DRAFT TECHNICAL SUPPORT DOCUMENT FOR THE EXPANSION OF THE COMPARABLE
FUELS EXCLUSION

May 2007

DRAFT:  Nov. 17, 2006

NOTE: This document shows new language changes resulting from OMB review
Iin RED), and new language changes made by EPA in BLUE) 

Table of Contents

  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc150685077"  1	Introduction	
 PAGEREF _Toc150685077 \h  1  

  HYPERLINK \l "_Toc150685078"  1.1	Background	  PAGEREF _Toc150685078
\h  1  

  HYPERLINK \l "_Toc150685079"  1.2	Motivation	  PAGEREF _Toc150685079
\h  3  

  HYPERLINK \l "_Toc150685080"  1.3	Summary of Proposed Expansion	 
PAGEREF _Toc150685080 \h  4  

  HYPERLINK \l "_Toc150685081"  2	Constituents of Emission-Comparable
Fuel	  PAGEREF _Toc150685081 \h  6  

  HYPERLINK \l "_Toc150685082"  2.1	Definition of Emission-Comparable
Fuel	  PAGEREF _Toc150685082 \h  6  

  HYPERLINK \l "_Toc150685083"  2.2	Properties of Oxygenates and
Hydrocarbons.	  PAGEREF _Toc150685083 \h  7  

  HYPERLINK \l "_Toc150685084"  2.3	Incinerability Index	  PAGEREF
_Toc150685084 \h  9  

  HYPERLINK \l "_Toc150685085"  2.4	Relative Hazard Categorization
Scheme	  PAGEREF _Toc150685085 \h  12  

  HYPERLINK \l "_Toc150685086"  2.4.1	Ranking Poly-Aromatic Hydrocarbons
(PAHs) as a Category	  PAGEREF _Toc150685086 \h  14  

  HYPERLINK \l "_Toc150685087"  2.4.2	Updating/Collecting
Constituent-Specific Data	  PAGEREF _Toc150685087 \h  15  

  HYPERLINK \l "_Toc150685088"  2.4.3	Subdividing the Ranked
Constituents into Categories	  PAGEREF _Toc150685088 \h  15  

  HYPERLINK \l "_Toc150685089"  2.4.4	More on Category B	  PAGEREF
_Toc150685089 \h  17  

  HYPERLINK \l "_Toc150685090"  2.4.5	Scoring Outcome	  PAGEREF
_Toc150685090 \h  17  

  HYPERLINK \l "_Toc150685091"  2.5	Restrictions on High Hazard
Compounds	  PAGEREF _Toc150685091 \h  22  

  HYPERLINK \l "_Toc150685092"  3	Industrial Boilers	  PAGEREF
_Toc150685092 \h  23  

  HYPERLINK \l "_Toc150685093"  3.1	Fuels	  PAGEREF _Toc150685093 \h  23
 

  HYPERLINK \l "_Toc150685094"  3.1.1	Fuel Oil	  PAGEREF _Toc150685094
\h  23  

  HYPERLINK \l "_Toc150685095"  3.1.2	Natural Gas	  PAGEREF
_Toc150685095 \h  26  

  HYPERLINK \l "_Toc150685096"  3.1.3	Coal	  PAGEREF _Toc150685096 \h 
26  

  HYPERLINK \l "_Toc150685097"  3.1.4	Non Fossil Fuel	  PAGEREF
_Toc150685097 \h  26  

  HYPERLINK \l "_Toc150685098"  3.2	Heat Transfer Configuration	 
PAGEREF _Toc150685098 \h  26  

  HYPERLINK \l "_Toc150685099"  3.2.1	Watertube	  PAGEREF _Toc150685099
\h  27  

  HYPERLINK \l "_Toc150685100"  3.2.2	Firetube	  PAGEREF _Toc150685100
\h  28  

  HYPERLINK \l "_Toc150685101"  3.2.3	Cast Iron & Tubeless Boilers	 
PAGEREF _Toc150685101 \h  29  

  HYPERLINK \l "_Toc150685102"  3.3	Burner Design	  PAGEREF
_Toc150685102 \h  30  

  HYPERLINK \l "_Toc150685103"  3.3.1	Suspension Firing -- Gas, Oil, and
Pulverized Coal	  PAGEREF _Toc150685103 \h  30  

