels that reflect the specific conditions at the time the data
were generated and do not include any of the operational variability.
The commenter suggests that EPA must look beyond its snap shots of
performance to make a reasoned evaluation and estimation of all
operating conditions and factors that might impact the level of actual
emissions from those kilns in practice, and adjust their reported short
term test data appropriately. EPA can and should adjust raw emissions
results to estimate sources’ achieved emissions levels when setting
MACT floors and standards. Since Brick MACT, EPA’s methodology now
must be able to reasonably estimate the impacts of variability
associated with both technological and nontechnological factors over the
full range of circumstances. 

Response: See response in preamble to final rule.

Comment: One commenter (2844) also stated that EPA adopted a floor
setting methodology that is based on using lowest reported emission
results with minimal variability adjustments to estimate emission
control achieved in practice by best performing sources. EPA considered
test-to-test variability, but did not consider the inherent variability
due to raw materials, product mix, fuels, operating conditions and plant
types. The commenter stated that EPA has not evaluated or validated
whether its methodology accurately estimates emissions control achieved
in real world circumstances at sources.

Response: See response in preamble.

Comment: One commenter (2844) stated that EPA cannot evaluate floors
using methodologies that focus exclusively on technology if the
resulting standards do not reflect actual average limitation[s] achieved
(Brick MACT, 479 F.3d at 882). The commenter concludes that Brick MACT
requires EPA to address the role of nontechnological factors that impact
emissions in setting floors and EPA must develop a methodology that
accurately estimates the actual emissions achieved in practice by the
best performing sources under a variety of operating conditions, taking
into consideration testing and technological and nontechnological
variability. As proof that EPA failed to properly account for sources’
variability in setting the standards, the commenter (and commenter 2845)
included a chart purporting to demonstrate that the kilns comprising the
pool of best performers for each HAP could not themselves meet the
proposed standard. 

Response: See response in preamble.

Comment: One commenter (2844) stated that the proposed standards cannot
be viewed as assessing the emissions control achieved by best performing
sources, given that there is no single existing source that meets the
standards simultaneously and for two of the pollutants, no existing
units meet the standards without operational changes. The EPA has an
obligation to devise appropriate variability methodologies and to
collect the information necessary to do so.

Response: EPA has responded above and rejected the argument that the
statute unambiguously compels that floors be established on a complete
facility basis.

Worst Reasonably Forseeable Circumstances

Comment: Several commenters (2844, 2846, and 2915) stated that EPA’s
floor methodology does not estimate emission control achieved under the
worst reasonably foreseeable circumstances. EPA must estimate emission
levels achieved under the worst foreseeable circumstances in order to
reasonably estimate achieved in practice control levels. The proposed
portland cement rule proposes that sources must achieve compliance with
the proposed floor standards at all times, including periods of startup,
shutdown and malfunction (SSM) events. Because the data to estimate the
appropriate variability factor for SSM events do not exist and are
technologically impracticable to develop, EPA cannot reasonably set a
floor based on compliance during SSM events. 

Commenter 2845 stated that the policy and common sense approach is
recognized and required by the D.C. Circuit. Once the similar sources
category is established, the EPA must identify the lowest emission
limitation that has been achieved in practice by these sources. In
another MACT case, the D.C. Circuit has held that the phrase achieved in
practice refers to the performance of the source under the worst
foreseeable circumstances (Sierra Club v. EPA, 167 F.3d 658, 664, D.C.
Cir. 1999). 

Commenter 2845 stated that EPA should follow its own common sense
approach and the directive of the D.C. Circuit and develop MACT
standards that consider the long term emission variations in portland
cement facilities.

Response: First, EPA has established separate standards for startup and
shutdown, so these operating conditions are not part of the performance
EPA is evaluating in developing the main MACT floors. EPA also explains
in the preamble to the final rule why it reasonably did not consider
malfunction events in developing those floors, which reasons include
that EPA is therefore adding to the final rule an affirmative defense to
civil penalties for exceedances of emission limits that are caused by
malfunctions. Since malfunctions, if properly demonstrated, would not be
considered to contribute to an exceedance of the standard (for purposes
of civil penalties), EPA surely need not consider malfunctions in
assessing best performing sources’ performance and establishing MACT
floors.

Second, EPA agrees that long term performance of sources – their
variability – is to be considered both in determining which sources
are best performing and what their level of performance is. EPA has
reasonably done so in establishing this NESHAP.

CAA requires emission reductions by 94 percent of sources

Comment: One commenter (2890) stated that the setting of a MACT floor
requires 94 percent of any population of sources, including those that
had never engaged in any emission reduction techniques, to either add
air pollution control devices or shut down processes, without the
consideration of costs. The floor should identify what the best actual
sources in the category are doing; presuming that if the average of the
top 12 percent was achieving this level of control in the absence of a
rule, achieving that level of control must have been at a reasonable
cost.

Response: The commenter is correct that costs cannot be considered in
establishing MACT floors and that by virtue of the floor process the
statute implicitly assumes that the great majority of plants will have
to upgrade their performance to meet promulgated standards of at least
the designated minimum stringency.

Exclusion of sources engaged in mining

Comment: Several commenters (2890 and 2916) stated that Congressional
intent in the establishment of MACT shows a clear intention to exclude
mined minerals: For categories and subcategories of sources of hazardous
air pollutants engaged in mining, extraction, beneficiation, and
processing of nonferrous ores, concentrates, minerals, metals, and
related in-process materials, the Administrator shall not consider the
substitution of, or other changes in, metal- or mineral-bearing raw
materials that are used as feedstocks or materials inputs, or metal- or
mineral-bearing materials processed or derived from such feedstocks or
materials in setting emission standards, work practice standards,
operating standards or other prohibitions or requirements or limitations
under this section for such categories and subcategories. The
prohibition of the preceding sentence shall not apply to the
substitution, modification, or changes of chemicals (not including
metal- or mineral-bearing materials) used in mining, extraction,
beneficiation, or processing of nonferrous ores, concentrates, minerals,
metals and related in-process materials which is necessary to reduce air
emissions of such chemicals and for which substitutes that are safe and
effective in performing the intended function of the chemical to be
substituted are reasonably available [From Joint Explanatory Statement
of the Committee of Conference, H.R. Rep No. 101-952, at 339 (1990)].
EPA used this same quote in a previous court case involving the
secondary copper MACT [Sierra Club v. EPA (No. 02 -1253)]. 

Commenter 2890 suggested that after mined minerals are eliminated from
consideration for MACT, several options for establishing the best
controlled sources are:

Control of the mercury and other pollutants that are not in materials
other than the limestone typically mined onsite.

Any emission control equipment or other practice that reduces the amount
of emissions that are ultimately emitted from the processing of the
mined material.

Response: See preamble to final rule for response.

Consideration of variability

Comment: One commenter (2845) stated that case law and policy dictate
that the EPA must consider variability in establishing MACT standards,
and the approach used by EPA in Prevention of Significant Deterioration
(PSD) permitting should also apply in establishing MACT standards. To
evaluate the emission limits achieved by existing sources, EPA is
required to develop methodologies for estimating the variability
associated with all factors that impact a source’s emissions,
including process, operational and non-technological variables (see
Nat’l Lime Ass’n v. EPA, 627 F.2d 416, 443, D.C. Cir. 1980). While
courts have affirmed EPA’s authority to choose a methodology designed
to estimate emissions in setting the MACT floor, the courts have also
made clear that EPA’s method must allow a reasonable inference as to
the performance of the top 12 percent of units (Cement Kiln Recycling
Coalition v. EPA, 255 F.3d 855, 862 (D.C. Cir. 2001) (citing Sierra Club
v. EPA, 167 F.3d 658, 663, D.C. Cir. 1999). Accordingly, the Court of
Appeals for the D.C. Circuit has stated that EPA must show not only that
it believes its methodology provides an accurate picture of the relevant
sources' actual performance, but also why its methodology yields the
required estimate (Cement Kiln Recycling Coalition, 255 F.3d at 862).

Response: See response in preamble.

Comment: One commenter (2845) further stated that EPA has a clear policy
of considering long-term variability when establishing enforceable
emission rates. In considering appropriate enforceable emissions rates
in PSD permits under the CAA, the Environmental Appeals Board has held
that a permitting authority is not required to set the Best Available
Control Technology (BACT) emissions limit at the most stringent
emissions rate that has been demonstrated in practice by a facility
using similar emission controls (In re Newmont Nevada Energy Investment,
12 E.A.D. 429, 441-42, EAB 2005 (citing Kendall New Century Dev., 11
E.A.D. 40, 53, EAB 2003). The Board has recognized the use of safety
factor[s] in the calculation of permit limits to take into account both
the variability and fluctuation in the expected performance of pollution
controls and the fact that setting emissions limitations for one
facility at the highest control efficiency demonstrated at another
facility would make violations of the permit unavoidable (Id. at 442,
citing In re Masonite Corp., 5 E.A.D. 551, 560, EAB 1994).

Response: Although BACT determinations under PSD are by no means
analogous to determining MACT floors (cost is considered in determining
BACT, for example), EPA agrees that consideration only of single data
points may not adequately represent sources’ performance over time,
and EPA consequently is not utilizing that methodology in this rule. EPA
also has taken variability into account during the process of
establishing the emission limits for this rule. 

Comment: Commenters 2844, 2845, and 2916 objected to EPA’s
interpretation of CAA Sections 112(d)(2) and 112(d)(3) and the Brick
MACT (Commenter 2845 provided a white paper as an appendix to their
comments for the HWIMI MACT proposal, dated December 01, 2008.). The
paper, entitled Implications of the Brick MACT Decision on EPA’s
Discretion in Setting MACT Floors, discusses variability at some length.
The paper’s main points were:

The Agency has chosen to focus on setting MACT floors based on lowest
emitting sources derived from limited test results that are not
appropriately adjusted to account for stack test variability. 

The Brick MACT decision holds that EPA must base MACT floors on achieved
emissions control rather than control technology, but it does not
require EPA to ignore operational variability in determining those
floors.

Variability methodologies must reasonably estimate or predict emissions
or variability through a demonstrated relationship between the data used
and the performance intended to be estimated. 

Non-technological factors (i.e. raw materials and fuel) must be
considered in determining emission control achieved by best performers.

It is within the EPA’s discretion to define the best performing
sources. 

EPA should estimate variability in determining achieved emissions.

The Agency can and must seek appropriate data from regulated entities
and other stakeholders, and to develop appropriate fact-based estimating
methodologies on the data available.

Response: See response in preamble.	

Brick MACT

Comment: Several commenters (2834, 2844, 2845, and 2859) stated that the
Brick MACT decision makes the expectations of what is expected by EPA in
regards to setting the floor very clear: 

Floors must be based on emission control achieved in practice by the
best performers

Variability methodologies must reasonably be calculated to estimate such
actual performance levels

Nontechnological factors that impact emissions must be considered in
determining the achieved emissions levels 

In devising new methodologies to set lawful floors, EPA must ensure that
compliance with §112(d)(2) general standard setting criteria is also
met.

EPA can use discretion to define the factors necessary to establish a
floor under §112(d)(3) that meet the CAA’s statutory intention to be
technology-forcing and require compliance with the best that has been
demonstrated to be achieved in practice while considering costs, other
non-air health and environmental impacts, and energy impacts when
defining achievable MACT standards. 

Based on these points, the Brick MACT decision requires that EPA’s
floor analyses include a review of the best performers and consideration
of variable factors that can affect emissions achieved in practice by
the pest performance sources. EPA’s analyses and methodological
choices must be based on facts and records that support its final
determination. It is appropriate that a full review of the existing
records and the development of new data be considered in making a final
decision.

Response: The case law makes overwhelmingly clear that the requirement
in section 112 (d)(2) that MACT standards be achievable after
considering various factors does not supersede the requirement of
section 112 (d)(3) that they be of minimal stringency as determined by
the criteria for establishing MACT floors. See, e.g. Cement Kiln
Recycling Coalition, 255 F. 3d at 861-62. As discussed in the preamble
and in other comment responses, EPA has carefully assessed best
performers’ short- and long-term variability for each HAP in
estimating their performance.

2.3.7	Format of Emission Limits

Use CEMS units as limits

Comment: One commenter (2638) recommends the EPA use plain language and
units that can be related to ambient measurements for its emission
limits. The units currently proposed require modifying the direct CEMS
output with inputs under the control of the plant, which make the
results un-interpretable by anyone wanting to make comparisons with
ambient measurements or health effects; or the CEMS concentration
readings should be required to be posted/reported.

Response: Emission limits are typically normalized to some type of
production or raw material input value because this allows comparison
(and ultimately the ability to set a single standard) for different
sized facilities. As we noted in the NSPS proposal, the current NSPS and
limits (and NESHAP limits before today’s amendments) for PM are
expressed on a pound of PM per ton (lb/ton) of dry feed input format.
See 73 FR at 34075-76. In this final NESHAP (and NSPS) we are adopting a
new normalizing parameter of lb/ton of clinker – i.e., normalizing
based on kiln output rather than input for both PM and mercury.

We noted at proposal of the NSPS that adopting an output-based standard
avoids rewarding a source for becoming less efficient, i.e., requiring
more feed to produce a unit of product, therefore promoting the most
efficient production processes. 73 FR at 34076.

EPA therefore proposed that all of the NSPS (for PM, NOx, and SO2) be
normalized by ton of clinker produced, and later proposed the same
parameter for the two standards in the NESHAP which are normalized,
mercury and PM. 73 FR at 34076; 74 FR at 21140.

In the final NESHAP (and NSPS) we are therefore adopting a new
normalizing parameter of lb/ton of clinker – i.e. normalizing based on
kiln output rather than input – for mercury and PM in the NESHAP, and
for PM, NOx, and SO2 in the NSPS.

We are currently developing a reporting mechanism for CEMS data. We will
consider these comments as we continue that effort. In concept, the CEMS
data base would be publically available. However, the reporting
frequency has not been developed.

See the preamble discussion under format of the normalized standards in
the NESHAP and the NSPS, for more information on EPA’s decision.

Use of kiln feed vs clinker in emission limits

Comment: Commenters (2830, 2844, 2845, 2859, and 2879) stated that they
supported emission limitations in pounds of pollutant per ton of kiln
feed rather than as pounds per ton of clinker produced. Commenters 2844,
2859, and 2879 noted that clinker conversion factors may vary from plant
to plant and may give erroneous results. According to the commenters,
clinker weigh systems are not common in the cement industry and will not
improve accuracy over a kiln-feed measurement system with an appropriate
clinker factor, as is commonly used. On a short-term basis, clinker
output from the kiln will fluctuate while the kiln feed rate remains
constant. The variable nature of clinker production creates more
inaccuracy in short-term (hourly) clinker weigh data than the +/- 5
percent specified by the EPA. If the standards are based on tons of
clinker, then the kiln owner should have an option to determine clinker
weight from feed that should not require EPA approval of the method and
the averaging time and accuracy requirements should reflect what is
practical, achievable and useful.

Commenters 2844 and 2845 asked for clarification from EPA on proposed
language regarding alternative methods to measuring clinker production
and the need to obtain Administrator approval for such methods. The
Proposal also stated that the use of a system that directly measures
kiln feed rate and uses a conversion factor to determine the clinker
production rate is an acceptable method. The commenter states that it is
unclear whether EPA believes that Administrator approval is required to
utilize such a method when EPA identifies in the regulations that the
method is acceptable. In any event, cement plants should be given the
option of complying with the regulation by either measuring the clinker
production or by measuring kiln feed and developing a clinker factor to
calculate clinker production. Almost all cement plants measure kiln feed
and develop a clinker factor. Such a method requires periodic
adjustments to the clinker factor and justifies a longer time-frame than
hourly for determining clinker production.

Response: See the response to the comment above, regarding the format of
the standards.

Commenters maintain that the measurement of kiln output is not as exact
as the measurement of kiln input, and that many kilns have not installed
clinker measuring equipment. These objections do not necessitate
normalizing by inputs. Most commenters also stated that kiln feed could
be accurately measured and also indicated that most facilities currently
derive reasonable feed-to-clinker conversion factors from these
measurements. Kilns already calculate clinker production in this way
when required to meet emissions limits normalized by clinker production,
as many NSR and PSD permits for cement kilns presently do. 

Since it appears from comments that the equipment to accurately measure
clinker is not typically installed in this industry, we must assume
these facilities use a feed-to-clinker conversion factor to calculate
clinker production on whatever time basis is necessary (e.g., daily,
hourly, etc.). Therefore, we have modified the rule language to more
clearly provide the option allowing facilities to measure feed inputs
and to use their site specific feed/clinker ratio to calculate clinker
production (and that no prior approval from a regulatory authority is
necessary to do so). Facilities would be allowed to use a constant
feed/clinker ratio in accord with their usual cycles for determining
such ratios, typically on a monthly basis when clinker inventories are
reconciled. 

Commenters were nonetheless concerned that because clinker/feed ratios
change somewhat and are only re-determined at the end of a cycle, a
slight change in clinker/feed ratio, determined at the end of the cycle,
would show lack of compliance without an opportunity to alter operation.
To obviate this concern, the rule provides that facilities are not
required to retroactively update clinker production estimates after
re-computing feed/clinker ratios. We would not expect that the
clinker/feed ratio will change significantly from month to month, so we
do not see this as creating a situation where facilities will be able to
have large amounts of excess emissions but still be considered in
compliance (especially since the 30-day standards are all rolling
averages).

See the preamble discussion under format of the normalized standards in
the NESHAP and the NSPS, for more information on EPA’s decision.

Accuracy of determining kiln input or output

Comment: One commenter (2845) stated that kiln feed accuracy should be
established on a longer timeframe than one hour:

Daily clinker production typically is estimated by measuring the kiln
feed rates and utilizing a clinker factor to convert kiln feed rate
(which includes kiln feed and kiln dust returned) to clinker production
rate. Several variables in the calculation are measured are not
determined more frequently than a weekly basis and can vary on a short
term basis. Another variable, dust loss, is estimated, not measured. It
cannot be expected to have a clinker factor in the short term with an
accuracy of better than +/-10 percent. Tracking these variables on a
short term basis is not practical. 

Plants typically measure clinker inventory daily, and conduct more
detailed physical inventories on a monthly and/or quarterly basis. 

Plants that utilize a single clinker handling system for multiple kiln
lines cannot confirm the clinker factor for each individual kiln.

Response: See the response to the comment above, regarding the accuracy
of determining kiln input or output, and the latitude for determining
site-specific clinker factors. See the preamble discussion under format
of the normalized standards in the NESHAP and the NSPS, for more
information on EPA’s decision.

Grain loading format

Comment: Commenter 2859 stated that as an alternative to basing the PM
standard on pounds per ton dry feed, EPA could use a grain loading
standard (i.e., gr/acf).

Response: See the preamble discussion under format of the normalized
standards in the NESHAP and the NSPS, for more information on EPA’s
decision regarding format of the standard.

2.3.8	Other

Central stack

Comment: Many commenters (2893, 2975, 2990, 2992, 3001, 3003, 3005,
3017, 3022, 3034, 3038, 3045, 3052, 3056, 3059, 3063, 3073, 3074, 3077,
3089, 3098, 3107, 3117, 3124, 3125, 3128, 3130, 3131, 3132, 3160, 3226,
and 3227) stated that EPA should require that all cement kilns have one
central stack. Without a central stack, the emissions from the kiln
cannot be adequately monitored, and the CEM required in the proposed
rule would not be viable for monitoring emissions and confirming
emissions reductions. In addition, necessary scrubbers and/or other
pollution control devices require a central stack for placement. 

Response: The majority of cement kilns are equipped with control devices
that are equipped with a central stack. However, those kilns using
positive pressure fabric filters may be equipped with roof monitors
instead of stacks. In the case of PM emissions, such devices can be
equipped with a bag leak detection system, which is a viable monitoring
alternative to PM CEMS and explicitly allowed in the rule as a
monitoring alternative. In other cases, the source may have to retrofit
the control device outlet to install one or more stacks. Finally,
sources can always request EPA approval under the part 63 General
Provisions (§63.7(f)) to use an alternative method to accommodate
site-specific conditions. For example, open-path monitoring may be
viable in some situations.

2.4	General Legal Issues

Comment: One commenter (2901) stated that National Lime II stands for
the principle that as long as the facility is controlling pollution, the
levels achieved can be considered, even if the facility was not
intending to control the specific HAP. The commenter noted that the
major issue in National Lime II was whether EPA could decline to set an
emission standard for HAPs that no facility in the source category was
controlling with technology. The National Lime II court was saying that
if a facility was already controlling its emissions, the fact that it
was not intending to control the HAP in question could not be a reason
for EPA to consider the pollutant uncontrolled.

Response: The commenter correctly describes this portion of the second
National Lime decision.

2.4.1	Subcategorization

Comment: One commenter (2844) stated that subcategorization based on
nontechnical factors may result in opportunities for reasonable beyond
the floor (BTF) standards. It would be appropriate for EPA to base the
NESHAP standard on BTF levels that reflect application of the
nontechnological factor as an intended control strategy under §112.
However, while Brick MACT requires EPA to take into account emissions
data irrespective of whether it was achieved through specific intended
operational choices or control technology, nowhere did the court say EPA
had to or even could simply assume that a source’s emissions in
practice would be invariably equivalent to a particular observed
emission level and that all operational and nontechnological factors
that happen to be in effect on the day of testing could be assumed and
required to be in effect, in the future.

Response: EPA agrees that sources’ variability is unlikely to be
encompassed by a bare single measurement. It is for this reason that EPA
developed all of the standards based on substantial data gathering
augmented where appropriate by statistical estimates of variability. 

Comment: One commenter (2844) stated that if EPA determines that raw
material inputs impact the emission control achieved in practice by the
best performing sources, EPA must assess that impact. Given that the raw
material input varies, the commenter stated that EPA’s variability
methodology must include that variability in estimates of emission
control achieved by the sources; further, if the variability in the raw
materials used by the differing units is significant, EPA may determine
that subcategorization is the most reasonable way to regulate the
industry to ensure that each source is subject to the most stringent
floor level of control reasonable to apply to sources utilizing raw
material sources of similar HAP input.

Response: EPA’s reasons for not creating a subcategory for the two
plants with exceptionally high mercury emissions due to limestone quarry
mercury content is set out in the preamble to the final rule.

Comment: One commenter (2846) stated that EPA has authority to establish
subcategories of sources when issuing standards. Subcategorization would
allow for a workable MACT standard and allow sources that are limited in
certain ways to apply the best available control technology for their
facilities and still comply with the standards. The commenter believes
that subcategorization for on the basis of feed material quality is
wholly appropriate and required by law.

Commenters 2832 and 2859 support subcategorization as permitted in the
Concurring Opinion in the Brick decision. This support includes but is
not limited to subcategorization according to process operations (type,
raw materials, chemistry/formulation data, associated equipment and
final products), emission characteristics (amount and type of HAP), and
air pollution control device applicability, opportunities for pollution
prevention, and existing regulations or guidance from states and other
regulatory agencies.

Commenters 2776, 2777, 2810, 2814, 2833, 2859, 2904, 2915, 2920, 2926,
3204, 3212, 3214, and 3215 supported the use of subcategorization in the
proposed rule. Commenter 2915 stated that in allowing EPA to create
subcategories, §112(d)(4) does not restrict subcategorization to only
factors that affect emissions. The provision requires EPA to establish
regulations for each category or subcategory on the schedules set out in
§112. EPA states in the preamble, EPA may create a subcategory
applicable to a single HAP, rather than to all HAP emitted by the source
category. While the commenter disagreed with EPA’s approach to MACT
standard-setting, if EPA continues to follow that HAP-by-HAP approach to
MACT standard-setting, then EPA should create different subcategories
for each HAP rather than one set for all HAPs.

Response: Because EPA is determining not to exercise its discretion to
create a subcategory for a subcategory for kilns with exceptionally high
mercury emissions due to high levels of mercury in their limestone
quarries, it is unnecessary to respond to these comments.

Comment: One commenter (2824) suggested that for the best protection of
human health, create an optimized mercury subcategory by evaluating
increased exposure from foreign plants in response to plants that could
shut down due to technology or economics while varying construction
demand. 

Response: This comment does not track the MACT standard setting process
prescribed by law, which requires technology-based determinations based
at a minimum on performance of certain best operated or controlled
domestic plants in the source category.

2.4.2	Fly Ash Ban

Comment: Several commenters (2816 and 2835) stated that it was
appropriate for EPA to remove the prohibition on kilns using fly ash
from utility boilers using activated carbon injection for mercury
control. Given that the mercury emission limit will apply regardless of
fuel or fuel substitutes, this prohibition is no longer needed.

Response: EPA agrees.

Comment: One commenter (2828) suggested that EPA retain a ban on cement
kilns pyroprocessing (or heating) of fly ash from power plants if the
mercury content of the fly ash has been increased above its normal
concentration by mercury capture systems at the power plant. Until there
is certainty on how coal- and oil-fired electric utility steam
generating unit (EGU) will be required to account for mercury
re-emissions from cement plants that acquire their fly ash, the cement
MACT rule should maintain a ban on the use of fly ash with captured
mercury. The commenter encouraged EPA to retain the proposed mercury
emission standard, and supplement it with a continued ban on fly ash
pyroprocessing if the ash has elevated mercury or if the power plant
producing the ash took credit for the reductions. The commenter
suggested that this rule, as well as the upcoming EGU MACT rule, require
coordination between fly ash producers and potential fly ash users so
that re-emission of mercury is properly accounted for, and we do not
have the problem of mercury being emitted to the atmosphere at one
location after being controlled at another. 

Response: The technology-based standard for mercury reflects the minimum
level of performance achieved by the average of the best performers, or
the best controlled similar source. If other kilns can meet these limits
while importing fly ash, there is nothing in section 112 (d)(3) which
prohibits them from doing so, since the provision is technology-neutral.
So long as the standards are satisfied, the statute does not dictate
how.

Comment: One commenter (2638) stated that coal ash is enriched in other
urban HAP metals besides mercury. These enrichments of coal are on the
order of 5 to 10 fold. With these enrichments, it will important that
emission limits be developed for each individual urban HAP metal and be
correspondingly monitored. Without this assurance, the EPA should not
allow coal ash to be used in the manufacture of portland cement.

Response: Given that the PM emission limit (as a surrogate for
nonvolatile metal HAP) will apply regardless of fuel, fuel substitutes,
or raw materials, this prohibition is not needed. The standard for PM
reflects the minimum level of performance achieved by the average of the
best performers, or the best controlled similar source. If kilns can
meet these limits while incorporating coal ash, there is nothing in
section 112 (d)(3) which prohibits them from doing so, since the
provision is technology-neutral. So long as the standards are satisfied,
the statute does not dictate how.

2.4.4	Other

Offsets

Comment: One commenter (2898) stated that any netting or offsets
pursuant to NSR must meet all CAA requirements. The CAA only allows
offsets for incidental emissions reductions that are not otherwise
required. 42 U.S.C. 172(c)(2). Federal regulations additionally require
that emissions reductions, in order to qualify as offsets, be surplus,
permanent, quantifiable, and federally enforceable. 40 CFR
51.165(a)(3)(ii)(C)(1)(i). Longstanding agency guidance, including the
General Preamble for State Implementation Plans (57 FR 13498, 13552-54,
dated April 16, 1992), imposes additional restrictions. Per the
foregoing, The EPA’s policy for NSR offsets is that since hazardous
emissions reductions required to meet the MACT standards at 40 CFR 60
and 61 … are required by the Act these emissions are not creditable as
NSR offsets. Memorandum from J. Seitz to B. Hannesschlager re: Crediting
MACT Emission Reductions for New Source Review (NSR) Netting and Offsets
(November 12, 1997).

Response: These comments relate to NSR, which is outside the scope of
this rulemaking.

Comment: Several commenters (2845 and 2874) agreed that EPA correctly
suggests that surplus sulfur dioxide credits could be used for offsets
and netting and commended EPA for this recommendation that surplus
sulfur dioxide emission reductions resulting from the installation of a
wet scrubber (used to satisfy sulfur dioxide offset requirements) can be
used for netting purposes under the CAA new source review program.
Commenter 2874 believed that this is an appropriate position and EPA
should repeat this statement in the preamble to the final rule, making
the point clear to regulators dealing with portland cement new source
review matters.

Response: See above.

2.5	Mercury Standards

2.5.1	Legal Issues

Subcategorization

Comment: Many commenters (2816, 2828, 2834, 2895, 2898, and 2923)
supported EPA’s approach of establishing a single national standard,
rather than weakening the proposed standard through the use of
subcategorization. According to several commenters (2816, 2834, and
2898), EPA properly rejected subcategorizing cement kilns for the
following reasons: 

Under §112(d)(1), EPA can only create subcategories based on classes,
types, and sizes of sources. Since limestone mercury content can vary
within a single quarry, kilns with proprietary quarries with sections
that have high mercury content are not sufficiently uniform to be
considered a class or type. Even for those high-mercury quarries
identified by EPA, the Agency is not aware of the mercury content of the
remainder of the quarry. Nor is there any suggestion that a quarry’s
mercury content is related to its size. Further, the lack of a causal
link between limestone mercury content and emission rates makes
subcategorizing by quarry content untenable.

Kilns with more mercury-rich limestone would be favored and those with
lower mercury levels disadvantaged. Kilns would be rewarded for
selecting the highest mercury limestone to test for categorization.
Thus, the category would create an incentive for kilns to concentrate in
areas with limestone formations that just exceed the mercury
concentration cut-off for the category. 

Subcategorization is unnecessary because it is not clear that any
outliers exist. EPA separated two facilities as having the highest
quarry mercury contents and identified those sources as potential
outliers. But EPA acknowledges that at one of those kilns it does not
currently have any data on the possibility that this site may have
portions of its existing quarry that have lower mercury content. They
may simply be in sections of their quarries that are particularly high
in mercury. Creation of a subcategory without further evidence it is
justified would be arbitrary.

Creating a subcategory for kilns with quarries that are high in mercury
would be a continuation of the pattern repeatedly rejected by the D.C.
Circuit as violating the CAA- setting emission standards based on a
concern about ensuring that its floor is achievable by all kilns in the
subcategory. If §112 is to meet its purpose of require the maximum
degree of reduction in emissions, then the standards created pursuant to
it cannot be set so as to accommodate the heaviest polluters. EPA must
heed the D.C. Circuit’s previous rejection of efforts to tailor
emission standards so as to ensure all sources can meet those standards.


Commenter 2898 agreed with EPA that none of the bases for
subcategorization discussed in the proposed rule meet the CAA’s
guidelines:

Any basis for subcategorizing must be related to an effect on emissions,
rather than to some difference among sources which does not affect
emissions performance. The commenter agreed that subcategorization on
the basis of the type of kiln or the presence of an inline raw mill is
inappropriate because none of these distinctions affects total mercury
emissions over time.

Subcategorization on the basis of geographic location is inappropriate
because mercury content can vary significantly from quarry to quarry and
even within a quarry, nor does a difference in geographic location
constitute a difference in class, type, or size. Moreover, EPA found no
correlation between geographical distribution of mercury emissions and
total mercury.

Subcategorization based on the practice of wasting cement kiln dust is
inappropriate as it can be considered a control technology and
subcategorization by control technology would defeat the purposes of the
MACT requirement, which is to control the worst HAP pollution.

Subcategorization by total mercury inputs is inappropriate because total
mercury inputs do not differentiate facilities by size, class, or type
because a particular set of inputs are not inherent to the operation of
the cement facility. Because subcategorization on this basis would
result in a situation where facilities with higher total mercury inputs
would have higher total emission limits, it would create the same
problems as subcategorizing by control technology. 

Subcategorization on the basis of mercury content of facility’s
proprietary limestone quarry is inappropriate would essentially be
subcategorizing by control technology. As Brick MACT confirmed, inputs
are a means of emissions control. As with total mercury inputs, the
mercury content of the proprietary limestone quarries does not
differentiate facilities by size, class, or type. There is nothing about
the limestone from the proprietary limestone quarry that is inherent to
the process of the cement facility. While it is true that all cement
plants have a limestone quarry located near the facility and that
facilities are often sited based on the location of the quarry, the
proprietary nature of the quarry does not diminish the fact that the
cement facility’s kilns will work equally well with limestone from
other sources. The only operational difference between limestone from
the proprietary quarry and limestone from other quarries is cost (for
transportation or purchase), an impermissible consideration in setting
the MACT floor. Separate standards would unfairly expose communities
near the dirtiest cement facilities to the highest levels of mercury
pollution from cement plants and leave them with the least protection
against such pollution. Subcategorization by the mercury content of the
limestone quarry runs counter to the holding of Brick MACT and other
decisions. Brick MACT holds that EPA must account for raw material HAP
contributions in establishing MACT floors and that the proprietary
nature or proximity of the raw material does not excuse EPA from
considering the HAP emission from the raw material in setting the MACT
floor.

Commenter 2898 disagreed with the Portland Cement Association’s
(PCA’s) (commenter 2845) arguments for subcategorization by mercury
content of proprietary quarries. The commenter provides counterarguments
to (PCA’s) citations and interpretations of case law.

Commenters 2842 and 2911 opposed the use of subcategories. Commenter
2911 stated that setting separate standards for kilns that use high
mercury limestone would be unlawful. 

The sentence in §112(d)(1) that allows EPA to distinguish among
classes, types, or sizes of sources within a category or subcategory in
establishing standards does not permit EPA to set separate standards for
cement kilns that merely have limestone quarries containing higher
levels of mercury. The distinctions between class and type must relate
to the source itself and not a distinction between the raw materials
used by the source. Under EPA’s own definition, a quarry is not any
part of an affected source.

EPA states inaccurately that kilns with mercury above a given level
might be considered a different type or class of kiln because their
process necessarily requires the use of that higher mercury input (74 FR
21145). EPA does not and cannot identify any kiln whose process requires
the use of high mercury input. At most, it is merely cheaper for these
kilns to continue using such input.

Although EPA stated that [t]he concurring opinion in Brick MACT supports
subcategorization in situations involving sources’ dependence on
high-HAP raw materials to avoid situations where a level of performance
achieved by some sources proves unachievable by other sources even after
application of best technological controls, viewing such sources as of a
different type than others in the source category (74 FR 21148 and
Sierra Club v. EPA, 479 F.3d 875, 884-885, D.C. Cir. 2007, and 74 FR
21140 [concurring opinion may provide a way out of the situation where
the floor standards are achieved for some sources, but the same floors
cannot be achieved for other sources due to differences in local raw
materials whose use is essential]), those statements are no more than
dicta in a concurring opinion without precedential value or authority. 

EPA itself states subcategorizing by control device … is not
permissible, 74 FR 21145. As the CAA and D.C. Circuit precedent both
make clear, the best performing sources are those with the lowest
emission levels, regardless of whether those emission levels are
achieved through the use of end-of-stack control technology or the use
of cleaner fuels or raw materials. Sierra Club (479 F.3d 879-880).
Because there is no relevant distinction for floor purposes between
controlling emissions with an end-of-stack control device and
controlling them through the use of cleaner fuels and raw materials,
setting separate and more lenient standards for sources with high
mercury limestone is just as impermissible as setting such standards for
sources with ineffective or outdated control technology would be.

Commenter 2911 stated that setting separate standards for kilns that use
high mercury limestone would be arbitrary and capricious. Nothing in the
CAA suggests that standards have to be achievable through the use of any
particular control technology or through control technology in general.
Creating a separate subcategory would create a perverse incentive for
other kilns to find high mercury limestone so as to avail themselves of
the same effective exemption from mercury emission standards.

Commenter 2842 states that the decision not to subcategorize is fully
consistent with the language in §112 for setting emission standards.
Congress left the decision of whether to subcategorize based on class,
type, and size of source entirely to the Agency’s discretion. Based on
the record before it, the Agency reasonably concluded that the mercury
emissions data do not readily support a subcategorization approach. The
commenter’s further argued that: 

EPA’s technical determination is worthy of deference. See Natural
Resources Defense Council, Inc. v. EPA, 489 F.3d 1364, 1375 (D.C. Cir.
2007) (rejecting industry challenge to EPA decision not to subcategorize
in light of the technical nature of the determination and the agency’s
reasoned explanation).

The argument advanced by certain industry representatives that
subcategorization is required because some kilns with high levels of
mercury in their feedstocks will not be able to achieve the standards in
a cost effective manner finds no support in the statute. The D.C.
Circuit has held that EPA must account for raw material HAP
contributions in establishing MACT floors, and the fact that raw
materials may be proprietary or otherwise not obtainable category-wide
does not relieve EPA of that obligation. 74 FR at 21148 (citing Sierra
Club v. EPA, 479 F.3d at 882-83).

Section 112 does not allow EPA to consider costs in establishing the
MACT floor for HAPs such as mercury. NRDC v. EPA, 489 F.3d at 1375; see
Sierra Club v. EPA, 479 F.3d 875 (D.C. Cir. 2007). Congress envisioned
that the establishment of the MACT floor would be a technical
determination, not an economic one. See Sierra Club v. EPA, 479 F.3d at
875 (rejecting argument that in establishing MACT floors based on what
best performing sources have achieved, EPA may redefine best performing
to mean those sources with emission levels that are achievable by all
sources). 

Judge Williams in the Sierra Club case noted, there is not necessarily
any guarantee, even with suitable subcategorization, [that] every source
will be able to achieve standards that meet a lawful application of
§112(d)(3) to reasonably defined subcategories. 479 F.3d at 885.

Response: Because EPA is choosing not to exercise its discretion to
create a subcategory for high mercury emitters with high mercury
limestone quarries, it is unnecessary to respond to these comments.

Comment: Several commenters (2845, 2863, and 2916) stated that case law
under the CAA §111 New Source Performance Standards (NSPS) made it
clear that EPA may establish subcategories that are essentially based on
a pollutant concentration in an input to the process. In its coal-fired
power plant NSPS, EPA essentially distinguished between sources based on
the concentration of a pollutant in an input – the sulfur
concentration in the coal. The D.C. Circuit upheld EPA’s approach,
holding that it had considerable latitude in differentiating between
source types. Sierra Club v. Costle (657 F.2d 298, 318-319, D.C. Cir.
1981). The power plant NSPS required greater sulfur dioxide removal
efficiency for sources with a higher concentration of sulfur dioxide
emissions than for sources with a lower concentration in the emissions.
When challenged, the D.C. Circuit quoted the §111(b)(2) language on how
EPA is authorized to distinguish among classes, types, and sizes of
sources, and concluded that [c]ertainly the text of the statute nowhere
forbids a distinction based on sulfur content. 

Examples cited by the commenters of other MACT standards using
subcategorization by pollution concentration in raw materials are:

Proposed National Emissions Standards for Hazardous Air Pollutants for
Refractory Products Manufacturing, 67 FR 42108, 42110, and 42119 (June
20, 2002);

Proposed National Emissions Standards for Hazardous Air Pollutants for
Plastic Parts and Products, 67 FR 72276 and 72288 (December 4, 2002);
and

Proposed National Emissions Standards for Hazardous Air Pollutants for
Chemical Manufacturing, 73 FR 58352 and 58365 (Oct. 6, 2008). 

Commenter 2845 stated that EPA should establish two subcategories for
mercury based on the concentration of mercury in the limestone, and
should set beyond-the floor standards for the high-mercury limestone
kilns requiring a 75 percent system removal efficiency. EPA’s Figure 1
and Table 3 of the proposal preamble clearly show two outliers with
mercury concentrations that far exceed those of any other sources in the
database. 

Commenter 2845 stated that as EPA further notes in the preamble, [The
Ash Grove Durkee kiln] has voluntarily entered into an enforceable
agreement to install activated carbon (the best control technology
currently available so far as is known) to control its mercury
emissions…Closure of this kiln and possibly other high mercury
emitting kilns is a possible consequence of a single standard without
subcategories. The same would be true for other kilns not in EPA’s
database that are similarly situated. This would be a result that
Congress clearly did not intend. As the House Report for the 1990 CAA
Amendments states, MACT is not intended to require unsafe control
measures, or to drive sources to the brink of shutdown. House Rep. No.
101-490, at 328 (1990).

. Creating a ≥225 ppb subcategory also is consistent with the
legislative history of the CAA. When 112(d) standards were developed,
the possibility of requiring substitution of cleaner ore stocks was
considered. In the Conference Committee report, Congress stated that the
Administrator shall not consider the substitution of, or other changes
in, metal- or mineral-bearing raw materials that are being used as
feedstocks or material inputs… Conference Committee Explanatory
Statement.

Response: Because EPA is choosing not to exercise its discretion to
create a subcategory for high mercury emitters with high mercury
limestone quarries, it is unnecessary to respond to these comments.

Best performers

Comment: One commenter (2911) stated that the holding that best
performers are those with the lowest emission levels is binding
precedent on the plain meaning of the CAA, and it reflects repeated
holdings of the D.C. Circuit to the same effect. To attempt to rank
cement kilns by control efficiency would not only contravene the CAA but
set bad policy and dangerous agency precedent. 

Response: EPA is choosing in this rule to rank best performers based on
emission levels (as normalized, for Hg and PM), and believes that this
is a reasonable and permissible approach for this rule.

2.5.2	Adequacy of Floor Determinations (Existing and New)

Overly lenient standards

Comment: Several commenters (2144, 2867, and 2879) stated that EPA’s
analysis of the floor resulted in overly lenient standards. They stated
that EPA should conduct simulations such as Monte Carlo and use other
statistical tools to handle truncated distributions. The commenter notes
that the proposed emission standard is greater than the 30-day
characteristics of any of the best performing 12 percent of the surveyed
kilns, resulting in a standard less stringent than the highest emitter
among the best performing kilns rather than equal to or more stringent
than the average emissions from the best performing 12 percent.
Commenter 2879 provided examples of alternative statistical analysis of
data that could be used to more rigorously determine the expected
average of the best performing 12 percent.

Commenter 2867 made the following recommendations:

Lower the proposed emission standard for existing kilns towards the
simple statistical average of 27.4 lb/MM tons of clinker consistent with
the plain reading of the law (average of best 12%) and leave the
proposed value for new kilns unchanged at 14 lb/MM tons feed (23.1
lb/ton clinker).

Resist attempts by industry advocates to introduce new variability
schemes in the determination of the final mercury standard.

Any subcategorization (if allowed at all) for the purposes of the
mercury limit should be in terms of time to achieve the standard and not
the magnitude of the standard.

Response: Monte Carlo simulation or bootstrap simulation has been used
to account for different types of uncertainties. The suggestion by the
commenters follows this recent tendency to use simulation techniques to
account for uncertainties in the data. However, for this rule, we
decided to use the Student’s t-test, which is called for when you have
only a relatively small, random sample of emissions data available for
analysis. EPA is not adopting any subcategories in this rule.

Analysis of variability

Comment: Many commenters (2830, 2836, 2841, 2845, 2858, 2859, 2863,
2864, 2874, and 2889) stated that the limited data used by EPA in
setting the proposed mercury limit does not adequately reflect the
significant variability in mercury emissions and EPA’s MACT floor
analysis resulted in overly stringent limits. They recommended that EPA
re-evaluate the data upon which the mercury standard for existing kilns
was derived, and use different statistical analyses. 

Commenters 2830 and 2845 recommended using the statistical analysis used
to set other NESHAP standards and programs under the CWA and the
Resource Conservation and Recovery Act (RCRA), which gives more accurate
consideration of the variability in certain data sets. 

Commenters 2859 and 2864 support commenter 2845’s discussion of an
alternative approach to analyzing the mercury data set used to determine
the MACT floor. This commenter recommended using the UTL method to
account for variability within the data set. The commenter’s UTL
analysis of the mercury data resulted in an alternative mercury floor no
less than 51.5 lbs/million tons of kiln feed for existing units and 14.2
lbs/million tons of kiln feed for new units. The commenter included
detailed statistical analysis in support of their recommended limits.
Commenter 2830 stated that the appropriate NESHAP standard for mercury
should be between 43 lbs/million tons of clinker and 50 lbs/million tons
of clinker for existing units. 

Commenters 2844, 2845, 2859, 2863, 2874, and 2889 stated that by using
data collection over only 30 days, EPA has failed to capture the
variability of the performance of those units. The commenters recommend
collecting data over a wide range of operating conditions or developing
alternative operating scenarios and the effects of alternative operating
conditions can be determined via mass balance calculations (by changing
the mix of materials and the mercury contents of those materials).
Commenters 2858, 2859, and 2874 noted that causes of variability such as
variations in raw materials and fuels, wasting of varying amounts of
kiln dust, and the production of different products were not adequately
addressed in the data collection process. Commenter 2863 provided three
years of results of analysis to highlight the variability of the mercury
content of their feed limestone, which all comes from the same quarry
and mineral formation. Limestone mercury concentrations ranged from 424
ppb to 1740 ppb. 

Commenters 2845, 2863, and 2864 noted that EPA considered only 30-day
samples of emissions data from various units in setting the mercury
standards, and in addition, EPA considered a kiln that was producing a
product during the 30-day data collection period that resulted in an
emission profile that was not representative of normal operations of the
kiln. If this kiln’s production of this product is weighted based on
the kiln’s actual production, its emissions increase and it is no
longer one of the best performing sources. 

Commenter 2845 conducted a site-specific variability analysis that
included interviewing plant personnel of the 11 MACT floor kilns.
Independently from the 30-day emissions data collected by EPA, the
analysis focused on normal and foreseeable operations of the kilns used
by EPA to propose the mercury standard. Each of the kilns in EPA’s
proposed MACT floor was reviewed for variations in products, raw
materials, fuels, operating conditions, and the effectiveness of
pollution controls. The commenter also considered EPA data regarding
mercury content of fuels that are used or could reasonably be used by
the facilities in the MACT floor. The analysis demonstrated that
foreseeable circumstances at units in the MACT floor justify an existing
source mercury standard of 106 lbs/million tons of kiln feed and 36
lbs/million tons of kiln feed for new sources. The commenter provided
detailed supporting information in an appendix. 

Commenter 2841 disagrees with EPA’s approach of mass balance tracking
the feed materials and assuming that all mercury entering the kiln
system would be emitted from the stack. Using a mass balance for such a
significant standard will result in significant costs with the real
possibility of little or no environmental benefit. It would be far more
appropriate to allow time for facilities collect more data and further
to allow each facility to determine the speciation of this naturally
occurring mercury in their limestone quarries and other sources of raw
materials. It may be appropriate to subcategorize by species of mercury
or other similar approaches, if enough data is collected.

Response: We carefully reviewed the comments on the analysis of
variability. (See the docket for the detailed review of these comments.)
It is clear that the commenters mistakenly believed that EPA intended to
set the MACT floor using the 99th percentile. The 99th percentile of a
data set, is the data value at which 99 percent of the data will be
found. They also state that because EPA was using a modified UPL, that
EPA was calculating a future probability rather than the 99th percentile
for the actual data, which they wrongly considered inconsistent with the
intention of the CAA. They stated that because of the small size of the
data sets for mercury, PM, and HCl that EPA’s choice of the 99th
percentile was faulty and they recommended that an upper tolerance limit
(UTL) for smaller data sets be used. They state that the UTL does
generate the 99th percentile. For larger data sets, such as the THC
data, the actual 99th percentile can be used. UTL is not for prediction
purposes, so it won’t say what the data will be in the future. The
prediction properties are a characteristic of the UPL. Both UPL and UTL
are based on the present data, but UPL take into account the future
observations variability. If the idea is to obtain an upper bound that
is greater than the 99 percentile, then the UTL works. The UTL will
generally be a larger number than the 99 percentile. This was not
EPA’s intention but we realize that our explanation at proposal was
not sufficiently clear. Based on comments and our reanalysis of our
methods, we realized that there was an error in the UPL equation at
proposal that has been corrected.

In response to comments, we re-evaluated the statistical analysis used
for the proposed rule and made changes to improve it. The changes we
made and why are discussed in the preamble to the final rule and in
supporting documentation in the docket and include why we continue to
believe the use of the UPL is the most appropriate statistical approach,
how we accounted for variability (including intra-quarry variability),
and other changes to the analysis since proposal. We disagree with the
commenters who said 30 days of data are insufficient, and we continue to
believe that the 30 data points provided from 30 days of sampling of all
kiln inputs for 89 kilns is a robust sample of the population of cement
kilns and adequate to characterize the variability encountered at each
individual kiln, especially as supplemented with the estimates of
intra-quarry variability. The top performing kilns used raw materials
and fuels from different sources, wasted varying amounts of cement kiln
dust, and included different types of kilns operating under various
operating conditions. Moreover, as discussed in the preamble, we
continued to refine our analysis and revised it to further refine
estimates of intra-quarry variability. We also corrected the case cited
by the commenter for the kiln that had an emissions profile that was not
representative of its normal product mix, and this kiln is no longer in
the top performing 12 percent.

We have also addressed in the preamble to the final rule the
commenter’s suggestion that, under “foreseeable circumstances,”
the MACT floor for mercury would be higher. In addition, we note here
that the 30-day rolling average form of the mercury standard allows an
owner or operator to detect trends in increasing mercury emissions over
days and to take corrective actions, whether by improving mercury
emission control devices or changing certain kiln inputs, before the
limit is necessarily exceeded. The 30-day rolling average provides much
more flexibility in dealing with foreseeable events than would a shorter
term averaging period (such as an hourly or daily average). 

Our use of mass balance tracking of mercury to develop the mercury
emission limit is reasonable because it is based on the knowledge that,
in the absence of a mercury emission control device, the mercury in the
kiln’s inputs is emitted. We see no reason why the use of a mass
balance approach rather than collecting data on speciated mercury
emissions (as the commenter suggests) would result in significant costs
with little or no environmental benefit, nor did the commenter provide
information that would support such a claim. There is also no reasoning
or information presented that would indicate that it might have been
appropriate to subcategorize by species of mercury or other similar
approach, and we do not believe such approaches are warranted.

Use of subcategories

Comment: Many commenters (2842, 2893, 2928, 2984, 2987, 2989, 2994,
3001, 3005, 3023, 3030, 3034, 3038, 3045, 3052, 3056, 3057, 3059, 3061,
3062, 3063, 3065, 3073, 3074, 3077, 3098, 3101, 3104, 3105, 3107, 3110,
3125, 3128, 3130, 3132, 3226, and 3227) agreed with EPA’s decision to
not create subcategories in establishing mercury standards. 

Commenter 2928 stated that because of the current effectiveness of
mercury controls available to cement plants, as well as a variety of new
control options available by the 2013 deadline these new proposed
emission standards set, it is not necessary to allow for subcategories
for cement plants that claim the mercury content of their limestone is
to blame for their emissions. 

Commenter 2893 stated that subcategorization based on limestone
composition unfairly depreciates the value of cleaner limestone sources
and undermines the intent of the proposed rule.

Commenter 2999 disagreed with the proposed subcategorization of plants
based on their limestone mercury levels, but sees the value of a
temporary subcategory for existing cement plants to allow those plants
more time to comply with mercury emissions rules if they have recently
made substantial investments to lower mercury emissions using a
reasonably close to state of the art technology.

Commenter 2828 stated that there are multiple options for the ultra-high
mercury emitting kilns:

Use of limestone with lower mercury levels, including transport of
limestone from sources within a 5 to 10 mile radius.

Installation of serial pollution control systems.

The combination of fuel selection (more natural gas, less coal), coal
and limestone washing, and multiple mercury control systems. The mercury
control systems could include ACI injection, followed by a baghouse, and
then if needed, another ACI/baghouse system.

Several commenters (2144, 2840, and 2845) argued for the creation of
subcategories. Commenter 2144 stated that EPA should create an interim
hyper mercury limestone subcategory only for the two kilns that
exhibited estimated emissions greater than 1,000 lb/yr and require the
least stringent of 90 percent Mercury removal and the final mercury
standard. The kilns in the hyper Mercury limestone category would
ultimately meet the final emissions standard and phase out the
alternative 90 percent removal requirement.

Commenter 2840 suggested that for a subcategory of cement plants with
elevated mercury levels in their limestone, EPA could require employment
of maximum achievable control technology to reduce mercury emissions and
limit the use of offsite materials containing elevated levels of
mercury.

Commenter 2845 stated that EPA should subcategorize emission sources by
mercury concentration in the limestone to address variability and
ability to comply issues. Given that cement kilns must use limestone
from a nearby quarry to produce clinker, they should be seen as
essentially co-joined with that quarry for purposes of
subcategorization. Mercury emissions characteristics are based in large
part on available limestone and it is unreasonable and arbitrary not to
establish subcategories where technical differences between different
types or classes of cement kilns result in different emissions
characteristics.

Commenters 2844, 2845, 2863, 2900, and 2917 stated that EPA should
provide a new subcategory for kilns that have a high concentration of
mercury or other HAP in the limestone quarries servicing those kilns. As
recognized in EPA’s proposal, these kilns likely would be unable to
meet the proposed standards even after installing state-of-the-art
controls. While there are a limited number of facilities in EPA’s data
set with this issue, commenter 2845 stated that there may be other
similarly situated kilns throughout the United States, not reflected in
the data set. Commenters 2844, 2845, and 2900 proposed a subcategory
that results in a high MACT floor emission rate and then proposed that
EPA use beyond-the-floor consideration to apply state-of-the-art
controls on those facilities. For sources in the proposed subcategory
for units with high mercury concentrations in the limestone, Commenter
2845 proposed a 75 percent removal standard as the beyond-the-floor
limit, which reflects the best percent removal of mercury that may be
obtained through technology at those particular sources.

Commenters 2298, 2483, 2502, 2503, 2506, 2507, 2651, 2653, 2654, 2660,
2661, 2667, 2670, 2671, 2672, 2676, 2811, 2812, 2813, 2830, 2832, 2833,
2840, 2841, 2845, 2863, 2874, 2886, 2900, 2903, 2904, 2973, 2974, 2979,
2982, 2983, 2996, 2997, 3012, 3013, 3014, 3016, 3031, 3036, 3039, 3046,
3052, 3058, 3066, 3071, 3075, 3079, 3081, 3083, 3085, 3090, 3091, 3093,
3095, 3096, 3097, 3100, 3102, 3111, 3114, 3115, 3116, 3119, 3120, 3122,
3123, 3125, 3126, 3127, 3129, 3134, 3135, 3141, 3144, 3147, 3153, 3155,
3157, 3158, 3159, 3166, 3174, 3175, 3177, 3178, 3180, 3204, 3208, 3209,
3210, 3211, 3213, 3217, 3218, and 3219 stated that EPA should provide a
new subcategory for kilns that have a high concentration of mercury in
the limestone quarries servicing those kilns. As specifically recognized
in EPA’s proposal, these kilns likely would be unable to meet the
proposed standards even after installing state-of-the-art controls. 

Commenter 2863 recommended the following actions on subcategorization
for purposes of the mercury standard:

No use of fly ash as an additive in the raw feed;

Implementation of dust shuttling to the maximum extent possible without
impacting product quality;

Implementation of a Mercury Minimization Plan to minimize mercury in the
raw materials imported from off-site; and

Implementation of a Control Enhancement Action Plan.

Response: The EPA reasons for declining to establish a subcategory for
high mercury emitting kilns with high mercury levels in their quarried
limestone are set out in the preamble to the final rule. We also
examined other actions on developing subcategories suggested by
Commenter 2863 and concluded that although these actions might help to
reduce mercury emissions at a given kiln, they do not provide a basis
for subcategorization, nor could they be used to represent MACT. We have
concluded that subcategorization for mercury emissions, which was the
subject of the comments provided above, is not necessary or warranted
because the mercury emission limits can be met with the application of
mercury emission controls and/or other practices, such as those that
reduce the amount of mercury in inputs to the kiln. 

We also disagree with a commenter’s statement that a 75 percent
removal standard should be the beyond-the-floor limit because ACI
technology for mercury will achieve higher removal levels than 75
percent, and for some kilns with high mercury inputs, a removal of only
75 percent would not even represent the MACT floor. We also disagree
with the development of “interim” standards and note that the CAA
provides the allowable schedule for compliance for existing sources in
section 112(i)(3) (with some discretion by EPA and the permitting
authority when extra time is needed to install emission controls).