  HYPERLINK \l "_Toc150685104"  3.3.2	Stoker Firing -- Solids	  PAGEREF
_Toc150685104 \h  31  

  HYPERLINK \l "_Toc150685105"  3.3.3	Fluidized Bed Boilers - Solids	 
PAGEREF _Toc150685105 \h  31  

  HYPERLINK \l "_Toc150685106"  3.4	Emissions from Industrial Boilers	 
PAGEREF _Toc150685106 \h  31  

  HYPERLINK \l "_Toc150685107"  4	CO as an Indicator of Incomplete
Combustion	  PAGEREF _Toc150685107 \h  36  

  HYPERLINK \l "_Toc150685108"  4.1	Combustion Failure Modes	  PAGEREF
_Toc150685108 \h  36  

  HYPERLINK \l "_Toc150685109"  4.2	CO vs. DRE	  PAGEREF _Toc150685109
\h  37  

  HYPERLINK \l "_Toc150685110"  4.3	Conclusions	  PAGEREF _Toc150685110
\h  40  

  HYPERLINK \l "_Toc150685111"  5	Organic Emissions & Qualitative Risk
Assessments	  PAGEREF _Toc150685111 \h  41  

  HYPERLINK \l "_Toc150685112"  5.1	Individual Organics	  PAGEREF
_Toc150685112 \h  43  

  HYPERLINK \l "_Toc150685113"  5.1.1	Industrial Boiler Benchmark	 
PAGEREF _Toc150685113 \h  43  

  HYPERLINK \l "_Toc150685114"  5.1.2	Comparative Data from Hazardous
Waste Boilers	  PAGEREF _Toc150685114 \h  45  

  HYPERLINK \l "_Toc150685115"  5.1.3	Exceedences of Industrial Boiler
Benchmark	  PAGEREF _Toc150685115 \h  47  

  HYPERLINK \l "_Toc150685116"  5.1.4	Nondetect Analysis—Limitations
of the Database	  PAGEREF _Toc150685116 \h  49  

  HYPERLINK \l "_Toc150685117"  5.1.5	Emissions of Other Toxic Organics
from Hazardous Waste Boilers	  PAGEREF _Toc150685117 \h  49  

  HYPERLINK \l "_Toc150685118"  5.2	Dioxins and Furan (PCDD/F)	  PAGEREF
_Toc150685118 \h  52  

  HYPERLINK \l "_Toc150685119"  5.2.1	PCDD/F Formation Mechanisms	 
PAGEREF _Toc150685119 \h  52  

  HYPERLINK \l "_Toc150685120"  5.2.2	Factors influencing PCDD/F
Emissions from ECF Fired Boilers	  PAGEREF _Toc150685120 \h  53  

  HYPERLINK \l "_Toc150685121"  5.2.3	PCDD/F Emissions from Boilers
burning Hazardous Waste	  PAGEREF _Toc150685121 \h  54  

  HYPERLINK \l "_Toc150685122"  5.2.4	Comparative Risk Assessment	 
PAGEREF _Toc150685122 \h  56  

  HYPERLINK \l "_Toc150685123"  6	Special Conditions for
Emissions-Comparable Fuel	  PAGEREF _Toc150685123 \h  62  

  HYPERLINK \l "_Toc150685124"  6.1	Storage of Emissions
Comparable-Fuels	  PAGEREF _Toc150685124 \h  62  

  HYPERLINK \l "_Toc150685125"  6.1.1	Above-Ground Storage Tanks	 
PAGEREF _Toc150685125 \h  62  

  HYPERLINK \l "_Toc150685126"  6.1.2	Underground Storage Tanks	 
PAGEREF _Toc150685126 \h  64  

  HYPERLINK \l "_Toc150685127"  6.2	Combustion of Emissions-Comparable
Fuel	  PAGEREF _Toc150685127 \h  64  

  HYPERLINK \l "_Toc150685128"  6.2.1	Combustor Type	  PAGEREF
_Toc150685128 \h  64  