Pounds of mercury per million tons of feed includes fuel in denominator

Comment: One commenter (2849) stated that EPA incorrectly calculated the
conversion from pounds per million tons of feed to pounds per million
tons of clinker when it developed its mercury floor levels. When EPA
computed its mercury balance for each of the 89 kilns it used in the
mercury MACT floor calculation, it used total mercury inputs from raw
materials and fuels, then subtracted out the mercury output from kiln
dust not recirculated to the kiln, to arrive at a calculated mercury
emission (in pounds). It then divided these pounds of mercury emissions
by the total quantity of raw material plus fuels inputs to arrive at
emission units that were entitled pounds of mercury per million tons of
feed. In fact, a more correct title for these emission units should have
been pounds of mercury per million tons of feed plus fuel.

Response: The commenter’s characterization of EPA’s approach to
calculating mercury emissions per ton of clinker is partially correct
with the exception that we did not, as the commenter states, include the
quantity of fuel along with the quantity of feed materials in
calculating mercury emissions in pounds per million tons of clinker. In
response to the comment, we reviewed the approach used as well as the
data to confirm that fuel materials were correctly categorized and not
included in the sum of raw materials for each kiln. 

To calculate mercury emissions per ton of clinker, all materials
reported in the responses to EPA’s May 8, 2007 information collection
requests (ICR) were entered into an Access database. All materials (feed
and fuel) were classified in a raw materials table and through database
queries into one of the following categories:

Limestone

Shale

Clay

Fly ash/bottom ash

Sand	Iron

Other raw materials

Coal/coke/fossil fuel

Tires/TDF

Wasted CKD



To be considered as a feed material, the reported raw materials for each
kiln could not also be in a fuel or CKD category. Mercury emissions per
million tons of feed were then converted to emissions per million tons
of clinker by multiplying by the conversion factor of 1.65 tons of feed
per ton of clinker. Although EPA’s approach for calculating mercury
emissions per ton of clinker was designed to exclude fuel in the
denominator, it is possible that a minor, uncommon fuel material used at
a kiln was incorrectly classified as a feed material. However, coal,
coke, other fossil fuels and tire derived fuel constituted nearly all of
the fuel for most kilns and a misclassification of a minor fuel material
as feed, or vice versa, would not have any significant effect on a
kiln’s final emission rate.

Data suggest a 12-month average standard

Comment: One commenter (2864) stated that although EPA collected 30 days
of mercury input data for 89 kilns, the Agency also collected annual
data on the mass inputs and the amount of material collected in the PM
control device (or alkali PM control device) that is wasted rather than
returned to the kiln. The annual average data were used to calculate the
mass-balance mercury emissions, which were then used to establish the
mercury floor. The commenter stated that this is an annual mercury
emission rate and EPA should establish a rolling 12-month average
mercury MACT standard.

Response: As stated previously, EPA’s legal obligation in establishing
MACT floors is to reasonably estimate the average performance achieved
by the best performing 12 percent of best performing sources, or for new
sources, the performance of the best controlled similar source. The
statute does not speak to the issue of permissible averaging times in
constructing standards, so it may be that in factually appropriate
cases, a choice of averaging time to account for best performing
sources’ variability could be legitimate. However, automatic (i.e., a
priori) determinations to adopt long-term averages for any type of data
set regardless of circumstances of sources’ performance and robustness
of the data set, would go too far. We chose a 30-day averaging period
for the mercury emission limit. The use of a 30 day average (as opposed
to hourly or daily averages) tends to reduce variability, and also best
reflects the nature of the data from which the floor was derived and
assures that several operating cycles of raw mill on and off are
included in each average. The mercury limit is based on a 30-day
averaging period to be consistent with the averaging periods for other
regulated kiln pollutants, and adequately allows additional operating
flexibility to meet the limit. Finally, EPA is setting the compliance
period using 30-day averaging periods for enforcement reasons in that
EPA can periodically determine compliance with the emission limits and
avoid long periods in which sources may be emitting at unacceptably high
levels.

Provide alternative removal efficiency standard

Comment: Several commenters (2845 and 2874) stated that EPA should set
an industry-wide MACT floor set on the basis of percent removal, or
system removal efficiency (SRE) to be used as an alternative to the
currently proposed mass-rate based emission limit.

Response: For the reasons stated in the preamble to the final rule (and
comment responses therein), EPA is not expressing standards, or
determining MACT floor, on a per cent reduction basis. No commenter
provided a convincing response to how such an approach, in this rule,
properly accounts for the role of HAP in kiln inputs. Nor did commenters
address how such a standard could even be constructed for mercury, where
there is essentially no system removal because kilns are uncontrolled
and mercury volatilizes.

2.5.3	Beyond the Floor Determinations

Comment: Several commenters (2845 and 2863) stated that EPA correctly
explains that if it did not subcategorize based on mercury concentration
in the limestone, it would not be appropriate to establish a
beyond-the-floor standard for mercury for either new or existing
sources. Doing so would not be cost-effective. As a result of the much
higher mercury emissions from kilns in a subcategory of sources with
high mercury concentration in the limestone; however, a beyond-the-floor
standard may be appropriate. The commenter states that such a
beyond-the-floor standard should be expressed in terms of system removal
efficiency (i.e., percentage reduction of mercury entering the kiln in
the raw materials and fuel). The commenter further states that a
75 percent system removal efficiency for mercury would represent a
stringent beyond the floor standard for both new and existing sources
that nevertheless could be achieved through the use of ACI, the best
mercury control available.

Commenter 2898 recommended that EPA address fuel-switching as a beyond
the floor option. EPA should explain what site-specific circumstances
might influence the substitution of low-mercury materials and fuels for
some facilities, and why the practice could not be applied uniformly on
a national basis. The commenter suggests that EPA fully explore
non-technology emission reduction options in the beyond the floor
discussion, including fuel-switching.

Response: EPA agrees that a beyond-the-floor standard is not appropriate
for mercury given the decision not to subcategorize. See preamble for
response to comment to consider fuel switching.

2.5.4	Compliance Assurance

Form of the Standard

Comment: Commenter 2879 stated that CEMS readings and clinker production
calculations can be unreliable. Rather than assuming that kiln operators
can accurately determine the mercury emissions per million tons of
clinker on a 30 day average basis, EPA should determine the mercury CEMS
concentration that corresponds to the proposed 43 lb/MM ton clinker, or
the appropriate value representing the average of the best performing
12%.

Commenter 2889 suggested that EPA conduct error analysis to evaluate the
ability of a source to comply with the mercury monthly block average
emission limit. To comply with the mercury limit in the proposed
amendments, a cement kiln operator will have to measure (1) the mercury
concentration in the gas stream, (2) the gas flow rate in the stack, and
(3) the clinker production rate. These three sources of data will be
combined to determine compliance with the monthly block average limit.
Considering the low emission rate in the proposed amendments, EPA should
evaluate the inherent measurement error associated with these quite
different measurements. Cement kilns could appear to be out of
compliance simply because of slight unavoidable variability of the
accuracy of the three measurements. This problem could be reduced
slightly by simply applying a mercury concentration limit in the gas
stream without any reference to the clinker production rate.

Response: Issues related to clinker measurement are addressed in the
preamble to the final rule.  Other facilities in the utility industry
and the cement industry are currently measuring mercury concentrations
at and below these levels using a CEMS.  

Data supports mass balance approach

Comment: Several commenters (2844, 2845, 2855, 2859, and 2901) stated
that EPA has provided no support for its assumption that the
mass-balance data used to establish the mercury MACT floors can be
correlated to continuous emissions monitoring data to demonstrate
compliance. The commenters stated that there is no basis for the
Agency’s assumption that conservation of mass data calculations made
for the cement industry correlate with the emission measurements that
may be made in the future with mercury CEMS. 

Response: The mercury emission standard is based on a MACT floor of
pounds of mercury emissions per ton of clinker, adjusted for
variability. We believe that the analysis described in the proposal
notice and in the docket does support a correlation between the MACT
floors and continuous monitoring data collected on a 30-day rolling
average. 

Comment: Commenter 2844 stated that CEMS will create compliance issues
that are not related to emissions compliance; CEMS represent a very high
added cost beyond any value; and EPA should adopt mercury limits that
allow the use of mass balance calculations.

Commenters 2855 and 2874 stated that EPA should reconsider a sampling
and analysis program involving kiln inputs as an acceptable option to
CEMS for demonstrating compliance with a mercury standard. Commenter
2855 stated that this approach should be allowed because EPA relied on
this approach to establish the mercury standards, and for low level
mercury emitting facilities that can comply with the standards without
add-on controls, the solid material sampling programs could be focused
to optimize accuracy and minimize costs.

Commenter 2855 stated that facilities that control mercury emissions
from cement kilns or kiln/in-line raw mills should be allowed to use
solid or liquid material sampling and analysis to demonstrate mercury
removal as an optional monitoring approach. Facilities that waste cement
kiln dust (CKD) or implement dust management programs to minimize
mercury emissions should be allowed to implement solid material sampling
and analysis procedures to quantify the mass of mercury removed. The
optimal sampling location and approach must be determined on a
site-specific basis. Similarly, sources with add-on controls should be
allowed to implement material sampling and analysis procedures to
quantify the mass of mercury removed and subtract the documented mercury
removed to determine compliance with mercury standards.

Commenter 2888 recommended that since the orientation of monovent kilns
does not provide a means for a stack flow meter (there is no stack) or a
CEM, these kilns should perform a mass balance of the material flows
into and out of the process, for purposes of demonstrating compliance
with the NESHAP mercury standard. 

Response: Mercury is a unique pollutant because the amount of mercury
input to the system equals the amount of mercury output. In the absence
of control devices, input equals output, and a material balance approach
is feasible. However, the cost of collecting such data on a daily basis
is comparable to the cost of a mercury monitoring system. In addition,
in order to meet the mercury standards, many sources will install
control devices. In order to conduct a material balance on this now more
complicated system, we would have to determine a reliable control device
efficiency value and require continuous monitoring of control device
parameters. Given these costs and uncertainties, we conclude that
continuous monitoring at the stack is the best method to determine
compliance. As always, however, sources are allowed to request
source-specific approval of alternative compliance methods under the
part 63.7 General Provisions.

CEMS not demonstrated in cement manufacturing

Comment: Many commenters (2830, 2845, 2846, 2855, 2858, 2859, 2879,
2887, and 2890) stated that CEMS are not a proven technology and should
not be required to determine compliance. 

Commenters 2588, 2844, 2845, 2846, 2858, and 2890 stated that EPA has no
data showing that mercury CEMS are feasible on cement kilns and that
emissions from cement kilns will likely be outside of the range of the
current CEMS technology. The commenters stated that EPA must evaluate
mercury CEMS through long-term field trials at cement plants in
accordance with the proposed performance specifications and quality
assurance procedures before imposing this regulatory requirement. The
commenters proposed a mass-balance approach for monitoring, which is
accurate and was used by EPA in setting the mercury standard.

Commenter 2855 stated that mercury sorbent trap monitoring systems have
not been evaluated through long term field trials at cement plants in
the United States (U.S.) in accordance with the proposed performance
specifications and quality assurance procedures, so the reliability and
performance of these measurement systems and the adequacy of the
technical specifications cannot be determined. 

Commenters 2855 and 2900 disagreed with EPA’s interpretation of the
operating experience with mercury CEMS in Germany. The commenters stated
that mercury CEMS are inaccurate and difficult to maintain. Further,
mercury CEMS operating in Germany are subject to monitoring regulations
that are different than the U.S. regulations and are used in a different
regulatory context than that proposed by EPA. The monitors used in
Germany, or those available from other European or Asian manufacturers
were not able to demonstrate acceptable performance in the Electric
Power Research Institute (EPRI) Trimble County Mercury CEMS study. 

Commenter 2855 stated that there is no legitimate technical basis on
which to establish detailed performance specifications or quality
assurance (QA) requirements for these CEMS. There is no legitimate
technical basis to conclude that these CEMS could meet such requirements
over any extended period when installed and operated at a cement plant.
The commenter recommended that EPA evaluate the performance of mercury
CEMS at cement kiln systems and acquire the information necessary to
serve as the basis for technical specifications and requirements. After
such information is available and analyzed, EPA should re-propose
appropriate and demonstrated performance specifications and quality
assurance procedures for mercury CEMS to monitor kiln and kiln/in-line
raw mill mercury emissions. 

Commenter 2855 disagreed with EPA’s interpretation that mercury CEMS
can be applied to the cement industry based on successful use on utility
boilers. The commenter evaluated the following issues:

Number of Installations in the Utility Industry – there are 35 – 40
continuous mercury monitors (CMMs) installed and certified to date (not
yet with a NIST traceable calibration source).

NIST Certification – In mercury CEM certification requirements
outlined in Performance Specification (PS)-12A, it states that all
calibration and span gases must be NIST certified. The draft protocols
were just released by NIST in July 2009. The major vendors of mercury
CEM are just now advertising NIST-certified calibration sources.
Therefore, none of the mercury CEMs that have been previously installed
are certified. NIST does not currently directly certify oxidized mercury
calibrations. The Interim EPA Traceability Protocols now in place
provide for certification of evaporative generators by certification of
the individual components of the calibrator. Therefore, the language
used in Section 7.0 that refers to a NIST trace oxidized mercury
calibrator needs to be clarified or changed.

Difficulties Encountered in the Utility Industry – The commenter gave
examples of power plants’ difficulties with mercury CEMS.

Installation on Wet Stacks – Installing a mercury CEMS on a wet stack
can result in problems: plugging, corrosion, and buildup of solids.
Although wet scrubbers are not currently common in the cement industry,
under the proposed rule, they may be required to a greater extent, and
many of these same problems with mercury CEMs potentially could occur
for the cement industry as well.

Data Output Requirements – There is no need for dry basis measurements
under the proposed rule and the language in either Subpart LLL should be
included to provide an exemption from this requirement for cement plants
or PS-12A should be revised. This language needs to be clarified by EPA.

Cost – The commenter provided information about CEMS costs, estimating
that if mercury CEMS were installed on all non-waste-burning U.S. cement
facilities, the total capital costs would be approximately $45 million,
with annual operating costs being about $25 million. 

Commenter 2901 stated that CEMS should not be used as a compliance
method for cement plants for the following reasons:

EPA reported in 1997 on an experiment where CEMS were installed on a
cement kiln burning hazardous waste. The Agency found substantial
problems regarding mercury CEMS measurement accuracy and precision,
deciding not to require Mercury CEMS at cement plants. The commenter
stated that the primary issue is whether there is a NIST traceable
standard that can be used to calibrate the unit. Because compliance is
based on the production rate and using a 30-day average, it is difficult
to know what range to calibrate these units.

The reliability of CEMS on cement kiln stacks has not been demonstrated
in the U.S., where standards and requirements are different.
Demonstrations in the U.S. at coal-fired power plants have different
conditions than those at cement kilns. 

There is no legal imperative for EPA to require CEMS. Under the CAA,
EPA’s monitoring requirements must provide a reasonable assurance of
compliance with emission standards Sierra Club v. EPA, 353 F.3d 976,
990-991 (DC Cir 2004) (Copper Smelters) citing Natural Res. Def.
Council, Inc. v. EPA, 194 F.3d 130, 136 (D.C. Cir 1999). 

Response: See response in preamble.

Comment: Commenter 2890 stated that CEMS represent a requirement that is
beyond the floor.

Response: Assuming the argument that measuring methodology can be
considered to be part of a standard, the use of CEMs to measure the
mercury standard does not make the standard more stringent. The mercury
standard was developed on the basis of 30-day data samples, and is
measured over the same duration. The standard was also developed on a
mass balance basis whereby essentially all input mercury was considered
to be emitted (except for the few kilns equipped with wet scrubbers).
This is essentially a worst case assumption (attributing no removal to
baghouses, for example). Given that the standard is measured over the
identical duration as was the underlying data, and that actual emissions
are measured so that kilns are no longer presumed to emit all inputted
mercury, we repeat that use of a CEM does not make the standard more
stringent and thereby constitute a beyond-the-floor standard.

Comment: Several commenters (2155, 2867, 2891, 2894, and 2898) supported
EPA’s decision to require mercury compliance monitoring using CEMS. 

Commenter 2867 reports that their own inquiry to the German Federal
Environment Agency, (Umweltbundesamt [UBA]) regarding the mercury-CEMS
features of the proposed cement MACT rule, which found the mercury CEMS
QA suitable to ensure quality measurements. According to the commenter,
UBA has the most experience in the world among environmental agencies in
the field of mercury-CEMS including the cement industry. The commenter
also cited the experience of the Ash Grove Durkee, Oregon facility. The
commenter noted that the results for this facility presented at the 2008
Air and Waste Management Association (AWMA) Conference in Oregon showed
that the mercury-CEMS registered the same average stack mercury
emissions reading (0.027 g/s) when the raw mill was on as the average
value of the Methods 324 and OH. The commenter stated that this
indicates that mercury-CEMS are feasible, precise and accurate compared
with the conventional in-stack sampling methods at low and at elevated
emission rates.

Commenter 2894 stated that as manufacturer of mercury monitoring
instrumentation, they are in the process of installing mercury CEMS on
U.S.-based cement kilns. There are no technical obstacles to the use of
mercury CEMS in this type of environment.

Response: We agree with these commenters that mercury CEMS should be
used for compliance monitoring.

Wet vs. Dry Concentration Format for CEMS

Comment: One commenter (2894) stated that in order to maximize accuracy
in mercury mass emission calculations, mercury CEMs should report
mercury concentrations as μg/m3 (at 20 ºC, 760 mm mercury pressure) on
a wet basis. No correction for O2, carbon dioxide (CO2) or any other gas
should be performed. Stack flow rate should also be measured on a wet
basis, in m3/min (at 20 ºC, 760 mm mercury) at the same point in the
process, with no correction for any diluent gas.

Commenter 2894 recommended calculating mercury emissions using a wet
concentration and wet flow rate in equation 5 (Ref Eq. 5, page 21187,
col 2) rather than the currently required dry concentration and dry flow
rate. A dry mercury concentration is not required to calculate mass
emissions and would actually reduce the accuracy. In the event that a
dry concentration is absolutely required, the rule should make it clear
that the same (BWS) factor must be used to correct both the mercury
concentration and the flow rate. This factor should be a calculated
value; a moisture analyzer is not required since the two occurrences of
the BWS factor cancel arithmetically. 

Commenter 2855 said that there is no need for dry basis measurements
under the proposed rule, and the language in either Subpart LLL should
be included to provide an exemption from this requirement for cement
plants or PS-12A should be revised. This language needs to be clarified
by EPA.

Response: We agree with the commenters that dry concentrations and flow
measurements are not required to calculate mass emissions. It is
important that the measurements are on the same basis, wet or dry, for
the calculations. The equations have been modified and clarifying text
added. We have revised to standard to state that dry measurements are
not required.

Comment: One commenter (2837) recommended that the period for
calibration drift check should be defined by system operator, rather
than defining this period on the basis of any one of the available
technologies. The commenter also recommended that EPA define how the
operator is to demonstrate the validity of the operator-defined
calibration check period.

Response: Requirements for calibration checks, including the frequency
(once daily), are specified in Appendix F, Procedure 5. The procedures
to be used are based on manufacturers’ recommendations.

Establish standard as CEM concentration

Comment: One commenter (2879) recommended that EPA should measure
mercury emissions using a CEM, reconciling CO2 emissions with mass
balance emissions and establishing the standard as a CEM concentration.

Response: Under the final testing and monitoring requirements, CEMS or
sorbent trap monitors are required to continuously measure mercury
emissions. See the preamble discussion under final testing and
monitoring requirements for more information on EPA’s decision.

Emission limits are typically normalized to some type of production or
raw material input value because this allows comparison (and ultimately
the ability to set a single standard) for different sized facilities. As
we noted at the NSPS proposal, the current NSPS and limits (and NESHAP
limits before today’s amendments) for PM are expressed on a pound of
PM per ton (lb/ton) of dry feed input format. See 73 FR at 34075-76. In
this final NESHAP (and NSPS) we are adopting a new normalizing parameter
of lb/ton of clinker – i.e., normalizing based on kiln output rather
than input for both PM and mercury.

EPA therefore proposed that all of the NSPS (for PM, NOx, and SO2) be
normalized by ton of clinker produced, and later proposed the same
parameter for the two standards in the NESHAP which are normalized,
mercury and PM. 73 FR at 34076; 74 FR at 21140.

In the final NESHAP (and NSPS) we are therefore adopting a new
normalizing parameter of lb/ton of clinker – i.e. normalizing based on
kiln output rather than input – for mercury and PM in the NESHAP, and
for PM, NOx, and SO2 in the NSPS.

See the preamble discussion under format of the normalized standards in
the NESHAP and the NSPS, for more information on EPA’s decision.

All mercury phases should be monitored

Comment: One commenter (2638) stated that all phases of mercury should
be monitored and PS 12A and Procedure 5 should be rewritten to include
total mercury. The commenter stated it is contrary to the intent of the
CAA to set emission limits for only vapor phase mercury and requested
that the EPA change its emission limit from one based on vapor phase
mercury to one based on total mercury that includes both PM and vapor
phases.

One commenter (2638) stated that all physical phases of all HAP metals,
including mercury, need to be measured during SSM events. For example,
if mercury is captured on powdered activated carbon, which in turn is
captured on a fabric filter, the percentage of mercury in the PM
fraction downstream of the fabric filter during malfunctions could and
would be expected to be significant.

Commenter 2638 stated that total mercury monitoring can now be
undertaken with proven methods, and will be consistent with Congress’
desire for the EPA to require enhanced monitoring where it is
appropriate to assure continual compliance with emission limits. The
commenter provided several reports describing alternative measurement
methods.

Response: The mercury standard is based on total mercury data in the
form of mercury inputs to the source. As part of this reconsideration
tests were performed at five facilities with wet scrubbers. Testing was
performed at the scrubber inlet and outlets. We believe tests at a wet
scrubber outlet would represent a situation where particulate mercury
would most likely be present due to the lower temperatures involved
(below 200OF). On average the amounts of particulate mercury measured at
both locations was below 1 percent. Given this fact, we believe that
measurement of total mercury would add unnecessary expense and
complexity to mercury monitoring. However, sources are able to request
approval for source-specific monitoring requirements from EPA under
section 63.7 (f) of the General Provisions. 	

Measurement bias not considered in achievability of standard

Comment: One commenter (2855) stated that EPA did not consider the
allowed measurement bias for mercury CEMS or sorbent tube monitoring
programs on the achievability of the emission standards. 

Response: We disagree that either the mercury CEMS or the mercury test
Method 30B include an inherent measurement bias. All measurement data
are subject to some degree of variability which would include
measurement error (e.g., interference) and imprecision (e.g.,
measurement repeatability). The procedures in Performance Specification
12A promulgated with this rule account for both measurement error and
imprecision in calculating the relative accuracy of the CEMS. Secondly,
we believe that any inherent variability of data due to measurement
imprecision is reasonably addressed in calculating the floor and
emissions limits. The data used to establish the limits represent
emissions tests from multiple facilities and we subjected the data to
statistical procedures to account for variance. 

Contradiction in performance test duration

Comment: One commenter (2855) stated that §63.1349(b)(5)(i) specifies a
mercury performance test duration of a calendar month, which contradicts
the averaging period of the proposed emission standard of a 30-day
rolling average. 

Response: We do not consider this a contradiction because both
monitoring periods have different purposes. The performance test for a
calendar month is to ensure that the control device and the CEMS are
both working correctly. The 30-day rolling average is the compliance
demonstration period. We are proposing additional clarifications such as
a definition of “kiln operating day” to further address questions
regarding the compliance demonstration period. In general, if the kiln
operates during a 24-hour period, the CEMS must operate too.

Initial performance demonstration recommendations

Comment: One commenter (2855) stated that the initial performance
demonstration should be based on a properly calibrated and compliant
CEMS or Sorbent Trap System (if such systems become available and are
demonstrated to perform acceptably), and the initial performance test
result must be based on the calculated average lb mercury/million ton
clinker produced for 30 kiln operating days. A definition for kiln
operating day must be established and must reflect a specified number of
hours of kiln or raw material dryer operation at normal production
conditions and a minimum number of hours of quality assured monitoring
data. Days with fewer operating hours or insufficient monitoring data
are not included in the 30 day average. 

Response: We agree that a definition of “kiln operating day” is
needed, and have added it to the final rule. We have not specified a
minimum number of hours of kiln or raw material dryer operation; a kiln
operating day is any day the kiln is operating, and all valid data (for
hours when the process is operating) for all operating days are included
in the 30 day average.	

Request daily instead of hourly mercury concentration data

Comment: One commenter (2855) stated §63.1349(b)(5)(i) requires
collection of hourly mercury concentration data which will not be
available for sorbent trap monitoring systems and is not required by
PS-12B. The commenter suggested that daily mercury emission rate
averages would be adequate.

Response: The final rule has been clarified to address this situation.
For example, if sorbent traps are collected every 7 days, each daily
average would be assumed to equal the 7-day average.	

Testing with raw mill on and raw mill off

Comment: One commenter (2855) stated that EPA should clarify that
separate performance tests are not required for mill on and mill off
conditions in the discussion in §63.1349(b)(5)(i) about demonstrating
mercury compliance.

Response: We have clarified this language in the final rule to state
that the raw mill off testing provision for mercury does not apply to a
kiln equipped with and demonstrating compliance using a mercury CEMS.

Calculation of monthly mercury emissions

Comment: One commenter (2894) disagreed with the calculations to
determine monthly mercury mass emissions (Ref Eq. 5, page 21187, col 2).
The commenter stated that this calculation could result in large biases.
The commenter suggested an alternative calculation that multiplies the
hourly average mercury concentrations by the hourly average flow rates
and totals these values for the month to determine mercury mass
emissions. 

Commenter 2894 stated that total monthly mercury mass emission must be
divided by the clinker production rate during the same calendar month in
order to determine compliance. Mercury monitoring systems of any type
are capable of determining only mercury concentrations. It appears that
even the most conscientious plant operators will thus not know until
well after the month has elapsed whether they were in compliance for
that month.

Commenter 2894 stated that in cases where plant operators go through the
effort to install systems capable of continuously measuring mercury
concentrations and stack flow rates, it seems only fair and reasonable
that the EPA devise a metric that allows plants to determine on an
ongoing basis whether they are complying with the regulation.

Response: The commenter misinterpreted compliance equation 5 (74 FR at
21187) to require a determination of monthly mercury mass emissions and
a monthly average flow rate. As recommended by the commenter, section
63.1349 (b)(5)(i) of the final rule requires the kiln owner or operator
to measure and record hourly mercury concentration data and stack gas
volumetric flow rates. These measurements are to be used to calculate
daily mercury emission rates. Owners or operators are also required to
operate a system to measure the hourly clinker production rates, either
directly or indirectly through measuring kiln feed rates and applying a
factor for converting feed rates to clinker production rates (section
63.1350(o)). They must record both the daily kiln feed rates and clinker
production rates. Using the daily mercury concentration data, daily
stack gas flow rates and daily production rates, the owner or operator
can track their performance on an ongoing daily basis rather than once a
month.

2.5.4.1	Performance Specification 12A, 12B, and Procedure 5

PS-12A

Comment: One commenter (2638) stated that PS-12A should not be
generalized to unspecified source categories other than portland cement.
Because emission standards and measurement methods are not written
independent of each other, defining the measurement method in effect
defines the emission limit/standard for other, unspecified sources. By
generalizing this measurement technology to vapor phase mercury, the EPA
appears to want to generalize the emission standards for other, as yet
to be specified sources to vapor phase mercury. By doing this, the EPA
greatly inhibits the development of new methods that might measure total
mercury.

Response: EPA will only apply PS-12A to other source categories through
specific notice and comment rulemaking, and only when it is appropriate
to do so. We are not opposed to the concept that PS-12A could also be
applied to monitoring systems that also incorporate non-gaseous mercury
in the mercury measurement for this or any other appropriate source
category (see §63.7 (f)). However, EPA does not have any evidence that
any appreciable non-gaseous mercury is present in the emissions from
this particular source category.

Comment: One commenter (2638) stated that PS-12A and ongoing quality
assurance should be re-written to replace contemporary
technology-specific requirements with generalized measurement-objective
driven requirements. The commenter provided comments about testing
linearity, calibration drift, accuracy, system integrity, reagents and
standards, and span value. 

Response: We disagree that PS-12A is technology-specific. We recognize
that some performance procedures of PS-12A may be more difficult for
some instruments and that some modifications to performance procedures
may be necessary. EPA has a program in place to deal with this situation
through the alternative test method request and approval/disapproval
provisions of §63.7(f).

Comment: One commenter (2638) suggested specific changes to PS-12A and
Procedure 5:

Subsection 6.1.1 Data Recorder Scale: Should be replaced to simply
require the operator to include the range defined by the zero and span
values. The EPA proposed requirement seems to be based on old technology
and defining this range with a specific numeric requirement (two times
the standard) serves no functional purpose and unnecessarily restricts
development. 

Subsection 6.1.2 CE determination: Change zero to low and upscale to
high.

Section 8.0 PS Tests: The EPA should re-write this section around
clearly defined analytical objectives of establishing the CEMS initial
accuracy, precision, linearity, stability and system integrity based on
some of the same principals used in EPA Method 301.

Subsection 8.1.3 Extraction point: This should be re-written to allow
for possible in situ monitoring systems that do not require sample
extraction.

Section 3.0 Definitions: A new subsection should be added defining zero
or the bottom of the range for valid results, and instructions as to
what value to use when a result falls below this defined concentration.
The commenter recommended that zero be replaced with detection limit as
defined by the EPA in Title 40, Appendix B Part 136 as the minimum
concentration of a substance that can be measured and reported with 99%
confidence that the analyte concentration is greater than zero. If the
concentration is less than this concentration, the reported value should
be replaced with a concentration equal to the detection limit.

Response: Subsection 6.1.1 of PS-12A has been revised in response to
comments received. Subsection 6.1.2 PS-12A has been revised to be
consistent with Section 7.1 of PS-12A on reference gases. The relative
accuracy test audit or RATA of PS-12A is designed to assess the overall
performance of the instrumental measurement system, including the
representativeness of the sampling (i.e., probe location). PS-12A is not
intended to address method detection limits.

Comment: One commenter (2837) provided analytical detail related to
initial and on-going quality assurance:

National Institute of Standards and Technology (NIST) certified
traceable sources of elemental and oxidized mercury vapors are not
generally needed to demonstrate CEMS compliance with performance based
quality parameters such as linearity, accuracy, precision, stability or
system integrity for either initial CEMS certification or on-going
quality assurance and should be removed from both PS-12A and Procedure
5.

The calibration check requirement currently in PS-12A and Procedure 5
allows the same mercury generator used to calibrate the CEMS to be used
to check the calibration. Thus, they are not independent and a drift in
the generator output would cause a drift in the CEMS calibration.
Spiking mercury into clean, conditioned, interference-free air does not
provide traceability to NIST standards for assurance of calibration
accuracy.

Spiking mercury into clean, conditioned, interference-free air does not
provide quality assurance of system integrity, analyte transportability,
sensor response, etc.

The current calibration drift check period is technology specific and
should be defined by the system operator with performance-based
protocols to demonstrate the applicability of the selected period.

Response: The NIST traceability and daily calibration requirements of
PS-12A are consistent with the Agency’s other compliance monitoring
requirements in 40 CFR Part 60 which require NIST traceable
“protocol” gases for CEMS quality assurance/quality control. This
approach has proven to be very effective for assessing overall system
performance (e.g., sample transport and analyzer accuracy) through
introduction of the reference gases at the CEMS probe. Dynamic spiking
with elemental and oxidized mercury gases through the probe would be
another way to accomplish this performance check and would certainly be
encouraged as a potential alternative test method procedural request
under §63.7(f). EPA believes that the requirements for NIST
traceability of calibration gas generators are sufficient to address
drift of the generators.

Comment: One commenter (2855) stated that PS-12A requires a system
integrity check using an oxidized mercury NIST calibration source, which
is designed to ensure the convertor portion of the mercury CEM is
converting all of the mercury to elemental mercury. The commenter stated
that there is a discrepancy between the oxidized and elemental mercury
calibrators of about 7% in both Tekran and ThermoScientific instruments
and that the cause of this discrepancy is still unknown.

Response: The commenter is correct regarding the discrepancy between the
oxidized and elemental mercury calibrators of about 7 percent for the
Tekran instrument; however, the Interim EPA Traceability Protocol for
Qualification and Certification of Elemental Mercury Gas Generators and
the Interim EPA Traceability Protocol for Qualification and
Certification of Oxidized Mercury Gas Generators (July 01, 2009)
recognizes this issue and provides a straightforward remedy. The
protocol specifies that until the time that NIST traceability of
evaporative oxidized mercury standards generators can be evaluated by
direct measurement, the generator manufacturers may use a model-specific
factor in their software to adjust the calibrated output of their HgCl2
generator to correspond with their NIST certified elemental generators,
thus negating the discrepancy. The generator manufacturer must then
provide the end user with the exact value of this correction factor and
instructions for its application, and have derivation data available to
the end user.

Comment: One commenter (2855) stated that EPA should not revert to 2004
CEMS performance specifications and should not require any
specifications that are more restrictive than the 40 CFR 75, Appendix A
and B requirements promulgated in January 2008.

Response: After consideration of the potential variability between the
learning curves of CEMS operators that new users of Hg CEMS may
experience, we have revised the relative accuracy specification in
PS-12A from 10 percent to 20 percent. We believe, however, that the
implication of this change is minor and short-lived because most
facilities that install mercury CEMS will find that they readily achieve
relative accuracies far better than 20 percent.

Comment: One commenter (2855) stated that after field validation, and
promulgation of instrumental CEMS and technical specifications and after
NIST traceable calibration standards at the necessary concentrations are
available, EPA should allow at least one year for purchase, installation
and start-up of CEMS and a one-year trial implementation period before
requiring the use of monitoring data to demonstrate compliance with
emission limits. 

One commenter (2144) recommended that EPA consider alternative QA
procedures for mercury CEMS but without delaying final rule
promulgation.

Response: The amendment to the Portland Cement NESHAP will not become
effective until 60 days following promulgation for new sources and
existing sources will have up to 3 years to install controls and
monitoring systems (see Section 63.1350(i)). The National Institute of
Standards and Technology (NIST) has recently completed certification of
a ‘NIST Prime’ elemental mercury gas generator at concentrations of
41, 68, 85, 105, 140, 185, 230, 287, and 353 µg/m3 and mercury gas
generator vendors may now submit elemental mercury gas generators for
certification to serve as ‘Vendor Primes’. Therefore NIST traceable
mercury gas standards can now be made available in concentrations that
exceed the equivalent mass standards for both existing and new kilns by
between one and two orders of magnitude, thus providing the capability
to accurately report excursions well beyond either standard. The rule
effective date should provide ample time for facilities to meet mercury
monitoring system installation and certification requirements.

Comment: One commenter (2855) provided the following specific comments
on PS-12A:

Establishing the span value at twice the concentration corresponding to
the emission limitation is acceptable for cement plants provided that
reasonable provisions are included for certification and quality
assurance of the mercury CEMS when emissions exceeding the span value
are periodically encountered (such as for plants with in-line raw mills
during mill off operation). The commenter provided suggested amendments
to the Subpart LLL monitoring provisions.

The definition of span value should be changed, explicit procedures for
its calculation should be added, and the value should be rounded to
established ranges. The definition could be modified by the addition of
the phrase Span value means the upper limit of intended measurement
range during normal ‘mill on’ operation which may be exceeded during
‘mill off’ operation or other short-term conditions lasting less
than 24 consecutive kiln operating hours. Or the first sentence in the
EPA proposed definition could be eliminated. Alternatively the span
value could be specified in the applicable Subpart (Subpart LLL) where
the specific requirements are more appropriately addressed.

Response: We have modified the definition of span in PS-12A to provide
additional flexibility and have adopted the commenter’s suggestion to
include source category-specific provisions regarding the span in
paragraph 63.1350(n) of Subpart LLL. We agree that the quality of
measurements outside the calibration range should be assessed and could
be further documented by extending the calibration range or other
appropriate means. The commenter provided the following three approaches
that EPA would consider appropriate:

If emissions exceed the span value for more than 5 percent of the kiln
operating time, the CEMS operator could conduct “direct
calibrations” of the Hg analyzer bypassing the dilution module.
Because the Thermo and Tekran Hg CEMS operate on a dilution ratio of 30
or 50 to 1, the calibrated range of the Hg analyzer will be 30 or 40
times the upscale calibration value. Hence, for a system calibrated to
measure 0-10 µg/m3 through the entire measurement system, the analyzer
would be calibrated accurate to between 0 and 300 µg/m3.

The ability of the CEMS to quantitatively transport high concentration
levels of oxidized Hg through the probe, filter, filter cake, sample
lines, and the ability of the CEMS converter to reduce the oxidized Hg
to elemental Hg should be confirmed initially and quarterly by injecting
a high level HgC12 calibration gas standard at the sample probe,
upstream of the PM filter. This high level system integrity test should
be performed at a concentration approximating the maximum hourly
emission concentration during the preceding three months or other
representative high level concentration. The high level system integrity
check is acceptable if: the stable Hg CEMS response is within ± (some
percentage to be determined based on future field evaluations) of the
reference value, and the system returns to normal operating level within
30 minutes of ceasing the calibration gas injection, or the time
integrated Hg CEMS response is equivalent to the total mass of HgC12
injected within ± (some percentage to be determined based on future
field evaluations).

As an alternative to the high level system integrity checks, the CEMS
operator could elect to demonstrate the accuracy of Hg CEMS measurements
exceeding the span level by conducting a three-run relative accuracy
audit (RAA), preferably when mercury levels are outside the calibrated
range of the CEMS . The three-run RAA would be performed using Method
30B as the reference method following the procedures described in
Procedure 5.

Comment: One commenter (2855) said that specific procedures should be
provided in Subpart LLL for the calculation of the emission
concentration equivalent to the 43 lb/million ton clinker (existing
sources) or 14 lb/million ton clinker (new sources).

Response: Facilities can rearrange Equation 3 of Subpart LLL in order to
calculate these quantities using their production rates and volumetric
flow rates. 

Comment: One commenter (2855) said that the calculated span value should
be rounded to the nearest 10 μg/m3 level as has been the practice for
mercury CEMS because of technical limitations associated with the
elemental Mercury gas generators and specific requirements to establish
NIST traceability.

Response: We have modified the definition of span in Section 3.5 PS-12A
to allow for additional flexibility in setting the span and have used
commenter’s recommendations and added a provision in paragraph
63.1350(n) of Subpart LLL to allow for rounding of the span value to the
nearest multiple of 10 µg/m3 of total mercury.

Comment: One commenter (2855) said that the PS-12A requirements at 6.1.1
Data Recorded Scale are misleading, ambiguous and obsolete and should be
replaced by a relevant specification for data recording and an
appropriate definition, (consistent with Span Value) for the high level
value.

Response: Subsection 6.1.1 of PS-12A regarding the data recorder scale
has been revised to respond to comments received. We have simplified the
specification to clarify that the scale must at a minimum include zero
through the span value, but the scale is not limited to that range.

Comment: One commenter (2855) said that the PS-12A requirements at 6.2
Reference Gas Delivery System are ambiguous, unlikely to be achieved
exactly in practice, and without means of verification. The commenter
recommends changing the section from The reference gas delivery system
must be designed so that the flowrate of reference gases introduced to
the CEMS is the same at all three challenge levels specified in Section
7.1, and, at all times exceeds the flow requirements of the CEMS. To The
reference gas delivery system should be designed so that the referenced
gas flow rate is approximately the same (i.e., ± 20%) at each challenge
level.

Response: We have revised the language of section 6.2 of PS-12A to
specify that the reference gas delivery system must be designed so that
the flow rate exceeds the sampling system flow rate of the CEMS and that
the gas is delivered to the CEMS is at atmospheric pressure.

Comment: One commenter (2855) said that PS-12A Section 7.1 requires the
use of NIST certified or NIST Traceable standards for both elemental
mercury and oxidized mercury, which are not currently available, and may
not be available in the future at the specified levels, depending on
EPA’s resolution of the span value.

Response: As noted above, NIST has recently completed certification of a
‘NIST Prime’ elemental mercury gas generator at concentrations of
41, 68, 85, 105, 140, 185, 230, 287, and 353 µg/m3 and mercury gas
generator vendors may now submit elemental mercury gas generators for
certification to serve as ‘Vendor Primes’. Therefore, as previously
stated, NIST traceable mercury gas standards can now be made available
in concentrations that exceed the equivalent mass standards for both
existing and new kilns by between one and two orders of magnitude, thus
providing the capability to accurately report excursions well beyond
either standard.

Comment: One commenter (2855) said that PS-12A Section 8.1.1 includes a
vague requirement to use a sampling location that has been shown to be
free of stratification for mercury or alternatively, sulfur dioxide
(SO2) and nitrogen oxides (NOx) through concentration measurement
traverses for those gases. This requirement should be changed to a
suggestion or guidance.

Response: We have revised language in Section 8.1.1 to specify that the
sampling location must be representative of the stack gas concentration
of mercury and, as the commenter suggested, have modified the
requirement to use a location shown to be free of stratification to be a
recommendation.

Comment: One commenter (2855) said that the PS-12A criteria for the
linearity test procedures are more restrictive than the Part 75,
Appendix A criteria, are inconsistent with the uncertainty of
calibration materials defined by Interim EPA Traceability Protocols, and
should be relaxed.

Response: By specifying the linearity test procedure measurement error
criterion as a function of span, this criterion is generally less
restrictive at the lower gas concentrations than were the prior Part 75,
Appendix A criteria. We have, however, reconsidered the linearity test
procedure specifications and, given the current quality of the available
oxidized mercury gas standards, we have decided to increase the allowed
measurement error for the oxidized mercury standards to + 10 percent to
be consistent with the target criteria in the EPA Interim Traceability
Protocol for Qualification and Certification of Oxidized Mercury Gas
Generators.

Comment: One commenter (2855) said that the PS-12A criteria for the
calibration drift (CD) test procedure: (a) omit the absolute CD
specification included in the Part 75, Appendix A criteria, (b) are
inconsistent with the uncertainty of calibration materials defined by
EPA draft interim traceability protocols, and (c) should be relaxed.

Response: With the exception of the alternative specification for a span
value of 10 µg/m3 or less, the PS-12A criterion for calibration drift
is essentially identical to the Part 75, Appendix A criteria for
calibration error.

Comment: One commenter (2855) said that the proposed revisions to
Subpart LLL and PS-12A fail to specify the conditions under which the
relative accuracy test is performed which will lead to inconsistent
interpretations and implementation requirements. Subpart LLL should
specify that the CEMS relative accuracy test be performed at the most
prevalent operating condition (i.e., mill on for plants with in-line raw
mills).

Response: We have revised paragraph 63.1350(n) of Subpart LLL to specify
operating conditions for relative accuracy testing.

Comment: One commenter (2855) said that sections 9.0 (Quality Control)
and 10.0 (Calibration and Standardization) are both reserved. Quality
Control procedures and Calibration and Standardization procedures must
be considered to determine the overall validity and reasonableness of
the performance specification. Without these integral sections, an
evaluation of the acceptability of this specification cannot be made.
When completed, EPA should re-propose PS-12A for public comment.

Response: These section headings are a result of using a standardized
method format originally specified by EPA’s Environmental Monitoring
Management Council [see http://www.epa.gov/ttn/emc/guidlnd/gd-045.pdf].
The substance of these reserved sections are covered elsewhere in
PS-12A, Procedure 5, and the Interim Traceability Protocols for
Elemental and Oxidized Mercury Gas Generators.

Comment: One commenter (2855) said thatPS-12A as proposed would make the
relative accuracy specification stricter for certain CEMS than is
required by Part 75, Appendix A and B.

Response: After consideration of the potential variability between the
learning curves for CEMS operators that new users of Hg CEMS may
experience, we have revised the relative accuracy specification in
PS-12A from 10 percent to 20 percent. We believe, however, that the
implication of this change is minor and short-lived because most
facilities that install Hg CEMS will find that they readily achieve
relative accuracies far better than 20 percent.

Comment: One commenter (2915) provided the following specific comments
on PS-12A:

Section 3.5–Span value is an obsolete term. Current technology has a
linear measurement range that covers several orders of magnitude.

d be a fixed number in terms of measurement units (μg/m3). Like this
proposed rule revision, many mercury emission limits are not expressed
in terms of μg/m3. The proposed cement MACT is expressed in pounds of
mercury per million tons of clinker. EPA should revise Section 3.5 to
state that the span should be set in accordance with applicable
regulations for the use of the Mercury CEMS.

The commenter suggested a revised definition of span: Span Value means
the intended upper limit of the mercury concentration measurement range.
The span value is specified in the applicable subpart.

Response: We have modified the definition of span in PS-12A to provide
additional flexibility in setting a span value including allowing for
rounding up to the nearest multiple of ten and clarifying that the
applicable subpart may contain more specific provisions regarding the
span value for specific source category. In this case, we have taken
another commenter’s suggestion to define span value for this source
category in paragraph 63.1350(n) of the subpart based on the intended
upper limit of the mercury concentration during “mill on” operation
and approximately equivalent to two times the emission standard.

Comment: One commenter (2915) stated that there are a limited number of
NIST traceable calibration levels available. Since the available
calibration values are all tied to specific span and linearity test
values, defining the span in terms of the emission standard will likely
result in spans that are inconsistent with the NIST traceable
concentrations.

Response: As noted above, NIST has recently completed certification of a
‘NIST Prime’ elemental mercury gas generator at concentrations of
41, 68, 85, 105, 140, 185, 230, 287, and 353 µg/m3 and mercury gas
generator vendors may now submit elemental mercury gas generators for
certification to serve as ‘Vendor Primes’. Therefore, as previously
stated, this capability means that NIST traceable mercury gas standards
can now be made available in concentrations that exceed the equivalent
mass standards for both existing and new kilns by between one and two
orders of magnitude, thus providing the capability to accurately report
values both below and well above either standard. Also, as described in
previous responses, we have revised the definition of span to provide
additional flexibility in setting the span value consistent with the
available NIST traceable concentrations. 

Comment: One commenter (2915) suggested that EPA review the Interim
Traceability Protocols for elemental and oxidized mercury generators so
that there will be regulatory compatibility.

Response: At the current time, we do not anticipate any issues with the
regulatory compatibility of the Interim Traceability Protocols used to
certify the NIST traceability of gases produced by mercury gas
generators; however, we do plan to review and update the protocols in
the future including clarifying that the protocols are not limited to
only providing procedures to qualify and quantify the performance of
mercury gas standards generators below 40 µg/m3.

Comment: One commenter (2915) said that in Section 3.7, EPA proposes to
define calibration drift in terms of a test performed on the CEMS after
a period of operation during which no unscheduled maintenance, repair,
or adjustment took place. The clause during which no unscheduled
maintenance, repair or adjustment took place should be removed. In the
commenter’s experience, unlike SO2 and NOx CEMS, mercury CEMS are
likely to require a significant amount of maintenance, repair and
adjustment. That same clause in PS-2 has been subject to considerable
debate and interpretation over the years and the commenter believes that
including it here will only result in further confusion. EPA also has
not explained why such a limitation is needed in this definition. 

Response: We have removed the term ‘adjustment’ from this phrase in
Section 3.7 of PS-12A as we clearly consider adjustment following the
daily drift check to be an appropriate option (see also Sections 4.2 and
4.3 of Appendix F, Procedure 2 and of proposed Procedure 5). Maintenance
or repair of the Hg CEMS during a 7-day calibration drift test is not
acceptable as the point of the test is to confirm that the system can
operate for 7 days without unacceptable drift.

Comment: One commenter (2915) said that in Section 6.1.1–EPA proposes
several requirements related to [t]he Mercury CEMS data recorder output
range. These requirements are obsolete. All recording is now done by
computer and virtually all data transmission and recording is digital.
Digital data transmission and management allows one to eliminate the
span and range language problem because digital data have no range. Such
a change will allow sources to utilize the full several order of
magnitude measurement capabilities of modern analyzers–especially
mercury analyzers.

Response: Section 6.1.1 of PS-12A regarding the data recorder scale has
been revised to respond to the comments received. We have simplified the
specification to clarify that the scale must at a minimum include zero
through the span value. This revision will allow sources to utilize
multiple orders of magnitude, as needed.

Comment: One commenter (2915) disagreed with EPA’s proposed
requirement that the measured value not exceed 95 percent of the high
value. The measured value should be allowed to be as high as the
calibration value. This section should be eliminated or modified to
accommodate modern analyzers.

Response: We agree with the commenter and this specification has been
removed from Section 6.1.1 of PS-12A.

Comment: One commenter (2915) said that in Section 6.2–EPA proposes to
require that the reference gas delivery system be designed so that the
flow rate of the reference gas is the same at all levels. The commenter
is not aware of any reason for this requirement and EPA offers none.
Several of the commercial mercury calibrators (including the market
leader) are designed to vary the flow rate of the reference gas, and the
gas sampling probes are designed to vent any excess flow. This proposed
requirement would render those calibrators ineligible for use under the
proposed rule, without any stated justification. In the commenter’s
view, the only pertinent requirement in this section is that the flow of
the reference gas exceeds that required by the CEMS sample probe design.

Response: We have revised the language of Section 6.2 to specify only
that the reference gas delivery system must be designed so that the flow
rate exceeds the sampling system flow rate of the CEMS and that the gas
delivered to the CEMS is at atmospheric pressure.

Comment: One commenter (2915) said that in Section 7.1.1–EPA proposes
to require that the zero level reference gas concentration be
0-20 percent of the span value. The zero-level should be made 0-30
percent of span instead of 0-20 percent of span to capture the best
available NIST traceable concentrations. For example, the available NIST
concentrations for a 0-10 μg/m3 span are 1.1, 1.3 and 2.7 μg/m3. Of
these, the 2.7 μg/m3 concentration has been the standard low-level
calibration/linearity concentration because its uncertainty, per NIST
traceability calculations, is much lower than the other two
concentrations.

μg/m3 gas as the zero level reference gas, but use a zero gas in
addition, or (3) request that the gas generator vendor provide a
generator that can deliver a standard at 2.0 μg/m3 or below using
guidance in the protocols.

Comment: One commenter (2915) said that in Section 8.1.1, because
mercury CEMS generally use a single point measurement, EPA proposes to
require use of a location that has been shown to be free of
stratification for Mercury or alternatively, SO2 and NOx. The provision
then states that if the RA test is failed because of the location, EPA
may require that the CEMS be relocated. The commenter stated that this
provision is not reasonable and explains why. 

Response: We have revised the language in Section 8.1.1 to specify that
the sampling location must be representative of the stack gas
concentration and have modified the requirement to use a location shown
to be free of stratification to be a recommendation rather than a
requirement. 

cification of ±0.8 μg/m3 absolute difference.

Response: We have reconsidered the linearity test procedure
specifications and, given the current quality of the available oxidized
mercury gas standards, we have decided to increase the allowed
measurement error for the oxidized mercury standards to + 10 percent to
be consistent with the target criteria in the EPA Interim Traceability
Protocol for Qualification and Certification of Oxidized Mercury Gas
Generators.

Comment: One commenter (2915) said that in Section 8.4.3, EPA proposes
to require that CEMS meet a calibration drift specification of 5 percent
of the span value. The problem in this section is the same as discussed
in the Section 8.3 above. The specification cannot be supported by field
data and should be changed to ±10 percent of span with a ±1.0 μg/m3
absolute difference, whichever is less restrictive. This specification
will be similar to the Clean Air Markets Rule (CAMR) rule specification
in Part 75, which was difficult to achieve.

Response: We believe that the calibration drift specification of 5
percent of the span value is reasonable and that many installations of
mercury CEMS have been able to achieve the specification.

Comment: One commenter (2915) said that in Section 8.5.4, EPA proposes
to require a minimum sample time of 30 minutes for Method 30A. The
sample time specification is inconsistent with the method. The sampling
time for Method 30A will vary depending on the number of traverse
points. The minimum sampling time for Method 30B is 30 minutes and that
is not included in Section 8.5.4.

Response: The commenter is correct. Method 30A sampling time depends on
the number of traverse points. The reference should have been to Method
30B and we have revised the section accordingly.

Comment: One commenter (2915) said that in Section 8.5.5, EPA proposes
to require use of paired reference method values when using certain
methods to perform a relative accuracy test audits (RATA), including
Method 30A. Method 30A should be removed from the method list in this
section because it does not provide paired reference method values.

Response: The commenter is correct; Section 8.5.5 of PS-12A has been
revised to remove the reference to Method 30A.

e criteria applicable to paired reference method values used in RATAs.
Method 30B provides for an acceptance criteria ±0.2 μg/m3 absolute
difference between data pairs in addition to the 20 percent relative
difference criteria. The absolute difference criteria are missing from
this section. The commenter does not believe it is appropriate to modify
a reference method in a Performance Specification. EPA should add the
alternative specification or state that alternative specifications in
individual reference methods may also be used.

Response: We have revised Section 8.5.6.2 of PS-12A to be consistent
with Method 30B.

Comment: One commenter (2915) said that in Section 13 the relative
accuracy, linearity and calibration drift specifications are too
stringent and will result in a significant number of failures. The
commenter suggested that the linearity, calibration drift and relative
accuracy specifications in this section mirror those that were
promulgated in Part 75 for use under CAMR. A large quantity of field
data show that present day mercury CEMS technology cannot routinely
achieve the specifications proposed in Section 13.

Response: These issues have been addressed in earlier responses.

ceable is currently 47.5 μg/m3. This is far too low to be of use for
cement kilns where the most common span value will be 260 μg/m3. The
commenter stated that this problem can be solved in one of several ways:

If the bracketing system used for certification is demonstrated to be
linear, there would be no obstacle to having a candidate calibrator
certified at concentrations several times the concentration of the
reference calibrator. This applies to both the NIST and vendors’
bracketing systems. A change to the 25% difference limit would be
required.

NIST could certify a NIST prime calibrator at the target concentrations
for the cement rule. Note that using the same NIST prime unit for both
high and low concentrations might impact the certification accuracy of
the prime unit for low concentrations due to residual contamination.
Thus a second NIST prime calibrator for high concentrations would likely
be required.

Response: As noted earlier, NIST has recently completed certification of
a ‘NIST Prime’ elemental mercury gas generator at concentrations of
41, 68, 85, 105, 140, 185, 230, 287, and 353 µg/m3 and mercury gas
generator vendors may now submit elemental mercury gas generators for
certification to serve as ‘Vendor Primes’. Therefore NIST traceable
mercury gas standards can now be made available in concentrations that
exceed the equivalent mass standards for both existing and new kilns by
between one and two orders of magnitude, thus providing the capability
to accurately report excursions well beyond either standard.

Comment: One commenter (2894) stated that many of the span value
concepts in this proposed rule date back to the days of strip chart
recorders, analog instrumentation, and low resolution A/D converters. In
these legacy systems, exceeding an upper output limit meant that all
information was lost as the pen pegged or the A/D overloaded. In
contrast, all modern CEM systems provide digital interfaces that
transmit actual numbers. These are capable of high resolution over an
unlimited dynamic range. Commenter 2894 suggested that vendors should be
required to provide test data on the dynamic range of their CEM systems,
particularly at high concentrations. The regulation should allow
temporary exceedences of the span range provided that they fall within
this linear range. (Currently, data is considered invalid if
concentrations exceed the span range.) This is particularly essential if
the rule requires the monitoring of mercury during SSM periods. This
wide dynamic range should be allowed only on installations that do not
artificially limit their dynamic range by using older 4-20 mA or other
analog signal transfer methods.

Response: As explained in response to earlier comments, the data
recorder scale of Section 6.1.1 of PS-12A has been revised to respond to
the comments received. We have simplified the specification to clarify
that the scale must at a minimum include zero through the span value.
This revision will allow sources to utilize multiple orders of
magnitude, as needed. In addition, we have modified the definition of
span in Section 3.5 to provide additional flexibility in setting a span
value including allowing for rounding up to the nearest multiple of ten
and clarifying that the applicable subpart may contain more specific
provisions regarding the span value for specific source category. For
these amendments to Subpart LLL, we have taken another commenter’s
suggestion to define span value for this source category in paragraph
63.1350(n) of the subpart based on the intended upper limit of the
mercury concentration during “mill on” operation, but at least
approximately equivalent to two times the emission standard.

Comment: One commenter (2894) suggested that the new regulation should
ensure that as few new span ranges as possible are defined. According to
the commenter, in the U.S. power plant environment, mercury levels are
substantially lower than the emissions levels at cement kilns. The span
ranges that have been defined for power plants are: 2 μg/m3 (future
use), 5 (future use), 10, 20, 30 and 40 μg/m3. This is more than are
actually required, given the performance of all current CEM systems. (In
particular, the span range of 30 μg/m3 is superfluous.)

Each of these defined span ranges requires three distinct and separate
concentrations of calibration gas: low (20-30% of span), medium
(50-60%), and high (80-100% of span). This requirement would pose an
undue burden on both NIST and the vendors if too many new span ranges
are defined.

The commenter stated that any requirements for dual span ranges for a
CEM installation are archaic and should be deleted.

Response: With the revisions to the definition of span in PS-12A and
Subpart LLL monitoring requirements described in the previous response,
we have provided ample flexibility in span selection and have eliminated
the need for dual spans.

Comment: One commenter (2894) suggested upgrades to PS-12A:

6.1.1 Data Recorder Scale. This section deals with obsolete strip chart
recorder technology and should be either deleted or redrafted.

6.2 There are actually four gas levels being used in CAMR power plant
installations, not three: Zero, Low linearity, Mid Linearity and High
span. There is an inconsistency here. (i.e., Should the Low Linearity
gas actually be allowed to be in the range of 0-20% as specified in
PS-12A or 20-30% as is current practice?).

8.1.3 Mercury CEMS Sample Extraction Point. Are most cement kiln stacks
sufficiently large to allow sampling 1.0 meter from the stack wall?

8.3 Linearity Procedure. Currently, power plant CEMS are required to
inject at three non-zero concentrations, not two concentrations plus
zero. Will replacement power plant regulations still require three
non-zero concentrations or will this requirement be relaxed as per this
section?

Response: Section 6.1.1 of PS-12A regarding the data recorder scale has
been revised to respond to the comments received. We have simplified the
specification to clarify that the scale must at a minimum include zero
through the span value. This revision will allow sources to utilize
multiple orders of magnitude, as needed.

We note that the vacated Clean Air Mercury Rule specifications are no
longer relevant. The low (“zero-level”) reference gas must be in the
range of 0 to 20 percent of span.

Based on the information we have, yes, most cement kiln stacks are
sufficiently large to allow sampling 1.0 meter from the stack wall.

Replacement regulations for mercury emissions have not yet been
proposed, so we cannot provide a definitive response to the question on
Section 8.3. However, PS-12A allows linearity tests to be conducted at
three non-zero concentrations or at two non-zero concentrations plus
zero.

Comment: One commenter (2915) disagreed with the proposed mercury
performance specifications and with the provisions for setting Mercury
CEMS span values and the stringency of the proposed performance
specifications for both Mercury CEMS and sorbent trap monitoring systems
(STMS).

Commenter 2915 stated that EPA’s proposed Mercury performance
specifications include setting Mercury CEMS span values equal to twice
the applicable emission limit and to use that span value to determine
the levels at which the Mercury CEMS must be calibrated. 74 FR at 21172.
EPA’s approach to span does not reflect the measurement capabilities
of Mercury CEMS or the limited number of potential National Institute of
Standards and Technology (NIST) traceable calibration gases available to
calibrate Mercury CEMS of various spans.

Commenter 2915 stated that EPA also should define span in a manner that
is consistent with the regulations and permits that are likely to
require use of PS-12A and Procedure 5. Because most Mercury limits,
including the ones EPA proposes in this rule, are expressed in terms
other than µg/m3 (e.g., lb per million tons of clinker produced, pound
per trillion British thermal unit [lb/TBtu], or pound per megawatt hour
[lb/MWh]), the commenter disagrees with defining the required Mercury
CEMS span (which is expressed in µg/m3) in terms of the Mercury limit.
Instead, PS-12A should defer to the applicable regulation or permit to
define the span (in mg/m3) that is most appropriate for the intended
use.

Response: We have revised the definition of span in Section 3.5 of
PS-12A as well as Subpart LLL to provide additional flexibility. The
definition of span in PS-12A now clarifies that the span is the
measurement range specified in the applicable regulation or other
requirement. If the span is not specified in the applicable regulation
or other requirement, then it must be approximately two times the
emissions limit. The definition also allows for rounding up to the
nearest multiple of ten. In Subpart LLL, paragraph 63.1350(n), we have
defined the span value for this source category as based on the intended
upper limit of the mercury concentration during “mill on” operation,
but at least approximately equivalent to two times the emission standard
and again allowed for rounding up to the nearest multiple of 10.

Comment: According to commenter 2915 EPA should provide more data to
support its proposed tightening of the relative accuracy standard for
both Mercury CEMS and STMS relative to the vacated CAMR standards. As a
result of its work on the Interim Traceability Protocols and other
research efforts, EPA has had access to a considerable amount of
additional field data. The commenter stated that the proposed linearity,
calibration drift, and relative accuracy standards for Mercury CEMS are
too stringent and may be frequently failed.

Response: As noted in prior responses, after consideration of the
potential variability among the learning curves that new users of Hg
CEMS may experience, we have revised the relative accuracy specification
in PS-12A from 10 percent to 20 percent. We believe, however, that the
implication of this change is minor and short-lived because most
facilities that install Hg CEMS will find that they readily achieve
relative accuracies far better than 20 percent.

We have also reconsidered the linearity test procedure specifications
and, given the current quality of the available oxidized mercury gas
standards, have decided to increase the allowed measurement error for
the oxidized mercury standards to + 10 percent. We have not changed the
calibration drift specification of 5 percent of the span value as we
believe is reasonable as many installations of mercury CEMS have been
able to achieve this specification.

PS-12B

Comment: Several commenters (2855 and 2915) provided specific comments
about PS-12B. 

Commenter 2855 provided the following comments:

EPA has mixed a PS with on-going quality assurance procedures making
this procedure unnecessarily complicated. EPA should develop an Appendix
F type procedure for on-going quality assurance (QA) and eliminate the
three section spike trap requirement for daily use.

Section 6.1.1’s substitution of three-section traps in place of
two-section traps is unnecessary. The addition of the third section is
used for QA purposes. This requirement should be moved to the Appendix F
procedure that EPA needs to develop for sorbent trap systems.

The third section spike trap should only be required when the duplicate
trains do not meet the RD specification, during the initial relative
accuracy test and for on-going QA such as quarterly audits.

Response: We recognize that the commenter is recommending that all
quality assurance (QA) procedures related to sorbent trap-based mercury
monitoring systems in PS-12B be moved to a separate procedure in 40 CFR
60, Appendix F; we agree that this approach is more consistent with our
prior performance specifications and we have revised PS-12B to remove
the on-going quality assurance procedures and have incorporated them
into Procedure 5. The three-section trap required by PS-12B provides a
routine quantitative indication of measurement performance analogous to
the daily span check performed on gaseous mercury CEMS.

Comment: Commenter 2855 said that when spikes are required, the
procedure should allow the use of liquid standard spikes that are
traceable to NIST.

Response: We do not allow the use of liquid standard spikes in PS-12B as
they are not representative of gaseous emissions of mercury.

Comment: Commenter 2855 said that this reference sentence in Section
1.2.2 should be deleted: Section 1.2.2. This specification is not
designed to evaluate an installed sorbent trap monitoring system’s
performance over an extended period of time nor does it identify
specific techniques and auxiliary procedures to assess the system’s
performance. The statement is incorrect. The fact that paired sampling
trains meeting a relative deviation (RD) specification are required is a
means for determining ongoing evaluation over extended period of time.

Response: We agree with the commenter and have removed Section 1.2.2
from PS-12B.

Comment: Commenter 2855 said that Section 8.1.1 should be changed to
read select a location that is representative of Mercury emissions.

Response: We agree with the commenter and have revised Section 8.1.1 of
PS-12B so that it is consistent with the corresponding section of
PS-12A.

Comment: Commenter 2855 said that Section 8.3 should be deleted because
the RATA procedure is redundant and unnecessary.

Response: It is agency policy to conduct relative accuracy testing of
mercury monitoring systems. These RATAs are used to confirm the
continuing accuracy and representativeness of compliance monitoring
system measurements.

Comment: Commenter 2915 said that their comments on Section 8.1.1 of
PS-12A apply to proposed PS-12B.

Response: Section 8.1.1 of PS-12B specifies only that the sampling
location must be representative of the stack gas concentration of
mercury with a suggestion to consider a location shown to be free of
stratification for mercury, SO2, or NOx.

Comment: Commenter 2915 said that in Section 8.2.3, for sources
calculating Mercury mass emissions, EPA proposes to require
determination of stack gas moisture content using a certified continuous
moisture monitoring system. This requirement is too restrictive. The
commenter suggested that EPA review the alternative moisture measurement
options provided in Part 75 at 75.11(b), 75.12(b) and Appendix K, 7.2.4.
Similar language should be incorporated into PS-12B, especially the
default moisture content and psychrometric chart look-up table language.
In addition to the alternative of using the moisture monitor and various
options defined in Part 75, the commenter stated that one or more of the
carbon trap monitor vendors has a technique for recovering the moisture
collected by the sampling train. That option should also be allowed. The
term certified continuous moisture monitoring system is undefined; the
commenter is not aware of a PS for certification of moisture monitors.

Response: We have revised Section 8.2.3 of PS-12B to remove the
requirement for continuous moisture monitoring system certification and
to allow for other moisture measurement options acceptable to the
Administrator such as those noted by the commenter.

Comment: Commenter 2915 said that in Section 8.3.2 – EPA proposes an
alternative relative accuracy specification of 0.5 μg/m3 if the
reference method value is less than 2.0 μg/m3. This alternative
relative accuracy specification is inconsistent with the alternative
specification in proposed PS-12A, Section 13.3 and there is no reason
for the inconsistency. In addition, there are considerable field data to
show that both PS-12A and PS-12B relative accuracy specifications are
unrealistic based on present day technology and will result in a
significant number of test failures.

Response: We have revised Section 8.3.2 of PS-12B to make it consistent
with the corresponding section of PS-12A and, as noted previously, we
have revised the relative accuracy specification in PS-12B to 20%.

Comment: Commenter 2915 said that in Section 11.2, this section of
PS-12B requires the analysis of the first glass wool plug in the trap,
while PS-12A states that Method 30B does not require the analysis of the
first glass wool plug (see Section 8.5.2 of PS-12A). If the first glass
wool plug is included in the analysis, particulate Mercury will be
measured not just vapor-phase Mercury. The commenter agrees with the
PS-12A requirement because the intent is to not measure Mercury that may
be on the particulate matter. It is clearly stated in Section 1.0 that
the intent is to measure vapor-phase Mercury emissions.

Response:  The first glass wool plug of the Method 30B sorbent tubes
must be analyzed. We have clarified Section 8.5.2 of PS-12A to clarify
that the sentence stating that the “filterable portion of the sample
need not be included when making comparisons to the CEMS results” is
only applicable to Method 29 and Method D6784-02.

Appendix F, Procedure 5

Comment: One commenter (2855) agreed with Section 4.1 requiring that
corrective action or adjustment shall be performed when the CD exceeds
twice the PS-12A limit, and Section 4.3 specifying out of control
criteria as CD checks results exceeding (a) twice the PS-12A limit for
five consecutive daily periods, or (b) CD check result exceeding four
times the PS-12A CD limit.

Response: EPA notes the commenter’s support.

Comment: One commenter (2855) said the use of oxidized calibration
standards for gas audits are redundant with System Integrity Checks
required as part of the quality control requirements in Section 3.0 and
should be deleted.

Response: We have provided some flexibility under Section 3.0 of
Procedure 5 in establishing the acceptance criteria for the weekly
system integrity check; we, therefore, on a quarterly basis require a
more extensive two level accuracy audit using oxidized Hg gases that
must meet specified criteria.

Comment: One commenter (2855) said that the language in Section 5.1.2
(2): “At a minimum, the audit gas should be introduced at the
connection between the probe and sample line” is not appropriate and
should be removed. 

Response: We agree with the commenter and have removed this sentence
from Section 5.1.2(2) of Procedure 5.

Comment: Commenter 2855 said that quarterly gas audits will be performed
using the elemental Mercury and oxidized Mercury calibration gas
generators of the CEM system. Section 5.1.2 (3) requires use of NIST
certified or NIST Traceable calibration materials for gas audits.
However, NIST traceable gas standards will not be available may not be
available at the required levels for several years.

Response:  As noted previously, NIST has recently completed
certification of a ‘NIST Prime’ elemental mercury gas generator at
concentrations of 41, 68, 85, 105, 140, 185, 230, 287, and 353 µg/m3
and mercury gas generator vendors may now submit elemental mercury gas
generators for certification to serve as ‘Vendor Primes’. Therefore
NIST traceable mercury gas standards can now be made available in
concentrations that exceed the equivalent mass standards for both
existing and new kilns by between one and two orders of magnitude, thus
providing the capability to accurately report excursions well beyond
either standard.

s ± 5.0 μg/m3.

Response: This was an inadvertent error; the specification in Section
5.2.3 (2) was intended to be ± 0.5 μg/m3 and the section has been
revised to reflect this.

Comment: Commenter 2915 said that in Section 1.1, in the second
paragraph EPA states that, Procedure 5 covers the instrumental
measurement of Mercury as defined in PS-12. It is not perfectly clear on
first reading that Procedure 5 is not applicable to Mercury measurement
systems using PS-12B. PS-12B is performance-based and is
self-validating. A sentence should be added making that fact clear.

Response: As noted in response to a previous comment, we have decided to
revise PS-12B to remove the on-going quality assurance procedures and
instead incorporate them into Procedure 5; thus, Procedure 5 will now be
applicable to both PS-12A Hg CEMS and PS-12B monitoring systems when
called for by the applicable regulation.

Comment: Commenter 2915 said that in Section 5.1.2(3), as proposed, the
section states, Use elemental Mercury and oxidized Mercury...... This
should be changed to read, Use elemental Mercury or oxidized
Mercury......... It is not necessary to use both elemental and oxidized
Mercury for this test because both are NIST traceable. 

Response: We disagree with the commenter’s recommendation; we believe
it is important to obtain information on system linearity and a
quantitative accuracy assessment for both species on a regular basis.

Comment: Commenter 2915 said that in Sections 5.1.3 and 5.2, EPA
proposes to require the analysis of performance audit samples, except
during audits following corrective action. However, EPA provides no
information regarding the availability of such samples or explanation
regarding how the results will be used. The commenter does not believe
such samples exist for Mercury and notes that audit samples are not
necessary for Methods 30A and 30B because they are performance-based
methods and are, therefore, self-auditing. This language also is
inconsistent with the statement in Section 1.2 that Procedure 5 also
requires the analysis of audit samples concurrent with certain reference
method (RM) analyses as specified in the applicable RMs. EPA should
revise this provision to include similar language that makes clear, for
example, that audit sample analysis is not required for Methods 30A and
B, because those methods do not require them. For other methods, EPA
should explain the requirement and make clear that it does not apply if
audit samples are not available. 

Response: We agree that this language on use of performance audit
samples is not necessary in this procedure and have removed those
references.

Comment: Commenter 2915 said that in Section 5.2.3, this section sets
out the criteria for excessive audit inaccuracy. Based on the
commenter’s experience, all of the specifications are too stringent.
For RATAs, EPA refers to PS-12A. For Gas Audits, EPA proposes an
alternate specification of ±5 ppm. Because Mercury CEMS do not read in
ppm (they read in μg/m3), this specification makes no sense. This
alternate specification should be changed to ±1.0 μg/m3. For RAAs, EPA
proposes the greater of ±15 percent of the three run average, or ±7.5
percent of the applicable standard. These specifications are too
stringent and EPA has provided no data to support them. EPA should
revise the standard to ±20 percent of the three run average or 10
percent of the applicable standard.

 μg/m3, and (3) we have revised the specification in Section 5.2.3(3)
for the relative accuracy audits to be ±20 percent of the three run
average or the equivalent of 10 percent of the applicable standard,
whichever is greater, as suggested by the commenter.

Comment: One commenter (2915) said that in Section 5.3, EPA proposes to
require a source to revise its QC procedures or modify or replace the
Mercury CEMS if excessive audit inaccuracy occurs for two consecutive
quarters. EPA provides no basis for this requirement or data to support
its criteria for excessive audit inaccuracy. The consequences of failing
performance specifications should be identified in the applicable
regulation. In addition, the requirement to, modify or replace the
Mercury CEMS is much too onerous. There are only two vendors that have
Mercury CEMS that are compatible with these proposed regulations and the
technology is what it is. The performance of both vendors is similar and
requiring a change from one vendor to the other makes no sense.

Response: We disagree; if the mercury CEMS was capable of passing the
initial RATA, then with modification, repair, or a change in the QC
procedures, the system should be able to continue meeting the QA
criteria.

Comment: One commenter (2837) stated that PS-12A and Procedure 5 appear
to be written around the perceived capabilities and limitations of a
specific CEMS technology. It is important to correct this because as
written, these two documents have the potential to fix future technology
and emission standards to the capabilities of the measurement technology
around which they were written. This is most apparent in those sections
where the specifications unnecessarily require certified NIST traceable
gas standards or generators. The commenter provided further detailed
discussion of this issue. The commenter requested that EPA modify both
PS-12A and Procedure 5 to remove requirements that are technology
specific and replace them with performance based requirements.

Response:  The use of NIST traceable mercury gases for system
calibrations, checks, and linearity determinations is not specific nor
exclusive to any particular type of CEMS technology. We disagree that
that their use is unnecessary; in fact, we believe that use mercury
gases (in particular NIST traceable gases) is the only way to routinely
confirm that the mercury in the stack gas matrix is being quantitatively
transported from the stack to the instrumental analyzer.

2.5.5	Impacts

Costs

Comment: One commenter (2855) states that the industry believes
installation and certification of HCl CEMS as currently proposed is
expected to cost between $225,000-$250,000 plus another $30,000 for
initial certification. The total costs for ongoing QA are unknown as
facilities cannot install and certify HCl CEMS using Performance
Specification-15. Additional costs for add on controls (e.g., wet
scrubbers) to meet the extremely low concentration levels proposed are
expected to cost upwards of $20,000,000.

Response: EPA estimated the cost of complying with the final HCl
emission standards as the cost of installing and operating add-on
control equipment and the cost of installing and operating continuous
emission monitoring equipment. Total capital and annual costs for HCl
CEMS were estimated to be $144,000 first time (or capital) cost, and
$43,000 per year annual cost. Total capital cost of a wet scrubber for a
new 1.2 million tpy kiln, including the cost of CEMS, was estimated to
be $25.1 million per kiln. Cost of control for HCl is discussed further
in, “Summary of Environmental and Cost Impacts for Final Portland
Cement NESHAP and NSPS.”

Plant closures

Comment: Many comments were received regarding the effects of the
proposed rules on plant closures.

Commenters 2298, 2483, 2506, 2507, 2651, 2653, 2654, 2661, 2667, 2671,
2672, 2811, 2812, 2813, 2824, 2832, 2833, 2863, 2886, and 3218 stated
that without the use of a subcategory within the MACT rule allowing EPA
to address the naturally occurring mercury in limestone of western
cement plants, then plants like the Durkee Cement Plant in rural Oregon
may have to close. The commenters explained that the loss of the Durkee
Plant jobs would have a substantial impact on the local economy. 

Commenter 2832 stated that the CAA does not give EPA the authority to
regulate, diminish, or steal away the financial and quality of life
advantages that a community may have because of local desirable natural
mined materials. Inclusion of emissions associated with such mined
materials would effectively be a ban on their use and would exact a
great toll on local communities. 

Commenter 2830 has determined that three of their existing plants could
be forced to close with many of their wet kilns at high risk of closure,
and high capital cost may also close long dry kilns should the 2009
proposed rule be implemented. 

Commenters 2830, 2840, 2841, 2863, 2886, 2916, and 3218 stated that
should cement facilities close because of the proposed rule, the lost
domestic supply of cement will be offset by imports, creating global
increases in mercury and greenhouse gas emissions if these foreign
plants are not regulated to the same degree, and increased
transportation emissions. Commenter 2841 added that there would be an
increase of NOx, volatile organic compound (VOC), PM, carbon monoxide
(CO) and CO2 emissions from the transportation of raw materials.

One commenter (2786) disagreed with comments suggesting cement
production will shift overseas due to weaker standards. The commenter
cites a United Nations effort to develop a global legally binding
instrument on mercury by the year 2013. 

Commenter 2867 also disagreed with the severity of industry projections
regarding plant shutdowns and a lack of capacity resulting from
implementation of the proposed rules. The commenter suggests that many
facilities will be able to meet limits with closer review of raw
materials or by implementing less-expensive dust shuttling. In addition,
it is possible for the industry to mitigate the claimed impact by
encouraging cement equipment designers to tackle these issues and
disseminating successful mercury reduction techniques throughout the
industry.

Response: As noted in the preamble, EPA carefully considered the
possibility of creating different subcategories of cement kilns with
respect to mercury emissions, including subcategories based on the
mercury concentration of the limestone. For the reasons stated in the
preamble, we decided not to create this type of subcategory.  We
anticipate an increase in imports of only10 percent in 2013, which
translates to a decrease in domestic revenues of 3 million metric tons. 


The preamble also notes that EPA is forbidden by law from considering
costs in determining MACT floors. NRDC v. EPA, 489 F. 3d 1364, 1376
(D.C. Cir. 2007); National Lime, 233 F. 3d at 640. Although one of the
overall goals of the Act is to protect and enhance the quality of the
Nation’s air and resources so as to promote the public health and
welfare and the productive capacity of the population,” CAA section
101(b)(1), this overall goal does not somehow authorize EPA to adopt
floors that either consider costs (overall or otherwise) or to base
floors on other than what best performers achieve. Further information
about the economic impacts of this rule can be found in the preamble and
the project docket.

Wasting CKD to reduce mercury emissions

Comment: Several commenters (2144, 2824, and 2845) commented on the
practice of shuttling or wasting CKD and its ability to reduce mercury
emissions. 

Commenter 2845 stated that EPA should not consider the removal, or
wasting of CKD a form of mercury control, nor should it be considered
akin to a mercury control device. The purpose of wasting CKD is to
ensure product quality. No plant wastes CKD with the intent of reducing
emissions. The level of mercury removed by wasting CKD also varies
significantly from plant to plant; to consider this practice as a form
of mercury control is highly questionable. Furthermore, as EPA
recognizes, wasting CKD generates solid waste, and cement companies
often have significantly reduced their solid waste generation by
decreasing their wasting of CKD. Accordingly, the Proposal should not
require companies to waste CKD or engage in dust shuttling.

Commenter 2845 stated there are other limitations to reducing mercury
emissions by removing or wasting baghouse dust. Instead of returning the
dust (i.e., uncalcined limestone), it can be continually bled off to the
finish mill and intermixed with cement. This is known as dust shuttling.
However, there are limitations to how much of this baghouse dust can be
intermixed with cement. Specifications on the quality of cement limit
the amount of baghouse dust that can be intermixed with cement. Dust
shuttling also results in additional carbon dioxide emissions from each
ton of cement produced because companies cannot utilize limestone
addition (the dust would use up all the insoluble residue and/or loss on
ignition available). The commenter includes expanded comments on this
topic as an appendix entitled Mercury Emission Control by Removing
Baghouse Dust.

Commenter 2144 disagreed with the reference at page 21150 of the FR
Notice that removal of all baghouse dust when the raw mill is down will
yield only 4% reduction in mercury emissions. The commenter states that
mercury control is actually negative during the raw mill-off condition
considering a plant-wide material balance.

Commenter 2144 agrees with EPA’s statement that the use of CKD wasting
to reduce mercury emissions is site specific. The commenter added that
with the construction specifications allowing 5 percent inorganic
processing additions, the CKD or filter dust will often not need to be
wasted and has been shown to improve or least not to degrade the cement
in areas where the raw materials are low in alkali.

One commenter (2824) stated that criteria for determining the
best-performing 12 percent of existing sources for mercury emissions
should either not include kilns wasting CKD and scrubber dust or the
contribution in the dust should not be subtracted from the input.
Otherwise, depositing high mercury waste in landfills is considered
best-performing. Wasting dust requires more fuel and feed which
generates increased CO2 emissions. This is an adverse cross-media
impact.

Response: The commenter is not correct in implying that intent is
relevant in determining if an activity which affects emissions can be
considered as a means of control. All that matters is the effect on
emissions. National Lime II, 233 F. 3d at 640. We realize that this is a
site-specific issue, and the percent removal obtained by shuttling or
wasting CKD ranges from very low to up to 80 percent. Some of the
best-performing kilns in the MACT floor waste CKD. It will take more
extreme facts than those presented here to justify invalidating a
practice that removes emissions from some of the best performers. 

Comment: One commenter (2844) stated that there is no demonstrated ACI
system that will simultaneously control mercury and hydrocarbons. In
light of that uncertainty, the industry is faced with taking a chance on
an unproven technology or using multiple technologies (ACI plus a
scrubber plus an RTO) that represent huge investments which have not
been analyzed fully by EPA.

Response: The reason EPA has long adopted the interpretation that the
existing and new source MACT floors are to be applied on a HAP-by-HAP
basis are that a whole plant approach likely yields least common
denominator floors, that is floors reflecting mediocre or no control,
rather than performance which is the average of what best performers
have achieved.  See the preamble to the final rule for more discussion.

Concerns over effectiveness of control technology

Many comments were received on the limited emission reductions
achievable with the installation of add-on control devices.

Effects of mercury speciation

Comment: One commenter (2844) stated that the standards are not
achievable, and EPA has used unrealistic assumptions about the control
effectiveness of individual technologies. Wet scrubbers are ineffective
controls for elemental mercury, and the effectiveness of ACI is a
function of mercury speciation. In the utility industry, the collection
efficiency ranges from less than 20 percent to greater than 90 percent.
If the concentration of oxidizing agents is low, a high percentage of
the elemental mercury is not effectively captured by the sorbent. There
have been only limited mercury speciation data collected at cement
plants, but what data there are indicates the mercury speciation at the
inlet of the particulate control device ranged from almost 100 percent
to near 0 percent elemental mercury. Depending on plant configuration
and flue gas composition, standard ACI will not provide a sufficiently
high level of mercury control to allow certain cement kilns to meet the
proposed rule. The cost of ACI is high, and may not be cost effective
for facilities where the pre-control emissions are very small, and the
emission reduction will be small.

Response: We set the standards based on what has been achieved in
practice in the industry, with MACT floors established based on actual
emissions data for the best-performing sources. We used statistical
analysis and a 30-day averaging period to account for variability, such
as the situations cited by the commenter.

Activated carbon injection

Comment: One commenter (2845) stated that EPA overestimates the emission
reductions from mercury controls. EPA states that most if not all cement
facilities will have to install some form of mercury control in order to
meet the proposed mercury rule. Currently, the best available control
technology for mercury is considered to be ACI. However, the commenter
argues that the effectiveness of ACI is site-specific, based on utility
industry studies and that should be the case for the cement industry as
well. The commenter provides discussion about the technical issues
surrounding the five primary factors that affect mercury collection
efficiency using ACI: mercury speciation; temperature profile; residence
time of the particulate control device; type of ACI used; and flue gas
composition (particularly sulfur trioxide (SO3)).

Commenter 2845 stated that the effectiveness of ACI for mercury control,
based on limited mercury speciation data, will be very site-specific. It
also appears that the concentrations of elemental and oxidized mercury
are very site-specific. As a result, as has been recognized in the
utility industry, ACI is not the magic bullet that will always control
mercury at a high level. Mercury control strategies will need to be
tailored for each plant. Such strategies may include one or more of the
following: installation of a polishing baghouse; use of treated carbons
or enhancement agents; change in the strategies for handling of CKD
(wasting or not wasting); installation of a wet scrubber; evaluation and
testing of more experimental types of mercury control technologies.

According to commenter 2845, all of these strategies will be expensive,
and there is no guarantee that ACI or any configuration will be adequate
to demonstrate compliance with the proposed mercury limit for a
significant number of cement facilities. 

Commenter 2845 stated that EPA does not consider the adverse impact that
ACI will have on the way CKD is handled by the cement industry. Industry
has been striving to utilize CKD by either returning the CKD to the
cement kiln, or beneficially utilizing it. To ensure the quality of CKD
and to maintain the ability to recycle the CKD back to the system, it
may be necessary to install a polishing baghouse. The polishing baghouse
would introduce new problems, including lower than expected air-to-cloth
ratios resulting in a larger footprint, self-ignition in the baghouse
hoppers, and an increase in fugitive emissions, all of which increase
installation and operating costs.

Response: The pilot test data we have available for ACI systems showed
mercury removal of 90 percent, and we conservatively used 80 percent
removal in our analysis. We considered the installation of a polishing
baghouse and included it in our analysis, as well as the use of a wet
scrubber. The standards do not require that facilities install a
particular type of emission control (e.g., ACI). Facilities will need to
decide for themselves which control strategy is best for them, that will
account for site-specific factors and allow them to obtain the level of
emissions control needed to meet the numerical standard. Furthermore, we
do believe that the test data from the electric utilities industry is
transferable to this industry and that the commenters have not provided
any technical basis for their comments other than a discussion of
potential issues.  EPA knows of nothing that would preclude these
technologies from operating effectively in sequence.  

Wet scrubbers

Comment: Several commenters (2844, 2845, and 2900) stated that EPA did
not fully consider the issue for wet scrubbers and mercury capture.
Commenter 2845 worked with EPA in 2008 to study mercury reductions
achievable through wet scrubbers, and includes the report in their
comment.

According to commenters 2844, 2845, and 2900, the only mercury captured
by a wet scrubber is oxidized mercury. Elemental mercury is insoluble in
water and, therefore, passes through the scrubber. Therefore, the
effectiveness of wet scrubbers for mercury removal is almost entirely
based on the form of mercury entering the unit. There is a wide range in
mercury speciation for the cement industry. As a result, the theoretical
level of mercury control obtained by already having or installing a wet
scrubber will be different for each plant. There is also a phenomenon,
known as mercury reemission, related to mercury/scrubber chemistry that
may reduce the effectiveness of a wet scrubber to capture even oxidized
mercury. Reemission is a result of mercury reduction chemistry in wet
scrubbers, where oxidized mercury is reduced to insoluble elemental
mercury. Mercury reemission is primarily a result of sulfate/sulfite
chemistry. Testing conducted at four cement kilns utilizing a wet
scrubber revealed that all four of the plants tested showed elemental
mercury concentrations at the outlet of the scrubber to be greater than
the inlet concentrations. Commenter 2845 stated that the pH in the
scrubber may affect reemission and data as to the impact of pH on
reemission are very limited and appear to be conflicting. 

According to commenters 2844, 2845, and 2900, EPA states that oxidized
mercury will be captured by the scrubber but, as has been found both in
the utility and cement industries, depending on the flue gas composition
and type of scrubber, this is not always the case. To make any blanket
statement as to the effectiveness of wet scrubbers to control mercury
would be premature and site-specific. There are proprietary additive
technologies that are available from companies such as Babcock & Wilcox
which may help mitigate this problem but will result in increased costs.
It should be noted that none of these additive technologies have been or
are currently being utilized or even tested in the cement industry.
Therefore, the effectiveness of these additives for the cement industry
is unknown. Also, these additive technologies are not relevant for use
in establishing the existing source floor because no such database for
the use of this type of technology exists. 

Commenter 2900 provided data from two stack tests, in order to show the
difficulty of achieving compliance with the new mercury limits, and the
variability in run to run emissions during stack testing. 

Commenters 2833 and 2840 stated that EPA’s analysis has not identified
add-on technology within existing kilns that is effective for all forms
of mercury. Only 3 of the 5 existing cement kilns equipped with wet
scrubbers are among the 11 lowest mercury-emitting kilns, primarily due
to low mercury concentrations in the limestone. One of the kilns
equipped with a wet scrubber provided no mercury control at all. 

Commenter 2863 stated that several approaches to reduce mercury
emissions had been considered at the commenter’s site:

Reduction of limestone mercury content by selective mining. The
commenter provided mining data showing the variability and lack of
predictability of mercury content in the commenter’s limestone quarry.

The addition of a wet scrubber to the commenter’s ACI system. The
commenter notes that wet scrubbers only control oxidized mercury as
elemental mercury is not water soluble. Given that testing at the
commenter’s kiln indicated that up to 70 percent of the mercury is in
the elemental form, a wet scrubber is not going to provide a material
amount of additional control. The ACI system is predicted to control
high levels of the oxidized mercury as well as promote the oxidation of
the elemental mercury. Therefore, no material additive benefit would
come from the use of a wet scrubber.

A potential small boost in control efficiency may be obtained by
bleeding a percentage of the dust collected in the primary baghouse into
the finished product. The concept behind dust shuttling is that a
percentage of the uncalcined limestone caught in the primary baghouse,
and the mercury condensed on it, is continually bled off to the finish
mill rather than the raw mill, and intermixed with cement. The
usefulness of this procedure is limited by the commenter’s product’s
performance standards. 

Another conceptual possibility for reducing mercury emissions from
affected sources that include quarries with elevated mercury levels is
to implement some sort of limestone scrubbing. The commenter is not
aware of any commercially available systems that could reduce limestone
mercury content.

Response: For mercury control, ACI systems are the most likely control,
not wet scrubbers. Our analysis did not rely on wet scrubbers as the
primary control for mercury emissions. In cases where wet scrubbers are
used for HCl control, the scrubber may provide additional control for
mercury. Wet scrubbers may also be helpful in reducing mercury emissions
for a couple of high-emitting kilns. We recognize the usefulness of dust
shuttling as a method of mercury control is site-specific, but it is one
of the control strategies to be considered. 

Support mercury control strategies

Comment: Several commenters (2144, 2786, 2867, and 2928) provided
comments in support of EPA’s assessment of the use of add-on controls
and other methods to reduce mercury emissions.

Commenter 2867 provided discussion of several instances of the use of
dust shuttling, ACI, wet scrubbers, and other Mercury control
strategies. The commenter recommends that EPA:

Help foster actual demonstrations of filter dust shuttling with
temperature management. 

Help foster the demonstration of catalytic oxidation or oxidant
injection to make mercury more collectable. 

Help foster demonstrations of ACI or lime injection with dust shuttling
and temperature control and determine ways to make this combination less
expensive.

Commenter 2928 supported the use of ACI for mercury control saying that
it has a long commercial track record in removing mercury from emissions
of coal-fired power plants and waste-burning facilities. 

Commenter 2928 provided information about a German cement
manufacturer’s patent for an ACI system based on powdered furnace
coke, coupled with dust withdrawal/shuttling, temperature control, and
mercury oxidation by sulfur doping. 

Commenter 2928 stated that the use of selective catalytic reduction
(SCR) in combination with scrubbers for removal of mercury is a
technically feasible and economically reasonable control technology for
cement plants. The commenter noted that the 2006 Assessment of NOx
Emissions Reduction Strategies for Cement kilns – Ellis County written
by five cement industry experts for the Texas Commission on
Environmental Quality concluded SCR was an available technology for
newer dry kilns and a transferable technology, for older wet kilns,
meaning it is commercially available and could reasonably be expected to
work on those wet kilns. 

Commenter 2144 recommended that EPA foster the demonstration of
catalytic oxidation or oxidant injection to make Mercury more
collectable. Technologies include SCR and potentially peroxide injection
into gas conditioning towers. The commenter has data demonstrating the
efficacy of mercury reduction in the coal-fueled power industry when SCR
is used to control NOx emissions. SCR technology has been commercially
demonstrated in the cement industry at Solnhofen, Germany, Monselice,
Italy and Sarche di Calavino, Italy. It has excellent multi-pollutant
possibilities including reduction of THC and dioxin/furan (another HAP
regulated under the Cement MACT rule). The commenter provided an
internet link to a report evaluating the use of SCR in the cement
industry.

Commenter 2928 stated that the use of scrubbers technically feasible and
economically reasonable control technology for mercury removal from
cement plant emissions. 

Commenter 2928 stated that the use of front end controls can also be
used to control mercury emissions. Technology that removes mercury from
coal that is being fired for fuel increases the efficiency of the coal
by 30 to 50 percent, reduces mercury by up to 70 percent, and decreases
SOx and NOx. 

Commenter 2786 stated that it is possible to achieve mercury reductions
in a cost effective manner, citing the Maryland Department of the
Environment’s agreement with Lehigh Cement Company at their Carroll
County facility to reduce emissions by 80 percent by 2012. 

Commenter 2142 stated support for EPA’s focus on good particulate
control as a way to achieve good mercury control. The commenter provided
two summary tables of mercury emission reductions by non-cement source
categories and stack test results for mercury emissions from coal-fired
power plants. 

Response: EPA acknowledges the support of the commenters of EPA’s
assessment of the use of add-on controls and other methods to reduce
mercury emissions and agrees that many of the suggested control methods
are cost effective for various installations. However, we should note
that MACT standards must be based on the best-performing sources in the
category or subcategory, not a specific control technology.   

Increased CKD waste as result of mercury control strategies

Comment: One commenter (2898) suggested that EPA address the potential
increase in cement kiln dust waste or reuse. EPA discusses the potential
for an increase in the amount of CKD wasting at cement plants as a
result of mercury emissions control practices. 74 FR at 21165. However,
the status of CKD that is not reused in the production process is
uncertain. EPA has yet to establish standards for the management of
cement kiln dust. EPA last published a Notice of Data Availability
concerning cement kiln dust in July 2002 and intimated that it would
propose new standards for the management of CKD, but has taken no
further action on the subject. In light of the potential for increase in
the practice of CKD wasting and the uncertainty surrounding the
treatment of CKD, the commenter urged EPA to consider adding
requirements to the rule to prevent or control the potential additional
solid waste or to initiate a separate rulemaking to address the proper
handling of CKD.

Response: EPA is making no determinations as to the status of waste
generated as a result of this rule, nor making any determinations as to
management practices for cement kiln dust. In assessing the costs of the
rule for purposes of EO 12866, EPA has conservatively assumed that
wastes will be landfilled and estimated the costs of doing so as part of
the estimated costs of the rule.	

Mercury standards not supported on public health basis

Comment: One commenter (3067) stated that EPA’s public health basis
for the proposed mercury requirements was not scientifically supported.
The commenter requested that EPA provide information about: what percent
of methyl mercury that is consumed by fish comes from cement plants; how
much of the industrial contribution of mercury is converted to methyl
mercury in soil; if there are studies suggesting that lowering mercury
emissions would lower methyl mercury levels in fish; and EPA’s
supporting evidence that these rules provide health benefits.

Response: According to the 2005 National Emissions Inventory, mercury
emissions from cement kilns represent 18% of the industrial
(non-electrical generating units) mercury inventory in the U.S. (U.S.
EPA, 2005). This final rule would reduce mercury emissions from cement
kilns by a substantial amount. The rate of mercury methylation is
spatially variable, depending on the specific aquatic chemistry of the
deposition location. 

For this final rule, EPA modeled mercury deposition from cement
facilities using the Comprehensive Air Quality Model with extensions
(CAMx) model. This model converted mercury emissions to ambient mercury
levels and mercury deposition. In the final RIA, EPA provides maps of
the reduced mercury deposition. Due to time and resource limitations,
EPA was unable to model the conversion from mercury deposition to
methylmercury to human health effects from consuming
methylmercury-containing fish. For more information regarding the health
effects from mercury, please consult EPA’s mercury website at 
HYPERLINK "http://www.epa.gov/hg/" http://www.epa.gov/hg/ .

However, EPA notes that standards under section 112 (d) are
technology-based, not risk-based. Thus, the decisions as to the level of
the standards is based entirely on the criteria set out in sections 112
(d)(3) and (d) (2), which do not include consideration of risk. EPA did
take into account that mercury is a potent neurotoxin, and the magnitude
of mercury emissions from particular cement kiln sources, in determining
not to create a subcategory for high mercury emitting sources with high
concentrations of mercury in their limestone. EPA believes it legitimate
to consider these factors in determining whether to exercise its
discretion whether or not to exercise subcategorization authority, since
these factors relate to the ultimate object of section 112 (d) of
reducing emissions of hazardous air pollutants.

Least cost control options may not yield expected SO2 emission
reductions

Comment: One commenter (2144) provided links to state air agency and
trade group presentations regarding control options. The commenter noted
that cement companies will gravitate to the least-cost options and it is
not a given they will choose mercury or HCl reduction strategies that
will actually yield the magnitude of reductions expected by EPA in SO2,
a pollutant not regulated under §112.

Response: EPA recognizes that any SO2 emission reduction obtained as a
result of this NESHAP rulemaking is a co-benefit of the control
technology selected either for Hg or HCl. EPA recognizes the multi
pollutant interactions and co-benefits of this rulemaking and how a well
planned integrated control strategy can maximize the environmental
results while optimizing the cost invested by a facility to comply with
such regulation. 

Restrict use of quarries with high mercury limestone

Comment: One commenter (2893) recommended that EPA should not allow any
new high mercury content limestone quarries to be mined, and that no new
high mercury content limestone quarries be allowed to be mined at kiln
locations in which an old limestone quarry has been mined out. 

Response: The new source standard for mercury reflects the performance
of the best controlled existing kiln. If a new kiln locates at a quarry
with high mercury levels but can still meet this limit, they are allowed
to do so under section 112 (d).

2.6	THC Standards 

Comment: Several commenters (2898, 2928) recommended that EPA adopt the
proposed MACT floor for THC. 

Response: EPA has revised the proposed THC floor and standard based on
more representative data. See “Development of the MACT Floors for the
final NESHAP for Portland Cement”.

2.6.1	Legal Issues

Use of CEMS data results in unlawful standard

Comment: One commenter (2901) stated that basing the THC on CEMs data
leads to an unlawful standard. The commenter’s concerns are that:

Using daily averages significantly reduces variability, leading to an
unlawful standard.

There is a flaw in using continuous data under normal operating
conditions to set a standard. 

The currently proposed 7 ppmv THC standard cannot be achieved by the top
performers. The commenter provided an analysis based on the daily data
for two sources that EPA used to set the floor, resulting in one
facility not meeting the standard 20 percent of the time.

Response: The standard in the final rule has been revised from proposal
based on additional data. However, all of the data on which the standard
is based comes from CEMs, so there is no basis for stating that use of a
CEM makes the standard more stringent. The data also are gathered from
periods of well over a year, and so can reliably be used to assess
monthly variability without use of statistical extrapolation.

Unlawful rescission of standards for Greenfield sources

Comment: One commenter (2864) stated that EPA is unlawfully attempting
to rescind the THC standards for greenfield sources in the proposed
rule. EPA established MACT standards over 10 years ago when it first
promulgated the NESHAP for the cement industry. 64 FR 31898 (June 14,
1999). The commenter stated that EPA may revise MACT standards under two
provisions: §§112(d)(6) and 112(f). EPA has not undertaken the
necessary analysis under either provision to conclude that new
greenfield sources should become subject to either of the THC standards
EPA has proposed.

Response: As noted in the preamble to the final rule in connection with
EPA’s decision to amend the PM standard, EPA does not read section 112
(d)(6) as the exclusive means of amending a MACT standard, particularly
when that standard does not properly represent MACT. The standard was
based on siting considerations only, and did not consider THC emission
levels of the best controlled similar source as required by law. The
standard is now anomalously higher than the existing source standard,
confirming that it does not represent a proper MACT standard. EPA
therefore sees no bar to amending it to establish a standard properly
based on section 112 (d).

2.6.2	Adequacy of Floor Determinations (Existing and New)

5 best performing sources vs 12 percent

Comment: One commenter (2911) stated that EPA correctly based its THC
floor on the 2 best performing sources.

Response: We agree that a literal reading of the stature results in a
floor based on the two best performing sources based on the THC CEM data
available at proposal

Comment: Many comments were received on the adequacy of the data and the
analysis used in the development of the MACT floor for THC.

Commenters 2841, 2858, 2859, 2863, 2864, and 2889 stated that EPA’s
use of 2 kilns as the basis for the THC limit for existing sources was
inappropriate. The commenters stated that EPA should not set the THC
MACT floor for existing cement plants using data from less than five
sources. According to the commenter, EPA proposed to use the limited
data set for two kilns due to a very narrow interpretation of
§112(d)(3). Because there are more than 30 sources in the source
category and EPA only has THC data for 14 kilns, it is apparent that EPA
seized on the parenthetical part of the language in §112(d)(3)(A) to
conclude that it can set the THC MACT floor based on only two kilns (12
percent of 14). According to the commenter, EPA made no attempt to use
its authority under §114 of the CAA to require cement plants to monitor
their emissions of THC in order to provide EPA with sufficient data to
establish the MACT floors based on the best performing 12 percent of
sources in the category. Congress clearly intended for EPA to obtain
data from an adequate number of sources to be able to set realistic and
achievable standards. Some of this commenter’s facilities comprised a
significant part of the THC MACT floor yet the proposed standards are
not achievable for most of the kilns owned by the company without the
installation of expensive controls such as regenerative thermal
oxidizers (RTO) to control THC emissions. 

Response: EPA and the industry agree that it is most appropriate to use
THC CEM data to set the floor because this data provided the most
representative information on long term emissions performance. EPA made
a concerted effort to obtain additional THC CEM data subsequent to
proposal of this rule. Unfortunately, the data was received too late in
the process to be thoroughly analyzed. In addition, the use of this
additional data would not have substantially changed the outcome of the
floor determination for most kilns. We continue to believe that the
literal reading of the statute required us to use the best performing 12
percent of the facilities of which we have emissions data, which, in
this case, is two kilns. 

Variability incorrectly analyzed

Comment: Commenter 2864 stated that EPA incorrectly analyzed variability
in calculating the proposed THC MACT floor. The commenter stated that
CEMS data cannot account for all of the long-term variation in a given
quarry, which extends over several hundred acres with a life of over 100
years. The commenter stated that EPA’s analysis using 30 days of data
results in a MACT floor that cannot be achieved over the long-term. EPA
must identify a MACT floor for each HAP and source category based on
emission levels achieved by the best-performing 12 percent of existing
sources, and must demonstrate with substantial evidence that the chosen
floors represent a reasonable estimate of the performance of the
best-performing units. EPA’s use of the UPL model to calculate the
MACT floors for THC violates this requirement because it does not
represent a reasonable estimate of the performance of the best
performing sources.

Commenter 2864 also critiqued EPA’s method for calculating
variability, and provided an alternative calculation after collecting
additional long-term data for the lowest emitting five sources. The
commenter stated that the UTL provides a much better estimate of the
99th percentile adjusted for variability than the UPL. The commenter
stated that EPA should have used the actual 99th percentile to establish
a MACT floor because the data set contains long-term data. 

Commenter 2889 requested that EPA review the additional data from the
recent §114 request or provide detailed information that demonstrates
that the 11 kilns for which data are available do, in fact, represent
the entire cement industry.

Commenters 2841, 2845, 2858, 2859, 2863, 2874, and 2889 stated that EPA
relied upon a data set that included a combination of long-term (470-880
data points) and short-term (31-35 data points) CEMS data, from 12
kilns. This resulted in data from the two lowest emitters being used to
establish the existing MACT floor. One of the two sources used in the
floor calculation had only 35 data points, or the equivalent of six
30-day rolling averages. To compensate for this lack of data, EPA used
the UPL statistical approach to project long-term variability. EPA is
now in possession of more than one year of CEMS data for 16 kilns. While
this still represents a rather small portion of the entire cement
industry, the commenters agreed that it is a statistically
representative database for purposes of setting existing and new source
floors. In addition, the commenters recommend that EPA use the
calculation of the 99th percentile rather than the UPL to establish the
MACT floor levels. 

Commenters 2859, 2845, 2863, and 2874 stated that with the additional
THC CEM data provided to EPA since proposal, EPA now has a sufficient
amount of data and may not need to apply any statistical analysis to
account for variability and simply use the available emissions data.
However, given the new data and considering that the new data sets
account for variability in the 5 best performing sources, the revised
THC standard should be 27 ppmv dry at 7 percent oxygen for existing
sources and 24 ppmv dry at 7 percent oxygen for new sources. However,
the data does not show whether the limestone reserves available to units
in the existing data pool or other units have significant variations in
the nature and amount of organics. Furthermore, where either coal mill
and/or raw material dryer exhaust are co-mingled with the kiln stack
emissions, it is important to determine whether either case warrants
subcategorization. Although the PM standard in the proposed NESHAP was
based on a limited data set of 3-hour stack tests, short-term test will
not address the variability issues. Commenter 2874 stated that EPA
should base the THC standard on substantially more data.

Response: In response to comments, we re-evaluated the statistical
analysis and methodology used for the proposed rule and made changes to
improve it. The changes we made and why are discussed in the preamble to
the final rule and in supporting documentation in the docket and include
why we continue to believe the use of the 99th percentile UPL is the
most appropriate statistical approach, how we accounted for variability
(including intra-quarry variability), and other changes to the analysis
since proposal. We disagree with the commenters who said 30 days of data
are insufficient, and we continue to believe that the 30 data points
provided from 30 days of sampling of all kiln inputs for 89 kilns is a
robust sample of the population of cement kilns and adequate to
characterize the variability encountered at each individual kiln,
especially as supplemented in the final rule to account for documented
inputs from outside the 30-day sampling period and with pooled estimates
of intra-quarry mercury variability (both of which are then put into the
UPL equation as part of the s squared term). 

Subcategorization

Comment: Many commenters (2844, 2845, 2859, 2863, 2864, and 2874)
suggested that EPA consider subcategorization due to the variations that
can be found in raw materials, source configurations, and processes in
which the current data from the 5 best performing sources does not
accurately illustrate how emissions may vary in different kiln
configurations and other sources. Commenter 2864 urged EPA to collect
data from all cement plants in order to properly subcategorize, and
delay setting THC MACT standards until that time. The commenters
suggested that EPA gather long-term THC CEMS emission data for the
following reasons:

For some kilns, high THC levels can be attributed to a combination of
the level of organics in the limestone, the form of the organics in the
limestone, and a kiln configuration that has one or more source of
organic releases downstream of the kiln (in particular in-line raw mills
and in-line coal mills venting to the main stack.). While a few of the
units with THC CEMS have high THC emissions, the data on raw material
(e.g., limestone organics) have not been collected or correlated for
these kiln emissions. Commenter 2845 submitted a white paper (Appendix
9) addressing these issues as part of their comment.

EPA has assumed that any raw material dryers should meet the kiln THC
limit; however, EPA has no data to show that this is an appropriate
extension of kiln THC measurements. There are limited data that suggests
that these dryers have lower volume and higher THC emissions than kilns.

Separate standards for standalone raw mills operated with supplemental
fuel firing, raw material dryers, and those kilns with high THC levels
due to organics in the co-located quarry may be appropriate. This cannot
be done without a strong database of THC CEMS readings and feed material
characteristics supporting the best possible presentation on the need
for one or more added subcategories for THC limits.

There is no data to suggest that organic HAP at standalone raw mills
that use supplemental fuels and coal mills that use kiln or cooler gases
are emitted.

The proposed organic HAP standards did not account for the various
volatiles that are added to the kiln exhaust as a result of contract
with the ground coal in a coal mill. Commenter 2863 testing documented
that a considerable quantity of organics are added to the stack exhaust
when the kiln gases are routed through the coal mill. Although, the
majority of the organics are methane, the non-methane, non-ethane
hydrocarbon, and the total hydrocarbon emissions exceeded 100 ppmvd on
average.

EPA should consider collecting additional data to determine if
subcategorization is appropriate to address THC emission variability
from different physical characteristic of kilns or different
characteristics of raw materials.

Response: See preamble to final rule.

Background THC levels not accounted for

Comment: One commenter (2889) stated that EPA has not taken into account
ambient background THC levels and the variability in ambient background
THC levels that inherently affect measured stack concentrations. Ambient
levels of methane and ethane will increase the measured THC by as much
as 15 percent. The commenter stated that additional data are needed
concerning the multi-day and multi-week increases in ambient air THC
levels due to forest and agricultural fires.

Response: We disagree with the commenter in that the MACT floor
determinations were based on demonstrated performance testing and THC
monitoring from several plants over extended time periods. These
long-term monitoring test data included ambient background THC levels
and their associated variability. 

THC standards are adequate

Comment: One commenter (2144) stated that the proposed 30-day emission
standards of 7 and 6 ppm for existing and new kilns, respectively,
appear to be adequate. These values equate to roughly half of the BACT
VOC limits when relying on THC as the surrogate. According to the
commenter, at many kilns the primary contributors to THC emissions are
additives acquired as wastes from other industrial sources such as steel
mill scale. Excessive VOC/THC emissions problems in Florida were
typically solved by acquisition of less oily mill scale. Another
solution is to introduce the oily additives directly into the kiln
system where they can be thoroughly burned thus avoiding low and medium
temperature THC evolution from colder parts of kilns and preheaters.
This commenter agreed with EPA regarding regenerative thermal oxidation
(RTO) or ACI particularly when naturally existing kerogen in native raw
materials contributes to THC emissions. The commenter also noted that
the Polysius Environmental Technology (POLVITEC) available since the
early 1990’s includes a coke filter that has been demonstrated to
reduce THC and SO2 while storing mercury until it can be removed to a
very high degree by filter dust shuttling.

Response: See the preamble to the final rule explaining the basis for
the final THC standard.

Shorter averaging period

Comment: One commenter (2142) supported a shorter averaging time (24
hours or less) with regard to hydrocarbon emissions, but agreed with
EPA’s proposed limits and 30-day averaging time.

Response: EPA acknowledges the commenter’s support.

POM and naphthalene not adequately addressed

Comment: One commenter (2828) expressed concern that the proposed MACT
rules do not establish MACT limits or properly evaluate control
technologies that could greatly reduce emissions of polycyclic organic
matter (POM) , naphthalene, or dioxins/furans. The commenter agreed with
EPA that limits should be established for organic hazardous air
pollutants, and the choice of VOC or total non-methane hydrocarbons
(TNMHC) seem like reasonable options as surrogates, but emissions of POM
and naphthalene are likely to be unrelated to emissions of the other
organic HAPs. The commenter stated that traditional organic HAPs are
primarily driven by hydrocarbons in the raw material mix. Variations in
the hydrocarbon composition of the limestone seemed to be the largest
contributor to variability in organic HAP and THC emissions.

Because the processes resulting in POM and naphthalene formation are not
the same as those resulting in VOC or benzene, toluene, ethylbenzene,
and xylene (BTEX) compounds, cement kilns that are low VOC and BTEX
compound emitters may be high emitters of POM and naphthalene. The
commenter encouraged EPA to move forward with establishing final MACT
rules proposed in this rulemaking, including rules for organic HAPs, but
suggested that EPA consider the difference in the formation mechanisms
when deciding how to establish MACT standards for all organic compounds,
and potentially separate POM and naphthalene from the other organic
compounds in establishing MACT standards and appropriate monitoring
protocols. Further, EPA should undertake additional study of POM and
naphthalene emission rates from cement kilns throughout the U.S., to
determine if there is tight correlation between emissions of these
compounds and the other organic HAPs. If not, EPA should then establish
POM and naphthalene MACT standards, independently of the standards for
the other organic HAP compounds.

Response: EPA reviewed test reports for emissions data on THC and
organic HAP. Based on those test data, we identified specific organic
HAP that were measured with sufficient frequency that we could and at
levels that would allow us to establish an emission limit for organic
HAP as an alternative to the THC limit. Naphthalene emissions were
included in the analysis of organic HAP as an alternative to the limits
for THC. It is regulated under the final rule in that the definition of
organic HAP specifies naphthalene as one of the organic HAP that must be
measured to determine compliance with the organic HAP limit. In
evaluating emission test reports, the test results for POM were not
sufficient to include as one of the organic HAP that must be measured.
However, we believe that the limits for THC and organic HAP serve as
surrogates for POM and that POM emissions will be reduced as a result.

Organic HAP alternative

Comment: Regarding the organic HAP alternative to THC, several
commenters (2845, 2859, and 2874) stated that the organic HAP standards
should be raised to reflect the commenter’s proposed THC standards of
27.4 ppmv and organic HAP standard of 6.6 ppmv (24 percent of 27.4) for
existing sources and 24 ppmv and organic HAP standard of 5.7 ppmv (24
percent of 24) for new sources. 

Commenter 2863 stated that the proposed THC standard of 7 ppmvd for
existing sources and 6 ppmvd for new sources is inadequate and
requested that EPA not revise the organic HAP standard until it has time
to better evaluate existing data and that a separate subcategory for
kiln exhaust gases passed through an indirect fired, inline coal mill.
Also, the volatility of the coal being used should be considered in
establishing standards for inline coal mills. As part of the final rule,
EPA should include a specific exclusion for coal mills from coverage
under subpart Y NSPS so plants will not have to deal with dual and
potentially conflicting regulations

Commenters 2859, 2845, 2863, and 2874 objected to the option of
demonstrating compliance with the alternative equivalent organic HAP
emission limit rather than the THC limit for companies that have
naturally low hydrocarbon emissions, very low organic HAP emissions, and
use Method 320 to demonstrate organic HAP compliance for the following
reasons:

The organic HAP standard does not account for variability in THC levels
when using Method 320 to demonstrate compliance. The organic HAP rate
may be substantially lower than the 2 ppm indicated in the proposed
rule; however, the concurrent THC concentration levels may be
substantially higher than the 6.7 ppm even if testing during typical
conditions.

Units should be permitted to extrapolate the same factor for organic HAP
level and THC compliance limit set during the organic HAP compliance
demonstration test.

As an alternative, kilns should be given an opportunity to establish and
take into consideration THC variability before the THC limit is set
during the organic HAP compliance demonstration test and be allowed the
option to install CEMS before the organic HAP compliance demonstration
test in which THC and organic HAP emission rates could be collected for
6 months prior to the demonstration test. The operator would have to
demonstration that operating conditions are representative of the
6-month period at the time of the stack test and once the compliance is
demonstrated EPA could set the application THC limit at 99 percent
variability for that 6-month period for the unit.

Facilities that choose to comply based on actual organic HAP should have
the option to demonstrate compliance whenever THC emission exceeds the
limit set during organic HAP compliance demonstration test because
limestone deposits are not homogenous creating the possibility that new
high THC seam might be encountered and thus raising the possibility that
THC will go up but organic HAP emissions may remain constant. The
facility should be given 6 months to retest for organic HAP if the
alternative THC limit is exceeded and if the kin complied with the
organic HAP limit, no violation should be found related to the high THC
emissions.

Commenter 2844 recommended that EPA adopt the proposed THC compliance
alternative for sources whose organic HAPs are a low percentage of their
total THC. This compliance alternative is important for kilns whose THC
emissions may be higher due to other factors but who have established
that HAPs are a low percentage of those emissions. 

Commenter 2889 stated that EPA does not reference the study that
provided the basis for the HAPS/THC ratio. Considering the importance of
this alternative standard, this study should be cited. It should also be
evaluated to determine if the 0.24 HAP/THC ratio is, in fact,
representative of the cement industry. 

Response: See discussion of THC and organic HAP in the preamble.

Emission limits for raw material dryers

Comment: Several commenters (2845, 2859, 2863, 2874, and 2910) stated
that it is inappropriate to set the same emission limits, and controls,
for kilns and raw material dryers and that EPA establish MACT floors for
new and existing sources of raw material dryers using data from the
dryers. Cement kilns and raw material dryers differ in the following
ways:

The heat source in the kiln is designed to reach high temperatures for
calcining and combing raw materials to make clinker; the heat source in
the raw material dryer is designed to remove free water from the raw
materials. of the heat source in the raw material dryer is for removing
free water from the raw materials. Because the temperature within the
kiln system and tower is above 600 degrees F, hydrocarbons present in
the feed volatilize. In contrast, the temperature in a raw material
dryer is significantly lower, resulting in a completely different
profile of hydrocarbons volatilized.

The volume of gasses drawn through kiln systems is 10 times greater than
the volume of gasses passing through the raw material dryer.

The proposed THC limits for dryers expressed at 7 percent oxygen are
equivalent to a 1 ppm actual emission concentration standard at the
material dryer. 

Facilities may intermittently use raw material dryers. Commenter 2910
stated that the cost to install the required monitoring equipment and
possible emissions control devices would prove to be more costly than
EPA has assessed.

Based on the substantial differences between kilns and raw material
dryers, there is no basis to simply assign the proposed kiln THC
standard to raw material dryers. EPA should withdraw the limit and more
closely evaluate actual raw material dryer emissions in order to confirm
that these units generate organic HAPS to a level necessitating
regulation and if they do establish the data necessary to support an
appropriate limit. Commenter 2845 has initiated a process to gather
additional THC emissions data, data to more thoroughly assess raw
material dryer THC emission performance, and organic HAP data for
standalone raw material dryers. These data will assist EPA is developing
more appropriate standards.

Commenter 2845 stated that when defining the THC limit for raw material
dryers at major sources in §63.1346(a), the proposed rule does not
include the alternate option of complying with a total organic HAP
limit. The commenter suggested that the language should be updated to
reflect the same option as allowed in (b) for area sources.

Response: We did not have sufficient data, information, or justification
to subcategorize raw material dryers as suggested by the commenter. We
were working with a limited data set of kilns and dryers and the MACT
floor is based on the performance of the two lowest emitting kilns. The
additional THC data was received too late to be considered in this
rulemaking. These data were not submitted to EPA until mid-June 2010,
virtually too late for any consideration, much less considered in the
MACT floor analysis and subcategorization.   

Regarding the alternate total organic HAP limit for raw material dryers,
we have corrected the final rule to state the alternate option of
complying with a total organic HAP limit applies to raw material dryers
at major and area sources.

Subcategorize based on coal grinding system

Comment: One commenter (3218) requested that EPA consider developing
subcategories that reflect the different configurations for coal
grinding systems. There are two main configurations, where gases are
vented from the coal system directly to atmosphere out of a separate
stack and where coal system gases are vented to atmosphere after being
comingled with exhaust gases from the kiln and vented out of a common
stack.

Response: EPA did consider subcategories based on the coal grinding
configurations suggested by the commenter but decided against creating
separate subcategories. See the preamble discussion under the THC limits
for more information on EPA’s decision.

2.6.4	Compliance Assurance

Relative accuracy test audit

Comment: One commenter (2622) stated that the PS included in the
proposal for certifying the VOC monitor does not require a Relative
Accuracy Test Audit (RATA). The proposal would also apply Appendix F to
the VOC monitor, which requires annual RATA. There is an inconsistency
in not requiring a RATA for certification, but requiring it annually
after certification.

Since the proposal calls for RATA on the VOC monitors when the emission
limits are low (emission limits less than 10 ppm), EPA will need to set
a low emission criteria for passing a RATA. The commenter expressed
concern that 20 percent of the reference method or 10 percent of the
emission limit will be too restrictive.

Response: The final rule has been revised to require the certification
of the CEMS using Performance Specification 8, which does require a
Relative Accuracy Test Audit.  The performance criteria have not been
changed.

Monitoring at low concentrations

Comment: Several commenters (2844 and 2845) stated that monitoring THC
at the 6-7 ppm level will present significant challenges, noting that
hydrocarbon CEMS have been used successfully on hazardous waste burning
kilns but at concentrations greater than three times those required by
this proposed rule (i.e., 20 ppm). Furthermore, when a hazardous waste
kiln approaches the limit of 20 ppm, it triggers an automatic waste fuel
cut-off, which ensures that the facility is in compliance with the
standard as there are no hazardous wastes in the combustion chamber
under higher hydrocarbon concentrations. This approach will not be
available for non hazardous-waste burning cement kilns as the THC
concentration approaches or exceeds 6-7 ppm. EPA should raise the limit
to account for the variability associated with the raw materials and the
inability to change raw material input to comply with the standard. Once
installed it is expected that the CEM up-time and data availability will
be less than the 90 percent required by EPA due to the learning curve
that will be associated with operating and maintaining these devices.
The commenter stated that hot/wet straight extractive systems are
difficult to maintain on a continuous basis due to moisture condensation
in heated sampling systems, often long calibration times, and monitor
wear and tear from the corrosive environment. Some of the challenges
that will need to be addressed include: the difficulty of monitoring
hydrocarbons at very low concentrations, the accuracy of calibration
materials as 1 ppm propane standards are the lowest available, and the
variability of hydrocarbon compounds in the limestone used to
manufacture cement.

One commenter (2889) stated that the variability in flame ionization
detectors (FID) response factors for organic compounds changes the
stringency of the THC standard for some kilns. It is well known that THC
monitors using FID have different response factors for different organic
compounds. Cement kilns having organic compounds with a higher set of
response factors will be subject to more stringent emission limits than
kilns having organic compounds with lower response factors. Considering
that there are very little data concerning the ranges of specific
organic compounds in the kiln effluent gas stream, the impact of the
variations in response factors cannot be determined. Additional data are
needed to determine if the THC standards of 6 ppmvd (new kilns) and 7
ppmvd (existing kilns) is equitable.

Response: Based on data received since proposal, we have revised the THC
limit to 24 ppmvd which should address the commenters’ concerns about
monitoring at very low concentrations. The standard is based on data
from different kilns with different constituents, which addresses
variability concerns. Finally, with the change to PS8, sources are not
restricted to use of FID is they do not work as well as needed in
certain applications.	

Stack testing for monovent facilities

Comment: One commenter (2888) requested that EPA allow monovent
facilities to stack test for THC, using EPA approved methods, on the
same 30-month frequency as required for Dioxin/Furan testing under the
current PC MACT.

Response: We remain convinced that CEMS are the best choice for
continuous monitoring. Sources with monovents can install stacks to
accommodate their use.

Testing to comply with speciated organic HAP alternative standard

Comment: Several commenters (2844 and 2845) stated that the Fourier
Transform Infrared Spectroscopy (FTIR) Method 320 and the American
Society for Testing and Materials (ASTM) Test Method D6348 are not
sufficiently sensitive to quantify the 1-2 ppm of total organic HAPs for
companies wishing to speciate organic compounds. Because a single test
method is not available for the organic compounds EPA has identified and
because of the low concentration levels, EPA must be flexible in test
method selection and allow use of appropriate methods to quantify and
sum these compounds. It is anticipated that at least 3 separate test
methods will be required to capture the low concentration of organic HAP
to sum to 1 or 2 ppm aggregate. 

Response: As stated above, we have raised the level of the THC limit
which should address concerns about low concentrations. We have
carefully considered detection limits when setting the standards; a more
detailed discussion of this issue is presented in the preamble. 

Comment: One commenter (2844) stated that a THC alternative option must
address THC variability for those units meeting the organic HAP
alternative limit. The commenter recommends stack testing for organic
HAPs at the same interval as that for dioxin/furans. Under the proposed
rule, a new THC limit would apply based on the level at the time of the
test. The commenter provided the following options for compliance
demonstration limits, with the suggestion that a kiln operator would be
able to choose between the three:

Demonstration of operating conditions consistent with test conditions
and good engineering practices. The ongoing compliance demonstration
would be based on two operating conditions that are indicators of
organic HAP emissions at the time of the stack test, with subsequent
continuous monitoring of those two parameters: the temperature at the
exit of the precalciner and the oxygen content of the exhaust at the top
of the preheater. 

Extrapolation to set the revised THC limit. The revised THC limit is
based on the total organic HAP measured during demonstration testing,
extrapolated to the level of compliance and then extrapolating the
average THC value recorded during the total organic HAP test to the
proportionate level. 

Accounting for measured THC variability when setting the revised THC
limit. The kiln owner installs the THC CEMS in advance of the stack
testing and gathers data for at least 6 months prior to the stack test.
At the time of the stack test, the operator must demonstrate that the
operating conditions are consistent with that 6 month period. Once
compliance is demonstrated, the applicable THC limit will be at the 99%
of variability for that 6 month period for that unit.

The commenter also suggested that if the alternative THC limit is
exceeded, the operator should be given six months to retest for organic
HAPs; if the organic HAPs are compliant in the retest, there is no
violation related to the high THC reading.

Response: The MACT floor determinations were based on demonstrated
performance testing and THC monitoring from several plants over extended
time periods. These long-term monitoring test data included ambient
background THC levels and their associated variability

Comment: One commenter (2845) stated that companies that have low
hydrocarbon emissions and low organic HAP emissions may elect to
demonstrate compliance with the speciated HAP portion of the rule and
that a new THC limit would apply based on the level at the time of the
test. However, because THC levels are variable, it could mean that 50
percent of the time the THC will be recorded over the limit even though
the performance of the unit is unchanged. According to the commenter,
there are at least 2 options, extrapolation or accounting for the THC
variability, that could be used to show practical THC limits along with
allowing the kiln operator to select the most appropriate option.

Response: Companies that elect to comply with the speciated HAP portion
of the rule should collect any background or previous test information
that would allow them to conduct their testing in a worst-case scenario
for their particular operations. The issue described by the commenter is
a key reason for having a 30-day compliance period – where any given
day the measured levels will be above or below the average. Sources also
have the option to retest if they believe the previous testing was not
representative.

Comment: One commenter (2844) requested that EPA defer promulgating a
final THC standard until more adequate data has been gathered, but urged
EPA to finalize the Alternative Compliance Option for THC as
expeditiously as possible.

Response: EPA has gathered significantly more data since proposal and
believes it ample to set the final standard.	

Measuring THC from raw material dryer

Comment: One commenter (2845) stated that the process and combustion gas
composition of the kiln system and dryer are completely different. In
the kiln, the main constituents are nitrogen and carbon dioxide. Oxygen
content of the kiln exhaust gas is typically below 2 percent at the kiln
exit and about 7-10 percent at the stack. In the raw material dryer the
main constituents are nitrogen and oxygen, with oxygen levels typically
above 19 percent. Also, if the THC concentration in the raw material
dryer is adjusted to 7 percent oxygen as proposed for the kiln, it would
require the actual emission concentration to be less than 1 ppmvd THC to
meet the proposed limit of 7 ppmvd. The accuracy of the THC analyzers
available for monitoring is not sufficient to determine compliance with
a 1 ppm standard. (1ppm THC at 19 percent oxygen = 7.3 ppm at 7 percent
oxygen). Calibration gases are manufactured at a lower concentration of
1 ppm propane, and demonstrating compliance at the level corresponding
to the lowest level calibration standard generates an unknown degree of
inaccuracy.

Response: We received no data indicating that using the same limit for
raw material dryers as used for kilns was infeasible, or that it would
otherwise allow us to set a different THC emissions limit for raw
materials dryers. Therefore, in these final amendments we are setting
the THC emissions limit at the same level as the cement kiln’s, which
is 24 ppmvd measured as propane.

However, because raw material dryers have high oxygen contents due to
their inherent operation characteristics (and not due to the addition of
dilution air), referencing the raw material dryer standard to 7 percent
oxygen would actually result in a more stringent standard than for
cement kilns. For example, given the typical oxygen contents of kiln
exhaust (7 to 12 percent), a kiln just meeting the THC limit of 24 ppmvd
would have an actual stack measurement of approximately 16 to 24 ppmvd.
If the raw material dryer standard is referenced to the same oxygen
level, they would have to meet a measured THC limit of approximately 3
ppmvd. For this reason, we are referencing the oxygen level of the
standard for raw materials dryers to 19 percent oxygen, which is the
typical oxygen level found in the exhaust of these devices. 

Flow rate monitoring

Comment: One commenter (2845) stated that EPA should not require
volumetric flow rate demonstrations during speciated HAP or THC testing.
The proposed standard is a concentration based standard, so no flow rate
monitoring data are needed for reporting in units of the standard. At 40
CFR 63.1349(b)(4)(iv), EPA incorrectly requires that the flow monitoring
system be evaluated according to procedures in Part 75, therefore this
requirement should be deleted.

Response: The commenter identified an error in the proposed standard,
which has been corrected in the final standard.	

PS-8A

Comment: One commenter (2845) stated that PS-8A should be modified for
low concentration measurements. PS-8A was developed for cement plants
burning hazardous waste. PS-8A has an explicit span requirement of 100
ppm as propane. The calibration drift and calibration error requirements
are 3 percent (3 ppm) and 5 percent (5 ppm) respectively, which are
very near the proposed compliance level. The proposed THC standard is at
6 and 7 ppmv dry at 7 percent oxygen. The criteria contained in PS-8A
will not evaluate the acceptability of the THC monitoring systems with
adequate accuracy to demonstrate compliance with the proposed standard.
Therefore, PS-8A must be completely re-worked to evaluate low
concentration level monitoring systems, without simply (and arbitrarily)
lowering the span and calibration drift specifications.

Response: The final standard has been revised to specify use of PS-8,
instead of PS-8A, because PS-8 is a better fit for this application. The
span for the RATA is specified as 40 ppm and the reference method is
established as EPA Method 25A. 

Need for Appendix F procedure

Comment: One commenter (2845) stated that EPA should develop an Appendix
F Procedure specific for hydrocarbon CEMS. Although an Appendix F,
Procedure 1 was written for SO2, NOx and diluent CEMS (O2, and CO2), it
is not directly transferrable to a THC CEMS monitoring for compliance at
extremely low concentration levels. For example, a key element of
Procedure 1 is an annual relative accuracy test. However, PS-8A does not
include a relative accuracy specification and instead correctly relies
on calibration gas audits because there is no independent reference
method. Furthermore, the cylinder gas audit levels for Procedure 1 may
not be appropriate for a revised PS-8A, especially if multi-range
analyzers are necessary. Requirements at Appendix to Subpart EEE for
hazardous waste combustors contain quality assurance procedures for THC
and other CEMS used at these facilities. These procedures should be
modified for low-level measurements and appropriate requirements should
be developed for Subpart LLL sources.

Response: The proposal identified PS-8A as the monitoring performance
specification for THC monitoring through an oversight, and PS-8 is now
specified. Also, the rule language now proscribes the THC CEMS span
value as 40 ppm and the reference method for the RATA as Method 25A.
With these revisions and in view of the THC standard concentration
value, EPA is confident that the performance requirements of Procedure 1
of Appendix F are appropriate and can be met.

CEMS alternatives

Comment: One commenter (2844) stated that EPA should not require CEMS
but instead look to other effective monitoring methods that are more
practical and efficient. The commenter’s experience with a VOC
Analyzer (CEMS) was that the unit requires a lot of maintenance,
expense, and frequent repairs, resulting in increased downtime compared
with priority pollutant CEMS.

Response: We believe that CEMS are the best alternative to demonstrate
continuous compliance and that they are a demonstrated technology, as
well as the source for all of the data used to develop the standard.
However, we are always interested in the evolution of monitoring
technologies and will consider source-specific alternatives proposed
under the General Provisions.

Contradictory, erroneous requirements and clarifications

Comment: Several commenters (2845, 2859, and 2863) stated that numerous
areas of the proposed regulations include contradictory, erroneous, or
otherwise impractical requirements relating to compliance demonstrations
in the performance testing requirements and or monitoring requirements.
Comments are as follows:

The initial compliance demonstration procedures for total hydrocarbons
appear contradictory and should be revised to be consistent with the
proposed emission standards and 30-day rolling average for THC
emissions. Specifically, §63.1350(h)(4) states that daily average THC
concentration must not exceed the THC limit for new or reconstructed
facilities which is inconsistent with §§63.1343, 63.1346, and
63.1350(h)(2) and (h)(3). The initial performance demonstration should
be based on a properly calibrated and compliant CEMS and the initial
performance test result must be based on the calculated average THC
concentration for 30 kiln operating days. A definition for kiln
operating day must be established and must reflect a specified number of
hours of kiln or raw material dryer operation at normal production
conditions. Days with fewer operating hours than a kiln operating day
should not be included in the 30 day average.

The initial compliance demonstration procedures for total hydrocarbons
specifies a performance test duration and conditions that are not
appropriate. Section 63.1349(b)(4)(i) addresses initial compliance for
THC specifying a performance test duration of 24-hours, and calculating
the daily average THC concentration from the hourly averages during the
performance test. Furthermore, it states that initial compliance shall
be demonstrated by conducting separate performance tests for in-line
kiln/raw mills with the mill operating and with the raw mill off. This
would require the facility to run in a mill-off condition for 24 hours,
which is not possible. There should be no specification for separate
initial compliance tests for mill-on and mill-off for in-line raw mills.

The THC monitoring requirements for facilities complying with total
organic HAP limits are inconsistent with the averaging period of the
standard. Section 63.1350(r)(2) requires calculation of 3-hour average
THC concentrations and stating that the 3-hour concentration may not
exceed the average established during the initial performance test is
not consistent with the basis of the limit. The site-specific THC
emission limit is based on a 30-day rolling average and EPA should
correct this apparent discrepancy.

The testing to establish THC emission limits for facilities with in-line
kiln/raw mills that choose to comply with total organic HAP limits
should be simplified to require testing at the operating condition that
is appropriate for the particular plant. Section 63.1349(b)(4)(iii)
requires performance tests to establish site-specific THC emissions
limits corresponding to compliance with the emissions limits for total
organic HAPs. For kiln/in-line raw mills, this section specifies that
tests are to be conducted while the raw mill is off and while the raw
mill is operating. This could potentially create different THC limits
for the two operating conditions which becomes completely unworkable
since compliance with the resulting standard is determined on a 30-day
rolling average and will obviously contain both mill-on and mill-off
operation. EPA should clarify that the test is to be inclusive of both
mill on and mill off conditions and the full range of plant operations
including feed materials and fuels/fuel combinations in order to develop
a single alternate limit. EPA should not specify separate tests under
mill-on and mill-off conditions. Furthermore, where total speciated
organic HAP emission tests demonstrate that emissions are below the
emission limit, EPA should explicitly allow proportional extrapolation
of the measured HAP and THC values to determine the corresponding THC
surrogate compliance limit.

Section 63.1349(b)(4)(vi) states that emission tests to determine
compliance with total inorganic HAP limits shall be repeated annually
and this is inconsistent with other sections of the proposed rule
regarding the parameters to be tested and the frequency of the test, and
this section should be changed. The word inorganic appears to be a
typographical error and EPA should confirm that the phrase should be
total organic HAP. There are no requirements for performance testing of
inorganic HAP. The annual test requirement in this section also
contradicts the proposal where it explicitly states that stack tests for
organic HAP would be required every five years for kilns and raw
material dryers meeting the organic HAP alterative to THC to determine
the organic HAP component of their THC emissions. The proposal states
that annual organic HAP tests are not necessary, will be very costly,
and could be disruptive to plant operations. EPA should specify that
testing should be conducted at least every five years. The facility
should be permitted to conduct the testing more often than every five
years and reset the alternative limit if changes in the facility
materials or operations warrant such changes. In addition, because such
testing for the organic HAP could result in THC levels higher than the
previously determined alternate limit, the time period of the testing
should not be included in the rolling average for compliance.

EPA needs to add a formula to enable sources to weight the relative
emissions contributions to determine compliance when emissions occur at
two points. Many plants with indirect fired, in-line coal mills are
designed to vent directly to atmosphere. Compliance with the organic HAP
standard for a plant so configured should allow the combination of both
stacks given that they both constitute sources of kiln gases. These
commenters requested that EPA add to the rule a formula equivalent to
Equation 2 of 63.1349 whereby an affected source configured such as the
one depicted below can demonstrate compliance with the THC standard by
combining the concentrations from the main stack and the coal mill
stack.

Response: We have made several clarifications and corrections in the
final rule to respond to these comments, including:

Clarifying that the 30-day average is a rolling average,

Adding a definition of “kiln operating day,”

Raw mill off testing is required for the initial compliance test to
establish the site-specific THC level.   

Removed paragraph 63.1350(r)(2), because this language is inconsistent
with the 30-day rolling average.

Replaced the reference to “inorganic” HAP emissions testing with
“organic” HAP emissions testing.

Clarified that organic HAP testing is to occur at least every 30 months
unless there are changes at the plant that warrant more frequent
testing.

Clarified that data collected during a performance retest need not be
incorporated into the existing 30-day rolling compliance average.

As described elsewhere, we agree with comments that the initial
performance test period should also be a 30-day average. Regarding the
comment requesting proportional extrapolation of the measured HAP and
THC values, we lack the data to establish such a provision. If a source
needs to expand its operating range, it should retest and establish a
new limit. Finally, the suggestion to add a formula to address combined
concentrations from the main stack and coal stack is not needed, because
of the concentration format of the standard. Such a formula is only
needed when the standard is in the form of a mass limit.

Comment: One commenter (2845) stated that the definition of total
organic HAP in section 63.1341 should be changed to add the reference to
test methods EPA Method 18 and ASTM D6420, and adsorbent based methods
or impinge train methods to test for total organic HAP for speciating
organic HAP.

Response: The final rule includes a higher THC value than proposed, with
the result that we do not think these additional methods are needed to
determine total organic HAP. However, a source is always able to request
an alternative method under the General Provisions.

Comment: One commenter (2845) provided the following suggestions and
clarifications to §63.1349(b)(4)(i):

Clarify that Alkali Bypass Stacks are exempted from CEM requirements for
THC.

A performance test for a source with a CEMS is duplicative and should
not be required.

The THC limit is a 30-day average (see §63.1350(h)(3)). Thus the
requirement for the duration of a performance test to be 24 hours
appears to be a typographical error.

This section appears to require a 24-hour test while the raw mill is
operating and a 24-hour test while the raw mill is not operating. A
typical downtime for a raw mill is 8 hours per day, one day per week. It
is not possible for the majority of in-line kiln/raw mill systems to
operate with the raw mill off for 24-hours.

Raw Material Dryers also have THC limits, and thus may be required to
use a continuous emission monitor and demonstrate compliance during raw
mill on/off conditions based on the proposed rule. It should be
clarified that the raw mill on/off condition is not needed for the
performance test of the Raw Material Dryer as it is not directly
associated with the kiln system.

This requirement should be removed or updated, according to language
provided.

Response: Alkali Bypass Stacks are subject to the standard.  However, we
have changed the final rule to allow the source to use the results of an
initial performance and subsequent compliance test as a surrogate for
continuous monitoring. We believe this is reasonable, given that the
majority of emissions are coming from the main stack.  We have clarified
that the initial compliance test for the CEMS is the initial 30-day
operating period.  For the raw material dryer, testing is not required
for the raw mill on/off conditions.

Comment: One commenter (2845) suggested that §63.1350(h)(2) and (3) be
combined into one requirement.

Response: We believe it is clearer to distinguish existing source
requirements from new and reconstructed sources in the format of the
standards.

Comment: One commenter (2845) suggested that the combustion system
inspection and reporting required under §63.1350(i), §63.1350(a)(3),
and §63.1354(b)(9)(iv)are no longer needed and can be removed. The
combustion system inspection was previously included in lieu of a THC
limit. Now that there is a THC limit, these inspection, recordkeeping
and reporting requirements are no longer needed.

Response: These requirements were actually part of the 1999 final rule
and were intended to help ensure compliance with the dioxin/furan
emissions limit. Because the dioxin/furan emission limit has not
changed, these requirements are still applicable.

Comment: One commenter (2845) included the suggestion that the EPA
should clarify that THC is on a propane basis and is a 30-day rolling
average at 7 percent oxygen. 

Response: We agree and the final rule contains this clarification.

2.6.5	Impacts

Performance of controls overstated, impacts understated

Comment: Several comments (2841, 2844, 2845, 2864, and 2889) were
received that question EPA’s impact analysis for THC including the
performance of ACI and RTO, and adverse impacts associated with the use
of RTO.

Commenter 2845 stated that EPA has overestimated the use of ACI and
underestimated the use of RTO. According to the commenter, the EPA based
the ACI information on one single kiln that is no longer in service with
emissions and a configuration that was unusual for the industry. EPA
stated that 75 percent control should be achieved for cement kilns using
ACI because it was achieved for dioxin/furans in the incinerator
industry. The commenter disagreed with the comparison for the following
reasons:

There are substantial differences between cement kilns and incinerators
in operations, plant configurations, and flue gas composition as well
being two different industries, which can affect the effectiveness of
ACI for THC.

THC cannot be compared with dioxins/furans because THC also consists of
low carbon-based VOC (such as methane, ethane, and formaldehyde), which
has a substantial impact on the effectiveness of ACI. 

Formaldehyde is not expected to be well controlled using ACI.

The ability of ACI to capture THC is total a physisorption-based
mechanism and therefore the effectiveness of ACI would be also entirely
mass-transfer-limited.

Lower THC concentrations would have to have an increase in the ACI feed
rate to match the sample percentage removed of THC at higher
concentrations.

The percentage of low-carbon THC is highly variable and the available
data indicates that low-carbon species can be <5 to >90 percent of THC,
which will impact the effectiveness of ACI.

Using ACI would have an adverse impact on the ability to reuse cement
kiln dust. Recycling the CKD would result in revolatilization of some of
the THC. Insufflation of a portion of CKD is possible for some kilns,
which would destroy the captured THC. It may be necessary to add a
polishing baghouse thereby increasing costs 10 to 15 times that of using
ACI alone.

Dioxins/Furans may be increased when using ACI to control THC for cement
plants that burn hazardous waste. 

Several commenters (2844 and 2845) noted that EPA has proposed using ACI
to provide simultaneous control for both mercury and THC. Although a
standard powdered activated carbon may control both mercury and THC
(greater than 90 percent removal), there is no data to support the
removal of THC when a halogenated carbon was used to control high
elemental mercury concentrations. 

Commenters 2864 and 2889 expressed concerns about operating problems
caused by a HCl wet scrubber ahead of a RTO. Commenter 2864 provided an
example of their experience with RTO with wet scrubbers at two wet
kilns. Although 89 percent THC removal was reported initially, the
system was plagued with problems and equipment failure. After six years
of repairs and upgrades, the system worked in the 2008 ozone season and
the plant was shut down permanently later that year. Commenter 2889
stated that EPA has not considered the costs due to kiln outages caused
by the unreliability of wet scrubbers and RTO required by the proposed
amendments. Both wet scrubbers and RTO in the cement industry and other
industries have been subject to frequent maintenance problems. Existing
kilns installing wet scrubbers in series with a RTO will compound the
reliability issue. Problems affecting either one of the control systems
will require the kiln to shut down. The cost for the lost production
should be included in these cost impact analyses.

Commenters 2844 and 2889 stated that EPA’s claimed activated carbon
injection system THC control efficiency of 75 to 80 percent is not
adequately supported and does not take into account the low control
efficiency for light hydrocarbons. As a consequence of the
unrealistically highly assumed ACI removal efficiency for THC, EPA has
underestimated the number of RTO that will be needed to achieve the
standard for existing kilns. The cost, environmental, energy, and
economic impacts of the proposed amendments described in Section VI of
the Preamble are inaccurate due to the probable need for numerous RTO. 

Commenter 2844 stated that RTO are not expected to provide for
continuous compliance with the limit. Vendors have not been willing to
guarantee levels below the proposed standards of 6 and 7 ppm. The one
commercial application of an RTO on a kiln cannot meet the standard, and
a number of units have such high THCs that RTO may not be able to lower
the emissions sufficiently to get close to the proposed limit. Where the
concentrations of THC in kiln stacks is low, but still over the
standard, the emissions reductions that an RTO can provide are small.

Commenters 2841, 2844, 2845, and 2889 stated that the requirement of
regenerative thermal oxidizers (RTO’s) for control of THC will result
in an increase of CO2 emissions, along with a host of other emission
impacts due to increased fuel and energy consumption. 

One commenter (2845) stated the following regarding the use of RTO’s
for THC control:

RTO will increase cost, energy use, and greenhouse gas emissions. 

RTO have performance limitations. Based on available data, RTO have a
performance limit of 10 ppm THC which restricts the emission control
capabilities. Therefore, the THC emission standards should be no less
than 10 ppm.

RTO are not cost-effective due to high maintenance costs (especially in
harsh weather). Maintenance includes component replacements and periodic
cleaning and washing.

Numerous cement plants lack access or have limited access to the natural
gas need to operate RTO.

RTO use will increase green house gas emissions with higher NOx
emissions due to the use of natural gas or propane used for firing. The
commenter included data concerning the increase greenhouse gas emissions
with their comment.

Commenter 2900 provided stack test data for THC emissions. The data
demonstrates that the plant may have three or more times the new EPA
limit in emissions and may be unable to meet compliance. 

Response: Hazardous waste burning cement kilns are not subject to the
subpart LLL standards. The proposed level of control for THC is exceeded
in practice by the only cement plant in the U.S. equipped with an ACI
system. With regards to the comment regarding dioxins/furans emissions
potentially increasing as a result of these rule requirements, hazardous
waste burning cement kilns are not subject to subpart LLL standards. For
those comments related to control device performance and the secondary
impacts of increased CO2 and GHG emissions from oxidizers, those impacts
were considered in our analysis and were based on the best information
available to us. If a given facility wants to avoid such secondary
impacts, the organic HAP limit is a viable compliance option for them.

Comment: One commenter (2928) stated that RTO are technically feasible
and cost effective options that can be used to meet the proposed
standards. TXI operates the largest cement kiln in the state of Texas in
Midlothian with an RTO acting as an effective afterburner for VOC, THC
and CO emissions. It is 80-90 percent effective in removing these
contaminants from the plume. As has already been noted, SCR technology
also significantly reduces VOCs/THCs from the exhaust.

Response: EPA does not anticipate that widespread use of RTOs will be
needed to comply with the THC standard. See Regulatory Impact Analysis:
Amendments to the NESHAP and NSPS for the Portland Cement Manufacturing
Industry, Final Report, August 2010.

Baseline THC emissions

Comment: One commenter (2845) stated that when evaluating the emissions
database for each kiln EPA used, the standardized format is problematic
from multiple standpoints. Here are several related examples:

In reviewing data for those facilities for which EPA had actual data,
there appears to be the following types of issues:

Projected emission rates were extrapolated from a particular time period
of production (2002 production extrapolated to 2005) that does not
necessarily reflect the actual emissions for a particular kiln

Kiln emission allocations for facilities with multiple kilns are not
consistent, with the following examples:

Main stack emission rates where 3 kilns feed a common stack are applied
to each of the three kilns individually for some pollutants; thereby
overestimating individual kiln emissions.

Emission rates are apportioned improperly between the multiple kilns.

Deriving factors to extrapolate HAP emissions from those facilities
where EPA has data to the remainder of the kilns, does not have a sound
justification as a basis for such detailed cost calculations. For the
THC factor, removing facilities with high levels of THC is not
warranted. With many facilities not having THC data, and THC emissions
being based on geographic location, other cases of higher THC emissions
is a realistic outcome. EPA’s factor of 21 ppm was compared to the
average of industry’s corresponding database. Removing only the
highest THC emission kiln (which is now closed), the average THC
emission rate was calculated to be 28.1 ppm.

The use of dry versus actual flow rates appears to be used
inconsistently.

The assumption that most plants burn coal is not accurate and will
affect flow rates, and therefore emissions.

The data includes some HWC (waste burning kilns) and kilns currently
shut down. The inclusion of this type of data skews the database.

Factors established by kiln type may not be as relevant as factoring by
geographic area, as the newly regulated HAPs generally are variable due
to the limestone and/or silica source quarried on-site.

Response: The database for this rulemaking has been updated to account
for the kilns active in 2009 that are subject to this rule. Hazardous
waste kilns are included in the database for the ISIS Cement Model but
are not subject to this rule. THC emissions for each kiln subject to
this rule were calculated using emission factors developed from the data
EPA received thru the Information Collection Request (ICR). For kilns
where we did not have specific emissions information, we developed an
average emissions factor. We removed the highest emitting kiln because
we believe it represents an outlier. We know of no other facility that
has emission even close to the highest THC emitting kiln. We recognize
that using an average emissions factor may overestimate or underestimate
costs. However, we know of no more accurate way to estimates emissions
and costs. 

 2.6.6	Other

THC emissions due to burning waste vs due to limestone

Comment: One commenter (2928) supported the proposed emission standard
for THC. The commenter noted that cement industry will argue that the
entireties of their THC and CO emissions are from the cooking of the
limestone that contains these chemicals naturally. EPA should be
skeptical of this claim without proof when so many cement plants are
burning wastes, including tires, asphalt shingles and oil filter fluff
that contribute directly to these kinds of emissions. Industry must
distinguish between those THC emissions legitimately caused by raw
materials and those cause by the burning of waste fuels.

Response: EPA acknowledges the commenter’s support and notes that the
expansion of the database for THC more accurately reflects performance
than the more limited information at proposal.

2.7	HCl Standards

Comment: One commenter (2898) recommended that EPA adopt the proposed
MACT floor for HCl. 

Response: EPA acknowledges the support of this commenter, although the
proposed standard did not properly account for variability in the
analytic detection method due to its quantitation limit.

2.7.1	Legal Issues

Health based vs MACT emission standards

Comment: Many commenters expressed their preference for EPA’s previous
health-based approach to setting limits for HCl. 

Commenter 2830 requested that EPA clearly articulate why it did not
reaffirm its position in the original proposed NESHAP where it found
that an HCl standard is not necessary given the low HCl emissions from
cement kilns and risk assessment findings from a national survey of the
industry showing a lack of risk to health or the environment from any
cement facility. There is no compelling environmental or public health
reason for EPA to change course at this time, and relying on §112(d)(4)
would be consistent with Congressional intent to not impose burdensome
HAPs regulations where no risk to public health or the environment
exists.

Commenters 2841, 2844, 2859, and 2890 stated that in previous MACT
rulemakings, the Agency determined that HCl’s nature as a threshold
pollutant and the clearly demonstrated lack of risk posed by HCl
emissions at the property boundaries of cement manufacturing facilities
precluded the need to set such a standard. EPA’s reversal of this
approach in this latest proposed rulemaking runs completely counter to
its approach in other rulemakings.

Commenters 2844, 2845, 2859, 2864, 2874, 2901, 2915, and 2916 stated
that to the extent that EPA believes it needs to address HCl in this
rulemaking, the standard should continue to be health-based. 

Commenter 2845 proposes that sources be allowed to set a site-specific
limit, ensuring that a national standard of a hazard index of 1.0 is not
exceeded, thereby demonstrating an ample margin of safety. This approach
was adopted in EPA’s hazardous waste combustor MACT rule and the
industrial boiler MACT rule (for HCl). There is no compelling
environmental or public health reason for EPA to change course at this
time, and relying on §112(d)(4) would be consistent with Congressional
intent to not impose burdensome HAPs regulations where no risk to public
health or the environment exists. 

Commenters 2845, 2874, and 2901 stated that EPA should reaffirm the
position it took in the December 2006 Portland Cement NESHAP rule that
in accordance with §112(d)(4) that it is not necessary to set a MACT
standard for HCl. In two MACT rulemakings for other industry sectors,
EPA has decided that an HCl standard was not warranted where EPA
determined that even the individuals with the greatest exposure to HCl
emissions from neighboring facilities would be safe with an ample margin
of safety. EPA now proposes to reverse course because controls installed
to reduce HCl emissions would supposedly result in emissions reductions
of sulfur dioxide and other non-HAP pollutants. MACT standards, however,
may only address HAPs. The commenter states that it is inconsistent with
CAA §112(d) and arbitrary for EPA to depart from precedents in the 2006
and other earlier MACT rulemakings to set an HCl standard for the
purpose of reducing criteria and other non-HAP pollutants.

If EPA does decide to set a standard, commenters 2845, 2874, and 2901
stated that EPA should follow the approach it adopted in its hazardous
waste combustor MACT rule (for total chlorine) and the industrial boiler
MACT rule (for HCl), where EPA established health-based compliance
alternatives. In those rulemakings, EPA set a national standard such
that for someone living at the fence line of the facility, the hazard
index could not exceed 1.0 (thus providing an ample margin of safety).
The facility then would be able to set a site-specific limit, ensuring
that the hazard index did not exceed 1.0.

Commenters 2845, 2874, and 2901 noted that EPA’s 23 ppm by volume
(ppmv) suggested not-to-exceed limit is an order of magnitude less
stringent than the proposed 2.0 ppmv existing source standard for HCl.
Thus, if EPA adopts the standard in the Proposal, it will be setting a
standard that is an order of magnitude more stringent than the level
(using extremely conservative assumptions) necessary to protect human
health. Commenters 2844 and 2901 suggested that if EPA decides that an
HCl limit is needed, the Agency should set a limit based on protection
of health at 23 ppmv. 

Commenter 2901 stated that although some might argue that HCl is not a
threshold pollutant because it has not been shown to be
non-carcinogenic, Congress stated that ammonia was a HAP with a
well-established threshold for which EPA could set a risk-based
standard. A comparison of the Integrated Risk Information System (IRIS)
information relating to carcinogenicity for ammonia and HCl shows that
the two chemicals have the same notation relating to carcinogenicity,
not undergone a complete evaluation and determination under U.S. EPA’s
IRIS program for evidence of human carcinogenic potential. 

Commenters 2845 and 2846 stated that the proposal solicits comment on
whether EPA should promulgate a health-based standard for HCl under
§112(d)(4) that would be in the form of a not-to-exceed standard, based
on risk analysis of 68 facilities using a screening level dispersion
model (AERSCREEN). EPA states that such a not-to-exceed standard would
be set at a level of 23 ppmv. 

Commenter 2845 disagreed that EPA must issue a standard.

Commenter 2846 agreed that a health-based approach will allow facilities
alternative compliance options so that capital and operating costs are
not incurred solely on controls for control’s sake, for example, where
there is no HCl health risk and where investing in a control technology
brings minimal or no co-benefit.

Commenter 2901 stated that EPA’s analysis for this proposed rule shows
that 23 ppmv emitted from any plant would result in an annual average
ground concentration of more than 20 µg/m3. Because the RfC for HCl is
20 µg/m3 an emission standard of 23 ppmv would be protective for any
facility. The commenter stated that a 2 ppmv technology-based standard
is unnecessary when EPA has already concluded that 23 ppmv is protective
with an ample margin of safety, and why Congress gave EPA the
flexibility to use a risk-based alternative. As Congress noted in the
legislative history, §112(d)(4) was enacted to avoid expenditures by
regulated entities which do not have public health or environmental
benefit.

Response: Most of these issues are responded to in the preamble in
section IV.A.1.d.iv. Given the enormous environmental benefits projected
from a MACT standard for HCl, EPA rejects the characterization that a
MACT standard would be ‘control for control’s sake’.

Comment: Several commenters (2816, 2898 and 2911) agreed with EPA’s
decision to not propose a health-based standard. The commenters agreed
that a health-based standard should not replace a MACT standard for HCl.
The CAA allows the EPA to consider the health threshold for a HAP in
setting emission standards, if one has been established. 42 U.S.C.
§7412(d)(4). The commenter noted that the use of the term has been is
retrospective, and shows that Congress intended that EPA may only use
the authority when an accepted threshold is already in existence.
Further, Congress specified the kind of proof that EPA needed in order
to invoke subsection 112(d)(4): direct evidence of no effects. The
Senate report on the 1990 CAA stated that the provision allows EPA to
use the no observable effects level (NOEL). Thus, unless there is clear
evidence of an existing threshold for the pollutants in question, EPA
may not rely on subsection 112(d)(4) at all. The Agency may not
speculate from inconclusive science; rather, it must be sure that there
are no effects from exposure at the level chosen for the emission
standard.

Commenter 2898 stated that the 2006 rule that EPA relies on for a
possible alternative health-based standard has been challenged in court
and is being held in abeyance pending the results of EPA’s
reconsideration of the rule, carried out in the present rulemaking (74
FR 21139-40, 21154). Thus, the 2006 rule cannot be said to have
established a health threshold. 

According to commenter 2911, HCl is not the only acid gas that cement
plants emit. Among the other acid gases that cement plants emit are at
least two listed hazardous air pollutants, hydrogen fluoride and
hydrogen cyanide. EPA has a clear statutory duty to set emission
standards for each HAP that a source emits and, even if EPA could set
§112(d)(4) standards for HCl, the agency does not and cannot claim that
there is an established threshold either for hydrogen fluoride or for
hydrogen cyanide. Further, because EPA must set standards for the latter
pollutants under §112(d)(2) and (3), a §112(d)(4) standard for HCl
could not possibly serve as a surrogate for either of them. To satisfy
the CAA, EPA must set individual standards for both hydrogen fluoride
and hydrogen cyanide or must use a valid surrogate standard for them.

Response: EPA thanks the commenters for their support on our decision to
not set health based standard for HCl. Regarding the suggestion that EPA
set limits for HG and HCN, EPA notes that a factor in not setting a
health-based standard was that setting a MACT standard for HCl not only
controlled HCl but also co-controlled other HAP (such as HF, Chlorine
(Cl2), and hydrogen cyanide (HCN)) and criteria pollutants yielding very
substantial environmental benefits. EPA agrees that HF and HCN emissions
from cement manufacturing should be reduced when possible and believes
that the final rule helps to accomplish that end.

Claiming benefits from SO2 reduction not permissible

Comment: Several commenters (2841, 2844, 2846, and 2916) stated that
MACT standards may only address HAP pollutants, but EPA is trying to
control sulfur dioxide and other non-HAP pollutants through application
of control technology used to reduce HCl.

Commenter 2846 stated that basing MACT limits on such considerations is
not permissible and limiting compliance flexibility by excluding other
options is not good policy. Section 112(d)(2) provides an express list
of factors that EPA may consider in setting MACT standards – including
the cost of achieving such emission reduction, and any non-air quality
health and environmental impacts and energy requirements. This list does
not allow consideration of non-HAP air quality benefits, such as the
co-benefits of reducing PM2.5 emissions.

One commenter (2898) stated that Congress intended EPA to consider
benefits such as a reduction in emissions of SO2 and other acid gases,
along with condensable PM, ammonia, and semivolatile compounds when
setting a MACT floor, and EPA has appropriately considered these
co-benefits in deciding not to set health-based standards for HCl. 

Response: Although MACT standards may directly regulate only HAPs and
not criteria pollutants, Congress did recognize, in the legislative
history to section 112(d)(4), that MACT standards would have the
collateral benefit of controlling criteria pollutants as well and viewed
this as an important benefit of the air toxics program.. See S. Rep. No.
101-228, 101st Cong. 1st sess. at 172. EPA consequently does not accept
the argument that it cannot consider reductions of criteria pollutants,
for example in determining whether to take or not take certain
discretionary actions, such as whether to adopt a risk-based standard
under section 112 (d)(4) or to adopt standards for HAP metals that truly
reflect MACT by amending the PM standard. There appears to be no valid
reason that, where EPA has discretion in what type of standard to adopt,
EPA must ignore controls which further the health and environmental
outcomes at which section 112 (d) of the Act is fundamentally aimed
because such controls not only reduce HAP emissions but emissions of
other air pollutants as well. Thus, commenter 2844’s invocation of the
decision factors enumerated in section 112 (d)(2) seems off point, since
the issue being addressed is not whether to regulate non-HAP (EPA is
not), whether to consider other air quality benefits in setting section
112 (d)(2) standards (EPA is not), but rather whether to make the
discretionary choice to regulate certain HAP based on the MACT approach
and whether EPA must put blinders on and ignore collateral environmental
benefits when choosing whether or not to exercise that discretion. EPA
knows of no principle in law or common sense that precludes it from
doing so. 

 With regard to the comments specifically addressing SO2, EPA is
regulating SO2 emissions for new sources and modified sources under the
NSPS rulemaking. EPA is not regulating SO2 in the NESHAP, although EPA
recognizes that any other SO2 emission reduction obtained from existing
sources is a co-benefit resulting from the control technology selected
to comply with either the mercury and/or HCl NESHAP limits. EPA
recognizes the multi pollutant interactions and co-benefits of this
rulemaking and how a well planned integrated control strategy can
maximize the environmental results while optimizing the cost invested by
a facility to comply with such regulation. 

SO2 as surrogate

Comment: Several commenters (2898 and 2911) agreed with EPA’s decision
not to propose using SO2 as a surrogate for HCl. HCl is not present in
SO2, and control of SO2 is not the only means by which cement kilns
achieve reductions in HCl. To the contrary some kilns achieve HCl
reductions through the use of raw materials and fuels that contain less
chlorine.

Response: We agree with the comments and are not using SO2 as a
surrogate for HCl.

2.7.2	Adequacy of Floor Determinations (Existing and New)

Adequacy of data

Comment: One commenter (2144) agreed with EPA’s HCl emission limit of
2 ppmvd for existing sources or a more stringent standard if the
statistical methodology biased the results on the high side. The
commenter agrees with the 0.1 ppmvd for new kilns unless the statistical
methodology justifies and the law allows a greater limit not to exceed 2
ppmvd. The commenter has not yet determined whether it agrees with the
method by which existing and new kiln emission standards were set.

Commenter 2898 agreed with the EPA proposal to base the MACT floor for
HCl for existing sources at the average of the lowest emitting 12
percent of the kilns for which it has data and at the emission rate of
the single lowest emitting source for new sources is consistent with the
CAA’s requirements.

Response: The final standards for HCl are 3 ppmvd on a dry basis at 7
percent oxygen, for both new and existing sources. A detailed discussion
of the MACT floor determination for HCl is included in the preamble
(section IV.A.1.e).

Comment: Many commenters stated that EPA’s MACT floor analysis for HCl
was inadequate or that EPA lacked adequate data.

Several commenters (2830, 2836, 2845, and 2859) stated that EPA should
reaffirm its position in the 2006 NESHAP where it found that a HCl
standard is not necessary given the low HCl emissions from cement kilns
and risk assessment findings from a national survey of the industry
showing a lack of risk to health or the environment from any cement
facility. The commenters further stated that EPA’s HCl emissions data
is not sufficient to establish a defensible MACT standard. The majority
of emissions data used by EPA is from area sources of HAPs and/or
suffers from data quality issues, such as data recorded below the
quantification limits of the test method or by using improper test
methods. After reviewing the data, the commenters believe EPA has
emission data from only one source that would be a defensible data
point. EPA should not base a nationwide existing NESHAP standard on data
from one facility. Clearly, no variability considerations are afforded
as a result of using one data point. 

Commenter 2841 stated that EPA gave limited notice to the industry to
respond and provide additional data, is inappropriate when considering
the development of any standard.

Commenters 2845, 2846, 2863, and 2889 stated that EPA should not use HCl
emission data below the current Method 321 practical quantification
limit. The commenters stated that it is unnecessary to set a floor or
standard for HCl, but if EPA decides to do so it should exclude all test
data below the practical quantification limit and exclude from the pool
of MACT floor sources any kiln with test data below that limit. Given
that the data were gathered when Method 321 identified a practical
quantification limit in the range from 1000 to 5000 parts per billion
for specific path length and test conditions, EPA should not use any
data below that range. If EPA believes it may be appropriate to revise
Method 321, it should conduct a formal review process for the Method,
and make available to the public any relevant supporting information.
But if EPA does amend Method 321, it should not apply the changes
retroactively.

Commenters 2836, 2844, 2845, 2859, 2863, and 2889 stated that the
existing source floor is improperly based on inadequate data. Commenters
2845 and 2863 stated that there are not enough data upon which to
establish an existing source MACT floor. One-third of the sources in
EPA’s database – 9 of the 27 sources listed in Table 5 of the
preamble – are true area sources. Commenters noted that test data for
so many of the 27 kilns are invalid that when true area sources and
sources with invalid data are excluded, only one kiln remains (i.e.,
only Kiln 20 in the table is a major source for which there are valid
HCl test data). Additional concerns presented by the commenters were:

EPA established the proposed HCl MACT floor for existing sources based
upon Method 321 test data from what it describes as the top 12 percent
of kilns for which it had such data. There were only 8 of 27 kilns in
EPA Table 5 of the preamble with Method 321 data and EPA examined the
top four kilns without respect to test method. The lowest emitters,
Kilns 1 and 2 used data from California Air Resources Board (CARB) 421,
which is specifically disallowed. 

Only one of the kiln sets from which EPA calculated the floor is a major
source. EPA must change the kilns from which it sets that existing
source floor. The proposal notes that it is appropriate to limit the
floor database to major sources. This is correct because area sources
are not in the same category or subcategory as the major sources for
which EPA is setting the floor and issuing standards. CAA §112(d)(3).

EPA should use data from 5 kilns, rather than 4, to establish the floor.
Where there are 30 or more sources in a category, but fewer than 30 with
data, CAA §112(d)(3) should be read to require the use of 5 sources to
establish the floor.

Much of the HCl data are based on invalid tests. EPA must exclude
results from HCl testing conducted with methods other than Method 321 or
ASTM D6348. Some methods were incorrectly applied. The commenter
provided an appendix with tables detailing the commenter’s concerns
about test data. The commenter recommended that EPA exclude 19 kilns
from its database due to invalid tests.

Response: In response to comments we received on the HCl MACT floor
analysis, EPA revised its analysis. We have removed those natural area
sources from the floor analysis and because many of the source tests
were not actually EPA Method 321 tests and others lacked important
quality assurance information, we used our authority under section 114
to collect additional data. We used this new data set to calculate a
MACT floor. See the preamble to the final rule for a discussion of the
revised analysis.

Comment: Several commenters (2830, 2844, and 2845) stated that EPA has
provided no support for its assumption that the stack testing data used
to establish the HCl MACT floors can be correlated to continuous
emissions monitoring data to demonstrate compliance. The commenter
states that EPA has provided no analysis to demonstrate that the
variability of data provided by the proposed emission monitoring methods
would be equivalent to the variability of the stack tests. Under the
Proposal, the HCl MACT floor was set based on the averaging of stack
test data. However, compliance emission measurements will need to be
made based on continuous measurements. The D.C. Circuit disfavors
standards that impose a method for demonstrating compliance that differs
significantly from the method used to arrive at the standard. This is
precisely the case here.

Response: As stated in the preamble to the final rule, EPA has
significantly increased the level of the standard to account for any
uncertainty associated with the analytic method detection level, and is
also expressing the standard as a 30-day limit. Since the level of the
standard is higher than the stack measurements, and is averaged over 30
days (which normally would result in a lower level due to the multiple
observations which dampen variability), EPA does not believe there is
any legitimate issue that use of a CEM to measure the standard makes the
standard more stringent than the underlying data would support. At the
same time, EPA believes that its approach is a legitimate way to gauge
best performers’ performance given the uncertainty associated with the
analytic method detection limit.. 

Subcategorization

Comment: One commenter (2863) suggested that EPA should subcategorize to
reflect arid sources. EPA cannot reasonably set a standard based on a
control technology type that cannot be employed in a large portion of
the western United States. EPA considered wet scrubbers as a basis for
meeting the HCl standard and no consideration was given for alternative
types of control. However, wet scrubbers are not a viable control
technology in vast parts of the western United States where water is not
available for use in a scrubber. The commenter stated that given the
unavailability of water, EPA should create a subcategory for arid areas
for those plants that cannot use wet scrubbers. Failure to do so would
break from well established precedent and arbitrarily impact plants in
one part of the country.

Response: As noted in the preamble, we are not mandating wet scrubber
control technology in this rule. Other control techniques are available
(hydrated finely ground lime, spray dryers, fuel and additive switching)
that we believe would allow a cement kiln to meet the HCl emission
limits in areas where sufficient water for a wet scrubber is not
available. Although we have no cost data available for these alternative
control techniques (some of which would be site-specific), we anticipate
that these control techniques would be no more expensive that a wet
scrubber. 

CISWI

Comment: One commenter (2889) stated that the MACT floor for HCl should
be re-evaluated after the data anticipated in response to the EPA §114
Letter CISWI request, to be submitted by October 2009 is evaluated. 

Response: EPA performed a separate date gathering for HCl emission data
as part of this rulemaking and used these data to recalculate the MACT
floor.

2.7.4	Compliance Assurance

Support proposed monitoring method

Comment: One commenter (2144) stated that EPA should require HCl-CEMS
for compliance.

Response: EPA acknowledges the commenter’s support

Comment: Commenter 2832 supports EPA’s proposal to reject EPA
Reference Method 26/26A stack testing data. Method 26/26A does not
accurately measure HCl.

Response: EPA acknowledges the commenter’s support

Compliance Based on Performance Testing

Commenter 2845 stated that EPA should require that compliance with HCl
limits should be measured by periodic stack tests. Because the HCl
floors were developed from HCl stack test data, the standard for HCl
should be based on periodic stack testing. EPA must evaluate valid data
from Method 321/ASTM D6348 stack tests instead of the data contained in
Table 5 of the proposal. Using CEMS to measure compliance effectively
makes the standard more stringent than what has been achieved by the
best-performing sources. If CEMS compliance demonstration is retained,
then the limit for CEMS compliance must be raised to reflect the added
variability that will be measured by the CEMS. While continuous
measurement will capture variability of emissions 24 hours per day, 7
days per week over the full range of process and control system
operating conditions over the life of the plant and its associated
quarry, the stack test is merely a snapshot in time. By definition, a
stack test contains no parameter related to variability other than that
obtained during the three hours of testing. In addition to the inherent
variability of HCl emissions, a CEMS standard must also consider the
inaccuracy of the CEMS as determined (and allowed) relative to the
required stack test methods, the uncertainty of calibration
standards/materials, and other factors affecting the sampling,
transport, and analysis of HCl which is a highly reactive compound.

Response: See response in preamble, as well as the response above to
commenter 2844.

Monitoring requirements for scrubbers

Comment: One commenter (2911) disagreed with EPA’s proposed
requirement to test sources with limestone scrubbers no more than every
five years. The commenter stated that monitoring must be sufficient to
assure continuous compliance with emission standards and to enable
sources to certify accurately whether their compliance with emission
standards is continuous or intermittent and to report any deviation from
those standards. Because both chlorine inputs and the performance of the
scrubber may change significantly over five years, testing just once
every five years cannot possibly provide the required assurance.

Response: Given the efficiency of wet scrubbers in removing HCl, we
believe that parameter monitoring is generally sufficient to demonstrate
compliance combined with periodic testing. However, we agree that a 5
year period between tests may be too long for the reason noted by the
commenter and have changed the testing frequency to 30 months. .

Issues with use of CEMS for HCl compliance monitoring

Comment: Many commenters cited difficulties in the use of CEMS to
determine compliance with the HCl emission limits.

Commenters 2622 and 2889 stated that there will be difficulties
conducting a RATA on HCl CEMS operating at HCl concentrations below 2
ppmvd. Commenter 2622 stated that because that proposal calls for RATA
on HCl monitors when the emission limits are low, EPA will need to set a
low emissions criteria for passing a RATA. The commenter expressed
concern that 20 percent of the reference method or 10 percent of the
emission limit will be too restrictive.

Commenter 2844 stated that that given the level of the standard, a CEMS
will require FTIR and the commenter is not aware of such CEMS available
for purchase. The commenter recommends that compliance be based on
periodic stack tests.

Commenters 2844 and 2846 stated that the new source HCl limit is below
the published Method 321 detection limit and it is not appropriate to
adopt a limit that is inconsistent with the existing standardized test
method.

Commenters 2845, 2846, 2858, and 2859 stated that HCl CEMS cannot be
used to demonstrate compliance at the levels proposed by EPA in this
rule. Although HCl CEMS are installed in some hazardous waste
incinerators, the concentrations being monitored at the incinerators are
much higher than the HCl detection limit based concentration standard
proposed by EPA for cement plants. In addition, calibration gases are
not commercially available at the concentration levels proposed by EPA
in this rule. Commenter 2845 expects a very high learning curve
associated with operating and maintaining these technically challenging
devices. This industry has little experience with their use and, since
HCl CEMS are not widely used in other industries, very little outside
assistance can be provided through industry workgroups. As with mercury
CEMS, EPA should resolve the reliability and performance issues with HCl
CEMS before imposing substantial risks and costs on the portland cement
industry. Although SO2 control is not the primary purpose of installing
a wet scrubber under the proposed rule, the scrubber would be designed
as if SO2 removal were the purpose. 

Commenter 2855 stated that FTIR based HCl CEMS cannot function at cement
plants at the concentration levels proposed in the Rule. HCl CEMS are
installed at some hazardous waste incinerators, but generally are
process instruments relied on to provide measurements of HCl levels from
the incineration process and are not FTIR based. The use of an
FTIR-based CEMS for monitoring compliance at a cement plant at such low
levels of HCl introduces a number of problems and concerns for the
cement industry:

Calibration gases are not commercially available at the concentration
levels proposed by EPA in this rule. The lowest concentrations of HCl
calibration gases commercially available are between 5 and 10 ppm, with
an accuracy of ±10% (0.5 ppm to 1 ppm). These calibration gases are 25
to 5 times greater than the proposed HCl standard for existing plants.

There is no independent anchor for HCl calibration gas certification and
the lowest concentration range available is over twice the proposed
standard with a +/- 10% accuracy, numerous uncertainties are introduced
to any data that are collected for the purposes of demonstrating
compliance.

The industry expects a very high learning curve associated with
operating and maintaining these devices. The industry has little
experience with their use and manufacturers have not sold many
monitoring systems into the cement industry. 

Commenter 2855 stated that the current state of the technology for FTIR
CEMS for HCl is such that providing precise accurate measurements below
2 ppm HCl will be a challenge and has not routinely been demonstrated in
full-scale industrial applications.

Commenter 2855 stated that a moratorium of 18 months is needed for any
HCl monitoring requirements that may prevail for new and refurbished
kilns having to meet the requirements of the standard on startup or
promulgation.

Response: Several comments relate to the low level of the proposed HCl
standards. We believe that these comments have merit and explain in the
preamble that we are adopting a standard of 3 ppmv to accommodate
issues relating to uncertainties of reliably measuring below that value.
We have also modified the RATA emissions criteria to allow measurements
up to 20 percent of the limit, as we have done in other relatively low
emissions concentrations situations. 

Regarding concerns about the use of FTIR, a number of vendors sell FTIR
instruments, and they will work with their customers to adapt the
instruments to the needs of particular facilities. In addition,
calibration gases are readily available at 5 ppm and even down to 2.9
ppm. Method 205 may be used for further gas dilution (yielding high
quality standards) down to the levels needed. We have also talked to
vendors who state they can measure HCl down to 2 ppm. (Email from
Jeffrey Lacoss to Rima Howell RE: Quick Question: PS-15.  March 5, 2010.
 As forwarded to Robin Segall on March 5, 2010.  Document in Docket).

Commenters also raise concerns about the learning curve associated with
HCl CEMS. We note that sources have 3 years to comply with the standard,
which will be sufficient time to implement a working CEMS.

Comment: One commenter (2855) provided comments related to detection
limit issues:

EPA erred when they focused on detection limits in setting the proposed
standard and mandated use of an FTIR based CEMS.

Accuracy is unknown at the detection level; therefore accuracy at the
proposed compliance level cannot be demonstrated.

EPA focused on the detection limits associated with Method 321 area
source tests contained in the docket for HCl in establishing the
proposed HCl standard. EPA must not establish standards based on the HCl
detection level of Method 321, nor should they establish standards for
any compounds at their corresponding method detection levels.

The commenter noted that there is a difference between the detection
limit and practical quantification limit. There is also a difference
between test methods and continuous monitoring systems, although EPA has
proposed a continuous monitoring requirement and cited test Method 321.

The commenter requested that EPA provide information regarding the
differences between current detection levels and those established more
than 15 years ago. 

Detection limits are only useful to determine if the measurement is
possible. Accuracy at the detection limit and values below is unknown.
Therefore, proposing a standard at the analytical noise level will
result in the inability to make accurate determinations of compliance.

Commenter 2855 stated that the current stated lower practical
quantification limit for Method 321 still holds true. EPA has proposed
to make changes to Method 321 without providing the information relevant
to this change. It is the belief of the commenter and its member
companies that the Practical Quantitation Limit (PQL) of Method 321 is
between 1 ppm and 5 ppmv wet depending on how well the analytical system
is optimized for HCl. These values are well within the guidance of 5-10
times the analytical detection limit, and any claims to the contrary
must be proven through laboratory demonstrations and placed in the
docket to this rulemaking.

Commenter 2855 stated that EPA had requested comment on which way to set
the emission detection limit and to handle emissions information below
the detection limit for use in this rule. The commenter stated in
response that it is inappropriate to set any standard at a level where
accuracy is unknown and thus demonstration of compliance is not defined.
Further, EPA should not make changes to a promulgated test method (see
changes to Method 321 in the Proposal at 21,192) without first providing
information in the docket for public review and comment.

Commenter 2855 stated that EPA should focus on measurement system
capability and the availability of quality calibration materials in
their standard setting. The commenter provided results of an analysis of
analyte spike data showing sampling and analytical system biases when
non-detect values were reported. The commenter stated that a
concentration based standard cannot be based on values such as detection
limit that do not represent the difficult nature HCl measurement. The
standard cannot be set at a level below the demonstrated system bias and
should account for uncertainties in the calibration materials and the
measurement system ability.

Response: We have provided additional information in the docket on
detection limits (list) and have reviewed test data below detections
limits. Sources can optimize their system to measure HCl and thus
satisfy the PQL of Method 321. However, given that the final standard is
3 ppmv (for both existing and new sources), we believe that these
concerns are addressed since the standard provides an adequate margin
above the method detection limit to account for uncertainties of
measurements near the detection limit. 

Procedure 15

Comment: One commenter (2855) stated that PS-15 for FTIR Based CEMS is
not appropriate for an HCl CEMS for portland cement applications. EPA
requested comments on whether a CEMS can meet the requirements of PS-15
to measure HCl emissions. The commenter stated that EPA should develop a
PS for HCl CEMS that does not mandate a specific technology, but is
directly applicable to measurement of HCl. Additional problems stated by
the commenter concerning PS-15 are:

PS-15 is outdated and does not reflect how CEMS function. 

PS-15 requires that an FTIR based CEM be used. It is unclear why an FTIR
must be used when other technology is also available (i.e., GFCIR and
TDL).

Performance Specification-15 requires that a spectroscopist or operator
be present on a nearly continuous basis to meet the performance
criteria. 

Performance Specification-15 is not specific for HCl and many of the
technical requirements contained therein cannot be met for the 0.1 ppm
to 2 ppm HCl concentration levels proposed by EPA. 

Performance Specification-15 contains no requirements for: a)
determining calibration error or drift; b) setting instrument span
range; or, c) determining relative accuracy such as those contained in
the Lilly documents.

There are no NIST traceable HCl calibration gases as suggested by PS-15
at any concentration level.

There are no HCl calibration gases that can meet the 2% accuracy
certification requirement contained in Section 9.1.1, particularly at
the 0.1 ppm to 2 ppm concentration levels. 

The highest quality HCl calibration standard available today is at a
concentration level of 5-10 ppm with a vendor certification of ±10%.
The certification of the gas is by an FTIR calibrated with HCl gas
standards prepared by the manufacturer. The certification has no
independent anchor to NIST or other certifying organization.

The requirements of PS-15 for data storage (one interferogram per hour,
all absorbance spectra, all sample spectra, all background spectra, all
CTS spectra, etc...) and spectral audits cannot be met with any
available FTIR Based CEM systems. 

PS-15 requires modification of the FTIR CEM analytical program when
analytical results are not within ± 5 percent of the certified audit
concentration for each analyte. FTIR CEMS programs are largely
proprietary and cannot be modified by CEM operators at cement plants.
Furthermore, FTIR CEMS cannot simply load outside software for analyses
of audit spectra. This process must be tested with manufacturers of FTIR
CEMS at the concentration levels proposed in this rule. 

Requirements included in PS-15 assume the presence of a full time FTIR
CEMS operator, which is impractical. 

Commenter 2915 disagreed with EPA’s proposed standards for the use of
HCl CEMS:

There is a lack of appropriate EPA performance specifications or quality
assurance procedures for HCl CEMS;

There are serious technical feasibility issues with implementation of
the available technology; and 

The proposed standards are at or below the CEMS practical quantification
limit and/or the minimum detection limit of the analyzer.

Response: PS-15 is an FTIR CEMS procedure, but it does not specify any
instrument. Any commercially available FTIR instrument may be used as
long as the performance specifications and quality assurance
requirements of PS-15 are met. We recognize, however, that there are
also non-FTIR technologies available to monitor HCl. We are planning to
develop an HCl monitoring performance specification to address these
additional technologies over the next year or so in advance of the
compliance date of this rule; however, in the interim, owner/operators
always have the option to petition the Administrator for an alternative
testing approach for HCl under §63.7 (f). 

The commenter is concerned that a spectroscopist or FTIR operator will
need to be present on a nearly continuous basis at a monitoring
installation. PS-15 has initial certification requirements that may
require the assistance of a spectroscopist, as do most monitoring
performance specifications, but we do disagree that PS-15 will require a
full time operator. This issue has never been brought to our attention
by users of this PS-15. 

Regarding the comment that there are no NIST traceable HCl calibration
gases, at least one specialty gas vendor (Air Liquide) sells high
quality “Protocol” gases for HCl which are NIST traceable. These
gases are currently available in concentrations ranging from 2.9 ppm to
5000 ppm. For levels below this range, Method 205 (40 CFR Part 51,
Appendix M) is available for operators to use prepare dilution gases in
lower concentrations from the higher level cylinder gases.

As the cost of computer data storage continues to decrease, we do not
expect the data storage requirements of PS-15 to be particularly
burdensome. We do not have and are not planning to develop spectral
audits for FTIR. Should we decide to develop them in the future, we will
have to address the issue of proprietary software systems at that time. 

Comment: Commenter 2855 stated that specific data storage requirements
and data submission requirements that are subject to outside party
interpretation should not be contained in any PS for a continuous
monitoring system. For example, PS-15 requires that the data be
submitted to EPA, and that EPA would interpret these data, and also
provide audit spectra. The commenter stated that they were unaware that
EPA had a spectroscopic branch that is developing these spectral audit
materials, or that they are prepared to handle such data submissions in
the future. If EPA is developing such materials, they should be provided
in the docket so that the cement industry can evaluate how such a
process could be handled. 

Response: As stated in the prior response, EPA has no plans to develop
audit samples or audit spectra for the FTIR. If no audit materials are
available, then an audit is not required. We believe the PS-15
requirements for spiking and other quality assurance requirements are
sufficient in the collection of quality data. The data storage
requirements are necessary due to the unique properties of FTIR. In
particular, for the data analysis and review, the interferograms are
critical in identifying and quantifying the target pollutants, and the
pollutant results are not transparent to the data reviewer unless the
raw data are available to allow reproduction of the results.

Comment: Commenter 2855 stated that EPA should develop an HCl monitoring
PS or work with ASTM in such an effort should the HCl monitoring
requirements prevail in this rule. The PS should employ analyte spiking
as a means to both certify and validate the system rather than
comparisons with Method 321 for Relative Accuracy testing. The commenter
recommended documents EPA OTM 22 and EPA OTM 23 submitted by Eli Lilly
prescribing the use of HCl CEMS at hazardous waste incinerators as a
starting point for a new HCl monitoring performance specification.

Commenter 2915 stated that they do not believe that an HCl CEMS could
meet PS-15, and EPA should develop appropriate, achievable performance
specifications that are applicable to all the available technology and
not dictate a monitoring technology such as FTIR. Once those are
established, EPA should collect data with those technologies to
determine whether they are appropriate for determining compliance with
the proposed limit. If a technology meeting appropriate performance
specifications is not capable of detecting or quantifying emissions at
the level of the standard, the technology is not appropriate for use as
a compliance method for that standard. The commenters provided a list of
difficulties in using PS-15 for continuously monitoring HCl.

Response: Facilities that desire to use monitoring technologies other
than FTIR may utilize the General Provision to request approval to use
an alternative method by providing the alternative specifications and
QA/QC measures they would use. We agree there may be benefits to an ASTM
process as well, and will work with them to evaluate such methods.

Comment: Commenter 2855 stated that EPA has specified use of Appendix F,
Procedure 1 along with PS-15. Appendix F, Procedure 1 was written for
SO2, NOx and diluent CEMS (O2, and CO2). It is not directly
transferrable to an FTIR CEMS monitoring HCl at extremely low
concentration levels. Appropriate on-going quality assurance procedures
are needed.

Response: We have revised the QA procedures to accept values up to 20
percent of the emission limit.

Comment: Commenter 2855 stated that the initial compliance demonstration
procedures for HCl emissions from kilns or kilns/in-line raw mills
appear contradictory and should be revised to be consistent with the
proposed emission standards and 30-day rolling average. Section
63.1349(b)(6)(ii) is inconsistent specifying a performance test duration
of 24-hours, and requiring calculations of the daily average HCl
concentrations from the hourly averages during the performance test.
This is inconsistent with the language in the preamble and in
§63.1343(a). 

Response: The 24-hour period is consistent with the RATA, which is used
to demonstrate that the CEMS is operating properly. In this case,
operation while the raw mill is on, is the only requirement and the
final rule has been revised to reflect this condition. Initial
compliance is determined by the first 30 days of data.

Comment: One commenter (2855) stated that the initial performance
demonstration should be based on a properly calibrated and compliant
CEMS (if such a system becomes available that is capable of accurate
measurements at the level of the HCl emissions) and the initial
performance test result must be based on the calculated average HCl
concentration for 30 kiln operating days. A definition for kiln
operating day must be established and must reflect a specified number of
hours of kiln or raw material dryer operation at normal production
conditions. Days with fewer operating hours than a kiln operating day
should not be included in the 30 day average. There should be no
specification for separate tests for mill-on and mill-off for in-line
raw mills.

Response: We agree that a definition of “kiln operating day” is
needed and have added one to the final rule. We also agree that a raw
mill off test is not required.

Comment: One commenter (2855) stated that the initial compliance
demonstration procedures for HCl emitted from sources not controlled by
a wet scrubber specifies a performance test duration and conditions that
are not appropriate. Furthermore, 40 C.F.R. §63.1349(b)(6)(ii)(2008)
states that initial compliance shall be demonstrated by conducting
separate performance tests for in-line kiln/raw mills with the mill
operating and with the raw mill off. This would require the facility to
run in a mill-off condition for 24 hours, which is not possible. There
should be no specification for separate initial compliance tests for
mill-on and mill-off for in-line raw mills.

Response: We agree that a raw mill off test is not required and have
changed the final rule to reflect this.

Comment: One commenter (2901) stated that EPA did not follow the
Agency’s own protocols in proposing to change the practical lower
quantification range of Method 321. 

Response: We believe we have adequately followed protocol in proposing
changes the M321 through this notice and comment rulemaking process.

Monovents

Comment: One commenter (2888) proposed that monovent facilities stack
test for HCI, using EPA approved methods, on the same 30-month frequency
as required for Dioxin/Furan testing under the current Portland Cement
Maximum Achievable Control Technology (PC MACT).

Response: Sources equipped with scrubbers are required to test every 5
years. Such sources would have to construct a stack for testing.
However, where sources use CEMS, we believe the current monitoring
methodology is appropriate. Sources may request alternative performance
testing requirements under the General Provisions to address site
specific circumstances.

Clarifications

Comment: One commenter (2845) stated that §63.1343(b)(6) and
§63.1343(c)(6) should be clarified to state that the HCl standard is
corrected to 7 percent oxygen based on a 30-day rolling average.

Response: We agree and have made the suggested clarifications.

Comment: One commenter (2845) stated that §63.1350(q)(2) should provide
specific requirements rather than reference to 40 CFR part 63, subpart
SS, which may change in a manner that could make it inappropriate for a
kiln system. The commenter provided detailed language to include in this
section.

Response: While there are merits to using cross references, in this case
we agree to add information specific to this rule to avoid the use of a
cross reference.

Coal mills

Comment: One commenter (2910) asked that EPA clarify in the proposed
revisions that in-line coal mills that vent directly to the atmosphere
be exempt from installing or operating an HCl CEMS. The commenter stated
that they knew of no basis for concluding that coal drying would
increase the HCL concentration in the in-line coal mill exhaust. The
commenter stated that monitoring the concentration of HCl in the main
stack would be an appropriate way to measure compliance with the HCl
standard. 

Response: We agree and have clarified the final rule accordingly.

2.7.5	Impacts

SO2 scrubbers and HCl control

Comment: Several commenters (2845 and 2900) stated that the scrubbers
that would be used to control HCl are generally designed to control
sulfur dioxide, not HCl. Vendors of HCl scrubber systems indicated that
the system they would install would be designed for SO2, not HCl, if SO2
is present in the flue gas. The limited number of wet scrubbers in the
U.S. on cement plants are designed to remove SO2, and are not intended
to be used to comply with and HCl limit.

One commenter (2844) stated that there is little experience with
scrubbers designed to control both SO2 and HCl. The commenter stated
that based on limited short term HCl data, many kilns have preliminary
HCl levels that will require much greater control.

Several commenters (2845 and 2900) stated that there are no data to
support the conclusion that, under various conditions, a wet scrubber
designed to utilize lime will provide the required collection efficiency
to meet a 2 ppm emission standard. One commenter (2845) provided
information about vendors’ HCl scrubbers, stating that the systems
that can provide 99 percent control of HCl use NaOH as the scrubbing
media. These systems are installed for applications that have been
producing little if any SO2, and the inlet HCl concentrations are
substantially higher than would be typical in cement plants. Because of
product quality/specifications and economic concerns, NaOH would not be
a reasonable reagent for cement kilns. Lime is easily obtained by cement
plants, but NaOH would have to be purchased. Neither the portland cement
industry nor the Agency have any data to support the conclusion that a
wet scrubber designed to utilize lime will provide the required
collection efficiency to meet a 2-ppm emission standard when SO2 is
present in the flue gas and the inlet HCl concentration is very low. 

Commenter 2889 disagreed with EPA’s estimates of wet scrubber
efficiencies of 99 to 99.9 percent. The commenter notes that at the
lower HCl concentrations found in cement kiln gas streams, it is
unlikely that a wet scrubber could achieve those efficiencies.

Response: We agree that the wet scrubbers currently on U.S. cement
plants are designed to remove SO2, rather than HCl. However, an EPA
review of HCl emissions from 21 plants indicated an average
concentration of 12 ppm, and a maximum concentration of 55 ppm.
Therefore, required HCl removal efficiencies to meet the 3 ppm emissions
standard are expected to be much lower than 99 percent for virtually all
kilns. A preliminary review of data gathered from a 2010 Information
Collection Request (ICR) for Electrical Generating Units indicates that
wet scrubbers utilizing lime will be able to provide the necessary
collection efficiency to meet the 3 ppm emissions standard. We also
think that test data from the secondary aluminum industry regarding the
HCl removal by lime-coated fabric filters (e.g., Docket A-92-61, Item
II-A-9) indicate that the 3 ppm emissions standard can be achieved by
alternatives to scrubbers.

Baseline HCl emissions

Comment: One commenter (2845) stated that for HCl, EPA assumed that all
kilns without recent test data would emit at the average rate observed
across those relatively few kilns for which test data were available.
While this may be true on average, this approach ignores the large
variability that undoubtedly exists across untested kilns in terms of
their actual HCl emissions rates. Some of the kilns to which EPA applies
the average HCl emissions rate will in fact have high emissions and will
need more intensive and costly HCl controls; EPA fails to accurately
represent total control needs across the entire industry by missing
these kilns that have emission rates well above the average.

Response: It is true that kilns may have emission rates above the
average, and, therefore, may require more costly controls. However,
there are also kilns that have emission rates below the average, and,
therefore, would require less costly controls. Data available to EPA
indicated that there were more kilns with HCl emissions lower than the
average rate utilized by EPA in cost calculations than there were kilns
with HCl emissions above the average rate. Therefore, we believe the
cost calculations are accurate, if not overestimating the total control
costs for the industry.

Achievability using wet scrubber

Comment: One commenter (2900) provided a summary of information about
one of their plant’s wet scrubbers. Information includes water and
power use, costs, and HCl stack emission data. Prior to the installation
of the scrubber, the plant conducted extensive tests in 2000 for
purposes of evaluating HCl emissions from the various stacks. The data
from 2000 demonstrates that the HCl emission rate as lb/hr was doubled
when the raw mill was shut down and indicates a scrubbing effect through
its operation. The data taken in 2006 using similar methods showed that
the plant experienced some reduction in HCl emission but not enough to
achieve the EPA limit at less than 2.3 ppmvd at 7 percent oxygen. A data
point was not found for the raw mill down condition. It is expected that
the results will be less with the operation of the scrubber than the
data point in 2000; but higher than the new EPA limit. Both the raw mill
running and down conditions will likely be noncompliant with the new
limit. The commenter noted that the scrubber at their plant was set at
5.5 to 6.5 for strictly SO2 control, and that the set point was well
below the 8 to 10 pH that the commenter stated is needed for HCl
control. As this scrubber is utilizing water, limestone, and some lime
concentration present in the stone, it can be surmised that the EPA’s
limit is difficult to achieve even with the best available scrubber
technology.

Response: We believe that a preliminary review of data gathered from a
2010 Information Collection Request (ICR) for Electrical Generating
Units indicates that wet scrubbers utilizing lime will provide the
necessary collection efficiency to meet the 3 ppm emissions standard.
Furthermore, because the HCl standard is established at a level higher
than all measured values (to account for the inability to reliably
measure any lower standard) and measured based on 30day averages, it
provides an ample compliance margin. The 30-day averaging period assures
that several operating cycles of raw mill on and off are included in
each average, thereby averaging them out over the 30-day period.

Water availability for scrubbers

Comment: Several commenters (2845 and 2941) stated that wet scrubbers
are highly water-intensive, and based on the locations of many of the
cement plants, obtaining additional water rights could be very
difficult, particularly for plants located in states where water rights
issues have been contentious. The commenter provided several estimates
of water use, including an estimate for a typical cement plant of 460
million gal/yr, and an estimate that a scrubber will require, on
average, 30 million gallons of water annually.

Response: As noted in the preamble, we are not mandating wet scrubber
control technology in this rule. Other control techniques are available
(hydrated finely ground lime, spray dryers, fuel and additive switching)
that we believe would allow a cement kiln to meet the HCl emission
limits in areas where sufficient water for a wet scrubber is not
available. Although we have no cost data available for these alternative
control techniques (some of which would be site-specific), we anticipate
that these control techniques would be no more expensive that a wet
scrubber. 

Scrubber costs are underestimated

Comment: Several commenters (2830 and 2845) stated that EPA has
underestimated the cost of wet scrubbers to capture HCl, and there will
be problems if a facility uses both a wet scrubber and a RTO in series.
If the proposal for HCl is implemented, wet scrubbers will need to be
installed at many of the cement plants. Wet scrubbers are designed such
that the flue gas exiting the scrubber is saturated. As a result, the
exiting flue temperature will be 160°–180°F. Therefore, plants that
install these scrubbers will either need to reheat the flue gas to
prevent condensation in the stack and provide flue gas buoyancy, or it
will be necessary to line the stack with a material to withstand
continuous water condensation. In either case, this will be an added
cost to the plant. If natural gas were used to reheat the flue gas from
160° to 225°F, the theoretical plant generating 200,000 dry standard
cubic feet per minute (dscfm) would generate an additional 7,500 tons
CO2/year.

According to commenters 2830 and 2845, if an RTO is required because of
THC emissions, the exhaust exiting the wet scrubber enters the RTO fully
saturated with moisture, meaning that as it is heated, it passes through
the temperature range with the highest corrosion potential (above the
water dew point and below the acid dew point). Thus, very costly nickel
alloys must be used at certain areas within the RTO. Also, in designing
the RTO, if the heat exchanger is designed to achieve high thermal
efficiency (low gas consumption), then the RTO outlet temperature will
be below the acid dew point. As a result, all outlet ducting and the
stack must be resistant to acid condensation and high levels of
corrosion. If the outlet temperatures are raised above the critical acid
dew point temperature, then the RTO will consume very large quantities
of natural gas. Therefore, the choice forced upon the kiln by the
proposed rule is either high energy costs or high material (alloy)
costs.

Response: As noted in a previous response, we believe that the wet
scrubber costs are accurate, if not overestimated, for the reasons
cited. The wet scrubber and RTO costs were calculated based on actual
costs from the industry for those systems, and, therefore, should
already account for the factors mentioned by the commenter. Furthermore,
as noted in the preamble, the wet scrubber/RTO combination has already
been demonstrated in cement kiln applications. 

Accounting for scrubber waste handling

Comment: Several commenters (2841, 2845, and 2889) stated that there
will be problems with the solids and liquid handling of wet scrubbers.
The commenter provided an overview of the wastes associated with wet
scrubber operation. The commenter’s concerns include: 

Plants will need to dispose of some or potentially all of the sludge.
EPA has assumed that the solids would be recycled to the finishing mill,
but the chemistry of the sludge will be site specific and may not allow
recycling. Disposal of solid waste would require additional permits.

Dewatered solids from the scrubber have a moisture content of 25% to
35%. This material is difficult to handle in the finish mill and its
associated classifier. Furthermore, the solids from a wet scrubber are
not composed strictly of gypsum (calcium sulfate). The waste material
includes sodium and potassium salts. There will also be some calcium
sulfite. There are limits to the concentrations of these contaminants
that can be included in the cement product.

Because of recirculation of the slurry, the concentration of chlorine in
the slurry liquid can rise to high steady-state concentrations,
requiring wastewater treatment.

The costs for water treatment imposed by the proposed rule would be at
least $50,000 per year, not taking into account total dissolved solids
(TSS) or biochemical oxygen demand (BOD) costs.

Response: Our wastewater and solid waste impacts estimates for wet
scrubbers were determined based on information from industry and, as
such, should already account for the situations cited by the commenter.
As noted in the preamble, all of the facilities currently using wet
scrubber mix the gypsum created in the scrubber with clinker in the
finish mill and were assumed to also account for other waste materials
that accumulate during the normal operation of the facility. As noted in
the preamble, there are four facilities (five kilns) that currently use
wet scrubbers that report no water releases at any time, such as that
cited by the commenter. We requested comment in the Portland Cement
NESHAP proposal on the potential for water releases due to wet scrubber
system purges and any regulations that might apply. Though commenters
raised concerns of the possibility of water impacts, they did not
provide a rationale of why it would be expected, when it is not
occurring at the four facilities that currently use wet scrubbers, due
to their on-site reuse of water. If discharges did occur, there would be
a potential for water quality issues. But given these facts, we believe
our estimate of no water quality impacts resulting from production of
wastewater by wet scrubbers is reasonable.  The sludge that is generated
is used again in the process, so there is no cost of disposal associated
with sludge.  

HCl emissions data, Table 5 of Preamble

Comment: One commenter (2863) stated that Table 5 in the preamble
outlines the available HCl data for the 27 kilns for which Method 321
data were available. Ash Grove’s Durkee plant is identified as source
#26 on that list with emissions of 34.68 ppm HCl. This is not correct.
The Durkee plant’s HCl emission rate is 3.2 ppm (7 percent
oxygen)—an order of magnitude lower than what is stated in Table 5.
Portland cement kilns are not significant sources of HCl. 

Response: As noted in the preamble, the analysis for the HCl MACT floor
was revised to remove errors identified in comments on the proposal,
including this error. 

Dry scrubbers

Comment: One commenter (2889) stated that there are problems associated
with spray dryer absorbers or lime injection-based dry scrubbers
recommended for arid regions. A review of recent boiler BACT analyses
summarized in the RBLC Clearinghouse indicates that the HCl removal
efficiencies are well below the 99 percent levels assumed by EPA.
Furthermore, both types of dry scrubbers generate significant quantities
of solid waste.

Response: EPA expects that the HCl removal efficiencies needed to comply
with the 3 ppm HCl emissions standard will be well below 99 percent for
virtually all kilns. Regarding the solid waste argument, as stated in
the preamble, the PM captured in the kiln fabric filter (cement kiln
dust) is essentially recaptured raw material, intermediate materials, or
product. Based on the available information, it appears that most
captured PM is typically recycled back to the kilns to the maximum
extent possible. Therefore, we estimate that any additional PM captured
will also be recycled to the kiln to the extent possible.

2.8	PM Standards

2.8.1	Legal Issues

MACT-on-MACT

Comment: Many commenters (2830, 2832, 2841, 2844, 2845, 2858, 2859,
2863, 2864, 2874, and 2914) stated that EPA lacks the authority to
promulgate new PM standard using the MACT floor approach because EPA
issued new and existing source NEHSAP for PM in 1999. Once the NESHAP is
established, the CAA §112(d)(6) requires any revision to the standard
to be accomplished using a technology review and not using the process
proposed by EPA in this rulemaking. The proposed EPA standard in this
rulemaking leads to a MACT-on-MACT approach which results in a floor
that is much more stringent than it would have been if EPA had used
emissions data from sources that had not already complied with the 1999
PM standards. 

Response: This is factually incorrect. If EPA properly established MACT
for PM using the 1999 data base, the standard for existing and new
sources would be higher than the standard adopted in today’s rule, as
explained in the preamble to the final rule (although EPA considers its
current data base to be more representative, and so is using it to amend
the standard).

Comment: Due to the inappropriateness of EPA’s proposal and the fact
that the 2000 remand did not address PM at all, the commenters urge EPA
to maintain the 1999 PM standard. Commenter 2864 argued that EPA does
not have the authority under the CAA to completely recalculate PM MACT
floors for cement plants where the initial floors were upheld. Although,
EPA has the authority to revise the PM MACT standards for the industry,
EPA must do so in accordance with §112(f) or §112(d)(6) without
recalculating the floors. Because EPA has failed to conduct the analyses
necessary under those provisions, it has no authority to revise the PM
MACT standards for either existing or new cement plants.

Response: As stated in the preamble to the final rule, MACT-on-MACT is a
misnomer because the 1999 standard simply does not represent MACT. For
the reasons stated in the preamble, EPA does not see that it is barred
from amending that standard here to make a proper determination of MACT.

2.8.2	Adequacy of Floor Determinations (Existing and New)

Support for PM standards

Comment: Several commenters (2898 and 2911) supported EPA’s proposed
standards for PM and EPA’s proposed standards for area source cement
kilns under §112(d)(2)-(3).

Several commenters (2786 and 2898) stated that particulate metal
emissions from cement plants may pose health risks at very low levels
and therefore stringent emission controls are needed. Cement facilities
are sources of the non-volatile metal HAPs antimony, cadmium, lead,
selenium, chromium, arsenic, nickel and manganese. 

Commenter 2928 supported the proposed emission standards for PM.

Response: EPA acknowledges the commenters’ support.

Comment: Two commenters (2845 and 2859) said that EPA has not
promulgated any regulations requiring PM CEMS at any source category due
to its inability to address fundamental technical and policy issues and
must resolve these issues through rulemaking before requiring PM CEMS at
any cement plants. Furthermore EPA has not performed a legitimate
technical analysis of emissions variability and compliance determination
uncertainty to allow the use of PM CEMS for determining continuous
compliance with a PM limit at cement plants.

Response: We reject the commenters’ assertions that PM CEMS have not
been required via rulemaking because of unresolved fundamental technical
or policy issues. Concerns about PM CEMS were identified and addressed
prior to the January 2004 publication of Performance Specification 11
and Quality Assurance Procedure 2 for PM CEMS (69 FR 1786, January 12,
2004). As mentioned in that rule’s preamble, “…we believe that the
PM CEMS field demonstrations completed to date encompass a range of
operating conditions and emission characteristics…” including those
exhibited by sources such as cement kilns. 

Moreover, we disagree with the assertion that our analysis of PM
emissions variability is not legitimate, yielding an overly-stringent PM
emissions limit. The PM limit is based on our analysis of PM emissions
from test data, adjusted from an hourly to a 30-day averaging period and
further adjusted for variability. As mentioned in the preamble to the
Credible Evidence Rule (62 FR at 8314, February 24, 1997), we have
addressed and continue to address concerns about perceived “…limited
number and distribution of test runs and the inherent variability in
levels of emissions…” by a number of approaches, including changing
emissions averaging periods. Further, a statistically-based adjustment
to account for emissions variability, and which, in this case, increases
the numerical value of the standard (and its longer averaging period) by
fifty percent, does not make the standard more stringent.

Variability

Comment: Several commenters (2830, 2844, 2845, 2859, and 2864) stated
that because the data in the proposed NESHAP is based on a limited data
set (3-hour stack tests) that short term tests will not address the
variability issues. The commenter analyzed EPA’s released PM data and
found that EPA did not use all available PM data and according to the
commenters, EPA used an inappropriate statistical analysis (UPL instead
of upper tolerance level UTL). The commenter suggested that when an
appropriate statistical analysis using all of the EPA’s PM data is
used, the PM MACT standard should be 0.092 lbs/ton of kiln feed for
existing sources and 0.0625 lbs/ton of kiln feed for new sources.
Commenters 2845 and 2859 submitted PM data as part of their comments on
the proposed NSPS rules for portland cement.

Commenters 2845 and 2864 stated that the data set used to set the MACT
floor was too small. The commenter recalculated the PM floors using
additional data, and presented data showing the relationship between bag
life and PM emissions. The commenter stated that it is important that
EPA includes as part of the MACT floor the operational variability
experienced due to filter life and the baghouse itself that EPA needs to
include as part of the MACT floor. The commenters provided graphs
showing the effects of age on baghouse performance.

Response: In its comments, commenter 2859 stated that EPA ignored one of
two data points. They stated that EPA selected 0.0108 lb/ton clinker and
ignored a data point of 0.032 lb/ton clinker. The commenter
misunderstood how EPA calculated PM emissions for the Sugar Creek kiln;
EPA did not select one data point over another. The PM data for that
kiln was obtained from the Missouri Department of Natural Resources
(MDNR) and was used to calculate average emissions for that kiln. The
Sugar Creek kiln is one of the six MACT floor kilns used in the analysis
of the MACT floor for PM. The data for the Sugar Creek kiln used in the
MACT floor analysis is presented in the following table.

Lafarge Sugar Creek, PM Emissions: MDNR Data for 2002 and 2007

Roller Mill On

Run	Date	PM (lb/hr)	Kiln Clinker Production (tph)	PM Rate (lb/ton
clinker)

103*	8/7/2002	0.044	125.33	0.0004

105*	8/8/2002	0.587	119.37	0.0049

101*	10/30/2002	0.967	113.82	0.0085

111*	9/5/2007	2.52	127.00	0.0198

112*	9/5/2007	1.7	126.93	0.0134

113*	9/5/2007	1.09	126.93	0.0086



	AVG (mill on)	0.0093







Roller Mill Off

Run	Date	PM (lb/hr)	Kiln Clinker Production (tph)	PM Rate (lb/ton
clinker)

101*	8/6/2002	8.118	124.42	0.0652

104*	8/7/2002	1.823	120.27	0.0152

106*	8/9/2002	1.746	120.72	0.0145

101*	11/6/2007	5.22	119.12	0.0438

102*	11/6/2007	7.24	123.59	0.0586

103*	11/6/2007	4.84	118.32	0.0409

 	 	 	AVG (mill off)	0.0397



	WTD AVG (Mill on & off)	0.0108



In their comment, commenter 2859 provided PM test results for 2002 and
2007 as well as new PM run data for tests conducted in 2004, which EPA
did not have when it was the developing the MACT floor for PM. The
commenter’s PM data for 2002 and 2007 is essentially the same as the
data set we used with minor differences due to differences in kiln feed
rates. The individual run data submitted by commenter 2859 are presented
below.

Lafarge Sugar Creek, PM Emissions: Lafarge Data for 2002, 2004, 2007

Roller Mill On

Date	Filterable PM (lb/hr)	Kiln Production (tph)	PM rate 

(lb/ton clinker)

8/7/02	0.04	127.87	0.0003

8/8/02	0.59	121.80	0.0048

10/30/02	0.97	116.18	0.0083

5/20/04	3.02	117.95	0.0256

5/20/04	2.51	125.77	0.0200

5/20/04	2.65	126.10	0.0210

5/20/04	2.27	121.14	0.0188

5/20/04	2.76	108.80	0.0253

5/20/04	2.69	108.80	0.0248

6/25/04	3.30	126.10	0.0262

6/25/04	3.52	126.10	0.0279

6/25/04	3.41	125.88	0.0271

7/14/04	3.79	125.99	0.0301

7/14/04	3.97	126.10	0.0315

7/14/04	3.52	126.10	0.0279

9/5/07	2.52	129.52	0.0195

9/5/07	1.70	129.41	0.0131

9/5/07	1.00	129.52	0.0077



AVG (mill on)	0.0200





	Roller Mill Off

Date	Filterable PM (lb/hr)	Kiln Production (tph)	PM rate 

(lb/ton clinker)

8/6/02	8.12	126.98	0.0639

8/7/02	1.82	122.69	0.0149

8/8/02	1.75	123.13	0.0142

5/18/04	4.90	128.86	0.0381

5/18/04	3.79	126.54	0.0300

6/29/04	9.07	126.10	0.0719

7/15/04	8.02	126.10	0.0636

7/15/04	8.95	126.10	0.0710

8/12/04	12.57	129.52	0.0971

11/6/07	5.22	117.06	0.0446

11/6/07	7.24	111.22	0.0651

11/6/07	4.84	117.17	0.0413



AVG (mill off)	0.0513



WTD AVG (Mill on & off)	0.0216



Adding in the 2004 test runs increases the weighted average (assuming
the raw mill is on 95 percent of the time and off 5 percent of the time
per the MDNR) for the kiln from 0.0108 lb/ton clinker to 0.0216 lb/ton
clinker. The kiln is still among the top six performing kilns.

One commenter (2845) stated that they collected additional PM data and
combined the new data with the EPA data set to create a new MACT data
set consisting of a total of 319 runs/test data points for 47 kilns. The
commenter’s data included the number of runs/data points, the minimum
and maximum PM values, and the mean PM emissions in lb/ton clinker.
Based on average emissions for each kiln in the new data set, the
commenter identified a new list of the top 6 kilns with average PM
emissions for the 6 MACT floor kilns of 0.024 lb/ton clinker. Commenter
2864 concurred with the reanalysis of the MACT floor using the new MACT
data. EPA evaluated the new data and found shortcomings that led us to
not include the data in our analysis for the PM MACT floor. The
shortcomings identified include the following: 

Individual run data from which the commenter calculated average emission
rates were not included preventing us from fully evaluating the
emissions data or the average emissions they presented. 

When EPA selected the top performing kilns, it used the average PM
emissions for each kiln and ranked them from lowest to highest. In the
case of kilns with raw mills, EPA calculated the weighted averaged
emissions using test run data during raw mill on and raw mill off
periods. Because a raw mill operates 85-95 percent of the operating
time, the PM emissions during the raw mill on period will dominate the
weighted average. It appears that the commenter calculated a simple
arithmetic mean rather than a weighted average for kilns that have raw
mills. This was determined because for two kilns with raw mills that
only reported two values for each kiln, the commenter calculated the
simple arithmetic mean of the two values. Because PM emissions tend to
be lower when the raw mill is on and because the raw mill is typically
on 85-95 percent of a kiln’s operating time, calculating a simple
arithmetic mean as the commenter did, will likely overstate emissions.
Because individual test runs were not supplied, it was not possible to
recalculate the weighted averages for those kilns with raw mills. Among
the six MACT floor kilns identified by the commenter, four operate raw
mills. 

No explanation was given for excluding the Lehigh, Union Bridge kiln
from its data set or for dropping it from the top 6 kilns used in
EPA’s analysis. The average emissions for Lehigh, Union Bridge would
put it in the top 6 in the commenter’s set of new MACT floor kilns.

The commenter’s data set for Lafarge kiln 2 at Whitehall consisted of
13 data points whereas we used 15 test runs.

Commenters 2845 and 2864 commented that EPA’s PM limits did not
account for declining fabric filter performance over time. Commenter
2864 provided graphical evidence for one of its kilns that it claimed
showed that PM emissions increased with aging bags. The graph indicates
low emissions during the initial performance test as well as several
years later after replacing all of the baghouse bags. Following
replacement, PM emissions increased after installationof new bags and
remained high even when the bags were again replaced. Contrary to the
commenter’s claim, performance declined even when new bags were
installed. Based on the information provided by the commenter, it
appears that there were factors other than aging bags that contributed
to declining PM performance. EPA has reviewed PM test data for several
kilns over time and have concluded that PM emissions can be highly
variable. This variability can even be evident in consecutive test runs
on some kilns. The first figure below shows the highly variable PM test
results from a single for nine test runs over a 3-day period. The second
figure plots the average test results for six different kilns over
varying number of years. From these data, there are no clear indications
that baghouse performance necessarily declines over time, rather it
indicates that performance is variable. Obviously it will be necessary
for cement plant owners and operators to operate and maintain their
baghouses in peak condition to comply with the new PM limits. But EPA
rejects the commenters’ arguments that baghouse performance
necessarily declines over time and that the PM limits should be less
stringent to account for this.

Condensable PM

Comment: Several commenters (2845 and 2874) stated that it would be
inappropriate to regulate condensable fraction of PM2.5 for the
following reasons:

Because there are known problems with the test method 202 which is used
to determine the emissions for PM2.5 and the little data EPA has is
based on questionable AP-42 factors which is based on several tests, EPA
used limited available data that were of very uncertain accuracy.

The EPA did not show that listed HAP can be found in the condensable
fraction and therefore cannot reasonable use condensable PM as a HAP
surrogate under the authority of CAA §112(d).

If HAP existed in condensable PM, they would be controlled by other
means which include PM control devices for filterable PM, mercury which
is already directly regulated, and organic HAP are regulated through the
THC surrogated. Therefore, there would be no need to include condensable
PM as part of the PM standards.

EPA would need to propose PM standards because the current proposed
standards are based only on filterable PM and not both filterable PM and
condensable fraction of PM.

Response: The PM limits address filterable PM, including PM2.5 and PM10,
but not condensable PM. We do not have information on emissions of
condensable PM from cement kilns sufficient to set emission limits and
the limited information we do have is highly uncertain. We also believe
that these emissions will be controlled to some degree via controls on
HCl in the NESHAP and SO2 in the NSPS. EPA has recently promulgated a
new test method for condensable PM (revised Method 202) which will allow
for more reliable assessments of condensable PM. We anticipate that data
more representative of condensable primary PM emissions will be
available at the time of the next review of the NSPS.

Combined stack gases

Comment: One commenter (2841) stated that the proposed standards are
based on a kiln stack or some estimated number based on the flow
apportioned from a kiln and clinker cooler. This does not address
situations, such as those where additional sources of particulate are
included into the same duct system leading to one stack. In order for
facilities to design more efficient manufacturing operations, other
configurations are likely and should be encouraged. In addition, using a
grain loading standard is easily measured and can be monitored more
effectively.

Another commenter (2845) stated that the MACT floor for combined stacks
(new or reconstructed kiln or in-line kiln/raw mill) should be
calculated based on all available data and should subcategorize PM
standards by different kiln types. The alternative PM standard should
provide for including other exhaust gas streams in the main stack such
as coal mills. Facilities should have the option of meeting a grain
loading based standard instead of the lbs PM/ton of clinker standard, if
that is retained by EPA. The rule should be worded such that it is clear
that for combined exhausts, the gas flow and PM testing only needs to be
measured at the stack when using the alternative PM limit. The rule
should be worded such that it is clear that PM testing will not need to
be conducted in mill-on/mill-off conditions.

Response: Regarding these two comments, facilities have the option to
petition the EPA Administrator for a site-specific emissions standard
for the situations described above (e.g., including additional sources
of particulate in the same duct system leading to one stack, meeting an
alternative grain loading based standard), as long as the facility can
demonstrate that the alternative standard will achieve an emissions
reduction equivalent to that obtained under the promulgated standard. 

Clinker coolers

Comment: One commenter (2845) stated that §63.1345(a) standards for
clinker coolers should not be changed, remaining at 0.10 lb per ton of
feed (dry basis).

Response: We disagree with this comment for the reasons discussed in the
in the preamble regarding the PM limit for kilns. Therefore, we are
setting the same PM emissions limits and compliance requirements for
both clinker coolers and kilns. 

Increased PM with use of wet scrubbers

Comment: One commenter (2889) stated that EPA has not adequately
considered the emissions of particulate matter from wet scrubbers due to
reentrainment of solids-containing droplets from mist eliminators.
Subcategorization is needed with respect to particulate matter emissions
from wet scrubber-controlled cement kilns. Based on similar units in the
lime industry, it is apparent that this will be a significant problem.
In fact, Subpart AAAAA includes a higher particulate matter allowance
for wet scrubber-controlled lime kilns than for fabric filter controlled
lime kilns. Without a similar form of subcategorization, there will be
an incentive not to install wet scrubbers and instead use a less
efficient dry scrubbing system.

Response: The commenter is confused about the structure of subpart
AAAAA. First, the separate higher PM limits in subpart AAAAA of part 63
are there to permit facilities with wet PM scrubbers installed before
2004 to continue using them but to require more effective PM control for
newer units. The rule recognizes that wet PM scrubbers are not as
effective as fabric filters for PM control. Secondly, wet scrubbers
designed and operated to provide that gaseous SO2 or HCl react with
caustic slurries are fundamentally different from wet scrubbers designed
to remove PM (e.g., Venturi scrubbers). Acid gas scrubbers are not
intended to be PM removers; any collection of PM in an acid gas scrubber
is incidental to the primary purpose and probably minimal. The commenter
is correct that a poorly operating mist eliminator downstream of an acid
gas scrubber may produce PM emissions that could affect overall PM
control. The source must provide effective mist elimination in order to
maintain compliance with not only the acid gas emissions limit but to
avoid increasing PM emissions. 

Commenter 2900 provided data on particulate emissions before and after
installation of a wet scrubber. The data indicated particulate emissions
decreased. EPA expects installation of scrubbers after fabric filters
will result in slightly lower filterable particulate emissions than with
fabric filters alone. We do not think that Subpart AAAAA cited by the
commenter provides evidence that installing a scrubber downstream of a
particulate filter results in higher particulate levels than a fabric
filter alone.

2.8.4	Compliance Assurance

Performance of bag leak detection systems and other PM monitoring
systems

Comment: Several commenters (2845, 2859, and 2887) stated that in
proposing PM compliance assurance provisions, EPA did not, but should,
conduct a technical evaluation of the applicability, installation,
design, performance, certification, quality assurance and alarm level
setting procedures for all PM monitoring approaches and then re-propose
appropriate technical requirements for monitoring kilns, kilns/in-line
raw mills, and clinker coolers using bag leak detectors (BLDs),
Particulate Matter Detectors (PMD) and Continuous Opacity Monitors
options. According to the commenters, EPA does not recognize that many
BLD designs are inappropriate and inadequate for use under the varying
effluent matrix and conditions, especially at kilns/in-line raw mills
nor does EPA acknowledge that technically superior and far more
practical PM monitoring approaches have been developed for cement plants
that burn hazardous waste. The commenters stated that the proposed PM
monitoring provisions are inconsistent with the promulgated PM
monitoring requirements for cement plants that burn hazardous waste (40
CFR 63, Subpart EEE-2008), which includes provisions for the use of PMD
on both Electrostatic precipitators (ESP) and fabric filter controlled
cement. EPA appeared to be unaware that recent work at a cement plant
with membrane coated fabric filters has demonstrated that continuous
opacity monitoring system (COMS) can also meet the rigorous ASTM
D7392-07 requirements for certification as a PMD. 

Response: At this point, we still find that BLD is an appropriate
technology for PM monitoring. We believe that sources will work with
system vendors to install and use appropriate systems for kiln-specific
conditions. We note that cement kilns that burn hazardous waste operate
under different conditions than ordinary kilns. However, we will
continue to review advances in PM monitoring technologies in general and
will be happy to consider source-specific requests for alternative
approaches. The rule already provides BLD as an alternative for PM CEMS.

Comment: Several commenters (2845 and 2859) stated that EPA has not
demonstrated acceptable performance of BLDs to monitor kiln, kiln/
in-line raw mills at cement plants and has not developed necessary
guidance for or technical requirements that will be necessary for a BLD
to meet the requirements in the proposed rule. EPA has provided no
guidance on appropriate alarm setting procedures, installation,
operation, and quality assurance procedures and corrective action
procedures and relevant validated procedures are not available from the
measurement system suppliers for installations on cement kilns and
kilns/inline raw mills. According to the commenter, in the selection of
BLD devices it is clear that the most widely used probe type BLDs (such
as triboelectric, or probe electrification, or charge transfer devices)
are unsuitable for use on kilns/in-line raw mills because the responses
of these devices are proportional to the square or to the cube of the
local gas velocity and are also affected by changes in effluent
temperature, moisture content, and resistivity of the dust. All of these
factors change when the raw mill operational status changes. 

Several commenters (2845 and 2859) stated that because EPA did not
provide technical specifications, installation procedures, operation and
maintenance procedures, quality assurance procedures, or alarm setting
procedures or any relevant guidance for these issues for BLDs, the
burden is on the user to develop this information. If EPA cannot provide
technical requirements, criteria or guidance, then it certainly has no
basis to review the site-specific plans for approval. With no
established criteria or guidance, review by different regional offices
or states as delegated authorities will likely result in different
requirements and different interpretations, and confusion. The deadline
for submission of the site-specific BLD plan is not explicitly stated.
There is no time limit for EPA’s approval, no obligation for EPA to
render an opinion as to the acceptability of the plan, no instruction to
the user as to monitoring prior to receipt of approval, and no
indication of the consequences in the event the plan is not approved.
Until legitimate guidance and criteria are developed, the commenter
stated that it is sufficient that site-specific monitoring plans be
available for review. The commenters stated that, for the above reasons,
the requirement to submit the site-specific plans for approval should be
removed.

Several commenters (2845 and 2859) stated that EPA needs to provide
appropriate design and performance specifications so that manufacturers
can certify that their BLD meets applicable criteria in order that the
regulated user of such devices can rely on the manufacturer’s
certifications. Without an appropriate design and performance
specifications, there are no means available to the cement plant
operator to verify claims made by the BLD vendor or manufacturer
regarding detection capability. Plant operators cannot be responsible
for the design or performance of these systems and cannot be expected to
able to determine the technical validity of vendor claims, especially
where there are in effect no regulatory or technical requirements.

Response: While there is not a formal performance specification, there
is existing guidance on BLD. We have written a performance-based rule
for sources to use to tailor operation of the BDL to meet the needs at
each plant including variations in operation with the mill on vs. mill
off. The final rule clarifies that the resulting plans must be kept on
site. Submittal is not required. There are several types of BDL and, as
stated above, we believe that sources will work with system vendors to
install and use appropriate systems for kiln-specific conditions.
Sources have 3 years to comply with the final standards and should use
this time to ensure they are using a system that works well for their
application.

Comment: Several commenters (2845 and 2859) had the following comments
concerning BLD and alarms:

The proposed requirements for recording the BLD output and setting the
activating alarm are inconsistent with procedures actually used for
contemporary BLD.

EPA should remove the out-of-date requirement that the BLD should
activate an audible alarm and use more up-to-date guidance which applies
to settling alarm levels for the full range of BLD and fabric filter
configurations that are found at cement kilns that burn hazardous waste.
The proposal reflects bias to a particular manufacturer and particular
approach which is not applicable to all cement kilns. The commenter
provided several examples of what the commenter considered to be
out-of-date requirements.

The requirements to initiate procedures to determine the cause of every
alarm within one hour and alleviate the cause of the alarm within three
hours are problematic. EPA must revise its requirements to accommodate
real-world conditions that cannot be avoided and EPA must recognize
site-specific BLD monitoring plans will need to be revised as actual
experience is gained from the use of these devices at kilns and
kilns/in-line raw mills. The commenter provided examples to support
their position. 

The time allowed for alleviating the cause of a BLD alarm is not
sufficient especially when a false alarm is due to malfunction of the
BLD or an in appropriate alarm setting procedure.

The requirement for audible alarms may be distracting and are
unnecessary. The type of alarm should be at the discretion of the plant
operator. The commenter stated that many control rooms are equipped with
visual alarms that the operator must acknowledge and respond to. These
alarms include situations with far greater safety and emergency priority
and consequences than BLDs alarms and are critical to operation of the
plant. 

Response: We agree that an audible alarm is not required and have
changed the final rule to reflect this. We have also provided additional
time for the operator to determine the cause of the alarm (3 hours
instead of 1 hour) and up to 24 hours to alleviate the cause of the
alarm (up from 3 hours) to better reflect some of the complex situations
that operators may face.

PM detectors

Comment: Several commenters (2845 and 2859) had the following comments
concerning PMD:

Because PMD are a technically superior and more practical alternative to
BLDs, EPA should allow the use of PMD as an alternative to BLDs as is
allowed for cement plants that burn hazardous waste under 40 CFR 63
subpart EEE.

PMD should be allowed as an alternative to BLDs and/or ESP predictive
monitors and the PMD alarm set-point should be established during PM
source tests demonstrating compliance with the applicable PM standard.

PMD avoid the problems with attempting to set qualitative alarms based
on nearly instantaneous shifts in peak instrument responses, baseline
shifts, or rates of decay, as is the case with BLDs. PMD provide an
integrated average and warn the operator of the need for corrective
action, before an exceedance has occurred. According to the commenter,
EPA’s failure to allow the PMD under Subpart LLL causes redundant and
unnecessary monitoring for cement plants that are not subject to Subpart
EEE when not burning hazardous waste. For sources where the PM
concentration can be varied, a series of tests should be allowed to
determine the PMD output corresponding to the highest PM concentration
which is demonstrated to meet the applicable standard. For sources for
which the PM concentration cannot be varied (certain fabric filters)
alternative procedures to determine the PMD output corresponding to the
level of the emission limitation should be established. If future work
demonstrates that a representative PM spiking approach can be applied to
increase the PM level, the regulations should then be revised to allow
such an approach.

EPA should acknowledge that particulate matter detections systems (PMDS)
and/or BLDs meeting ASTM D7392-07 are acceptable for monitoring because
it was developed to fill the void in EPA requirements so that reliable
instruments could be used at cement plants that burn hazardous waste.
The practice provides a comprehensive set of design and performance
specifications for a broad range of monitoring technologies specifically
intended for monitoring of PM at cement plants. In addition to technical
requirements applicable to various cement plant configurations, the
Standard Practice includes detailed approaches for setting PMD and BLD
alarm set points that are more appropriate than those contained in EPA
guidance. It is also important to recognize that:

EPA was obligated to consider this ASTM standard under National
Technology Transfer Act, but apparently did not do so.

EPA Office of Air Quality Planning and Standards (OAQPS) and/or Office
of Solid Waste (OSW) representatives participated in the development of
the voluntary ASTM D7392-07 Standard Practice for use at cement kilns
that burn hazardous waste, offered comments during participation at
meetings and through correspondence on various drafts, and raised no
negative votes during balloting.

PMDS are superior to BLDs and are required by ASTM D 7392 to meet more
rigorous specifications, do not suffer from subjective and interpretive
alarm setting procedures, and should be allowed as an alternative.

PMDS can be used on kiln stacks with co-mingled emissions (e.g.,
kiln/in-line raw mill and coal mill emissions) where as separate BLDs
would be required for each baghouse.

PMDS are far more practical and far more cost effective than PM CEMS and
should be allowed as an alternative.

Teledyne Monitor Labs (the major American manufacturer of COMS used at
cement plants) has demonstrated that a contemporary LightHawk Model 560
transmissometer meeting ASTM D 6216-07 was able to meet the rigorous
ASTM D 7392-07 PMD requirements based on field tests at a cement kiln
burning hazardous waste equipped with high efficiency membrane coated
fabric filters.

Response: PMDs were not considered in the proposal. We will consider
requests to use PMDs on a case-by-case basis as an alternative to BLDs.
An owner or operator who wishes to use a PMD as an alternative
monitoring procedure must submit an application to the Administrator.

COMS

Comment: Several commenters (2845, 2859, and 2887) had the following
commenter concerning COMS:

COMS should be allowed as an alternative to PMDS and BLDS for both ESP
and fabric filter controlled sources as an option that can be selected
by the plant operator because they are a highly evolved, mature
technology that is well understood by both cement plant operators and
agency personnel. These devices are capable of detecting changes in PM
concentrations downstream of both ESP controlled sources and fabric
filter controlled sources and the need for inspection or corrective
action of the PM control device. The selection of the most appropriate
monitoring approach should be left to the plant operator and should
include the use of COMS, where applicable, as an option.

EPA’s attempt to dismiss COMS based simply on its opinion that they
lack sufficient accuracy at the proposed PM emission limit levels is
technically wrong and contrary to current work and other recently
promulgated regulations. EPA provides no technical support for its
dismissal of the accuracy of COMS and EPA appears to be completely
unaware of other developments that should have been considered. 

EPA’s dismissal of COMS for units equipped with fabric filters is
based on an incorrect assertion about measurement accuracy in favor of
BLDs which are a purely qualitative measurement technique making
absolutely no claim to accuracy and relying on subjective, interpretive,
and un-documentable procedures to set alarm set points. EPA attempts to
dismiss the long established and well understood use of COMS which
include:

Highly evolved ASTM D 6216-07 design and performance specifications for
manufacturer certification. 40 CFR 60, Appendix B, PS-1 (2008) and the
initial field certification procedures;

Well known quality assurance procedures and external audit capabilities
using NIST traceable calibration materials to facilitate audits by users
and control agencies;

Daily zero and upscale calibration checks, lens cleaning requirements
and other operation, maintenance, data recording, record keeping and
reporting requirements contained in the general provisions of 40 CFR 60
and 63;

A technology that is well understood by plant operators, plant
technicians, control agencies, and instrument manufactures;

Existing COMS requirements included in virtually every state operating
permit for cement kilns, kilns/in-line raw mills and clinker coolers.

Response: We disagree with the commenter that the accuracy of a COMS
meeting the ASTM D6216-07 and Performance Specification 1 requirements
is at issue in this decision. A COMS so verified will reproduce opacity
measurements satisfactorily (e.g., within 0.2 percent opacity) for
levels of 3 percent and higher. The concern about monitoring the
operation of a fabric filter control device is the need to discern
between very small changes in opacity that are already at very low
levels that would indicate a leak that may indicate a compliance
problem. Such leaks will produce changes in opacity of less than 0.1
percent out of an average opacity of 0.5 to 1 percent. These are values
that would be comparable to the instrumental output noise in the data
from COMS making the data insensitive to the leaks. PM CEMS and bag leak
detectors are much more sensitive to these levels of change necessary to
confirm compliance.

Comment: Several commenters (2845 and 2859) stated that the proposal
does not remove the COMS requirement for kilns and kiln and/or inline
raw mills for sources that would use BLDs. More important, EPA cannot
remove COMS requirements that already exist in state operating permits
for cement plants because it is expected that States will not be willing
to reopen permits and other States would refuse based on previous and
continuing reliance on COMS and Method 9 as the primary performance
tools that are more proven than BLDs. The commenter argued that
redundant BLDs and COMS are not wanted and are not needed. Due to these
complexities and overlapping requirements, the plant operator should
have the final decision as to the selection of the most appropriate
monitoring option.

Response: We agree that this was an oversight in the proposed rule and
we have corrected it in the final rule.

PM CEMS

Comment: Two commenters (2845 and 2859) said that EPA has not
promulgated any regulations requiring PM CEMS at any source category due
to its inability to address fundamental technical and policy issues and
must resolve these issues through formal rulemaking before requiring PM
CEMS at any cement plants. Furthermore EPA has not performed a
legitimate technical analysis of emissions variability and compliance
determination uncertainty to allow the use of PM CEMS for determining
continuous compliance with a PM limit at cement plants.

The use of PM CEMS in Europe and other countries does not constitute a
valid basis for application of PM CEMS at cement plants in the United
States. Light scattering, light transmission, and extractive beta
attenuation instruments are all inferential measurement devices and a
correlation must be established to relate the device output to the
actual PM concentration, then the accuracy and bias of the reference
test and the uncertainty of the statistical correlation, as well as the
stability of the correlation must be considered. Under the German TUV
and the European monitoring standard (EN 14181) these uncertainties are
considered; emissions are not considered to exceed the allowable limit
until the lower bound of the confidence interval and/or tolerance
interval exceeds the emission limit; emission standards may contain
different averaging periods requiring different levels of conformance;
and when a problem is encountered, the emphasis is on resolving the
emission problem rather than direct enforcement and collection of
financial penalties. All of these considerations place the European
monitoring program in an entirely different regulatory context than the
proposed PM monitoring requirements.

Response: We reject the commenters’ assertions that PM CEMS have not
been required via rulemaking because of unresolved fundamental technical
or policy issues. Concerns about PM CEMS were identified and addressed
prior to the January 2004 publication of Performance Specification 11
and Quality Assurance Procedure 2 for PM CEMS (69 FR 1786, January 12,
2004). As mentioned in that rule’s preamble, “…we believe that the
PM CEMS field demonstrations completed to date encompass a range of
operating conditions and emission characteristics…” including those
exhibited by sources such as cement kilns. Moreover, we disagree with
the assertion that our analysis of PM emissions variability is not
legitimate, yielding an overly-stringent PM emissions limit. The PM
limit is based on our analysis of PM emissions from test data, adjusted
from an hourly to a 30-day averaging period and further adjusted for
variability. As mentioned in the preamble to the Credible Evidence Rule
(62 FR 8314, February 24, 1997), we have addressed and continue to
address concerns about perceived “…limited number and distribution
of test runs and the inherent variability in levels of emissions…”
by a number of approaches, including changing emissions averaging
periods.

Certainly, lengthening the averaging period beyond the duration
associated with conducting a performance test (typically three hours) to
30 days without changing the value of the emissions limit does not
increase stringency of an emissions limit. As mentioned in the
description of the Salo and Pederson memoranda, “…(t)he effect of
the change from a 3-hour averaging time to a 30-day averaging time is to
make the standard more lenient…” Further, a statistically-based
adjustment to account for emissions variability, and which, in this
case, increases the numerical value of the standard (and its longer
averaging period) by fifty percent, does not make the standard more
stringent.

Comment: Two commenters (2845 and 2859) had the following comments
concerning technical issues associated with application of PM CEMS.

EPA has not addressed nor resolved the primary technical issues limiting
the effective application of PM CEMS at cement plants including:

inability to generate a sufficiently wide range of PM concentrations to
establish an acceptable correlation (i.e., calibration),

accuracy and precision limitations of reference method at PM levels
necessary to generate valid correlation, and

subsequent changes in effluent matrix and/or PM (i.e., particle size
distribution, refractive index, particle density, etc.) that influence
the stability of the correlation and hence, the relationship between the
output of the inferential measurement device relative to actual PM
concentration.

Valid PM CEMS correlations cannot be established for PM CEMS at cement
plants due to limitations of process operation and control equipment in
conjunction with the proposed emission limitation.

The requirements in Appendix A, PS-11 for the PM CEMS correlation and in
Appendix F, Procedure 2 do not provide a sufficiently reliable means to
determine compliance with emission limitations.

Response: We have not identified problems cited by the commenters at
existing installations. PS-11 establishes the acceptable range for the
concentrations for the correlation. This range is site specific. 

The acceptable accuracy and precision is specified in PS-11 by
statistical criteria. If the reference method accuracy and precision is
poor, a poor correlation would result and the statistical criteria
limits would not be met. PS-11 strongly suggests the use of paired
trains to insure that accuracy and precision is obtained. 

Section 2.4 of PS-11 stresses the importance of planning for subsequent
changes in effluent matrix and/or PM that might influence the
correlation and describes the process required to enhance the
probability of obtaining a successful correlation and the selection of
the most appropriate CEMS for the installation. 

We note that PS-11 has several features to address correlation issues
due to any limitations of process operation and control equipment. PS-11
provides for the addition of a zero point. For example, if control
equipment operations cannot be varied adequately to achieve higher PM
concentrations resulting in a cluster of data points at a very low
level, making it difficult to achieve PS-11 criteria, an artificial data
point may be selected at zero that allows the correlation curve to be
developed, meeting correlation criteria. 

PS-11 provides the steps in obtaining a reliable successful correlation
with reference methods, thereby insuring compliance data that is
sufficient to determine compliance with PM emission limits. Procedure 2
describes the required audits to insure that subsequent measurements are
stable and within acceptable limits, thereby insuring reliable
compliance data. 

Comment: Two commenters (2845 and 2859) had the following comments
concerning PS-11 and Procedure 2.The requirements at §63.1349 for PM
CEMS are incomplete and ambiguous and EPA has failed to specify
important QA frequencies and other information relevant to the
implementation of PM CEMS in accordance with PS-11 and Procedure 2. The
proposed Subpart LLL revisions fail to address critical elements
including the following sections of PS-11 and Procedure 2:

PS-11 3.20, species reference method as method defined in applicable
regulations (Method 5 with 250ºF filtration temperature) but this is
inadequate for low concentrations where Method 5I should be used, and is
inapplicable to sources with PM that condenses between the stack
temperature (mill on and mill off, if applicable) and 250ºF where
Method 17 should be used or ASTM D 6831.

PS-11, 6.2 You must ensure that the averaging time, the number of
measurements in an average, the minimum data availability, and the
averaging period for your CEMS conform to those specified in the
applicable regulation - but none are specified.

When using PS-11, 6.5 Your CEMS must sample the stack effluent such that
the averaging time, the number of measurements in an average, the
minimum sampling time, and the averaging procedure for reporting and
determining compliance conform to those specified in the applicable
regulation - but none are specified.

Procedure 2, 10.3 You must conduct an response correlation audit (RCA)
and an relative response audit (RRA) at the frequency specified in the
applicable regulation… but none are specified.

Procedure 2, 10.3, You must perform an RRA at the frequency specified in
the applicable regulation… but none is specified.

When using Procedure 2, 10.3(7) You must perform an RCA at the frequency
specified in the applicable regulation… but none is specified.

When using Procedure 2, 10.9 You must report the accuracy results for
your PM CEMS at the frequency specified in the applicable regulation…
but none is specified.

Response: We recognize that PS-11 does not specify a reference method;
we have revised the final rule to specify Method 5 or Method 5I (40 CFR
60, appendix A) as the reference method. Facilities with issues in
application of these reference methods, may petition the Administrator
for alternatives or modifications under §60.8(b) or §63.7(f). The
averaging times and data reduction specifications have been added to
§§60.63(c) and 63.1350(b) of the rule. There is no specific data
availability requirements, §§60.63(g) and 63.1348(b) require that
monitoring be conducted at all times the affected source is operating
except for periods of monitoring system malfunctions, repairs, or
quality assurance/quality control activities. The language of the final
rule has been revised to specify the frequency of the Relative Response
Audits (annually) and the Response Correlation Audits (every three
years), for specifics, see §§60.63(c) (2) and 63.1350(b)(2). Absolute
Correlation Audits are required by Procedure 2 on a quarterly basis.

Comment: Two commenters (2845 and 2859) said that the monitoring of PM
emission concentrations downstream of wet scrubbers is not possible at
cement plants due to the presence of water droplets and, during scrubber
upsets, gypsum slugs. Heated extractive systems (either Beta gage or
light scattering) are used at monitoring locations downstream of wet
scrubbers at electric utilities. However, none of these devices can
operate reliably in the presence of gypsum spray or very large
particles. During malfunctions, upsets of the PM APCD, or during period
of bypass, the wet scrubber will collect a large portion of the PM so
that monitoring upstream of the wet scrubber may overstate actual PM
emissions.

Response: We agree that PM CEMS would not work some of in these
situations. However, the types of wet scrubbers that would be used for
acid gas control are typically spray towers, which are not effective in
controlling PM. Therefore, in these situations the source could elect to
monitor for PM upstream of the wet scrubber without significantly
affecting the PM measurements during these malfunction situations. In
cases where this is not feasible, the source can apply for an
alternative monitoring on a source specific basis.	

Format of PM limit

Comment: Two commenters (2845 and 2859) stated that in the current
rulemaking, EPA proposed an alternate PM MACT standard for sources that
combine kiln and cooler exhaust streams to a single stack. The commenter
agreed that an alternate PM MACT standard is required for sources that
combine exhaust streams. However, the commenter argued that EPA failed
to recognize all potential exhaust streams and has overly complicated
the formula to account for the various exhaust gases. These commenters
recommended that the PM standard should be based on units of pounds PM
per ton of dry feed. This eliminates the need for a typical feed:clinker
conversion factor and have also recommended that EPA revise the PM MACT
standards for existing and new sources to be more in line with the
values derived by Commenter 2845. Using the values of 0.092 lb/t of dry
feed and 0.062 lb/t of dry feed for existing and new sources,
respectively and EPA’s assumption of a standardized kiln flow rate of
54,000 dscf/t of feed, the applicable grain loading rate for each case
would be 0.012 gram per dry standard cubic feet (gr/dscf) for existing
sources and 0.008 gr/dscf for new sources. Lastly, it is not necessary
to measure the volumetric rate from each exhaust stream separately. If
two or more streams are combined into a single exhaust point, the
measurable volumetric flow rates will already be added. A simplified
formula can be found as part of the comments for Commenter 2859.

Response: As we noted at proposal, some kilns combine the clinker cooler
gas with the kiln exhaust and send the combined emissions to a single
control device. There are significant energy savings (and attendant
greenhouse gas emission reductions) associated with this practice, since
heat can be extracted from the clinker cooler exhaust. However, there
need to be different conversion factors from concentration to mass per
unit clinker in these cases to allow for the increased gas flow, which
result in a different PM emissions limit. We proposed adjustment factors
that would account for these differences and create a site specific PM
emission limit of this situation. See 74 FR 21155-56 and 2184. We
received no comments on these factors and are thus adopting them as
proposed, except that the factors have been changed to account for
changes n the underlying kiln and clinker cooler emissions limits.

A site-specific compliance method can be requested for situations where
additional sources of particulate are included in the same duct system
leading to one stack. 

See the preamble discussion under format of the normalized standards in
the NESHAP and the NSPS, for more information on EPA’s decision
regarding format of the standard.

Comment: One commenter (2844) requested that sources be given the option
of compliance with an equivalent grain loading figure that would be in
the rule and the rule automatically approve the current methods for
converting feed to clinker. This should be done on a monthly rather than
daily basis. The commenter ask that there be an ongoing compliance
demonstration (between stack tests) alternative based on retaining the
use of COMS as the indicator of acceptable operating conditions.

Response: See the preamble discussion under format of the normalized
standards in the NESHAP and the NSPS, for more information on EPA’s
decision regarding format of the standard.

We are removing all opacity standards for kilns and clinker coolers
because these sources will be required to monitor compliance with the PM
emissions limits by more accurate means. Although some commenters
requested retention of opacity as a backup standard, and others as an
alternative, none of these comments offered any convincing information
or other justification for perpetuating a less reliable compliance
methodology.

Clarifications and corrections

Comment: Several commenters (2845 and 2859) stated that EPA has several
areas in their proposed regulatory language that must be revised. In
particular, EPA has language requiring opacity monitoring for kilns and
coolers that no longer have opacity standards. Also, EPA has mandated
performance testing where CEMS are required. Performance testing is
typically required for sources that must conduct stack testing
periodically to demonstrate that compliance is maintained. If CEMS are
mandated performance testing is not necessary when compliance is
demonstrated on a continuous basis.

One commenter (2845) stated that the current language for
§63.1343(b)(2) limits opacity from kilns to no greater than 20 percent
and in §63.1343(a)(2) limits opacity from clinker coolers to no greater
than 10 percent, but the preamble of the proposed rule states We are
proposing to remove all opacity standards for kilns and clinker coolers
because these sources will be required to monitor compliance with the PM
emissions limits by more accurate means. The commenter suggested that
EPA should remove the opacity limits from the rule text for sources
demonstrating compliance with a BLDS, PMDS or PM CEMS unless opacity
monitoring is used as an alternative to the proposed options for
monitoring PM emissions. EPA should provide an option for sources to
comply with an opacity limit of 10 percent for kilns and 5 percent for
clinker coolers using a COMS in lieu of the requirements for a BLD, PMD,
and/or PM CEMS.

One commenter (2845) stated that §63.1349(b)(1) should be updated to
account for clinker coolers that use PM CEMS. The commenter suggested
the following alternative language: §63.1349 (b)(1) .The owner or
operator of a clinker cooler subject to limitations on PM emissions that
is not equipped with a PM CEMS shall demonstrate initial compliance by
conducting a performance test as specified in paragraphs (b)(1)(i)
through (b)(1)(iii) of this section. The owner or operator shall
determine the opacity of PM emissions exhibited during the period of the
Method 5 (40 CFR 60, Appendix A–3) performance tests required by
paragraph (b)(1)(i) of this section as required in paragraphs (b)(1)(v)
through (vi) of this section, if the unit is not equipped with a BLDS,
PMDS or PM CEMS.

One commenter (2845) stated that the opacity limits in §63.1349(b)(1),
§63.1350(c), §63.1350(d) need to be removed from the rule text for
source demonstrating compliance with a BLDS, PMDS, or PM CEMS unless
opacity monitoring is used as an alternative to the proposed options for
monitoring PM emissions. Example language for each section is provided
by the commenter.

Response: We have corrected the rule to remove duplicative requirements
related to opacity monitoring for units that no longer have opacity
standards and performance testing requirements for sources that are
using CEMS. Regarding the comment that COMS be allowed as an
alternative, EPA’s experience is that facilities can significantly
exceed PM limits without exceeding opacity limits. In contrast, BLD
systems are much more sensitive to the point they are considered to be
CEMS. Similarly, we believe that PM CEMS are a superior technology as
well. However, we would be happy to review evidence supporting more
advanced COMS systems.

Comment: One commenter (2859) stated that EPA has not specifically
stated in the rulemaking what the averaging period for the PM MACT
standard will be. EPA has specified that the averaging periods for
mercury, THC, and HCl are 30-day rolling averages. EPA has proposed CEMS
for compliance demonstration for each of these pollutants. For
consistency purposes, EPA should designate that the PM MACT standard
should also be a 30-day rolling average if a PM CEMS is used to
demonstrate compliance. If sources choose to use BLDS or other approved
compliance demonstration methods, then a performance test consisting of
the average of three one-hour stack testing periods would be the
appropriate compliance averaging period for the PM MACT standard.

One commenter (2845) stated that there is no reference in §63.1343(a)
to the averaging period for the PM limit if PM CEMS are used. The
commenter suggested that the last sentence of §63.1343(a) be changed to
read, If using a CEM to determine compliance with the PM or HCl
standards, these standards are based on a 30–day rolling average.

Response: We have corrected the rule to provide for a 30-day rolling
average when using PM CEMS.

Comment: Two commenters (2845 and 2859) stated that EPA should clarify
whether an ESP predictive model may be used in lieu of other PM
monitoring approaches and what requirements would apply if this option
is selected. No mention of an ESP predictive model is found in the
performance testing or monitoring sections of the proposed rule. If EPA
allows use of a predictive ESP model as an alternative to other
monitoring, then the minimum parameters and technical criteria for
determining the acceptability of the predictive model should be
specified.

Response: We will consider source-specific requests to use an ESP
predictive model

Comment: Two commenters (2845 and 2859) stated that it should be
clarified that kiln/raw mill and cooler stack(s) that have installed
BLD, PMD, or PM CEMS would not be subject to monthly 20-minute visible
emissions test of each affected source in accordance with Method 22
under §63.1350(a)(4)(i).

Response: We have corrected the rule to remove this duplicative
requirement.

Comment: Two commenters (2845 and 2859) stated that EPA has not provided
any technical basis for changing the duration of the Method 22 test in
§63.1350(a)(4)(i) and that changing the duration from 1 minute to 20
minutes for numerous sources places an unnecessary burden on the plant
operators. An increase in the duration of visible emissions from 1
minute to six minutes would be a more reasonable change and would be
more consistent with the minimum observation period for Method 9 visible
emission observations.

Response: While we do not think this is an unreasonable requirement once
per month, we have reduced the observation period in half to 10 minutes,
which should be sufficient to observe any problems. There is no reason
to be consistent with a Method 9 observation period, because the tests
are different.

Comment: One commenter (2845) stated that Equations 3 and 4 apply a
metric conversion factor to the emission rate. The commenter stated that
the formula does not work in metric units, and suggested alternative
wording for these equations that had the emission rate in only lb/ton.

Response: We have not been able to identify the problem raised by the
commenter. The units in the equations are those that are commonly
measured.

Comment: One commenter (2845) stated that §63.1349(b)(1)(vi) should
either be removed or updated to account for clinker coolers. The
averaging period for sources using PM CEMS for compliance with the
particulate matter standards should be based on a 30-day rolling
average. The commenter noted that there is no need to conduct a separate
performance test under raw mill on/off conditions when using a PM CEMS
for measuring compliance as the PM CEMS will measure compliance and thus
account for both conditions.

One commenter (2845) stated that §63.1350(a)(4)(v) indicates that
visible emissions monitoring requirements do not apply to totally
enclosed conveying system transfer points. However, §63.1349 does not
currently indicate that initial performance testing is not required for
totally enclosed conveying system transfer points. Section 63.1349
should be revised to clarify that initial performance testing is not
required for totally enclosed sources. 

Response: We have corrected the rule to remove the inconsistencies
identified by the commenter. We have also clarified the rule to specify
a 30-day rolling average.

Comment: One commenter (2845) stated that §63.1350(a)(4)(vii) indicates
that you must test visible emissions from each side, roof, and vent of
the building. However, this is not included in the initial performance
testing section. Section 63.1349 should be revised to be consistent with
the wording in §63.1350(a)(4)(vii). Further, a 6 minute observation
should be sufficient to demonstrate compliance with visible emissions
standards. The 6 minute duration would also be consistent with the
minimum observation time for a Method 9 data set (40 CFR 60 Appendix A,
Method 9, Section 2.5 Data Reduction). The commenter provides example
alternative language.

Response: We have corrected to the rule to include the performance
testing requirements. Regarding the length of the observation period, we
have reduced the period to 10 minutes, which should be sufficient to
observe any problems. There is no reason to be consistent with a Method
9 observation period, because the tests are different.

Comment: One commenter (2845) stated that in regard to §63.1349(b)(2),
the proposed rule should specify test requirements for sources with
limited operation. An option would be to use language similar to that
used in NSPS Subpart OOO (see 40 C.F.R. §60.675(c)(2)(ii): The duration
of the Method 9 (40 C.F.R. Part 60, Appendix A–4) observations may be
reduced to the duration the affected facility operates (but not less
than 30 minutes) for baghouses that control storage bins or enclosed
truck or railcar loading stations that operate for less than 1 hour at a
time.).

Response: The observation does not need to be consecutive, which we have
clarified in the final rule to address this situation.

Comment: One commenter (2845) stated that the proposed requirement to
conduct five 6-minute Method 9 tests is unnecessary. No rationale was
provided for the increase in the number of tests. 

Response: This test is only required when visible emissions are
identified, which should be a limited occurrence. However, when it does
occur, we think that conducting a valid Method 9 test would take at
least 30 minutes.

Comment: One commenter (2845) stated that §63.1350(k) does not
currently address PMDS. Multiple portland cement plants currently use
PMDS for compliance with Subpart EEE, and language used in
§63.1206(c)(9) (NESHAP Subpart EEE) provides PMDS requirements for
Hazardous Waste Combustors that utilize PMDS to monitor compliance with
PM standards which could be used as a template for inclusion in Subpart
LLL.

Response: We will consider source-specific requests to use PMDS

Comment: One commenter (2845) stated that an exclusion installing,
operating, and maintaining a BLDS for sources using a PM CEMS should be
included in §63.1350(k). The commenter also noted that the rule only
applies to sources with fabric filters, which is not a defined term. For
clarification, the rule should specifically state that it does not apply
to sources with gravel bed filters and ESPs.

Response: We have clarified that sources using a PM CEMS are not
required to use a BLDS. We have clarified in the final rule that fabric
filters do not include gravel bed filters and ESPs.

Comment: One commenter (2845) suggested that EPA delete
§63.1350(k)(1)(i) or develop design and performance specifications for
manufacturers to certify their BLDS to meet the criteria required in
this section.

Response: The proposed requirements have been used successfully in
several other rules. EPA is developing a BLD performance specification
that will be put in place when it is available.

Comment: One commenter (2845) stated that in §63.1350(k)(2), the phrase
At a minimum you must retain records related to the site specific
monitoring plan and information discussed in paragraphs (k)(2)(i)
through (k)(2)(vi) of this section for a period of 2 years on-site and 3
years off-site implies that off-site record storage is mandatory rather
than optional. The commenter suggested that EPA should only specify the
minimum period of time required for overall record retention and the
minimum period of time required for onsite storage, leaving the option
for offsite storage to the discretion of the owner/operator.

Response: We agree with the clarification suggested by the commenter and
have added it to the rule.

Comment: One commenter (2845) stated that §63.1350(k)(4) references
kilns and clinker coolers, but only the emission limits in §63.1343,
which is for kilns. To be consistent with §63.1350(p), the clinker
cooler emission limit reference would need to be added. However, this
paragraph is also a duplicate of §63.1350(p) introducing the PM CEM
requirements and it would be more appropriate to delete it altogether to
prevent confusion.

Response: We agree and have made the change suggested by the commenter.

Comment: One commenter (2845) stated that the reference to new or
existing sources should be removed from §63.1350(o). This should apply
to all sources; new, existing, or reconstructed.

Response: We agree with the clarification suggested by the commenter and
have added it to the rule.

Comment: One commenter (2845) stated that opacity standards only apply
to sources that elect to utilize a COM for PM compliance demonstration,
therefore, correlations with opacity will no longer be needed and should
be deleted from portions of §63.1357.

Response: We agree with the clarification suggested by the commenter and
have added it to the rule.

Transmisometers

Comment: One commenter (2879) stated that the European cement industry
has used transmisometers to report PM emissions for many years. The
commenter stated that opacity readings underreport PM emissions, and
provided emission test data using transmisometers. The commenter also
noted that BLDS are not adequate to measure noncompliance, and
recommended that EPA should adopt regulations similar to those
established in Europe, the parent companies of most American cement
companies, that require particulate emissions and associated metals
emissions be determined using monthly averaged transmisometer readings. 

Response: We have clearly established the value of PM CEMS for this
industry, and we believe PM CEMS are superior to BLDS for this industry.
 	

Advantages of opacity standards

Comment: One commenter (2786) stated that EPA should not eliminate
opacity standards in the proposed rule. The commenter stated that there
are benefits to having an opacity standard in conjunction with a
particulate matter standard. Opacity measurements can be made by anyone
who is trained to measure opacity, which can include members of the
public and not just inspectors, and opacity measurements are a cheaper
method of getting more frequent measurements.

Response: See response in preamble

Monovent compliance demonstration

Comment: One commenter (2888) requested that monovent kilns be allowed
to continue compliance demonstration as outlined in the current PC MACT.
The monovent design orientation does not provide a means for either BLD,
a PM CEM or and opacity monitor.

Response: The rule contains the necessary and appropriate monitoring
requirements. Monovent sources can add a stack. Or, they can apply for
alternative monitoring requirements on a source-specific basis.	

Require PM CEMS and retain opacity monitors

Comment: One commenter (2898) stated that EPA should require PM CEMS and
retain the opacity monitoring requirements. EPA is proposing
installation and operation of a BLD system, along with stack testing
using EPA Method 5 conducted at a frequency of five years for
demonstrating compliance with the proposed PM emissions limit. As an
alternative a PM CEMS that meets the requirements of PS–11 may be
used, and EPA is proposing to eliminate the current requirement of using
an opacity monitor. The proposed rule solicits comment on making the use
of a PM CEMS a requirement. The commenter stated that EPA should both
require CEMS and retain the use of opacity monitors.

EPA should abandon the BLD system requirement outlined in the proposed
rule and mandate the use of PM CEMS instead. The agency previously
concluded that PM CEMS is a superior monitoring technology that can be
implemented at a reasonable price. EPA has found that BLD systems,
standing alone, are inadequate to verify compliance and has also found
that COMS operate as a useful check on PM emissions and proper operation
of PM CEMS. 

Providing a superior level of compliance assurance is not the only
benefit of PM CEMS. EPA has acknowledged that the assumptions to assure
compliance are fewer and less conservative (direct measure of the
standard is the top of the monitoring hierarchy), CEMS mean facilities
need to monitor only one emissions parameter to assure compliance rather
than multiple operating limits, often relevant to more than one
standard, and that the cost of installing PM CEMS technology is
reasonable. 

Response: See response in preamble. 

HAP metals monitoring

Comment: One commenter (2638) stated that proven HAP metals monitoring
methodology is now commercially available.

Response: We are aware that multi-metal CEMS have been tested and would
support their use should they become readily available. However, at this
time we do not have sufficient data to set individual metals standards
for this source category, which would be necessary of we were to use
multi-metals CEMS to demonstrate compliance. .In addition, we believe
these systems are not sufficiently demonstrated for this source category
to be the basis of compliance measurement

30-day rolling average compliance period for PM CEMS

Comment: One commenter (2859) recommended that EPA adopt a 30-day
rolling average standard for sources that utilize PM CEMS. Sources that
opt to use an alternative compliance demonstration methodology (stack
testing) would demonstrate compliance by conforming to the relevant test
method.

Response: We agree with the clarification suggested by the commenter and
have added it to the rule. 

2.8.5	Impacts

Scrubber performance on PM control

Comment: One commenter (2900) provided stack test data for PM before and
after a wet scrubber was installed. The data shows that the emissions
would still exceed EPA requirements even after optimization of the
scrubber through the last five to six years. The commenter stated that
this scrubber will require some additional equipment or modification in
its operation to achieve the EPA target and it remains uncertain if this
is achievable. However, it can be reported that based on the current
data, it would appear that the emissions were significantly reduced
after installation of the wet scrubber.

Response: Most if not all cement manufacturers are using fabric filters
or electrostatic precipitators in order to reduce PM emissions. We
believe that in order to comply with the final limit for PM for both
existing and new sources, fabric filters using membrane bags will be
required. 

Benefits of PM2.5 reductions

Comment: One commenter (2916) stated that EPA should not use a
no-threshold model for estimating the benefits of reducing PM2.5
emissions. EPA’s proposal to adopt a no-threshold model as a part of
this rulemaking will obscure the greater uncertainty associated with
calculated premature mortality at PM concentration levels down to
background (or even below background). The treatment of uncertainty is
one of the key issues associated with conducting benefit/cost analysis.
A number of National Research Council reports have found that the proper
characterization of uncertainty is essential. The commenter recommended
that if EPA adopts a no-threshold model to develop estimates of
mortality, it should present its premature mortality estimates within
concentration ranges and provide a complete discussion of the
uncertainty in the estimates within each concentration range. EPA should
discuss the plausibility and the uncertainties of the estimates within
each concentration category in the body of its analysis and in its
Executive Summary.

Several commenters (2845 and 2849) stated that for the RIA, EPA
abandoned their traditional assumption that there may be a threshold in
the concentration-response function relating exposure to PM2.5 and
premature mortality and EPA should not change its approach on this
important issue until such a shift has been recommended by Clean Air
Scientific Advisory Committee (CASAC) and adopted as final policy. If
EPA were to assume a threshold at 10 ug/m3 as has been done in recent
RIAs, estimated benefits would be some 20–40 percent lower than what
the agency now calculates in the Portland Cement NESHAP RIA.

One commenter (2845) identified the following as serious shortcomings in
EPA’s estimate of benefits that include the following:

EPA has overstated the benefits of PM2.5 reductions by removing the
assumption of a threshold in the dose-response relationship for
premature mortality. This is a significant methodological that deserves
a discussion in the context of the ongoing review of the PM NAAQS,
rather than in the context of a NESHAP for a single industry that is not
even focused on PM2.5. This commenter is concerned that a premature
decision on this issue as a part of this rulemaking will serve as a
precedent for decisions in the PM NAAQS review.

EPA’s use of a no-threshold model in this rulemaking will obscure the
greater uncertainty associated with calculated premature mortality at PM
concentration levels down to or below background. The treatment of
uncertainty is one of the key issues associated with conducting a
benefit/cost analysis. 

In setting the NAAQS, EPA has recognized that there is greater
uncertainty in the relationship between the risk of premature mortality
and PM exposure at low concentrations. In the quantitative risk
assessment in the 2006 final rule for the PM NAAQS, EPA presented
several estimates based on different cut points intended as surrogates
for alternative assumed population thresholds. In reviewing the Staff
Paper and risk assessment, the CASAC panel commented that for the
purpose of estimating public health effects, it favored the primary use
of an assumed threshold of 10 ug/m3.

Several factors can contribute to the uncertainty in estimating
premature mortality at low PM concentrations. EPA has generally assumed
either a proportional or quadratic rollback to represent the
effectiveness of control strategies designed to meet the NAAQS, but in
areas with PM levels below 10 ug/m3 there may be little or no change in
air quality. If EPA adopts a no-threshold model, it should address these
several factors in its discussion of the uncertainties in the premature
mortality estimates. Commenter 2845 stated that this discussion on
whether to adopt a no-threshold model and the uncertainties associated
with EPA’s premature mortality estimates should include a separate
discussion for those estimates associated with exposure at lower ambient
concentration levels.

Commenter 2845 noted that EPA had adopted a no-threshold model in the
development of PM benefits for the Tier 2 rule and on-highway heavy duty
diesel engine rules in 1999 and 2001. However, the 2002 National
Research Council report estimating the Public Health Benefits of
Proposed Air Regulations suggested that EPA’s approach in these two
rules failed to provide information on the plausibility of estimates at
different PM concentration levels. The committee suggested that
alternative approaches that reflected expert judgment on the
plausibility of thresholds would have represented an improvement in
EPA’s primary analysis. EPA has recently carried out a multi-year
effort using expert elicitation to assess the relationship between
exposures to fine PM and the incidence of mortality. If EPA adopts a
no-threshold model, EPA should incorporate the results of this study in
presenting its PM benefits analysis for different concentration levels.

Response: The weight of scientific evidence strongly supports modeling
PM-related mortality and morbidity by using concentration-response
functions that do not incorporate an assumed threshold. In 1999, the SAB
Advisory Council on Clean Air Compliance Analysis (ACCACA) concluded
that there was currently no scientific basis for selecting any specific
threshold (EPA-SAB-Council-ADV-99-012). In 2004, the Health and
Environmental Effect subcommittee of ACCACA concluded, “it is
reasonable for EPA to assume a no threshold model down to, at least, the
low end of the concentrations reported in the studies”
(EPA-SAB-COUNCIL-ADV-04-002). In 2005, CASAC indicated that “[t]he
available epidemiological database on daily mortality and morbidity does
not establish either the presence or absence of threshold concentrations
for adverse health effects” (EPA-SAB-CASAC-05-007). 

In 2006, EPA conducted an expert elicitation on the relationship between
PM2.5 exposure and premature mortality. Eleven of the twelve experts in
this study did not specify a threshold in their concentration-response
function, whereas one expert specified a 50% probability of
no-threshold, 40% probability of a threshold below 5 µg/m3, and 10%
probability of a threshold between 5 and 10 µg/m3. EPA presents these
expert-derived results in the RIA independently of whether thresholds
are included in the core benefits results or not. 

In 2009, the final Integrated Science Assessment for Particulate Matter
indicated that “[o]verall, the studies evaluated further support the
use of a no-threshold log-linear model…” (U.S. EPA, 2009). In 2010,
the Health Effects Subcommittee of the Science Advisory Board “fully
supports EPA’s decision to use a no-threshold model to estimate
mortality reductions. This decision is supported by the data, which are
quite consistent in showing effects down to the lowest measured levels.
Analyses of cohorts using data from more recent years, during which time
PM concentrations have fallen, continue to report strong associations
with mortality. Therefore, there is no evidence to support a truncation
of the CRF ” (EPA-SAB-COUNCIL- -10-001).

Our approach to estimating PM-related health impacts is consistent with
the key findings in these reports. In addition, our approach is
consistent with the on-going PM NAAQS review. 

The commenter incorrectly suggests that using a no-threshold
concentration-response function would erroneously model benefits down
to, or below, background. Instead, we applied benefit per ton estimates
that were calculated using previously generated air quality scenarios
projecting air quality changes well above background levels. For the
final analysis, we estimated benefits on the basis of modeled air
quality changes from CAMx, where we again considered air quality changes
well above background levels.

In conjunction with the commenter’s concern, and consistent with
recent scientific advice, we have replaced the previous threshold
sensitivity analysis with a new “lowest measured level” (LML)
assessment. While an LML assessment provides some insight into the level
of uncertainty in the estimated PM mortality benefits, EPA does not view
the LML as a threshold and continues to quantify the PM-related
mortality impacts using the full range of modeled air quality
concentrations. The results of this analysis are provided in Section 6
of the RIA. 

2.8.6	Other

PM as surrogate for HAP metals

Comment: Several commenters (2832 and 2859) stated that the use of PM as
a surrogate for HAP metals is an important issue in this proposal;
however, EPA has failed to explain its selection of using PM as a
surrogate for HAP metals and how that approach was applied specifically
to the data in this administrative record. EPA is required to
specifically detail how this approach was used and what other approaches
were taken and rejected and why as part of the Preamble to the proposal.

Response: Use of PM as a surrogate for non-volatile HAP metals – in
this very industry – has already been upheld. See National Lime II,
233 F. 3d at 639.

Comment: One commenter (2638) stated that it is no longer appropriate to
use PM as a surrogate for HAP metals. The commenter provided citations
of studies, and summarized the main points supporting this statement.
The commenter further argued that the courts have questioned whether the
use of PM as a surrogate for HAP metals is scientifically reasonable,
that HAP metals are not invariably present in cement kiln PM, PM control
technology does not indiscriminately capture HAP metals along with other
particulates, and PM control is not the only means by which facilities
achieve reductions in HAP metal emissions.

The commenter stated that direct measurement of HAP metals is essential
for the EPA to comply with the spirit and intent of the Clean Air Act,
and direct measurement of HAP metals is required to evaluate residual
risks.

Response:    The EPA disagrees with the commenter.  We continue to
believe this is an appropriate surrogate.  Particulate matter controls
do in fact capture HAP metals along with other particulate matter, with
the exception of mercury, for which we have set separate emission limits
under this rule making.  We have shown that particulate matter is an
appropriate surrogate for non-volatile metal HAP in numerous previous
rulemakings, including  the NESHAPs for Iron and Steel Foundries (69FR
21912, April 22, 2004) and the NESHAP for Taconite Iron Ore Processing
(68 FR 61885, October 30, 2003).

NESHAP and NSPS PM limits

Comment: Several commenters (2845 and 2874) agreed that if EPA is
proposing PM limits under both the NESHAP and NSPS, the source should
comply with the most stringent emission limit. The commenters suggested
that EPA explicitly exempt the source from complying with all associated
requirements (monitoring, recordkeeping, and reporting) relating to the
less stringent limit, not just from that limit itself. 

Response: The standards are now identical. 

Volatilized metal HAP emissions

Comment: One commenter (2898) stated that under the high temperature
conditions of the cement kiln non-volatile metal HAP emissions could
occur outside of the particulate form. Therefore, EPA should assess the
emission levels and associated public health threats remaining after
successful implementation of the proposed standard. 

Response: EPA already has a mechanism in place for conducting an
assessment of public health threats remaining after implementation of
MACT. The Risk and Technology Review (RTR) is a combined effort to
evaluate both risk and technology as required by the Clean Air Act (CAA)
after the application of maximum achievable control technology (MACT)
standards. Section 112(f) of the CAA requires EPA to complete a Report
to Congress that includes a discussion of methods the EPA would use to
evaluate the risks remaining after the application of MACT standards.
These are known as residual risks. Section 112(f)(2) directs EPA to
conduct risk assessments on each source category subject to MACT
standards, and to determine if additional standards are needed to reduce
residual risks. Section 112(d)(6) of the CAA requires EPA to review and
revise the MACT standards, as necessary (but no less often than every 8
years), taking into account developments in practices, processes and
control technologies.

Thallium

Comment: One commenter (2144) stated that thallium metal (Tl) and its
compounds are highly toxic materials and are emitted from certain cement
plant on the finest dust particles emanating from the stacks. The
commenter stated that concentrations in certain cement filter dusts can
build up to levels greater than 1 percent (10,000 ppm) and that proper
regulation of PM coupled with the required mercury reduction will
concurrently address such Tl concentration buildup and emissions. In the
long run, consider regulation of Tl as a HAP consistent with present
European practice.

Response: Thallium is not currently listed as a hazardous air pollutant
(HAP) under Section 112(b) of the CAA. As such, EPA does not have the
authority to regulate emissions of thallium under the provisions for
establishing MACT standards.

2.9	Compliance Dates

Arbitrary compliance dates

Comment: Many commenters (2832, 2845, 2859, 2863, 2864, 2874, 2890,
2914, and 3193) stated that EPA should follow the intent of the CAA
sections 111 and 112 which established the requirement that new and
existing sources recognized in the proposed rule use the published date
of May 06, 2009, as the new source trigger date for mercury, HCl, and
THC emission standards instead of the proposed retroactive date of
December 2005 when the amendments for mercury, THC, and HCl were first
proposed. The commenters stated that despite the direct acknowledgements
in the proposed rule, EPA arbitrarily decided to apply a proposal date
for mercury, THC and HCl that reaches back and applies new source
standards to any plant constructed since December of 2005 and admitted
that the 2005 proposal did state EPA’s desire to establish floors for
THC, HCl and mercury. According to the commenters, this is irrelevant
because sources constructed in the past four years cannot have
reasonably anticipated their need to comply with the 2009 proposed
floors based on the 2005 proposed rulemaking. The 2005 rulemaking
resulted in a final rule with floors for THC and mercury, and no HCl
floor (due to its classification as a threshold pollutant), which are
significantly different than the 2009 proposal and the 2005 proposal did
not reference a desire to establish a floor for PM, nor one as stringent
as that set forth in the current proposal.

Response: As explained in the preamble, EPA agrees with these comments
and has adopted a May 6, 2009, compliance date for determining when
sources come into existence for purposes of determining which sources
are subject to new source MACT.

Reduced compliance period

Comment: One commenter (2891) disagreed with EPA’s decision to grant 3
years from the date of the rulemaking for existing sources to come into
compliance. Given the serious impacts of these emissions on public
health and the fact that the industry has had ample warning of these
standards, a shorter compliance timetable is reasonable. The commenter
supported a compliance date of no more than 2 years.

The commenter agrees that new sources should comply upon commencement of
operation. The commenter agreed with EPA’s selection of December 5,
2005, as the proper date for determining when a source is deemed to be
new for THC, mercury and HCL. The commenter stated that the
determination for PM ought to be March 24, 1998, which is the date that
EPA first proposed PM restrictions on this sector, and is therefore a
date consistent with the language and intention of the Act.

Response: The commenters do not address EPA’s analysis at proposal
indicating that the phrase “first proposes” in section 112 (a)(4) is
ambiguous, nor other provisions (e.g. sections 112 (i)(2)) indicating
that the phrase is ambiguous. Nor did the commenters address the similar
language in section 111 (a)(2) for NSPS, which has never been
interpreted as stretching back in time for decades to match an initial
proposal date. For the reasons stated in the preamble to the final rule,
EPA believes the proposed date of these amendments is a proper
compliance date for purposes of determining the applicable date for new,
reconstructed, or modified sources. 

Comment: One commenter (2898) recommended that EPA consider a faster
timeline for existing sources implementation and to finalize its
proposal on compliance dates for new sources. EPA has proposed to use
the date of December 5, 2005, as the date for determining if a source is
existing or new for the purposes of the mercury, THC, and HCl standards,
and has proposed May 6, 2009, as the relevant date for PM standards. The
commenter supported EPA’s proposed dates as a reasonable
interpretation of the statute’s definition of a new source as a
stationary source the construction or reconstruction of which is
commenced after the Administrator first proposes regulations under this
section establishing an emissions standard applicable to such source. 42
U.S.C. §7412(a)(4). We agree that the statutory definition should be
read to apply pollutant-by-pollutant and to the rulemaking record under
which a standard is developed to avoid subjecting sources in existence
for 10 years to the new standards which they could not have anticipated
at the time. This is consistent with the statute’s structure requiring
more stringent standards for new sources.

Response: As explained in the preamble, EPA has adopted a May 6, 2009,
compliance date for determining when sources come into existence for
purposes of determining which sources are subject to new source MACT.
For the reasons stated in the preamble to the final rule, EPA believes
the proposed date of these amendments is a proper compliance date for
purposes of determining the applicable date for new, reconstructed, or
modified sources.

For existing sources we proposed a compliance date of 3 years after the
promulgation of the new emission limits for mercury, THC, PM, and HCl to
take effect. This is the maximum period allowed by law (subject to a
one-year extension determined on an individualized basis to install
technology). See section 112(i)(3)(A) and (B). We continue to believe a
3 year compliance period is justified because most facilities will have
to install emissions control devices (and in some cases multiple
devices) to comply with the proposed emissions limits. Therefore, we
have retained a 3 year compliance data in this final rule. 

For new sources, the compliance date will be the effective date of this
final rule or startup, whichever is later. Because this is a major rule
as defined by the Congressional Review Act, the effective date of the
rule is 60 days after publication of the Federal Register. 

In determining the proposal date that determines if a source is existing
or new, we have decided to select the proposal date of these final
amendments, which is May 6, 2009, for all the standards. 

At proposal, we considered three possible dates, including March 24,
1998; December 5, 2005; and the proposal date of these final amendments,
which was May 6, 2009. As we noted at proposal, Section 112 (a)(4) of
the Act states that a new source is a stationary source if “the
construction or reconstruction of which is commenced after the
Administrator first proposes regulations under this section establishing
an emissions standard applicable to such source.” “First proposes”
could refer to the date EPA first proposes standards for the source
category as a whole, or could refer to the date the agency first
proposes standards under a particular rulemaking record or first
proposes the particular standards at issue. The definition is also
ambiguous with regard to whether it refers to a standard for the source
as a whole, or to a HAP-specific standard (so that there could be
different new source standards for different HAP which are regulated at
different times). At proposal we chose the date of December 5, 2005, as
the proposal date that determines if a source is new or existing for the
mercury, HCl, and THC, and the May 6, 2009, date for PM. 

After consideration of comments on the selection of the date for
mercury, THC, and HCl, we believe that the May 6, 2009, date for all
pollutants is more in keeping with the evident intent of Section 112
(a)(4) that source should have sufficient notice that new source
controls requirements can be considered in the initial design. We accept
commenters’ argument that sources coming into existence between the
proposed date of the 2006 standards and the May 6, 2009, proposal date
of these amendments would have no reasonable means of ascertaining the
standards’ final content and so lacked notice of what controls and
strategies to adopt. Since this is antithetical to the policy underlying
new source standards, EPA is adopting May 6, 2009, as the date which
determines if a source is existing or new.

We note that there are currently sources subject to new source limits
for mercury and THC contained in the December 20, 2006, rule. However,
the mercury the new source standards in this final rule are
significantly different than the limits in the December 20, 2006, rule,
and we do not see how the affected sources could have anticipated this
change prior to proposal of these amendments. Accordingly, we have
selected a date that allows these facilities to design and install the
required control equipment.

 See the preamble discussion under compliance dates for more information
on EPA’s decision.

2.10	Startup, Shutdown, Malfunction (SSM)

Limit emissions during periods of SSM

Comment: Several commenters (2155, 2891, 2895, 2898, and 2911) supported
emissions standards for startup, shutdown, and malfunction equal to
normal operating conditions. Commenter 2155 stated that if less
stringent standards are considered, they should be for only a very short
time period around startup and shutdown. To keep malfunctions at a
minimum, we would not recommend less stringent standards for such
occurrences, lest they become the norm. Commenter 2898 requested EPA to
clarify if it intends to remove the exemption for all sources at cement
facilities and stated that EPA may deviate from the MACT floor standards
only in very limited circumstances. EPA may establish work practice or
operational standards, different than the established MACT floor
standards, only if it is not feasible to prescribe or enforce an
emission standard on the basis of either a design or source-specific
basis. 

Response:  EPA is adopting separate standards for startup and shutdown.
These standards are explained in the preamble to the final rule. As
explained in the preamble, the standards for these operating modes is
expressed on a concentration basis (not normalized to production) and
use different averaging periods than the main standards. Malfunctions
may in fact occur, and when malfunctions result in emission that cause
the standard to be exceeded, they can contribute to violations, although
sources have a n affirmative defense to civil penalties if they are able
to establish the affirmative defense set out in 63.1344 of the final
rule.. During malfunctions, sources are also obligated to limit
hazardous emissions to the extent possible, following good engineering
practices protective of safety and the environment.

MACT emission standards for periods of SSM unreasonable

Comment: Many commenters (2832, 2841, 2844, 2845, 2859, 2863, 2887,
2890, 2901, 2916, 2929, and 3218) opposed setting MACT emissions
standards for periods of SSM. The commenters opposed setting MACT
emissions standards or limits that fail to reflect differing processes
and process equipment conditions, including air pollution control and
monitoring equipment, during each of these SSM timeframes. It is
physically impracticable, if not impossible, for the process units in
question to shut down immediately in the event of a malfunction. These
emissions conditions during SSM are not the same as the conditions
during regular or routine operations. Commenter 2887 provided examples
of opacity data collected during SSM events. Commenter 2901 stated that
EPA’s proposed requirement that facilities meet steady-state standards
during SSM events is not lawful, citing decisions previous to the
Court’s decision in Sierra Club v. EPA, 551 F.3d 1019 (DC. Cir 2008)
that held that technology-based standards must contain exemptions or
less stringent standards during SSM periods. The commenter stated that
because all pollution control technologies will occasionally malfunction
and take time to get to their steady-state conditions, achievable
technology-based standards must contain provisions excusing
noncompliance or making compliance easier during such unavoidable
events. 

Commenter 2916 stated that EPA knows that emissions during periods of
SSM are likely to be different than emissions during normal operations
but the Agency ignores these facts and, instead, simply points to an
absence of emissions data from SSM periods as a reason to propose that
standards based on data from periods of normal operation are
appropriate. The absence of emissions data creates a duty for the Agency
to gather the data needed to set specific standards for SSM or, if
developing such information is not technically feasible, setting work
practice standards for SSM periods.

One commenter (2844) stated that if EPA applies the proposed emission
limits at the time of SSM events, then such events constitute reasonably
foreseeable operating conditions, and certainly the range of such
events, including their worst impacts on emissions achievements, needs
to be taken into account when estimating the floor units’ emissions
variability achieved in practice.

Response: As discussed in the preamble, EPA has taken into account
cement kilns’ operating properties during startup and shutdown periods
and has established different standards for those periods where
appropriate. Therefore, commenters concerns regarding startup and
shutdown periods have been considered and addressed in the final rule.
For the reasons explained in the preamble to the final rule, EPA is not
setting separate standards for malfunctions and during a malfunction,
the otherwise applicable standard applies. However, as explained in the
preamble to the final rule, in the event that a source fails to comply
with the applicable CAA section 112(d) standards as a result of a
malfunction event, EPA would, of course, determine an appropriate
response based on, among other things, the good faith efforts of the
source to minimize emissions during malfunction periods, including
preventative and corrective actions, as well as root cause analyses to
ascertain and rectify excess emissions. EPA would also consider whether
the source's failure to comply with the CAA section 112(d) standard was,
in fact, “sudden, infrequent, not reasonably preventable” and was
not instead “caused in part by poor maintenance or careless
operation.” 40 C.F.R. § 63.2 (definition of malfunction). See,
Weyerhaeuser v Costle, 590 F.2d 1011, 1058 (D.C. Cir. 1978)( “ In the
nature of things, no general limit, individual permit, or even any upset
provision can anticipate all upset situations. After a certain point,
the transgression of regulatory limits caused by "uncontrollable acts of
third parties," such as strikes, sabotage, operator intoxication or
insanity, and a variety of other eventualities, must be a matter for the
administrative exercise of case-by-case enforcement discretion, not for
specification in advance by regulation.”) 

As also discussed in the preamble, in light of comments raising concerns
with EPA’s proposed approach to malfunctions, the final rule includes
an affirmative defense to civil penalties for exceedances of emission
limits that occur during malfunctions. See 40 CFR 63.1341 (defining
“affirmative defense” to mean, in the context of an enforcement
proceeding, a response or defense put forward by a defendant, regarding
which the defendant has the burden of proof, and the merits of which are
independently and objectively evaluated in a judicial or administrative
proceeding.). This approach is not inconsistent with the Pre-Sierra Club
cases, Portland Cement Ass’n v. Ruckelshaus, 486 F.2d 375 (D.C. Cir.
1973) and Essex Chemical Corp. v. Ruckelshaus, 486 F.2d 427 (D.C. Cir.
1973), cited by one commenter. Although the relevance of those cases is
not clear in light of intervening caselaw, EPA’s overall approach to
malfunction events in this rule is consistent with the approach set
forth in EPA’s 1972 proposed rules that were cited favorably in those
decisions in that both “impart a construction of "reasonableness" to
the standards as a whole and adopts a more flexible system of regulation
than can be had by a system devoid of "give." Id. at 399. Those
decisions did not require exemptions or less stringent standards for
malfunction events as the commenter suggests. 

EPA’s overall approach to malfunctions is also not inconsistent with
the NRDC v EPA, 859 F. 2d 156, 207(D.C.Cir. 1988) and   HYPERLINK
"http://www.lexis.com/research/xlink?app=00075&view=full&searchtype=get&
search=564+F.2d+1253%2520at%25201272" \o "Clicking this link retrieves
the full text document in another window" \t "x"  Marathon Oil Co. v.
EPA, 564 F.2d 1253, 1272 (9th Cir. 1977)  cases cited by the same
commenter. Although neither case is directly on point and the NRDC
case’s discussion of technology-based standards is dicta, these cases
don’t require exemptions or less stringent standards for malfunctions
either. The regulatory affirmative defense included in today’s final
rule is consistent with Marathon Oil’s holding under the Clean Water
Act that EPA must formalize its approach to upsets. But see,
Weyerhaeuser v Costle, 590 F.2d 1011, 1057-58 (D.C. Cir. 1978)(holding
that an informal approach is adequate). 

One commenter maintains that the absence of emissions data creates a
duty for the Agency to gather the data needed to set specific standards
for SSM or, if developing such information is not technically feasible,
setting work practice standards for SSM periods.  However, the statute
permits EPA to use available information and imposes no affirmative duty
to gather information.  Cf. Cement Kiln Recycling Coalition v. EPA, 255
F. 3d at 867.  Consequently, EPA does not accept the commenter’s
suggestion.

Work practice options

Comment: Many commenters (2841, 2844, 2845, 2874, 2887, 2890, 2929, and
3218) stated that EPA should propose either work practice standards as a
viable alternative to the current ongoing SSM litigation (General
Provisions, 40 CFR 63) or suspend efforts to establish SSM limits
during this rulemaking and consider a separate regulation. According to
the commenters, work practices would be appropriate given the
technological and economic limitation involved with applying a valid
measurement method. Commenters 2841, 2844, and 2845 included a work
practice plan suggestions and a sample SSM outline and plan as part of
their comments. As an alternative, EPA should initial a separate
rulemaking effort to address SSM emissions once more information on SSM
emission is available. Reasons stated for a work practice standard
include the following:

EPA, in setting a MACT floor, must take into account the variability of
the best performing source by including emissions during periods of SSM.

The proposed compliance requirement means that mercury emission cannot
be practically measured during SSM.

Accurate quantification of the mass of mercury emitted per million tons
of clinker produced cannot be assured during SSM.

During SSM, low kiln feed which in turn causes clinker production to be
very low will reduce mercury emission levels to the point they are below
the point of accurate measurement.

During startup, mercury emissions are divided by a very small clinker
production rate leading to increased mercury emission rates per unit. 

Also during startup, it is anticipated that mercury emission rates will
spike as ductwork heats and any absorbed mercury compounds are released
or due to other transient affects.

During shutdown, kiln feed and clinker production is decreasing to zero
and mercury emissions are greatly exaggerated which also leads to
increased errors of mercury rates per unit.

Units equipped with wet scrubber or activated carbon control devices are
not operated during startup until specified kiln operating parameters
are met in order to protect the control device from damage and therefore
active controls are not functioning during the start-up and shut-down
period for some kilns.

For mercury CEMS used in the electric utility industry, during SSM, the
sample gas molecular weight changes due to the reduction in CO2
concentrations which affects the accuracy of mercury measurements.

For HCl and THC, changes in moisture and oxygen concentration, kiln
volumetric flow rate, kiln feed, and fuel firing will result in
unreliable emissions measurements. 

PM CEMS monitoring data will not accurately reflect actual emission
during SSM because of technical difficulties, and excessive testing
costs; stability of the PM CEMS; and affects on CEMS operations due to
differences in fuels burned, firing rate, and feed rate.

Response: The commenters indicated that certain features of the
standards could not be measured (those which are normalized to
production), but not that the HAP itself was incapable of measurement.
Thus, EPA does not see that the criteria of section 112 (h) have been
satisfied here. The standards in the final rule are not normalized to
production for the reasons provided in the public comments.  SSM
conditions should not be contained in correlation tests because they are
usually short lived, and erratic and probably not repeatable. that the
correlation is developed under normal operating conditions because that
is typical of the majority of time and SSM emissions are then quantified
based on that correlation

Comment: One commenter (2845) stated that EPA should address SSM through
the use of work practice standards that include:

Preventive maintenance activities intended to minimize malfunctions and
unplanned shutdowns of air emissions control systems

Procedures to be followed to minimize HAP emissions during kiln startup
and during planned shutdowns

Procedures for detecting and remedying malfunctions that result in
excess emissions that include a determination of the cause and
corrective actions.

The commenter has provided a sample SSM plan as an attachment to their
submitted comment.

Response: See previous response.

Consideration of SSM in mercury standards

Comment: Several commenters (2845 and 2855) stated that EPA’s 30-day
rolling average limit for mercury does not consider the periodic
maintenance shutdown of the raw mill. The raw mill is generally shut off
for scheduled maintenance about 15 percent of the time. However, there
are times where unscheduled shutdowns do occur. In addition, the raw
mill operation is interrupted when feed storage silos are full.
Depending on the specific plant, the mill downtime could have a major
impact on the average emission during that period. This will have an
impact on mercury emissions during that month, resulting in substantial
outliers in the 30-day data sets. To ensure these outliers do not result
in the facility being out of compliance for the proposed rule, any
mercury control strategy to be considered by the plant would need to
take this possibility into account. As a result, the cost of the mercury
control strategy is greater than would be the case if these types of
maintenance issues did not occur. An alternative that could be
considered by EPA would be to provide some form of exception for
shut-down, maintenance, and start-up.

Response: The comment does not give EPA a sufficient view of the
predicament faced by the source to indicate that compliance during
startup, normal operation, and shutdown periods is impracticable. In
fact, the description of the scenario presented by the commenter leaves
ample room for the Agency to interpret that compliance depends on
source-specific conditions, operator discretion, and a variability in
emissions with indeterminate causes among shutdown and restart events.
For instance, an unscheduled shutdown may or may not share the emissions
profile of a scheduled shutdown. A prior comment alluded to
“impracticable immediate shutdowns.” Given the averaging periods
built into the standard which accommodate spikes, and the generally less
challenging conditions obtaining during startup and shutdown conditions
(when less mercury-containing material is input to the kiln), EPA does
not accept, in the absence of any data, that unscheduled shutdowns,
impracticable immediate shutdowns, and scheduled shutdowns all, somehow,
make compliance impossible. 

Clarifications, corrections

Comment: One commenter (2845) stated that §63.1342 indicates that
63.6(f)(1) Compliance with Emission Standards, is not applicable to
Subpart LLL. If this is the case, §63.6(e)(3) Startup, Shutdown
Malfunction Plans, should not be applicable as the SSMP provisions are
no longer needed if the facility cannot utilize the SSM exemption
provided for in Subpart A. The commenter suggested that in Table 1 to
subpart LLL of Part 63 – Applicability of General Provisions,
63.6(e)(3) Startup, Shutdown Malfunction Plan will not be applicable to
subpart LLL.

Response: The EPA agrees that if a source may no longer avail themselves
of an exemption for startup, shutdown, or malfunction, then a startup,
shutdown, and malfunction plan is no longer necessary. The applicability
of the requirement to have an SSM plan has been deleted.  . 

Emission standards and SSM exemptions

Comment: One commenter (2846) disagreed with EPA’s proposal to not
apply the SSM exemptions to the emissions standards in this proposed
rule. The commenter provided two options to address emissions during SSM
events: 

SSM emissions could be treated as de minimis emissions; and

EPA needs to establish work practice or operational standards that would
apply during SSM periods.

Response: See preamble to final rule and earlier responses as to why no
showing has yet been made as to why work practice standards are
permissible here.

Area sources

Comment: One commenter (2887) stated that with respect to area sources,
the EPA could utilize its discretion under §112(d)(5) to impose
management practices to limit emissions during SSM events.

Response: EPA interprets management practices to mean actions homologous
to work practice standards. See the preamble to the final rule and
earlier responses as to why no showing has yet been made as to why work
practice standards are permissible here.

Continuous monitoring of PM and vapor phases of all HAP metals

Comment: One commenter (2638) stated that with the inclusion of SSM
emissions subject to the standards, PM and vapor phases of all HAP
metals including mercury should be continuously monitored. CEMS would be
required to measure emissions during randomly occurring malfunction
events. PM is not expected to be a good surrogate for HAP metals during
SSM events in which emissions, control of these emissions, the chemical
state of the emissions and the partitioning between phases could and in
realistic circumstances would be expected to be different from those
during normal operation.

The use of multi-metal CEMS methods would allow all of the urban HAP
metals including mercury and all phases to be continuously monitored
with a single CEMS. This option would be simpler, more reliable and have
lower initial and on-going costs than the only other option consisting
of two CEMS, one for PM and one for mercury vapor.

Commenter 2638 recommended that SSM events be included in the standard,
and that HAP metals emissions be monitored using CEMS.

Response: EPA sees no reason to believe that the surrogacy relationship
would be altered by startup, shutdown or malfunction events.

Use of PM as a surrogate for non-volatile HAP metals – in this very
industry – has already been upheld. See National Lime II, 233 F. 3d at
639. See the preamble discussion under final testing and monitoring
requirements for more information.  

2.11	Economic Impacts

Control technology performance uncertainties

Comment: In order to estimate the economic impact of the proposed rule,
commenter 2845 examined the need for controls and the likely controls
and identified the following as issues with the level of performance
that each emissions control technology will achieve when applied to meet
the proposed standards (additional details are presented in Appendix 12
of the commenter’s submittal):

There is only one kiln with an RTO and we do not know how well an RTO
will work on kilns with differing profiles of organic emissions.

It was assumed that wet scrubbers will reduce HCl emissions in the
presence of high SO2 levels but there are technical concerns with the
successful control of HCl in that circumstance.

No one has applied activated carbon to a kiln exhaust to achieve either
the THC or the mercury control levels needed to comply with this
standard. In addition, there is an expectation that a single carbon
system can control both mercury and THC but that has not been
demonstrated. At present, carbon systems for THC are designed
differently from systems for mercury.

Wet scrubber performance in controlling mercury is variable and some of
the mercury captured is reemitted as operating modes change. There is
evidence that some units will gain very little mercury control with a
scrubber due to the form of the mercury in their exhaust.

There is strong evidence that RTO are not capable of reducing emissions
to below the proposed THC standards.

There is significant uncertainty as to whether a wet scrubber can reduce
HCl levels to below the proposed HCl standards.

Not all kiln locations have the natural gas service needed to fire an
RTO.

Not all kiln locations will have access to the water needed to operate a
wet scrubber.

Commenter 2845 stated that there are key uncertainties associated with
the control performance assumptions used by EPA but in order to develop
their own analysis of the cost of controls, has developed its own
control performance assumptions: 

An RTO will cut THC emissions by 95 percent and a scrubber will precede
the RTO. Doubts that an RTO can be operated consistently below 7 ppm of
THC without regard for the starting concentration due to the THC
associated with the combustion of the natural gas and the emissions
during switching of the thermal beds. The commenter assumes for their
analysis that RTO will be able to consistently reduce emissions below
the 7 ppm limit even though the one commercial application of an RTO on
a kiln has not done so.

A scrubber will control mercury, on average, by 50 percent. It is clear
that actual performance is a function of the mix of oxidized and
elemental mercury which can be highly variable.

Any ACI system will used to control both mercury and THC. ACI will
reduce organics by 50 percent due to their poor performance in
adsorbing light hydrocarbons. Any ACI system will include a dedicated
bag house to capture the carbon ladened dust. That bag house will be
50-66 percent of the size of a new bag. ACI with a bag filter will
reduce mercury emissions by 75 percent.

A unit that adds both ACI, with a bag house, and a scrubber is assumed
to have overall mercury control of about 85 percent.

The new PM standard will require that all units have a bag house. For
this analysis, the exact performance of the bag house is not addressed
but simply it is assumed that all kilns must have one. However, if a
unit has an ESP and adds ACI with a bag house, it is assumed that the
existing ESP will not need to be replaced.

To the extent a unit has baseline stack test HCl emissions over the
standard, it is assumed that a wet scrubber will be needed. The exact
performance of the scrubber is not known.

Another compliance requirement that will need to be consider for its
cost will the installation and operation of CEMS for THC, HCl and
mercury for each regulated stack.

Scrubbers installed to meet the limits under this rule will reduce SO2
to some degree. To assess that impact, we have surveyed scrubber
suppliers and test data and developed the following assumptions:
scrubbers will provide between 90 percent and 95 percent control for SO2
but will not control below 15 ppm of SO2.

Response: EPA is aware that the actual performance of any control device
will depend on a number of factors and will likely vary to some unknown
degree among kilns. EPA made based its estimates of performance on
available performance data including data for the cement industry as
well as other relevant industries. In some cases that data were limited.
For example, EPA’s estimate of scrubber performance for wet scrubbers
on mercury emissions was based on the test conducted at 5 kilns. ACI
performance was based on the experience of the utility industry, as well
as a pilot test conducted at one portland cement kiln. In estimating
cost impacts and emission reductions, we calculated the removal
efficiency that would be required to comply with the final limits and
assigned a control device based on the reduction that would be
necessary. The cost impacts are conservative in that for every kiln, we
assumed that one or more control devices would have to be added. We
believe that we have overstated what actual costs will be for the
industry. While we may not accurately predict the controls that each
kiln will eventually install to comply, we have provided an estimated
cost of compliance and accounted for the fact that the emission limits
will require expenditures by the industry to comply. The actual cost
associated with compliance is not factored in the decision of what the
level of the standard should be. EPA must set the standards based on the
best performers, regardless of the means by which they are the best
performers.

The commenter took issue with the controls and level of performance used
by EPA in its analysis of impacts. Regarding the ability of RTO to
adequately control organic emissions, a currently operating RTO with
several years of operating experience is controlling THC on a long-term
basis to an average of about 5 ppmvd at 7 percent O2 measured as
propane. Furthermore, as a result of additional long-term THC data
received after proposal, EPA has increased the THC limit to 24 ppm for
promulgation. In addition, sources have the option of complying with the
alternative limit for organic HAP. We acknowledge that a few sources may
find it necessary to add RTO because of the level of control necessary,
although we believe that most kilns can be adequately controlled using
ACI systems that get an estimated 50 percent control of THC (due to the
presence of lighter hydrocarbons such as methane and ethane that are not
as effectively removed by RTO as heavier compounds) and an estimated 80
percent control of other organic HAP. Where RTO are installed, they will
be expensive, due in part to the fact that the RTO must be preceded with
a wet scrubber in order to remove SO2 that can foul the RTO. However, we
have found that only 10 kilns nationwide may need to install RTO.

We also agree with the commenter that RTO use may not be feasible in
areas that lack access to natural gas supplies. However, as stated
above, the number of sites where we expect RTO to be necessary is
limited. Based on information available at the U.S. Energy Information
Administration, all six states where kilns potentially needing RTO are
located appear to have access to natural gas. We do not believe that
water availability will be a limiting condition in these locations.

The commenter had concerns about the ability of wet scrubbers to remove
HCl in the presence of SO2. We agree that the wet scrubbers currently on
U.S. cement plants are designed to remove SO2, rather than HCl. However,
an EPA review of HCl emissions from 21 plants indicated an average HCl
concentration of 12 ppm, and a maximum HCl concentration of 55 ppm. 
Therefore, required HCl removal efficiencies to meet the 3 ppm emissions
standard are expected to be much lower than 99 percent for virtually all
kilns. A preliminary review of data gathered from a 2010 Information
Collection Request (ICR) for Electrical Generating Units indicates that
wet scrubbers utilizing lime will be able to provide the necessary
collection efficiency to meet the 3 ppm emissions standard.

EPA believes the commenter’s concern over the performance of ACI
systems are not well–founded, based on the experience with the use of
ACI systems for both mercury and organic emission controls in other
industries, including medical waste incineration and coal fired boilers.
The ACI system installed at Ash Grove in Durkee, OR is expected to
achieve over 80 percent reduction in mercury emissions. Based on pilot
tests, removal efficiencies of over 90 percent were achieved, depending
on the type of carbon and the injection rate.

As noted by the commenter, scrubber performance on mercury can be
variable and depends on several factors, such as the mercury species
present, the presence of chlorine in exhaust gases, and the liquid
injection rates. Tests conducted on the scrubbers currently in use in
the cement industry showed removal efficiencies from 50 percent for one
scrubber to over 80 percent at two scrubbers. These removal efficiencies
were the average of tests conducted during raw mill on and raw mill off
conditions. Usually during the raw mill off period, removal efficiencies
are higher, due to the presence of greater concentrations of oxidized
mercury than during the raw mill on phase, when most of the mercury is
in the elemental phase. We agree that wet scrubbers may not perform as
well as ACI systems on mercury, and we believe that ACI systems, which
have higher removal efficiencies and are less expensive to install and
operate than scrubbers, will be the predominant control method used. Our
impact and economic analyses are based on this assumption.

Where both ACI and wet scrubbers are installed, the combined efficiency
of the two systems on mercury control will vary. We expect that ACI
systems with removal efficiencies of 80 percent or better will remove
most of the mercury and that, where mercury concentrations in the
exhaust gas stream are low to moderate, wet scrubbers will not impact
final mercury emissions significantly. Our impacts analysis uses this
assumption. However, in situations involving high mercury
concentrations, we believe that scrubbers after ACI systems will have
significant impacts on the final emissions. In our analysis, this was a
factor for only two kilns that have extremely high mercury emissions.

We agree with the commenter’s statements that, where a facility
currently operates an ESP and is expected to install an ACI system
(including polishing baghouse), in most instances they would not need to
replace their ESP. Our impacts analysis uses this approach. We also
considered the costs of CEMS in our analysis as suggested by the
commenter.

Regarding the commenter’s statement that scrubbers will not control
SO2 below 15 ppm, outlet test data from one cement plant report SO2
levels well below the 15 ppm level. (Memo, Bahner, Laney, Barnett,
Summary of Cement Kiln Wet Scrubber and Lime Injection Design and
Performance Data, May 29, 2008, Docket EPA-HQ-OAR-2007-0877)

Comment: One commenter (2891) stated that the goal of MACT standards in
general is to achieve 90 percent or better reductions in the targeted
hazardous air pollutant emissions. Some estimates place the likely
reductions from this proposed regulation as low as the low to mid-80
percentile for reduction. If that is correct, the commenter requested
that EPA describe in plain terms the Agency’s justification for
obtaining reductions below 90 percent.

Response:  MACT standards must be of minimum stringency reflecting
performance which is the average of a select group of best performers,
or the performance of the best-controlled similar source, and may be
more stringent. There is no specific emission reduction percentage,
however, which constitutes MACT.

Alternate impact analysis

Comment: One commenter (2845), in estimating the impacts of the proposed
rule, stated that there was insufficient data to perform a kiln-by-kiln
analysis of controls. Instead they used available emissions data for
subsets of kilns and extrapolated to the total population assuming 170
kilns needing to comply. In order to estimate the total number of RTO,
scrubbers, ACI, and bag filters needed to comply with the proposed
limits, the commenter listed the steps taken to profile the entire
population of kilns (see Appendix 12 to the commenter’s submittal for
details).

One of the key differences between EPA’s assessment and the
commenter’s assessment is that the commenter expects a larger number
of RTO because of the limited performance of activated carbon on
controlling light hydrocarbons and the need to design the carbon system
to control both mercury and THC. According to the commenter, a number of
kilns will be at risk of non-compliance even with the installation of
the best available controls. 

The commenter (2845) took issue with EPA’s assumption of a substantial
secondary benefit in the form of SO2 reductions from the installation of
wet scrubbers under this rule. The commenter stated that there is
significant uncertainty about the nature of a scrubber designed to
reduce HCl or mercury or both and whether it will operate in a manner
that provides the same level of SO2 control as a dedicated SO2 scrubber.
To assess the impact of wet scrubbers on SO2 emissions, the commenter
used the SO2 emissions from the ISIS model and average flow volumes to
calculate the stack outlet SO2 concentration in ppm for each kiln. The
commenter then calculated the control efficiencies that would be
required and control efficiencies that could be achieved for each kiln
assuming that the lowest achievable scrubber outlet concentration of SO2
is 15 ppm based on communications with a scrubber vendor. The commenter
estimated SO2 control ranging from 72-78 percent over the entire SO2
inventory. 

One commenter (2845) did not agree with EPA’s assumptions for control
technology effectiveness impacts, controlled emissions estimates, and
costs for installing/operating the various control technologies and are
listed as follows: 

The costs for limestone wet scrubbers as estimated using EPA’s (ISIS)
model and production assumptions were compared with best available
control technology (BACT)/best available retrofit technology (BART) cost
calculations using site-specific gas flow and SO2 concentrations for
these same facilities. The results of the comparison of the two methods
can be summarized as follows:

The capital and annual cost projections using the BACT/BART cost
estimates were several times greater than those calculated by EPA.

The SO2 concentration in the kiln gas and the presence of contaminants
in the synthetic gypsum (syngyp) produced will determine whether a
gypsum credit or a sludge disposal cost will be incurred for the
scrubber sludge. The ISIS model always assigns a credit, and the
BACT/BART analysis assigns a cost for disposal. These costs/credits will
be site-specific.

Two key capital and operating costs that are not included in EPA’s
cost analysis are a limestone preparation plant to grind limestone to
the fineness required for the LWS and the handling equipment used to
handle the syngyp created by the LWS. The syngyp has high moisture
content and cannot be handled by the equipment utilized to handle
natural gypsum. The additional costs for these two systems were provided
by the commenter.

Other observed discrepancies impacting annual variable costs include
costs for natural gypsum and costs for onsite or offsite wastewater
treatment. If onsite treatment is utilized, additional capital costs
must also be included.

The ISIS model does not include the potential for additional capital
costs at those plants that will require the kiln ID fan to be replaced.
This is particularly likely where other control devices such as an RTO
or ACI are also required by the new NESHAP requirements. The need for a
new ID fan will be site-specific.

Response: As a result of comments, we examined our costs (including
scrubber costs) and made several adjustments, including revisions to the
exhaust gas flow rates, power demand, and water usage. As an
illustration of the effects of these revisions, at proposal, the capital
cost for a 1.2 million tpy kiln was $9.3 million; the revised estimate
is $25 million. Annualized costs increased from $1.6 to $3.6 million.
Earlier responses in this section address other issues raised by the
commenter.

Comment: One commenter (2844) stated that the exhaust flow rates
reported and utilized in the cost of control section underestimates the
exhaust flow rates to be controlled by using inconsistent factors,
missing certain sources of flow, and using dry instead of actual (wet)
flow rates in the analysis.

Response: In response to this comment, EPA investigated its gas flow
rate estimating approach. Upon looking at additional data that had been
supplied by industry, it did appear that stack gas flow rates generally
exceed the original flow rates used. From the data, it appears that
other sources of gas flow dilute the kiln exhaust and increase the total
flow rate beyond what is expected at the kiln exit. Based on a review of
the additional data and review of operating cost estimates specifically
for scrubbers, changes were made in the operating cost equations.
Changes in the exhaust gas flow rates also affected costs of all control
equipment costs as equipment costs are generally a function of exhaust
gas flow rates. The details of this analysis of exhaust gas flow rates
and changes in operating cost equations is documented in memorandum from
Staudt, Andover Technology Partners, to Srivastava, Torres, Barnett, US
EPA, “RE: Wet Scrubber Cost Algorithms,” February 26, 2010.

Comment: One commenter (2844) stated that EPA’s analysis confuses
capacity of a plant and the actual production of a plant. This approach
overstates the emissions, and understates the control needs and costs.

Response: In our analysis, we chose capacity because actual production
can vary substantially over time. We disagree with the commenter’s
assertion that using capacity maximizes emissions and costs.

Comment: One commenter (2844) stated that for concentration-based
limits, EPA unreasonably assumes that exhaust is at 7 percent oxygen
when sizing and costing out controls. Higher oxygen levels occur in kiln
exhaust due to others sources of flow to the kiln stack. As a result,
the overall flow to the stack will be greater than has been assumed in
the EPA cost analysis. With higher flow will come larger control systems
and added cost. 

Response: As noted in a previous response, as a result of comments, we
investigated our gas flow rate estimating approach and made changes that
affected the control equipment costs. The details of this analysis of
exhaust gas flow rates and changes in operating cost equations are
documented in a memorandum from Staudt, Andover Technology Partners, to
Srivastava, Torres, Barnett, US EPA, “RE: Wet Scrubber Cost
Algorithms,” February 26, 2010.

Comment: One commenter (2844) stated that in the assessment of HCl and
THC impacts, EPA assumes each kiln is running at full capacity 365 days
per year, which is impossible. A more reasonable number for days of
operation is 330 days.

Response: EPA assumes 330 days of operation in the cost and impacts
analysis.

Comment: One commenter (2844) stated that scrubber power demand is
presented at 3kWh/ton clinker, which is too low. There appears to be a
minor error in the calculation that is done in converting units but the
main problem is that duct work pressure drop has not been accounted for
and the exhaust flow is too low. Thus, in Section 2.3.5 secondary
impacts are based on unreasonable assumptions. The commenter provided
examples of power costs that are higher than those assumed in EPA’s
assessment. 

Response: As noted in a previous comment, as a result of comments, we
examined our scrubber costs and made several adjustments, including
revisions to the power demand, resulting in a change in our control cost
estimates.

Comment: One commenter (2844) stated that other environmental impacts
(Sections 2.4 and 2.5) include errors. The commenter provided a list of
specific errors in the evaluation of other environmental impacts:

Scrubber make-up water is given as 30 gal/ton clinker. In NSPS
rulemaking, however, it was given as 105 gpm for a 1,200,000 tpy plant,
i.e. 42 gal/ton clinker (see EPA-HQOAR-2007-0877-0008, table 9, p.17),
which is 12 gallons more than the new number.

ACI solid waste is given as 3 lb/ton clinker in Exhibit 7 but later a
figure of 6 lb/ton clinker is given.

Scrubber solid waste is omitted. It is assumed by EPA that all the
scrubber waste can be incorporated into cement, but EPA has not
demonstrated that this can be done while meeting product standards.

EPA assumes that CKD can also be incorporated into the cement but EPA
has not demonstrated that this can be done while meeting product
standards.

Response: As noted in a previous response, we examined our scrubber
costs in response to comments and made several adjustments, including
scrubber water usage, resulting in a change in our estimates. In
response to this comment, we have corrected our ACI solid waste estimate
in section 2.5 to indicate 3 lb/ton clinker, consistent with the
estimate provided in Exhibit 7. Our assumption regarding scrubber solid
waste was based on a discussion with plant personnel, who indicated that
scrubber slurry is dried and added to the clinker as gypsum.

Comment: The following are comments from commenter 2845 regarding
EPA’s determination of costs associated with the installation and
operation of ACI for the control of mercury and THC:

A more refined economic analysis performed for an ACI system with a
polishing baghouse resulted in a capital cost factor of $13.33 per ton
clinker and an annual cost factor of $3.91 per ton clinker, both of
which are more than four times the values used by EPA.

The estimated levels of emissions control for ACI used by EPA are overly
optimistic; achievable industry average levels of control for ACI would
be 50 percent for THC and 75 percent for mercury. Depending upon the
form of mercury in the exhaust gases, an ACI system used in conjunction
with a wet scrubber is expected to provide a combined control efficiency
for mercury of 85 percent. 

EPA overestimated the amount of solid waste generated from ACI systems.
Based on the vendor-suggested injection rate of 12 lbs activated carbon
per million (actual cubic feet per minute [ACFM]), an ACI system will
generate approximately 1.68 lbs solid waste/ton clinker versus EPA’s
estimate of 3 lbs/ton. 

Response: Without any supporting data from the commenter to substantiate
their estimates for ACI costs, mercury emission reductions, and ACI
solid waste, we are unable to evaluate the commenter’s assertions. We
estimated the costs of using an ACI system and polishing baghouse to
control mercury emissions from cement kilns using control costs recently
developed for actual electric utility boilers with exhaust gas flow
rates comparable to those for cement kilns. We estimated the mercury
emissions control based on testing conducted on an ACI system at a
cement plant, which showed mercury removal efficiencies over 80 percent
at carbon injection rates of 10 lbs per million acf. The ACI solid waste
estimates were based on data provided in Docket item II-B-67, Docket
A-92-53.

Comment: One commenter (2845) provided the following comments regarding
EPA’s determination of costs associated with RTO systems for control
of THC.

The pollutant control matrix shown in ISIS incorrectly lists PM,
mercury, and HCl as pollutants addressed by the RTO control option

EPA derived a capital cost factor of $9,598 per dscfm0.6 to estimate the
capital cost of installing an RTO system on a cement plant from
information available on the only currently operating existing
installation within the cement industry. However, the correlation
required flow in dscfm, and EPA used flow in units of standard cubic
feet per minute (scfm), which includes moisture in the gas stream
resulting in a substantial level of error for EPA’s estimation. 

To estimate annual operating costs, EPA utilized an empirical formula
with several variables, some of which are dependent upon kiln type.
Based on the information used as EPA’s basis, the empirical formula
vastly underestimates the annual operating cost of the RTO system. The
typical annual maintenance materials costs for an RTO system would be
around $503,000 per year for the 1.2 million ton per year example
facility.

Using a more recent quote from the vendor for an existing RTO a
projected capital cost factor of $11,034 per dscfm0.6, which is about 15
percent greater than that utilized by EPA was estimated for a 1.2
million tpy kiln. The analysis projected an annual operating cost for an
RTO on a precalciner kiln of $4,574,531 as opposed to EPA’s estimate
of $1,199,996, over four times that estimated by EPA. 

The estimated levels of emissions control for the RTO used by EPA are
overly optimistic. Recommended, and achievable, industry average levels
of control for an RTO would be 95% for THC although an RTO system is not
capable of maintaining a THC emission level below 10 ppm. Therefore,
many kiln systems with elevated, but not extreme, levels of THC in the
exhaust would not be able to comply with the proposed standard of 7 ppm
through installation of an RTO system.

EPA underestimated the electrical power and natural gas fuel
requirements for RTO systems. Based on a manufacturer’s quote, the
power demand is 46.7 kWh/dscfm and the natural gas requirement is 1.26
million British thermal units (MMbtu)/dscfm. EPA’s estimates were 30.9
and 0.73, respectively. 

EPA underestimated the amount of secondary impact emissions from
operating the RTO systems due to the underestimated power demand and
fuel requirements for the installations.

Response: EPA acknowledges commenters’ input regarding RTO control
costs. However, EPA expects 10 or fewer RTO’s to be installed as a
result of this rulemaking. Further refinements to the RTO costs would
not change the underlying MACT floor determinations or emissions
standards. Revised RTO costs may change the estimated cost burden to the
industry, but RTO costs are expected to be relatively small in relation
to the total control costs to the industry.

Comment: One commenter (2916) stated that EPA should consider indirect
effects of the Cement Kiln NESHAP, including increased energy
consumption, increased water consumption, and increased dependence on
foreign supplies of a strategic commodity. The commenter suggests that
the EPA inadequately addresses these indirect effects when explaining
the justification for the MACT standards set in the proposed Cement Kiln
NESHAP.

Response: In considering whether to adopt beyond the floor standards,
EPA did take notice that additional controls would necessitate more
energy use, and could lead to increased water usage and waste
generation. These are additional reasons for not adopting beyond the
floor standards (along with high costs for limited reductions with
consequent poor cost effectiveness). EPA does not consider non-air
quality impacts at the floor to be so extreme as to invalidate what
otherwise is best performance, however.

Comment: One commenter (2916) provided the following cost issues
comments regarding EPA’s determination of costs associated with
installation and operation of baghouses and membrane bags:

Based upon information received from National Cement on a recent
baghouse installation, the commenter believes that capital costs for
baghouses have been understated by EPA by nearly a factor of two.

The pollutant control matrix shown in ISIS incorrectly indicates HCl and
THC being associated pollutants for the baghouse membrane bag (BHMB) and
membrane bag (MBR) control options. 

In the monitoring cost section for PM, it is stated that the CEMS Cost
Model assumes each baghouse has 5 compartments which seems off by a
factor of probably 2 or more and raises a question as to the size of the
baghouse EPA has assumed would be necessary for the kilns. In fact, EPA
assumes that the clinker cooler baghouse has the same number of
compartments as the kiln baghouse. Kiln baghouses are usually at least
50 percent larger than cooler baghouses, depending upon the type of
cement process.

EPA assumes that an existing baghouse would need a simple bag
replacement to comply with the PM limit and assumes that the costs to
replace only membrane bags is the difference in cost between the new
reverse air baghouse with fabric bags and a new baghouse with membrane
bags. Many older baghouses will need major repairs/retrofitting or even
outright replacement to use newer membrane bags and still meet the PM
standard. The number of facilities that would fit this category and the
cost to repairs/retrofit on a per facility basis is unknown. EPA’s
cost estimate assumes reverse air baghouses would be the standard
replacement, but most new baghouses on kiln systems are pulse-jet, which
may have higher costs than what EPA assumes.

For PM, EPA used their Air Pollution Control Cost Manual Sixth. The
background documents contain no information on what assumptions EPA used
to generate the costs.. Without the detailed background information on
the specifics of the size and other details regarding the particulate
control device used for the cost estimate, it is impossible to know if
the cost estimates are representative of real world installations or
whether the size of the baghouse used by EPA was based upon ACFM or
scfm, or, if it was based upon scfm, what assumed operating temperature
was used. There are additional costs associated with membrane bags as
follows:

Higher cost of membrane bags (same area as conventional bags).

In most cases, lower emissions will also require lower air to cloth
ratios to limit cleaning and emissions.

Reduced bag life to avoid operating with failed bags.

Response: First, it should be noted that the final rule is not requiring
bag leak detectors, but PM CEMS. Second, in our cost analysis, we used
the reverse-air baghouse design as the basis for our costs in order to
be conservative, based on our understanding that reverse-air designs
would be more expensive than pulse-jet designs. Third, the OAQPS Control
Cost Manual has been the standard reference used by EPA in costing
emissions control equipment, where actual applicable cost data are
unavailable. We assumed membrane bags were about four times more
expensive than standard bags. We used vendor information regarding the
superior performance and life of bags.

Comment: One commenter (2844) stated that Section 2.7 includes error in
the evaluation of cost impacts in the environmental analysis:

Andover 2008a at p27 provides figures for gypsum produced. In a wet
scrubber it is not plaster (or anhydrite CaSO4 M=136) that is made, but
bihydrated gypsum (CaSO4.2H2O M= 172). In addition, that gypsum is wet
(about 12%). So, 64 tons of SO2 removed should make 172*1.12 = 193 tons
of gypsum.

Capital cost for ACI seems very low ($2.41/ton clinker) based on the
commenter’s discussions with vendors.

Response: As noted previously, wet scrubber costs have been revised
based on comments received.

Comment: One commenter (2845) provided the following comments concerning
the CEMS cost modeling approach:

The commenter will need access to EPA’s completed cost model to fully
assess EPA’s assumptions.

The useful life of CEMS for purposes of estimating annualized costs is
unclear and confusing in the EPA background documents. 

All costs should be expressed in 2005 dollars for ISIS and for EPA’s
other cost and benefits estimates for the NESHAP. However, it appears
that the CO CEMS costs (that EPA evidently uses to represent costs for
THC CEMS in the NESHAP costing) was completed using 2003 dollars. Based
on the documentation, the commenter cannot determine for what year the
CEMS Cost Model provides costs.

It is anticipated that every facility will need to install a new CEMS
building; even if a building already exists, location issues associated
with the new control technology will dictate that a new building be
constructed. Also, all new CEMS for mercury and other pollutants are
built at the manufacturer or integrator place of business and shipped in
the shelter as a unit. The cost for on-site assembly is prohibitive and
greatly diminishes quality.

Commenter 2845 has more specific CEMS System Comments for mercury, HCl,
THC, and BLD as follows:

For mercury CEMS, 

It is expected that the capital cost for purchase and installation of a
single mercury CEMS is in the range of $300,000 – $500,000. The
initial testing for certification adds $35,000.

The extra elemental mercury calibration gas generator that is needed to
meet the QA criteria in the recently published EPA Traceability
Protocols for Elemental Gas Generators is estimated to cost $45,000.
This field reference calibrator could be used to QA generators at
multiple plants.

Upgrades to plant compressed air systems are required for the nitrogen
dilution systems (most widely used Thermo systems).

Using an average capital cost of $500,000 for contemporary mercury CEMS
is more realistic.

Mercury CEMS on-going operation and QA will require one dedicated plant
person per mercury CEMS. 

The annual cost to perform the required RATA is $25,000.

The vendor supported maintenance agreement is estimated to cost
approximately $30,000 per system.

Although, the useful life of the system is yet unknown; it is assumed to
be a maximum of 10 years; but, certain system components may have a
shorter life.

For HCl CEMS,

The industry believes installation and certification of HCl CEMS as
currently proposed is expected to cost between $225,000-$250,000
(including FTIR, oxygen monitor, data handling system, heated sampling,
calibration gases, and other ancillary equipment) in system capital
expenses and installation.

There are no current monitors that can measure at levels required by
EPA, and PS-15 is impossible to meet (requires spectroscopist to review
the FTIR data frequently and includes numerous requirements no CEMS in
the world could meet).

Due to special sampling requirements, an FTIR system as prescribed by
EPA could not be part of another CEMS; it would need to have separate
dedicated sampling equipment (special super hot sampling lines) and
separate data acquisition system (DAS).

Estimated capital costs of about $280,000 just to get started on a
system are expected.

Another $30,000 for initial certification of the HCl CEMS is
anticipated.

The total costs for on-going QA are unknown as facilities cannot install
and certify HCl CEMS using PS-5.

Estimated cost includes one full-time person, plus $30,000 for annual
testing, plus $20,000 for an annual service agreement.

For THC CEMS, the Summary of Environmental and Cost Impact document
lists $137,800-$138,900 for installed CEMS costs, and $41,672 for
ongoing costs. However, these are for CO monitors as this document was
prepared some time ago and EPA figured that industry would go with CO
monitors versus THC monitors because they were less costly. An FID
analyzer with dual range with auto switching to get to the low levels
and still measure spikes would be estimated at $25,000 as compared to CO
analyzer $10,000. 

For BLDS:

The commenter (2845) does not understand the basis for EPA’s number
for CEMS and BLDS monitoring costs. An estimate for the cost of a
transmissometer (also used as COMS) that meets ASTM D7392-07 for PMDS or
a good quality light scattering instrument is $35,000. According to our
expert on CEMS, a cost of $93,700 is sufficient for any of the
recommended options on the table: BLDS, PMDS, or COMS, assuming
monitoring only at a single location.

A BLDS is needed downstream of each baghouse (kiln, coal mill, by-pass,
etc) whereas a PMD may be able to monitor co-mingled emissions (e.g.,
kiln and coal mill). We are not clear on the assumptions EPA made.
However, on page 17 of the Summary of Environmental Costs Impacts of
Proposed Revisions to Portland Cement NESHAP, EPA assumes 1 sensor per
compartment and 5 compartments for kiln baghouse, and five more of the
same sensors for the clinker cooler baghouse, for a total of 10 sensors.
Thus, the EPA assumed capital cost is approximately $94,000 for 10
monitors. However, the assumption of only 5 compartments for each kiln
baghouse is understated by a factor of at least two times. Thus, EPA’s
estimated capital cost for BLDS on the basis of 1 sensor per compartment
is at least 50 percent low.

Response: The cost model is based on vendor supplied data and has been
placed in the docket for this rulemaking.

RIA and ISIS projections

Comment: One commenter (2845) stated that EPA’s RIA wrongly credited
all projected SO2 emissions reductions in 2013 to the NESHAP by assuming
all kilns operating in 2005 will be in operation in 2013 and will be in
compliance. The RIA also assumes 20 new kilns in operation by 2013. The
ISIS model projects a lower reduction by taking into account natural
plant closures and modernizations. The commenter stated that ISIS offers
significant improvements over EPA’s traditional cost approach. The
commenter stated that the ISIS results are based on outdated projections
regarding future U.S. cement consumption and capacity projections that
were developed before the severity of the current recessions became
apparent. According to the commenter, when EPA re-runs ISIS using
consumption and capacity projections that accurately reflect the impact
of the recession, EPA will predict a reduction in kiln emissions of SO2
in 2013 and the monetized value of the SO2 reduction attributable to the
proposed NESHAP will decline as will the cost estimated for compliance
with the proposed NESHAP.

One commenter (2845) stated that although ISIS offers the potential to
eventually provide a significant improvement over EPA’s previous
economic modeling for the cement industry, presently, ISIS is subject to
numerous shortcomings in data, assumptions, structure and validation.
The commenter does not believe that the current version of ISIS provides
believable estimates of likely economic impacts from the proposed NESHAP
or from other potential environmental requirements. ISIS is so complex
and its deficiencies so pervasive that, until it is substantially
improved, other methods should be used to assess the likely impacts of
the proposed NESHAP. The commenter provided an initial technical review
of the current version of ISIS that EPA has used to estimate the costs
and economic impacts of the proposed rule and also provided comments and
criticisms regarding the model and suggestions for improving and
eventually develops an acceptable version of ISIS. The commenter
submitted additional details and data as part of their comment. In
contrast to EPA’s previous approach, ISIS:

No longer models the U.S. cement industry unrealistically as a
collection of 20 completely independent regional cement markets, with
each cement producer selling product in only their own home market;

Models cement producer behavior as competitive rather than
oligopolistic, dropping the assumption that producers restrain output in
order to raise prices;

Allows a greater potential role for imports, consistent with historical
experience;

Simulates endogenously the full range of potential responses to
regulation instead of assuming many of them as exogenous inputs to the
model; and

In many respects uses updated and better data.

However, notwithstanding its commendable modeling objectives and the
improvements over previous modeling efforts, the current version of ISIS
that is used for the NESHAP analysis needs substantial improvement
before its outputs can be considered to realistically forecast the
impact of EPA regulatory options. The commenter submitted detailed
discussions of its recommendations for improvement to ISIS (see Appendix
13 of the commenter’s submittal).

Response: The ISIS Cement Model has been enhanced by updating cement
consumption and capacity projections, import modeling and transportation
matrix. EPA employed both a partial-equilibrium economic model and the
enhanced version of the ISIS Cement model to analyze the impact on the
industry and the economy. The final documentation of the enhanced
peer-reviewed ISIS model has been placed in the docket for this
rulemaking.

Comment: One commenter (2845) evaluated the potential economic impacts
of the proposed rule and estimated that as many as 30 existing plants
could be forced to close, with an additional 12 at high risk of closure
which will have significant employment and economic impacts as well as
collateral environmental impacts as well including a substantial
increase in imports from non-U.S. plants that are not subject to
stringent environmental standards. This would undermine the mercury
reduction benefits and increase greenhouse gas emissions from the
transportation of cement (the commenter provided additional details in
Appendix 10 to their comments).

Response: Neither our RIA economic analysis or our ISIS analysis
estimated impacts this large.   Based on our analyses, we believe that
up to 10 plants may have reduced capacity utilization during 2013 as a
result of the impacts of this rule.  (Regulatory Impact Analysis:
Amendments to the NESHAP and NSPS for the Portland Cement Manufacturing
Industry, Final Report, August 2010).

Comment: Several commenters (2844, 2846, 2859, and 2919) stated that the
proposed standards will have adverse economic impacts. The proposed
standards will make it cost and technology prohibitive for some plants
to achieve compliance, leading to forced closure of domestic plants that
will create job losses. The reduction of domestic cement production will
require imports, elevate costs, eliminate jobs, and constrain
infrastructure development.

Greater imports of cement will increase use of fossil fuels used for
transportation, and foreign sources of cement are manufactured under
conditions that are not as environmentally protective as in the United
States.

One commenter (2919) noted that the EPA should evaluate the potential
emissions from the production and transport of replacement cement for
those markets affected by plant closures. It does not appear that EPA
has addressed this issue in the preamble to the proposed rule, so the
commenter recommended that EPA evaluate this before finalizing the rule
to ensure that the final rule results in a net environmental benefit in
the event that it causes closure of cement facilities. If EPA chooses to
establish a subcategory for facilities with extremely high mercury
content ore that might otherwise be forced to close, we recommend that
it require the highest level of control achievable for that subcategory.

Response: In the RIA analysis, EPA modeled price, production, import,
and employment changes in a partial equilibrium model. In the benefit
analysis, EPA estimated disbenefits of increased energy consumption
quantitatively and the disbenefits of increased production in foreign
countries qualitatively.

Comment: One commenter (2846) stated that in the current economic
crisis, it is imperative that U.S. industry maintain sufficient
flexibility to improve operations while simultaneously continuing to
improve its record of emissions reductions. EPA’s estimate of the cost
of this proposed rule – over a quarter of a billion dollars annually
– could likely preclude emissions control technology development and
stifle economic growth, rendering the U.S. cement industry incapable of
competing effectively in the global marketplace. Given the increased
infrastructure development likely to take place as a result of the
stimulus package, Council of Industrial Boiler Owners (CIBO) believes
that EPA’s capital cost estimate for the cement industry’s
compliance with the proposed standards is likely inadequate. In imposing
MACT requirements, it is essential that EPA have realistic cost
estimates for its proposed standards, so that the impact of all proposed
standards can be adequately ascertained by all interested parties.

Response: EPA employed both a partial-equilibrium economic model and the
enhanced version of the ISIS Cement model to analyze the impact on the
industry and the economy. In the ISIS Cement model, EPA used the latest
projections on cement consumption from the Portland Cement Association
(October, 2009). The cement consumption projections take into
consideration both the economic crisis as well as the economic recovery
maintaining sufficient flexibility in the analysis of this rulemaking.
The Economic and Impact Analysis for this rulemaking using both of these
can be found in the rulemaking docket.

SO2 reductions

Comment: One commenter (2845) stated that EPA has overestimated
uncontrolled SO2 emissions by about 25 percent as a national total
because emissions were estimated as the tonnage of SO2 emissions abated
by the regulation. As a result, benefits were also overestimated because
SO2 emission reductions were multiplied by a dollar value of benefits
per ton providing benefits that are a linear multiple of the estimated
emissions reduction. It is not clear how EPA estimated the total
national SO2 emissions in either the RIA or Industrial Sectors
Integrated Solutions (ISIS) and it is believed that all of the estimates
for the 2005 SO2 emissions are inaccurate for the following reasons:

Inaccuracies in the data obtained for the 2002 NEI;

Inaccuracies in collapsing individual plant data into regional averages
or medians by kiln type and then into national averages or medians by
kiln type;

Potential inaccuracies in using medians rather than averages;

The error in multiplying emissions intensity (pounds SO2 per ton of
clinker) by clinker capacity in contrast to actual clinker production;

Errors in updating 2002 emission information for individual kilns to
2005 by inflating each kiln’s emissions by the national average
production increase from 2002 to 2005 instead of by each kiln’s actual
production increase over this period;

The potential error in including SO2 emissions from hazardous
waste-burning kilns in the inventory of 2005 baseline emissions that
might be affected by the proposed NESHAP. It is not clear whether EPA
made this error or not;

The commenter has access to a more up-to-date and accurate database
which when compared with EPA’s estimated results, the commenter’s
estimated results were 20 percent lower; therefore, because EPA
overestimated the baseline SO2 emissions by cement kilns in 2005 by
about 25 percent, this led EPA to overestimate the projected SO2-related
monetized benefits by a similar 25 percent. The commenter can provide
more up-to-date data to EPA.

Response:  We did collect additional information to supplement what was
available in the NEI, and we conducted quality control and quality
assurance audits on these data.  EPA believes this is a reasonable
estimate of SO2 reductions.    SO2 emissions from cement kilns are the
product of sulfur in the fuel as well as sulfur in the feed materials.
SO2 emission intensities (lb/ton clinker) were calculated for each kiln
type in each cement market using the SO2 emission data in the National
Emission Inventory (NEI). SO2 emission intensities by kiln type in each
cement market were used to calculate national SO2 emissions in the
traditional engineering analysis as well as in the ISIS model. The
detailed methodology used calculate SO2 emission intensities is
documented in memorandum from Staudt, Andover Technology Partners, to
Srivastava, Torres, Barnett, US EPA, “RE: NOx, SO2 and CO2 emissions
from Cement Kilns (Emissions Memo),” September 23 ,2008.

Comment: One commenter (2845) stated that EPA assumed or estimated an
unrealistically high 95 percent SO2 removal efficiency for all limestone
wet scrubbers (LWS) without regard to the reason for the scrubber (i.e.,
HCl control vs. protecting an RTO) and without regard to the inlet
concentration of SO2. According to the commenter, scrubber manufacturers
indicate that 15 ppm is the lowest SO2 outlet concentration consistently
achievable by wet scrubbers designed for SO2 capture at cement kilns.
The commenter estimated a lower removal efficiency assuming the lowest
achievable outlet concentration of 15ppm.

Commenter 2845 provided an MS Excel workbook showing the calculations as
part of their comment. The commenter projected that the proposed NESHAP
would result in 82 percent of all kilns needing to install scrubbers.
Since it is not known which specific kilns will need scrubbers, the
projected control efficiencies would need to be reduced by in order to
estimate the percentage reduction in SO2 that will likely be achieved in
practice, in contrast to the potential percentage reduction that would
occur if all kilns installed scrubbers. 

The commenter stated that it is not clear that the wet scrubbers
installed at cement kilns to meet the proposed NESHAP
requirements–either to control HCl or to protect an RTO installed to
control THC–will achieve the performance that is expected of scrubbers
intended to address SO2. The industry has no experience with wet
scrubbers intended specifically to control HCl to very low levels and
very little experience with scrubbers intended to protect an RTO.
Scrubbers designed specifically for HCl control would scrub using Sodium
Hydroxide (NaOH) and may not be suitable for scrubbers aimed at removing
low concentrations of HCl from flue gas that has a high SO2
concentration. SO2 scrubbers in cement industry generally use limestone
as a scrubbing medium (lower-cost than NaOH, results in recoverable
gypsum and little in the way of solid waste disposal cost). It is not
known how a scrubber would best be designed for removal of low
concentrations of HCl in the presence of high SO2 concentrations, or
whether the SO2 removal performance of such an HCl-optimized scrubber
would differ significantly from the typical performance of SO2-optimized
scrubbers.

Response: EPA has updated the SO2 removal efficiency of limestone wet
scrubbers (LWS) to 90%. The NSPS adopted concurrently contains a 90%
removal efficiency alternative standard for SO2 based on use of a
limestone wet scrubber, and this alternative garnered significant
industry support in the public comments. Consequently, EPA considers a
90% removal efficiency to be a reasonable estimate in assessing
reductions of SO2 in the NESHAP as well. We also disagree with the
statement that wet scrubber optimized to control SO2 will not control
HCl to the low levels required by the rule. Because HCl is more reactive
than SO2, we would expect that a scrubber optimized for 90 percent SO2
control would achieve even a higher level of control for HCl. In fact,
data collected from municipal solid waste incinerators confirms that
spray dryers (a similar SO2 control technology) remove HCl at levels
higher then those for SO2, and achieves level of 90 percent more even
when SO2 removal levels are in the 30 percent range. 

Comment: One commenter (2845) stated that using scrubbers to control
acid gas emissions from kilns with already low concentration of SO2 are
not cost-effective for the following reasons:

A majority of the monetized benefits that EPA estimates for the entire
proposed regulation derive from co-control of SO2 by kilns that are
projected to install scrubbers to meet the proposed HCl standard.

The SO2 co-benefits from controlling HCl cannot provide a valid reason
for HCl control requirements unless controlling SO2 itself is shown to
be cost-effective.

Emission control using LWS is not cost-effective for the cement kilns
that already have low concentrations of SO2 because, the commenter
estimated, it would cost more than $36,000 per ton of SO2, which is far
above a reasonable figure of cost-effectiveness for SO2 control. The
commenter provided cost and emissions reduction figures from a BACT
analysis for a single kiln as support.

Response: In the analysis of this rulemaking SO2 emission reductions are
co-benefits of HCl reductions. HCl is a hazardous air pollutant
regulated under section 112 of CAA. As noted in the preamble of the
final rule we are required to set limits for HAP based on maximum
achievable control technology. That is the reason for apply HCl
controls. We do note that SO2 co-benefits are achieved, and are very
significant. It is not necessary that we justify the cost of these SO2
co-benefits (and indeed, since they result from floor standard, EPA is
precluded from considering costs).

For these rules the SO2 reductions represent a large fraction of the
total benefits from reducing PM2.5 , but it is not possible to isolate
the portion if the total benefits attributable to the emission
reductions of SO2 resulting from the application of HCl controls. 

Comment: One commenter (2844) stated that there are challenges and
uncertainties in meeting the new limits including the following:

Facilities face extensive control investments because the limits are low
and the controls methods are largely undemonstrated in kiln
applications. 

Activated carbon injection (ACI) and scrubber performance is dependent
on having oxidized mercury to treat. The commenter does not know the
distribution of their plants’ elemental and oxidized mercury
emissions. 

The proposed standards are cost-ineffective in dollars per ton of
reduced emissions and the risk that even those significant, costly
investments will not assure compliance.

Response: EPA acknowledges that there are challenges and uncertainties
in attaining compliance with new limits under any MACT standard. See the
preamble and cost and environmental impacts documents for discussion of
control costs and estimates of cost effectiveness. 

Benefits Analysis

Comment: One commenter (2845) stated that EPA does not estimate the
benefits from control of the HAP whose emissions would be reduced by the
proposed NESHAP. Instead, EPA estimates the monetized benefits
associated with co-control of the criteria pollutants, PM2.5 and SO2 (a
PM2.5 precursor). The commenter stated that EPA performed no dispersion
modeling, exposure assessment or risk analysis for cement plant
emissions. Instead, EPA relied on the benefit per ton estimates
developed in previous RIA. The benefits per ton figure that EPA accesses
and uses for SO2 for the NESHAP benefits analysis reflect work done in
previous RIA (see EPA’s Regulatory Impact Analyses (RIA) for the PM2.5
National Ambient Air Quality Standards (NAAQS)) to quantify the
relationships between SO2 emissions, sulfates, and eventual PM2.5
concentrations attributable to SO2.

One commenter (2845) stated that EPA should not use PM2.5 reductions
that are not required by the rule to justify the cost of the rule for
the following reasons:

Even though the proposed rule was developed to regulate HAP emissions,
EPA estimates that most of the benefits, specifically the monetized
benefits, will come from reducing PM2.5.

Actions taken to reduce HAP emissions will not reduce emissions of PM2.5
and PM2.5 precursors to the extent that EPA has projected. It is
improper for EPA to take credit for PM2.5 benefits that may not be
achieved and that have not been realistically analyzed. EPA should
quantify the benefits of HAP reductions to compare to the costs of the
rule.

EPA estimated that the annualized domestic social cost of this
regulation is $694million and concedes that it does not have sufficient
information to estimate benefits from reductions in emissions of
mercury, HCl, or coarse PM (total hydrocarbons are not mentioned at
all). Nor does EPA discuss the qualitative measures of these benefits,
which must be considered essential to consider per Executive Order
12866. It is not possible to justify the cost of the potential
reductions in the four HAPs without discussion of the benefit of the
proposed reductions.

EPA’s discussion of the secondary benefits does not remedy this defect
in the cost-benefit analysis. Such regulatory intent is contrary to the
CAA State Implementation process. States may mandate PM2.5 controls and
allocate PM2.5 emissions based on unique local considerations or
legislative prerogatives, so long as the national standards are met.
States have virtually absolute power in allocating emission limitations
so long as the national standards are met and so long as the national
standards are met, the implementing state may select whatever mix of
control devices it desires. By setting MACT standards for the reduction
of PM2.5, EPA is usurping the role of the States under the CAA’s NAAQS
program. 

Response: As explained in the discussion of EPA’s decision to adopt a
MACT standard for HCl, Congress was aware that there would be
co-benefits from controls on air toxics in the form of co-control of
criteria pollutants, and Congress viewed these co-benefits as a positive
part of the air toxics control effort. It would be ironic indeed if EPA
were precluded from considering these very real benefits in those
instances where it is making discretionary choices under section 112
(d), notably in determining in whether to set a technology-based or
risk-based standard for HAPs with a threshold. Nor does EPA view
consideration of these co-benefits as somehow undermining the SIP
process. Indeed, States are customarily eager for federal controls of
criteria pollutants since otherwise States must adopt localized controls
to attain air quality standards. See e.g., 75 FR at 35570-71 (June 22 ,
2010) (noting importance of forthcoming MACT rules in helping areas
attain primary SO2 NAAQS). In addition:

Data, resource, and methodological limitations prevented EPA from
estimating the monetized benefits from reducing HAP emissions. For the
final rule, EPA provided qualitative descriptions of the health effects
associated with the top HAP emissions from cement kilns. In addition,
EPA provided maps of the modeled reduction in mercury deposition from
cement kilns. Despite our inability to provide monetized benefits of HAP
emission reductions, the total monetized benefits exceed the estimated
costs of the rule by a substantial margin, even when taking uncertainty
into account.

 The goal of an RIA is to provide a comprehensive estimate of the
anticipated costs and benefits associated with a regulatory action.
Co-benefits that occur as a result of a regulatory action are
appropriate to include in the RIA, and it is appropriate to compare the
total monetized benefits with the costs. In many instances, it is not
possible to quantify and monetize all of the benefits associated with a
regulatory action. An RIA is not prepared to justify the costs of an
action, but only to estimate what the likely impacts would be. 

The PM2.5 emission reductions that would occur as co-benefits of this
rule do not affect the States’ roles and responsibilities under the
NAAQS program. These emission reductions would occur regardless of an
area’s attainment status under the NAAQS, but these emission
reductions might help those areas that are in non-attainment to reach
attainment. This is a positive feature of the rule, not a negative.

Comment: One commenter (2890) stated that EPA must consistently evaluate
positive and negative impacts of a rule. The commenter’s specific
points were:

EPA cannot monetize the benefits of PM2.5 (or any other pollutant)
reductions to justify action at the floor. As EPA has repeatedly
claimed, the floor level of control shall not consider costs. If you
cannot consider costs to industry in assessing whether a floor is
reasonable, you also cannot use benefits (monetized or otherwise) from
other pollutants in establishing the floor. The floor must stand on its
own and follow the requirements of the Clean Air Act.

EPA cannot monetize the benefits of PM2.5 and ignore the monetized
detriments that would result from the controls. The EPA monetizes the
benefits of PM2.5 control but does not monetize the detrimental impacts
of the energy and other secondary emissions that are included in the
secondary impacts section. This is arbitrary and creates an unfair
depiction of the true impacts of this rule.

Response: The costs and the benefits presented in the RIA do not
influence the MACT floor determination. Consideration of cost (positive
or negative) is absolutely prohibited at the MACT floor, and EPA
necessarily did not consider cost or benefit information in developing
the floors. The floors reflect the average performance of the best 12 %
of existing sources, and the performance of the single best controlled
similar source (for which EPA has emission data), as required by law.
EPA did consider cost information, as well as non-air quality impacts
and energy considerations, in determining not to adopt standard more
stringent than the floor. The RIA describes impacts to energy in the RIA
in section 3.5 and the energy disbenefits in section 6.4. 

Comment: One commenter (2845) stated that methodological limitations
should not prevent EPA from estimating the monetized benefits of HAP
emission reductions from the proposed rule. Techniques exist for
conducting multipathway risk assessments for cement kilns for HAP
emissions that EPA and industry have applied many times. For example,
EPA’s worst case screening analysis concludes that an HCl emissions
limit of 23 ppmv would result in no health risk to surrounding
populations from current cement kiln emissions. In addition to risk
modeling for HAP, well-accepted methods for estimating the monetary
value of changes in HAP risks also exist and have been applied.
Executive Order (EO) 12866 requires that EPA must conduct full analysis
of benefits and costs (monetized to the extent possible) for any
economically significant regulations. EPA must analyze the benefits from
HAP emissions reductions as well as criteria pollutant emission
reductions. According to the commenter, analysis of the benefits of the
proposed NESHAP will likely show small to near-zero benefits from the
reductions in HAP emissions. It is believed that hazardous waste-burning
cement kilns will provide a worst case representation for cement kilns
and HAP risks associated with the industry will generally be less than
or equal to those from other hazardous waste-burning kilns.

Response: Data, resource, and methodological limitations prevented EPA
from estimating the monetized benefits from reducing HAPs emissions. For
the final rule, EPA provided qualitative descriptions of the health
effects associated with the top HAP emissions from cement kilns. In
addition, EPA provided maps of the modeled reduction in mercury
deposition from cement kilns. Despite our inability to provide monetized
benefits of HAP emission reductions, the total monetized benefits exceed
the estimated costs of the rule by a substantial margin, even when
taking uncertainty into account.

Comment: One commenter (2845) stated that EPA’s benefits analysis
incorrectly assumes that exposure to and risk from fine particle and
precursor emissions from the portland cement plant emissions is similar
on average to exposure and risk from fine particle and precursor
emissions from all other point sources. The commenter stated that cement
kiln emissions are likely to cause less exposure on average and are less
potent. For both SO2 and directly emitted PM2.5, EPA takes value-per-ton
benefits estimates for nationwide aggregates of these pollutants that
were developed in the Ozone NAAQS RIA and applies these values
specifically to the quantities of cement plant emissions that will be
controlled by the NESHAP. This represents an instance of what economists
term benefits transfer; taking a benefit estimate developed in one
context and applying it in another. Commenter 2845 stated that these
situations are not similar and that the benefits transfers are not
appropriate, for reasons having to do with both exposure and potency. In
support of their comment, the commenter provided analyses of exposure
and potency as factors affecting benefits.

Response: For the final RIA, EPA used the CAMx model to better reflect
the geographic distribution of the emission reductions from cement kilns
on ambient PM2.5 concentrations, which provides a superior
representation compared to the national average benefit-per-ton
estimate. 

The benefits models assume that all fine particles, regardless of their
chemical composition, are equally potent in causing premature mortality
because there is no clear scientific evidence that would support the
development of different effects estimates by particle type. The
recently released Integrated Science Assessment for Particulate Matter
concluded that “many constituents of PM can be linked with differing
health effects, but the evidence is not yet sufficient to allow
differentiation of those constituents or sources that are more closely
related to specific health outcomes.” In addition, the Health Effects
Subcommittee of the Science Advisory Board recently stated that “ the
evidence at this time does not currently support this sort of
[differential toxicity] assessment” (EPA-SAB-COUNCIL- 10-001).

Comment: One commenter (2845) stated that EPA made two other smaller
changes in adapting previous benefit-per-ton values for application to
the cement NESHAP; however, did not give a detailed explanation of what
the changes were so that the public could assess them. 

Response: EPA concurs that the two technical changes were not
sufficiently described in the proposal RIA. However, this language was
removed because benefit-per-ton estimates were not used as the basis of
the benefits estimates for the final RIA. 

In addition to the changes described in the Portland Cement NESHAP
proposal RIA, EPA’s Office of Air and Radiation has also recently
revised the value of a statistical life (VSL) used to calculate the
human health benefits of avoided premature mortality. The final RIA
contains a detailed explanation of this revision. 

Comment: One commenter (2845) stated that the various comments provided
regarding EPA’s benefits analysis would sum to a significant reduction
in EPA’s calculated values perhaps down to near the level of costs
that industry is estimating.

Response: EPA has reassessed the monetized benefits and costs for the
final rule, including the energy disbenefits. For the final rule, the
monetized benefits still outweigh the costs by a substantial margin,
even when taking uncertainties into consideration.

Small business impacts

Comment: One commenter (2887) stated that the final rule should minimize
impacts on small business. The Regulatory Flexibility Act (RFA), as
amended by the Small Business Regulatory Enforcement Fairness Act
(SBREFA), 5 U.S.C. §601 et seq., imposes procedural requirements on EPA
that must be followed in promulgating a rule. The commenter requested
that the EPA minimize the economic impact of the final Portland Cement
NESHAP on small entities. EPA’s analysis of the economic impact of the
proposed rule on small business cement plants suggested that the highest
single impact based on a cost-to-sales ratio is on a small entity. 

The commenter respectfully disagrees that the EPA minimized the impact
on small entities in the proposed rule. Small entities are subject to
the same MACT floor standards as larger businesses. To the extent that
the rule includes any provision that minimizes the impact on small
entities as compared to large businesses, the commenter requests that
the EPA specifically identify those provisions and explain the
difference in economic impact between small businesses and larger
companies. More importantly, the commenter does not agree that the MACT
floor is the least stringent level allowed by law.

Response: To the extent the commenter suggests that EPA should have
altered the floor methodology to accommodate small businesses, the
comment is misplaced because the requirements for MACT floors are
non-discretionary and dictated by statute. EPA is also certifying that
the final rule does not have a significant impact on a substantial
number of small businesses.

We estimate that 3 of the 26 existing Portland cement entities are small
entities and will not incur any impacts under the final amendments for
NSPS subpart F unless an affected facility is constructed, modified, or
reconstructed. Although this final NSPS will not have a significant
economic impact on a substantial number of small entities, EPA
nonetheless has tried to reduce the impact of this rule on small
entities by the selection of an emission level based on highly
cost-effective controls and specifying monitoring requirements that are
the minimum to ensure compliance. In the case where there are
overlapping standards between this NSPS and the Portland Cement NESHAP,
we have exempted sources from the least stringent requirement thereby
eliminated overlapping monitoring, testing and reporting requirements by
requiring that the source comply with only the more stringent of the
standards. 

For subpart LLL (NESHAP), EPA performed a screening analysis for impacts
on the three affected small entities by comparing compliance costs to
entity revenues. EPA’s analysis found that the ratio of compliance
cost to company revenue for one small entity (a tribal government) will
have an annualized cost of less than 1 percent of sales. The other two
small businesses will have an annualized cost of between 1 and three
percent of sales. Although this final NESHAP will not impact a
substantial number of small entities, EPA nonetheless has tried to
reduce the impact of this proposed rule on small entities by setting the
final emissions limits at the MACT floor, the least stringent level
allowed by law. In the case where there are overlapping standards
between this NESHAP and the Portland Cement NSPS, we have exempted
sources from the least stringent requirement thereby eliminating the
overlapping monitoring, testing and reporting requirements by requiring
that the source comply with only the more stringent of the standards. In
addition, we applied MACT for HCl emissions to major sources only. The
reduced compliance costs for two of the three small entities by a factor
of 4.

More discussion is provided in the preamble.

Comment: In extrapolating from available data, one commenter (2845)
asked if EPA has applied the correct basis for factors, such as using
the correct percentage of raw mill on versus raw mill off when applying
the emission rate to additional facilities (can affect mercury, HCl, and
SO2)? Have operational considerations across the various kiln types been
taken into account?

Response: When appropriate, EPA did take into consideration the effects
of the process on emission rates. For example, in identifying the top
performing kilns for PM emissions, we recognized that PM emissions were
generally higher during raw mill off conditions compared to raw mill on.
As a result, we calculated a weighted average PM emission level for
kilns with raw mills where we had test data for both raw mill on and off
conditions. When we collected additional HCl data, we specified that for
kilns with raw mill, HCl testing was to be done with the raw mill on and
off. The mercury emissions limit w as based on 30-days of raw material
sampling and analysis for mercury content so that the raw mill status
was not an issue. Also, the THC limit was based on long-term (hundreds
of days) CEM data which would account for the raw mill. In addition, all
of the compliance periods are set as 30 day averages which allow for the
changes in the raw mill operations. 

2.12	Environmental Justice

Comment: One commenter (2786) stated that mercury pollution more
significantly impacts poor and minority populations, and EPA should
regulate mercury emissions from the portland cement industry to the
greatest extent possible because of their adverse health impacts on poor
and minority communities. Executive Order 12898 states that to the
extent practicable and permitted by law, each federal agency shall make
achieving environmental justice part of its mission by identifying and
addressing, as appropriate, the disproportionately high and adverse
human health or environmental effects of its programs, policies, and
activities on minority populations and low-income populations in the
United States. The commenter stated that EPA has historically devoted
little attention to evaluate the impact of decisions made under the
Clean Air Act on low-income and minority communities. The commenter
provided discussion about pathways for human ingestion of mercury and
how poor and racial minority group are exposed to higher levels of
mercury through their livelihood and culture. In addition to the impact
of mercury on poor and minority communities, the commenter also listed
the following HAPs as matters of potential regulation: lead, chromium,
arsenic, and beryllium.

Response: See discussion of Environmental Justice in Section VII of the
preamble to the final rule.

Comment: One commenter (2863) stated that the proposed rule has serious
environmental justice issues. 

The first issue is that two plants are likely to have to close as a
result of the rule. The populations around these plants rely heavily on
the plants for income. The commenter conducted a comprehensive risk
assessment showing that the mercury emissions do not result in
unacceptable risks even before controls. The adverse impact of closure
would be to the health of the families who depend on this facility for
income and health insurance, with no commensurate improvement related to
the reduced emissions associated with closure as the plant will be well
under the acceptable risk thresholds when the controls currently being
constructed are operational.

Closure of the two plants will mean that more cement will be imported
from overseas. As Asia and India dominate imports to the west coast, the
gap left by California and Oregon plant closures will almost certainly
occur in nations with far less stringent controls and significantly
higher direct and indirect emissions of criteria pollutants, HAPs and
greenhouse gases. This is particularly relevant in relation to mercury
and greenhouse gases as these are the most obviously global pollutants.
However, moving production offshore will also result in localized
impacts and so will disproportionately impact impoverished areas.

Response: In this final rule, we are not projecting any facility
closures. To assess whether this rule's environmental and health impacts
would have a disproportionate effect on vulnerable populations, EPA
looked at demographic data in aggregate for census block groups in near
proximity to cement facilities. EPA analysis showed that within the
radii examined (0.5 mi, 3 mi, and 5 mi from sources subject to this
rule), the percentages of minority and low income populations were
similar to national averages. EPA's demographic analysis can be found in
the docket for this rulemaking.

In addition, as stated in the final rule preamble, EPA is reaching out
to the two high mercury limestone quarry sources to investigate the
possibilities of early aggressive reductions of mercury emissions with a
negotiated longer compliance period.

Comment: One commenter (2898) stated that EPA has failed to undertake an
actual analysis of the consequences of the rule on environmental justice
populations. The commenter stated that EPA should undertake a more
thorough analysis of the risks posed by the cement industry and the
benefits of more stringent standards.

Response: EPA analyzed each standard to determine if beyond the floor
control was warranted, and concluded that further controls were not
warranted due to high cost, poor cost effectiveness, non-air quality
impacts in the form of waste generation and water usage, and increased
energy use. Nothing in this comment undermines those conclusions. EPA
does note that it took two significant discretionary decisions in the
rule – adopting a MACT standard for HCl and not subcategorizing for
two high mercury emitting sources – which result in very significant
reductions of acid gases, secondary fine PM, and mercury.

Comment: One commenter (2908) stated that the proposed rule may
constrict cement supply and increase pressures on the availability of
affordable housing. 

Response: In this final rule, we are not projecting any facility
closures. To assess whether this rule's environmental and health impacts
would have a disproportionate effect on vulnerable populations, EPA
looked at demographic data in aggregate for census block groups in near
proximity to cement facilities. EPA analysis showed that within the
radii examined (0.5 mi, 3 mi, and 5 mi from sources subject to this
rule), the percentages of minority and low income populations were
similar to national averages. EPA's demographic analysis can be found in
the docket for this rulemaking.

Comment: One commenter (3222) stated that the commenter’s local cement
plant creates environmental and social justice problems.

Response: In this final rule, we are not projecting any facility
closures. To assess whether this rule's environmental and health impacts
would have a disproportionate effect on vulnerable populations, EPA
looked at demographic data in aggregate for census block groups in near
proximity to cement facilities. EPA analysis showed that within the
radii examined (0.5 mi, 3 mi, and 5 mi from sources subject to this
rule), the percentages of minority and low income populations were
similar to national averages. EPA's demographic analysis can be found in
the docket for this rulemaking.

2.13	Open Clinker Piles

Comment: Several commenters (2845, 2864, and 2874) stated that in the
past, EPA considered regulating open storage piles; however, in the 2001
settlement agreement and subsequent follow-up letter from EPA to
American Portland Cement Alliance that settled litigation over the 1999
NESHAP final rule, EPA agreed not to regulate fugitive emissions from
open or unenclosed material stockpiles or haul roads. Although, the
proposal asserts that open clinker storage piles located at two
facilities had measurable emissions of hexavalent chromium, the
emissions and the ambient concentrations in the area surrounding cement
facilities were low. The ambient concentrations are also miniscule when
compared to benchmark regulatory standards. Also, several California air
quality district have decided against regulating hexavalent chromium
emissions from cement plants due to the lack of risk. For these reasons,
it would be inappropriate to define open clicker storage piles as
affected sources under the NESHAP.

Response: At proposal we noted that open clinker piles were currently
unregulated, and that hexavalent chromium emissions had been detected in
fugitive dust from these piles. See 74 FR at 21163. We requested comment
and information as to how common the practice of open clinker storage
is, appropriate ways to detect or measure fugitive emissions (ranging
from open-path techniques to continuous digital or intermittent manual
visible emissions techniques), any measurements of emissions of
hexavalent chromium (or other HAP) from these open storage piles,
potential controls to reduce emissions, or any other factors we should
consider. 

Commenters did not provide data on this practice. Industry commenters
stated emissions were de minimis and should not be regulated. Other
commenters noted that the fact that we know these sources emit HAP is
sufficient to necessitate regulation. 

We agree that these operations do emit HAP and that regulation of these
sources is necessary. See National Lime, 233 F. 3d at 640 (upholding EPA
position that de minimis exceptions are not to be read into the MACT
standard setting process). Because the emissions in question are
fugitive dust for which measurement is not feasible since (by
definition) the emissions are not emitted through a conveyance or other
device which allows their measurement (see section 112 (h) (1) and
(2)(A)), we are incorporating the work practice standards and opacity
emissions limits contained in Rule 1156 as amended by the South Coast
Air Quality Management District on March 6, 2009 and incorporating them
into this rule. There are only two plants which EPA can state
definitively have open storage piles and are complying with Rule 1156,
so these existing regulatory standards would constitute a floor level of
control (and EPA does not believe beyond-the-floor controls are needed,
since utilizing some type of enclosure should well control fugitive
emissions). 

Industry commenters also maintained that regulation of open storage
piles would violate a 2001 settlement agreement between EPA and the
industry in which EPA agreed that the 1999 rule did not apply to
fugitive emission sources. But nothing in that settlement agreement
prevents EPA from amending its regulations if it is appropriate to do so
(nor could EPA legally bind itself in such a way). The agreement in fact
states that “[n]othing in this Agreement shall be construed to limit
or modify EPA's discretion to alter, amend, or revise, or to promulgate
regulations that supersede, the regulations identified in section III of
this Agreement.” Consequently, EPA’s action today amends the current
regulation, and does not violate any provisions of the settlement
agreement. 

See the preamble discussion for more information on EPA’s decision. 

Comment: Several commenters (2144, 2816, 2898, 2911, and 2913) supported
EPA’s efforts to investigate and address the effects of hexavalent
chromium from clinker storage. 

Response: EPA acknowledges the commenter’s support.

Comment: Several commenters (2816, 2911, and 3162) agreed that open
clinker piles should be included in the regulation. Commenter 2816 noted
that the Agency is already regulating enclosed clinker piles, and the
Agency has noted measurable emissions of hexavalent chromium due to
fugitive emissions. The commenter stated that while it may be more
difficult to control fugitive emissions, it does not relieve either the
industry or the Agency from reducing to the maximum degree feasible
emissions from this source. Commenters 2816 and 3162 recommended that
EPA should review the rules adopted by the South Coast Air Quality
Management District and consider adopting similar requirements for open
clinker piles.

Response: EPA has adopted this approach in the final rule.

Comment: One commenter (2864) stated that EPA should set work practice
standards in the event that EPA is compelled to regulate open clinker
piles.

Response: EPA has done so in the final rule.

Comment: One commenter (2898) stated that the rule should cover all
emissions of hexavalent chromium from a cement manufacturing facility
that constitutes a major source, including emissions from storage units,
conveying system transfer points, and fugitive emissions, including
those from open clinker piles. 

Response: The final rule includes provisions for clinker piles. See
preamble for more discussion.

2.14	Crushing Operations

Comment: Several commenters (2845 and 2874) stated that the 2001
settlement agreement that settled litigation over the 1999 NESHAP final
rule specified that the final rule should not apply to primary and
secondary crushing operations located at portland cement plants with
nonmetallic mineral processing operations. On April 05, 2002, EPA
amended §63.1240(c) to implement the language from the 2001 settlement
agreement. However, with no discussion in the preamble, the proposed
revision to §63.1240(c) would delete language concerning equipment of
the on-site nonmetallic mineral processing plant that precedes the raw
material storage not being subject to the NESHAP. The proposed rule only
leaves language stating that crushers are not covered by this subpart
regardless of their location without any explanation given in the
preamble. Commenter 2845 opposed the proposed breach of the 2001
settlement agreement negotiated 7 years ago. The rule would become
unclear as to whether subpart LLL of the NESHAP would apply to quarry
sources such as conveyance systems and exterior stockpiles associated
with crusher systems. There was no reason to eliminate the certainty of
the current language in §63.1340(c) which will only add to confusion
and needless applicability of the NESHAP to sources already regulated
under subpart OOO NSPS; thus, EPA should not amend §63.1340(c).
Commenter 2845 also stated that a word has been left out of the proposal
preamble language discussing §63.1340(c) where it states that crushers
were part of the affected source for this rule instead of crushers were
NOT part of the affected source for this rule. Thus, there appeared to
be a conflict with the current proposal, the current preamble, and also
the current preamble language is inconsistent with the 2006 portland
cement final rule. The commenters requested EPA to clarify in the final
rule that crushers were never covered by the NESHAP and should also
retain the existing language for §63.1340(c).

Response: We have not made the change requested by the commenters. See
the preamble to this final rule. The reason for this change was
explained in the proposed 2005 rule and final 2006 rule (Cite to 70 FR
72341-42 and 71 FR 76532).

2.15	Coal Mills

Comment: Several commenters (2845 and 2910) stated that portland cement
plants grind their coal using a stand-alone coal mill or in-line coal
mill and utilize two configurations of an inline coal mill. The first is
where air pollution control devices from the inline coal vent directly
to the atmosphere and the second configuration is where exhaust gases
from the inline coal mill are directed back to the main kiln stack,
where they are commingled with the kiln emissions. The commenter argued
that stand-alone coal mills that vent directly to the atmosphere should
remain regulated under Subpart Y whereas stand alone and inline coal
mills where the gases vent out the main kiln stack should be regulated
by the NESHAP. When inline coal mill gases are vented through the main
kiln stack, more THC emissions will exist than if there were no
commingling of kiln and coal mill exhaust gases. Thus, establishing a
subcategory for this particular configuration may be appropriate and the
commenter proposed that additional THC testing be conducted in order to
examine possible subcategorization for THC.

Response: EPA did consider subcategories based on coal grinding
configurations as suggested by the commenter but decided against
creating separate subcategories. See the preamble discussion under the
THC limits for more information on EPA’s decision.

Comment: Several commenters (2845 and 2864) stated that since coal mills
operate with a blend of kiln gases and ambient air, the concentration of
pollutants exiting the coal mill would not be expected to be higher than
the concentration exiting the kiln. In addition, coal mills are
generally equipped with some sort of temperature alarm or indicator to
avoid safety hazards related to the explosion potential of coal dust. If
the temperature of the gases leaving the coal mill is below the
temperature at which pollutants are formed, the concentration leaving
the coal mill would not exceed the concentration exiting the kiln and
thus performance testing and pollutant monitoring for coal mills should
not be necessary. The commenter suggests that language be added to the
rule specifying that although no in-line coal mills are covered by
subpart 63.1340(e), they must adhere to the standards in 40 CFR part 60
subpart Y, as applicable. In-line coal mills with exhaust that is
commingled with kiln exhaust are covered by this subpart, exempt from
subpart Y, and are exempt from monitoring and testing under this
subpart.

Response: If a coal mill is heated with kiln exhaust, it is subject to
all the emissions limits, and associated monitoring, reporting, and
recordkeeping requirements of this subpart.

Comment: One commenter (2863) stated that EPA should clarify that
indirect fired, in-line coal mills need not employ mercury or HCl CEMs
or separately demonstrate compliance if they vent through a separate
stack. Indirect fired in-line coal mills receive their sole source of
air and heat from the associated kiln (or in some cases, the clinker
cooler). Therefore, consistent with numerous EPA opinions, these units
are subject to the kiln exhaust standards. The proposal requires that
affected sources install and operate either a mercury CEM or sorbent
trap-based integrated monitor. There is no basis for concluding that the
drying of coal will increase the mercury or HCl concentration in the
kiln exhaust that is vented through the coal mill. In order to avoid
explosion hazard, indirect fired, in-line coal mills must be maintained
at temperatures below 200°F at all times and so the temperature is too
low to volatilize the mercury in the coal.

Response: We have not covered coal mills with separate stacks. They are
not listed as an affected source, so there is no need to mention them in
the rule. They were not originally covered in this subpart and are not
covered in the final rule.

If a coal mill is heated with kiln exhaust, it is subject to all the
emissions limits, and all associated monitoring, reporting, and
recordkeeping requirements of this subpart.

2.16	Compliance Assurance

Support for CEMS

Comment: Several commenters (2144, 2622, 2828, 2895, and 2911) supported
the provision for CEMS for portland cement kilns.

Response: We appreciate the commenters’ support.

2.16.1	Flow Rate Monitoring

Comment: One commenter (2845) stated that EPA has not developed and
promulgated appropriate performance specifications for effluent flow
rate monitors in 40 CFR 60 or 63. PS-6 is inappropriate for use with
either PM CEMS certified under PS-11 and/or mercury CEMS certified under
PS-12A or mercury sorbent trap monitoring systems certified under
PS-12B. Appropriate performance specifications and quality assurance
procedures should be developed for the proposed monitoring applications
in Subpart LLL.

Response: Part 75 contains flow rate specifications which are now
referenced on this rule.

2.16.2	Kiln Feed/Clinker Production

Comment: One commenter (2845) stated §63.1349(b)(1)(vi) requires
continuous measurement of kiln feed rate and clinker production during
the initial performance test for sources that use PM CEMS which is
unnecessary and should be changed. In addition to the PM concentration
measured by the PM CEMS, the effluent flow rate and clinker production
rate are needed to determine emissions in units of the standard (lb/ton
clinker produced). The measurement of both clinker production rate and
kiln feed rate are redundant. Measurement of one or the other process
parameter should be required, but not both.

Response: We have modified the rule language to make it clear that
separate direct measurements for kiln and clinker production are not
required. A facility always has the option of calculating clinker
production from the feed rate and their site specific clinker to feed
ratio..

Comment: One commenter (2845) stated that if an area source directly
monitors clinker output and emission limits are expressed as mass of
pollutant per ton of clinker produced, kiln feed does not need to be
recorded. If final standards are expressed as mass of pollutant per ton
of kiln feed, clinker production does not need to be recorded. Keeping
records of both parameters (daily clinker production rates and kiln feed
rates) as required by §63.1349(b)(3)(iii) is redundant.

Response: The emission limits for mercury and PM are based on clinker
production. Limits for HCl and THC are concentration-based. Most
facilities record kiln feed rates and apply a conversion to get clinker
production. If kiln feed is measured, records of the kiln feed and the
conversion factors should be kept. If clinker production is measured,
there is no need to keep records of kiln feed rates.

Comment: One commenter (2845) stated that very few cement plants in the
United States physically weigh clinker as it is produced. Those that do
weigh clinker are typically the newer precalciner kilns systems. In
older existing kilns, maintenance and calibration are difficult. Due to
the nature of the kiln process, at least on a short-term basis, clinker
output from the kiln will fluctuate while the kiln feed rate remains
constant. For those kilns that do continuously weigh clinker, the ±5%
accuracy requirement is not achievable for an hourly averaging period,
as is suggested in the proposed rule language. The ±5% accuracy is
achievable on a monthly basis where these short-term fluctuations and
upsets average out. Therefore, the commenter proposed that a monthly
averaging period for accuracy of clinker weigh scale systems for those
kilns that are able to employ clinker weigh scale systems.

The commenter agrees with EPA’s proposed alternative to measuring
clinker, but noted that procedures for development of the clinker factor
and verification of its accuracy need to be developed. The commenter
recommended that a standardized procedure should be developed through
ASTM.

The commenter provided suggested language for §63.1350(o)(1) to reflect
these suggestions.

Response: See response in preamble to final rule indicating that clinker
need not be weighed but can be estimated from the facility’s feed to
clinker ratio.

Comment: Several commenters (2845 and 2874) stated that there has been
confusion about the timeframe for the clinker weigh system. Some
interpret it as 30 days from when the rule is promulgated (i.e., 35
months before an existing source must comply with the rule), and others
interpret it as 30 days from the time existing facilities must comply
(i.e., 3 yrs + 30 days). The rule needs to specify that existing
facilities must comply on the compliance date for the revised PM or
mercury Standard (whichever may occur sooner), which is three years from
the effective date of the revised rules.

Response: For existing sources we proposed a compliance date of 3 years
after the promulgation of the new emission limits for mercury, THC, PM,
and HCl to take effect. This is the maximum period allowed by law. For
new sources, the compliance date will be the effective date of this
final rule or startup, whichever is later. Because this is a major rule
as defined by the Congressional Review Act, the effective date of the
rule is 60 days after publication of the Federal Register. Accordingly,
we have selected a date that allows facilities to design and install the
required control and other ancillary equipment (e.g., clinker weigh
systems) prior to the compliance date.

Comment: One commenter (2874) stated that EPA should clearly state that
the use of kiln feed monitoring does not require Administrator approval
and should modify the proposed regulatory text relating to compliance.
EPA’s acknowledgement in the regulations that this alternative method
is acceptable should act as Administrator approval and no owner or
operator should need to request Administrator approval to utilize a
methodology deemed acceptable by the regulations. Commenter 2874
suggested that the last proposed sentence appearing in both 40 CFR
63.1349(b)(1)(ii) and 40 CFR 63.1350(o)(1) be revised to state that The
use of a system that directly measures kiln feed rate and uses a
conversion factor to determine the clinker production rate is an
acceptable method and a method that does not require further
Administrator approval.

Response: Since it appears from comments that the equipment to
accurately measure clinker is not typically installed in this industry,
we must assume these facilities use a feed-to-clinker conversion factor
to calculate clinker production on whatever time basis is necessary
(e.g., daily, hourly, etc.). Therefore, we have modified the rule
language to more clearly provide the option allowing facilities to
measure feed inputs and to use their site specific feed/clinker ratio to
calculate clinker production (and that no prior approval from a
regulatory authority is necessary to do so). Facilities would be allowed
to use a constant feed/clinker ratio in accord with their usual cycles
for determining such ratios, typically on a monthly basis when clinker
inventories are reconciled.

2.16.3	Electronic Reporting

Comment: One commenter (2816) supported EPA’s decision to improve
accessibility of data from stack tests and include monitoring
information in the WebFIRE data base. The commenter urged EPA to make
electronic data submission a requirement of this rule. An electronic
database that is accessible to the public is an important accountability
and enforcement tool. Communities surrounding these industrial
facilities deserve to know detailed emissions data. Furthermore, as the
EPA points out in the preamble, electronic submission of data actually
reduces the burden on industry, and is a cost‐saving measure for the
Agency as well.

One commenter (2834) stated that requiring electronic submission of
emissions will allow EPA to bypass the laborious and cumbersome process
of collecting information from a variety of sources in a variety of
formats. Given that nearly every one of the categories in which EPA has
requested specific comments includes a request for additional data,
requiring electronic submission of emissions tests should be required.
Further, electronic submission of emissions results will facilitate
public to access those results. Standardized, electronic reporting of
emissions test results to EPA would facilitate the public’s access to
these emissions records.

Response: Our intent is to make electronic reporting a requirement
starting in 2012 when the system is fully capable of accepting data. The
rule language has been modified to reflect this.

2.16.4	Other

Detection limit

Comment: Several commenters (2832 and 2859) argued that EPA’s proposed
revision to an approved EPA stack testing method, specifically to
promulgate a downward revision of a method detection limit in an
approved stack testing method is arbitrary and reflects that EPA has
failed to consult and take action based on direction from its own
Emissions Measurement Center. In addition, EPA has failed to explain its
numeric approach to the actual lower detection limit selected. The
detection limit should be set at a level that all sources that are
subject to stack testing requirements would be able to achieve uniformly
in practice. The CAA requires that all stack testing detection limits
take all factors that can or do influence results into consideration
when proposing a detection limit.

Response: We have taken the appropriate steps to address the detection
limit issue, including consulting with EPA test experts and taking
comments on the detection limit in the proposal notice. Data obtained to
date conforms that detection limits using EPA Method 321 for HCl
emissions measurement are 0.2 to 0.3 ppmv in the field.

Alternatives to CEMS

Comment: Several commenters (2832 and 2859) opposed the proposed
requirement to install CEMS in order to satisfy compliance assurance
monitoring (CAM) for selected pollutants. Instead, the commenter
proposed that CAM requirements be satisfied using periodic stack testing
to the extent that stack testing is requested or require by state air
permits. According to EPA’s proposal, the MACT floor for new and
existing sources in this industry will be determined by stack testing
results of sources within the MACT pool. If EPA were to finalize a
numeric emissions limitation based on this approach to setting the new
and existing MACT floors, that limitation will be based on the same
stack testing data. CEMs will have played no role in this process. It
stands to reason that compliance assurance should be based on stack
testing results, and not a CEMs data that has played no part in this
process.

One commenter (2916) stated that EPA can achieve a reasonable assurance
of compliance without the use of CEMS. The requirement to use CEMS is
unreasonably costly and unnecessary, given that other reliable means of
showing compliance are available for all relevant pollutants. Raw
material sampling and kiln parametric monitoring, in conjunction with
periodic testing, would work well for THC and HCI. The sorbent trap
method for mercury is a good alternative to mercury CEMS and should be
retained in the final rule. EPA should refrain from requiring PM CEMS in
the final rule. Bag leak detection systems and parametric monitoring of
ESPs are proven methods for assuring ongoing compliance with PM limits.

Response: See response in preamble 

Modified visible emission requirements

Comment: According to one commenter (2874), EPA has modified the visible
emission requirements of 40 CFR 63.1350(a)(4) without adequate
explanation for the proposed revisions. The commenter stated that EPA
originally set a 1-minute standard as it was believed by the industry
and EPA that a longer test would be unduly burdensome–especially in
light of the fact that tests must be conducted for each affected source.
Nothing has occurred during the past 10 years that warrants a 20-fold
increase in the manpower and cost required for the testing. The
commenter suggested that EPA retain the current 1-minute testing
requirement.

Response: As explained above we have changed the observation time to 10
minutes.

Clarifications

Comment: One commenter (2845) stated that §63.1349(b)(3)(iii) required
that hourly average temperatures be calculated for each dioxin/furan
performance test run which is inconsistent with §63.1349(b)(3)(iv) and
should be corrected. The language of the existing §63.1349(b)(3)(ii)
states that one minute average temperatures must be calculated for each
minute of each run of the test and was replaced by language which states
that hourly average temperatures must be calculated for each run of each
test. This is inconsistent with §63.1349(b)(3)(iv) which correctly
states that the run average temperature must be calculated and the
average of all runs is used to determine the applicable temperature
limit. This is easily corrected and clarified by re-inserting a slightly
modified version of the original language which states that one minute
average temperatures must be calculated and recorded for each minute of
each run of the test.

Response: We agree and have made the chances suggested by the commenter.

Comment: One commenter (2832) stated that EPA needed to ensure that the
regulatory rule analysis statements in the preamble are reflected in the
regulatory text, including proposing specific regulatory text to make
the regulation clearer and easier to implement.

Several commenters (2864 and 2874) stated that the regulations proposed
at 40 CFR 63.1349(f) and (g) are so vague and contain numerous
inappropriate cross references to other regulations that it impossible
to provide a meaningful comment. The commenters provided several
examples of what they considered inappropriate cross references.
Commenter 2874 suggested that EPA withdraw the rule and perform a
complete review for these types of errors, fix the errors and re-propose
the rule for comment.

Response: We agree that the proposed rule contained some errors, and we
have made several corrections and clarifications to the final rule to
ensure it is consistent with our stated intent in the preamble. These
corrections and clarifications are not significant changes to the
proposed rule, however, and reproposal is not needed. We disagree
however that §63.1349(f) and (g) are confusing, and have retained these
cross references as they are accurate and ensure consistency.

Mercury, HCl and PM CEMS not demonstrated

Comment: One commenter (2858) stated that while THC CEMS have been used
by cement plants for years, CEMS monitors for mercury, HCl and PM have
not been demonstrated to be reliable and sufficiently accurate to
satisfy EPA’s performance specifications when operating on cement
kilns.

Response: We disagree. Mercury CEMS have been widely applied in the
utility industry and there are no technical reasons that would imply
that do not work equally well on a cement kiln. PM CEMS have also been
demonstrated in utility boilers. Hydrochloric Acid CEMS have been used
on source categories such as municipal waste combustors and utility
boilers. 

Publicly available monitoring data

Comment: Several commenters (2893 and 2895) requested that EPA make all
CEMS and other monitoring information readily available to the public,
and when produced in real time, available in real time on the web site.
Commenter 2893 stated that mechanisms to enforce the rules need to have
enforcement provisions that not only document agency compliance, but
also make the agency reports readily available to the public. Where
possible, emissions information should be posted on the EPA or Air
District’s web sites in a timely manner. This request is driven by the
commenter’s experience with inadequate responses from enforcement
agencies to potentially excessive emissions from local plants. 

Several commenters (2816 and 2928) recommended the use of CEMS as the
better monitoring method, and suggested a specific mercury monitor
suitable for use. The commenter suggested that to further earn public
confidence, the data should be reported every thirty days and the
monitoring results be made available to the public online. Commenter
2816 urged EPA to adopt CEMS requirements for all of the pollutants for
which an emission limit will be adopted.

Response: We are currently developing a reporting mechanism for CEMS
data. We will consider these comments as we continue that effort. In
concept, the CEMS data base would be publically available. However, the
reporting frequency has not been developed.

Potential to emit to determine applicability

Comment: One commenter (3138) stated that EPA should determine whether a
facility must comply with the standard based on their potential to emit
under the worst case operating scenario. The commenter stated that it
should make no difference if the source is an area or a major source.
The potential to emit should be based on the facility’s permitted
fuels, hours, and potential emissions during SSM events. The
determination for compliance with the new standard should include the
potential emissions from grace periods allowed by state regulators.

Response: Questions of how to calculate Potential to Emit are not part
of this rulemaking, and therefore this comment is beyond the scope of
this proceeding.

2.17	Other – Miscellaneous Comments

Failure to address greenhouse gases

Comment: Several commenters (2832 and 2859) opposed the approach taken
by EPA in its beyond-the-floor MACT analysis. Among other things, EPA
failed to consider the creation of incremental greenhouse gas emissions
associated with the construction, installation and operation of new
emissions control equipment, and the minimal incremental environmental
benefit associated with those controls. Also, EPA failed to consider the
cost of carbon credit purchases by the industry.

Response: See preamble for response.

	Counting multiple production units in MACT pool

Comment: One commenter (2832) concurs with EPA’s approach to
determining the number of sources that comprise the so-called MACT pool
to the extent that EPA has counted multiple production units whose
emissions are ducted to a single control device as one source. Any
attempt to apportion emissions between or among multiple production
units in the foregoing configuration would be arbitrary. If EPA
determined to count each production unit as a separate source, it would
be obligated to attribute for MACT purposes, all of the emissions from
the single control device to each source because there is no
scientifically demonstrable methodology to apportion emissions. The
legal consequence of EPA’s approach to determining the number of
sources that comprise the so-called MACT pool is to establish an
emission standard based on air emission control device removal
efficiency and as required by §112. Failure to act in a manner
consistent with the foregoing would be arbitrary.

Response: EPA acknowledges the commenter’s support.

Clarifications	

Comment: One commenter (2845) stated that the definition of clinker
should include that clinker can be ground with either gypsum and/or
other materials to form cement. Commenter 2845 suggests the definition
could be changed to read: clinker means the intermediate product of the
process in which limestone and other raw materials are heated in the
kiln and are then ground with gypsum and/or other materials to form
cement.

Response: See the final rule for the definition of clinker.

Comment: One commenter (2845) stated that since crushers vary in the
size of rock they produce and because crushers and their associated
conveying systems are already regulated under 40 CFR part 60 subpart
OOO, the definition of a crusher should be deleted and instead add a
definition describing a crusher system that reads as a crusher system
means equipment used to reduce the size of large raw material for
further processing and the associated material handling sources
including, but not limited to, screens conveying systems, storage bins
preceding the raw mill, and stockpiles or if §63.1340(c) is not changed
then no definition of crusher or crusher system will be needed. 

Response: See the final rule for the definition of clinker.

Comment: One commenter (2845) stated that the definitions for greenfield
kilns, greenfield in-line kiln/raw mills, or greenfield raw material
dryers and new brownfield kilns, new brownfield in-line/raw mills, or
new brownfield raw material dryers are no longer needed because all
references to them have been remove from the rule.

Response: The definitions identified by the commenter were removed from
the final rule.

Comment: One commenter (2845) stated that a totally enclosed conveying
system transfer point is defined within §63.1350(a)(4)(v) as a
conveying system transfer point that is enclosed on all sides, top, and
bottom. This definition should be expanded and moved to §63.1341, and
should read: Totally enclosed conveying system transfer point means a
conveying system transfer point that is enclosed on all sides, top, and
bottom. This includes, but is not limited to, pneumatic transfer
systems. 

Response: A definition of “totally enclosed conveying system transfer
point” was added to the final rule.

Comment: One commenter (2845) stated that §63.1343(b)(3) and
§63.1343(c)(3) should be clarified to state that dioxin/furan standards
are on a dry basis corrected to 7 percent oxygen.

Response: The rule was clarified to state that the standards for
dioxin/furan are on a dry basis at 7 percent oxygen.

Comment: One commenter (2845) stated that an additional option should be
included in §63.1344(a): §63.1344(a)(4) The owner or operator may
choose to set one applicable temperature limit for the main in-line
kiln/raw mill exhaust, specified in paragraph (b) of this section and
established during the performance test when the raw mill was not
operating. If the owner or operator chooses this option, the limit would
apply when the raw mill of the in-line kiln/raw mill is operating and
when the raw mill is not operating. As this is a more restrictive limit,
additional justification should not be necessary.

Response: The suggested revision was considered to be outside the scope
of this rulemaking; no change was made.

Comment: One commenter (2845) stated that §63.1344(e) should be
clarified to clearly state that the requirements only apply if using
carbon injection specifically for dioxin/furan control.

Response: The language was clarified as suggested by the commenter.

Comment: One commenter (2845) suggested that EPA should provide an
option for one dioxin/furan compliance limit set by using performance
test when raw mill was not operating. This is similar to the option
provided in 40 CFR 63 Subpart EEE. As it is a more restrictive limit,
additional justification should not be necessary.

Response: EPA did not believe the suggested change was within the scope
of this rulemaking.

Comment: One commenter (2845) stated that §63.1349(b)(3)(iv) should
include an option for maintaining the temperature limit if a repeat test
shows a lower value. The commenter noted that when conducting the repeat
performance test every 30 months for dioxins/furans, the resulting
temperature limit can often vary arbitrarily above or below the previous
temperature limit (set during the previous performance test). If no
change is operations has occurred, as described in 40 C.F.R. 63.1349(e)
as otherwise requiring the establishment of a new temperature limit, the
facility should be allowed to maintain their current temperature limit
even if a new repeat performance test average temperature is lower than
the current limit. The 30 month test should be allowed to be viewed as a
confirmatory type test that is demonstrating compliance with a limit
that is not changing, rather than to continuously change the actual
limit. This would be consistent with §63.1349(e) which requires
establishing new temperature limits only when a change in operations
occurs that would adversely affect compliance with a dioxins/furans
standard. The commenter provided suggested language.

Response: EPA did not believe the suggested change was within the scope
of this rulemaking.

Comment: One commenter (2845) stated that in §63.1349 (c) both existing
and new kiln and in-line kiln/raw mills are required to have CEMS for
PM, THC, mercury, and HCl. This rule appears to indicate that none of
the performance testing requirements in §63.1349 apply. If this is
indeed the correct interpretation, it should be made clearer in the
individual sections requiring performance testing that the CEMS are
sufficient to meet the initial and ongoing performance testing
requirements. The commenter suggested that EPA provide clarification on
the intent and extent of §63.1349 (c).

Response: Section 63.1349 (c) specifies that where CEMS is used for
monitoring, performance testing is not required.

Comment: One commenter (2845) stated that the final rule should clarify
if each vent must demonstrate compliance for situations where there are
multiple vents since there is no language in the current proposal
regarding sources with multiple stacks for a common source(s).

Response: We believe the rule is sufficiently clear that each source
must demonstrate compliance, regardless of the number of stacks or
vents.

Comment: One commenter (2845) stated that §63.1349(f) and (g) appear to
be direct excerpts from proposed HMIWI NSPS – 40 CFR §60.50, and
stated that these references do not apply to cement plants. Section
63.1349(a) already states what is required to be submitted as part of
the report required under the PC MACT regulation. The commenter noted
that there is a typographical error between bullets 3 and 4, and section
(g) appears to have been added in error.

Response: EPA has revised the section by removing paragraphs (f) and
(g).

Comment: One commenter (2845) stated that in §63.1350(g), the rule
should clarify that carbon injection requirements only apply to sources
utilizing carbon injection for dioxin/furan control.

Response: The rule is clear regarding monitoring requirements that apply
to sources using sorbent injection for D/F control.

Comment: One commenter (2845) stated that in §63.1350(g)(6), the
reference to mercury and THC should be removed to be consistent with
regulatory text.

Response: EPA has removed the reference to mercury and THC.

Comment: One commenter (2845) stated that Table 1 to §63.1350 was not
updated in the proposed rule. 

Response: Table 1, Applicability of General Provisions, is updated.

Comment: One commenter (2845) suggested that EPA delete all sections in
63.1353, 63.1354, and 63.1355 that reference subpart A and provide a
paraphrase of subpart A. The commenter noted that because Table 1 will
contain a listing of the sections of subpart A that apply and those that
do not apply, sections that reference subpart A and provide a paraphrase
of subpart A should be removed to prevent multiple interpretations.

Response: Sections 63.1354, 63.1355, and 63.1356 are being retained
because many of the recordkeeping and reporting requirements specific to
this rule, are not contained in the General Provisions.

Comment: One commenter (2845) suggested that §63.1355(d) is no longer
needed, and can be deleted. 

Response: See previous response.

Comment: One commenter (2845) suggested that §63.1355(e) can be
deleted, because it is a duplicate of §63.1350(o)(2). The commenter
noted that the reference to only area sources in §63.1355(e) may be a
typographical error, as the recordkeeping requirement applies to both
major and area sources. 

Response: The rule requires both area sources and major sources to keep
clinker production records.

Comment: One commenter (2845) stated that §63.1350(o) no longer
requires fly ash certifications so §63.1355(f) should be deleted.

Response: This is correct.

Develop waste disposal requirements

Comment: One commenter (2816) stated that EPA’s sector-based approach
should include consideration of solid waste impacts and should
reinvigorate its efforts to develop stringent regulations governing the
disposal of wastes generated by portland cement kilns

Response: Comments on regulatory measures for cement kiln dust waste are
beyond the scope of this proceeding.

Comment: One commenter (2895) stated that site specific regulation of
CKD should be in each facility’s Title V permit. The commenter asked
how CKD will be disposed of without creating water and solid waste
problems if it is not recycled into the kiln.

Response: Site specific permitting issues relating to waste management
are beyond the scope of this proceeding.

Comment: One commenter (2844) stated that the technical, economical and
environmental implications of landfilling or blending cement kiln dust
rather than recycling it to the kiln are significantly negative and
should be avoided.

Response: EPA took this issue into account in determining not adopt
beyond the floor standards as part of its consideration of non-air
quality impacts.

More stringent dioxin/furan standards

Comment: One commenter (2828) suggested that EPA set more stringent
standards for dioxins and furans. The commenter stated that the cement
industry represents a high and disproportionate cancer risk. The science
of dioxin/furan monitoring has improved to the point that commercial
dioxin/furan CEMS are available, and analytical laboratories can measure
dioxin/furan at significantly lower levels thanks to technologies such
as gas chromatography-mass spectrometry (GC-MS-MS). The commenter
suggested that EPA should establish stringent dioxin/furan emission
standards, based on a comprehensive study of dioxin/furan emissions from
the current fleet of kilns, and recognizing that the RTO technologies
already in place at some kilns, and likely to be installed at other
kilns in the future are bound to be extremely effective (99%+) at
destroying dioxin/furan. EPA should require dioxin/furan CEMS for
monitoring (or at the very least quarterly monitoring) for any cement
kiln not using an RTO to control emissions. EPA should consider the
beyond-the-floor possibilities for dioxin/furan control that are now
possible with the advent of SCR systems on cement kilns. If EPA
continues recent practice of trying to integrate MACT and NSPS
rulemaking for particular industries, then the establishment of
SCR-driven NOx standards in the Portland Cement NSPS would integrate and
support the establishment of SCR-driven dioxin/furan standards in the
Portland Cement MACT rules.

Response: EPA did not reopen or reconsider the 1999 standard for CDD/CDF
and so is not addressing this comment.



]



h

É

	

	



*

+

⓿+

,

-

2

3

:

;

<

V

W

X

Z

[

\

]

^

_

{

|

}

~

ƒ

„

Š

‹

Œ

¦

§

¨

ª

«

¬



®

¯

Ë

Ì

⓿Ì

Í

Î

Ñ

Ò

ÿ

-



E

F

G

a

b

c

e

f

g

h

i

j

†

‡

⓿‡

ˆ

‰

Ž

¦

§

¨

Â

Ã

Ä

Æ

Ç

È

É

Ê

Ë

ç

è

é

ê

ï

ð

j¸

j¬

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kdw

ÿ£	

ÿ£	

ÿ£	

kdÞ

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd·

ÿ£	

ÿ£	

kdQ

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd*

ÿ£	

kd÷

ÿ£	

ÿ£	

kd‘

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

b

ÿ£	

b

c

h

{

š

ÿ£	

š

›

 

³

Ð

ÿ£	

Ð

Ñ

Ö

é

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

*

ÿ£	

*

+

0

C

f

ÿ£	

f

g

l



Ÿ

ÿ£	

Ÿ

 

¥

¸

Ù

ÿ£	

Ù

Ú

ß

ò

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

kd

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

ÿ£	

r

{

 

 

ç

è

ñ

r

è

Ѐ(摧ᆜ²Ѐ&摧ᛀ^Ѐ(摧献

Ѐ(摧䥒KЀ(摧ᘔ

,

Ž

º

º

º

ᬁ

ñ

ú

û

A

I

ñ

A

ᔏ콨눽ᘀ쵨⬑㔀脈㐀

h

h

:

h

摧猻4Ѐ(摧ᇍ+Ѐ&摧猻4ᨀission level achieved by the best
performing 12 percent of sources for which EPA has emissions
information. The commenter provided additional arguments in support of
this statement. The commenter petitioned EPA to revise its dioxin
standards for cement kilns under the authorities described in Kennecott
Utah Copper Corp. v. Department of Interior, 88 F.3d 1191 (D.C. Cir.
1996).

Response: EPA did not reopen or reconsider the 1999 standard for CDD/CDF
and so is not addressing this comment here. EPA will consider the
commenter’s petition request separately from this rulemaking.

Comment: One commenter (2144) stated that consideration should be given
to more frequent measurement of dioxin and furan (D/F) from cement kilns
or to the possibility of requiring D/F-CEMS. One D/F test per 30-month
period followed by continuous temperature measurement and control is not
sufficient to ensure year round compliance at some kilns. In the short
run, increase the minimum frequency of dioxin/furan testing to once
every 12 months and begin assessment of D/F-CEMS.

One commenter (2867) emphasized the need for more dioxin/furan testing
to a minimum of once every 12 months and recommended that EPA begin
assessment of dioxin/furan-CEMS.

Response: EPA did not reopen or reconsider the 1999 standard for CDD/CDF
and so is not addressing this comment. 

Regulatory approach allows flexibility in fuels

Comment: One commenter (2846) supported EPA in developing a regulatory
approach that maximizes the usefulness of secondary materials. The
commenter recommended using a straightforward qualitative
principle-based approach to defining solid waste that recognizes the
inherent advantages of useful energy recovery and materials handling
practices under §112 regulation. Such an approach can be more
universally applicable to not only existing materials, but also for new
materials and processes not yet developed. Trying to be too prescriptive
could inhibit new technology development and implementation and work
against high level national goals. 

Response: EPA acknowledges the commenter’s support. On the matter of
defining solid waste, that involves issues beyond the scope of this
rulemaking.	

Solid waste as fuel

Comment: One commenter (2816) supported EPA’s inclusion of cement
kilns that combust secondary materials in the proposed rule. One
commenter (2895) stated that EPA should provide a conclusive definition
of solid waste, and in particular, define the HAP emitted when these
solid wastes are used as fuel.

Response: This issue is beyond the scope of this rulemaking. The
definition of solid waste for purposes of section 129 is part of a
different regulatory proceeding. See 75 FR 31844, June 4, 2010. EPA is
also performing a concurrent rulemaking, the Commercial and Industrial
Solid Waste Incinerator rulemaking under section 129 of the CAA that
will regulate kilns that burn non-hazardous solid waste. At the time
that rule becomes final, kilns burning solid waste (as defined in the
to-be-completed rulemaking) will be regulated under to-be-adopted
standards for commercial and industrial solid waste incinerators
whenever they are burning solid waste. 

Require natural gas

Comment: One commenter (2893) suggested that EPA require the use of
natural gas where available as less-polluting alternative to other
fossil fuels or other hazardous fuels.

Response: EPA acknowledges that the use of natural gas would reduce
emissions of mercury and other HAP in comparison to coal. We have
considered this approach but rejected it as an approach that would be
available to all cement plants on a nationwide basis. (71 FR 76520,
December 20, 2006).

Alternatives to cement and concrete

Comment: One commenter (2893) suggested volcanic ash as an alternative
to some conventionally manufactured cement, and alternative materials as
substitutes for the use of concrete in certain applications.

Response: Investigating alternatives to conventionally manufactured
cement, and exploring substitutes for concrete that may be used in
certain applications is beyond the scope of this rulemaking.

Benefits to industry of sector approach

Comment: One commenter (2898) agreed with EPA’s discussion of its
sector-based approach and of the impacts of the rule, stating that it is
a useful reminder for the portland cement manufacturing industry that
complying with the emission standards may help the industry meet other
pollution requirements. The commenter suggested that EPA discuss changes
in production practices or processes that may help facilities comply
with the emission standards. While this discussion should in no way
influence the final standards promulgated as the MACT floor, a fuller
discussion of these alternatives to incorporate consideration of
potential co-benefits will make EPA’s discussion a more useful tool
for the facilities trying to meet the standards and for the communities
living near these facilities as they evaluate the actions of the
facilities in their midst. Moreover, these co-benefits are important
considerations in any evaluation of the technologies for
beyond-the-floor standards. The commenter described aspects of energy
efficiency, adjusting the composition of cement, and cleaner fuel as
examples of actions that could provide these co-benefits.

Response: EPA acknowledges the commenter’s support of EPA’s
sector-based approach.

CO and combustion practices

Comment: One commenter (2928) stated that high CO levels are very often
an indication of poor combustion practices and should also be kept as
low as practical.

Response: A discussion of the EPA’s decision not to regulate CO can be
found in the preamble. 

	At proposal, EPA also presented long term mercury data from an Essroc
kilns showing relative standard deviations similar to the range of the
relative standard deviations of the best performing kilns, based on
30-days of data, again suggesting that EPA had properly assessed
variability of mercury inputs. 74 FR at 21143.

 EPA notes the support of commenter 2898 in this regard.

	See Section 4.1.2.1 of the Credible Evidence Rule Response to Comment
Document, available at <  HYPERLINK
"http://www.epa.gov/ttncaaa1/t1/fr_notices/certcfin.pdf" 
http://www.epa.gov/ttncaaa1/t1/fr_notices/certcfin.pdf >.

Version 2	  PAGE   \* MERGEFORMAT  2 	EPA Internal Review Draft:xxJA2010



	  PAGE   \* MERGEFORMAT  198 	

Preliminary Summary Subject to Verification and Additional Editing

		