  HYPERLINK \l "_Toc150685129"  6.2.2	Operating Conditions	  PAGEREF
_Toc150685129 \h  65  

  HYPERLINK \l "_Toc150685130"  6.2.3	More on Atomization	  PAGEREF
_Toc150685130 \h  66  

  HYPERLINK \l "_Toc150685131"  6.3	Other Conditions	  PAGEREF
_Toc150685131 \h  68  

  HYPERLINK \l "_Toc150685132"  7	Engineering Costs and Savings	 
PAGEREF _Toc150685132 \h  70  

  HYPERLINK \l "_Toc150685133"  7.1	Regulatory Options	  PAGEREF
_Toc150685133 \h  70  

  HYPERLINK \l "_Toc150685134"  7.2	Analysis of ACC Survey	  PAGEREF
_Toc150685134 \h  71  

  HYPERLINK \l "_Toc150685135"  7.3	Estimate of Qualifying ECF
Quantities	  PAGEREF _Toc150685135 \h  73  

  HYPERLINK \l "_Toc150685136"  7.4	Preliminary Information Used in
Costs/Savings Estimate	  PAGEREF _Toc150685136 \h  73  

  HYPERLINK \l "_Toc150685137"  7.4.1	Value of Fuel Oil and ECF	 
PAGEREF _Toc150685137 \h  73  

  HYPERLINK \l "_Toc150685138"  7.4.2	Hazardous Waste Disposal Costs	 
PAGEREF _Toc150685138 \h  75  

  HYPERLINK \l "_Toc150685139"  7.4.3	Labor Costs	  PAGEREF
_Toc150685139 \h  75  

  HYPERLINK \l "_Toc150685140"  7.4.4	Estimation of the Number of ECF
Shipments	  PAGEREF _Toc150685140 \h  75  

  HYPERLINK \l "_Toc150685141"  7.4.5	Estimation of Tank Size for
Generators and Burners	  PAGEREF _Toc150685141 \h  76  

  HYPERLINK \l "_Toc150685142"  7.4.6	Generator Types and ECF burn
Scenarios	  PAGEREF _Toc150685142 \h  77  

  HYPERLINK \l "_Toc150685143"  7.5	Costs/Savings Estimation	  PAGEREF
_Toc150685143 \h  77  

  HYPERLINK \l "_Toc150685144"  7.5.1	Group 1 Sites:	  PAGEREF
_Toc150685144 \h  78  

  HYPERLINK \l "_Toc150685145"  7.5.2	Group 2 Sites:	  PAGEREF
_Toc150685145 \h  82  

  HYPERLINK \l "_Toc150685146"  7.5.3	Additional Assumptions for Option
C for Offsite Combustion of ECF	  PAGEREF _Toc150685146 \h  83  

  HYPERLINK \l "_Toc150685147"  7.6	Results and Discussion	  PAGEREF
_Toc150685147 \h  83  

  HYPERLINK \l "_Toc150685148"  Appendices	  PAGEREF _Toc150685148 \h 
86  

  

DRE vs. CO/HC- Supplement to 1997 TSD

Analysis of Nondetect Organic Emissions

Outlier Analysis for Hazardous Waste Boiler Organic Emissions

Calculations for DF Comparative Risk Assessment

ECF Tank size and Number of Shipments

Assumptions used for Savings Estimate

Cost Model for Economic Analysis

Generator Savings by Option

LIST OF FIGURES

2-  SEQ Figure \* ARABIC  1 : Overview of the WMPT Scoring Method

3-1: Watertube Boiler

3-2: Firetube Boiler

3-3: CO Emissions from industrial boilers (a) Gas/liquid fired units,
(b) Solid/sludge fired 

        units.

3-4: Methane vs. CO emissions for Industrial Boilers

4-1: POHC Fraction Remaining vs. CO (reproduced from Staley et al,
1989-lines for 

       100% CO, 99.99% DRE added to original plot)

6-1: Commonly-Used Atomization Systems for Liquid Fuels

LIST OF TABLES

2-1: Current Comparable Fuels Specification Limits and Properties of
Oxygenates and Hydrocarbons.

2-2: Incinerability Rank and Thermal Stability Class for Oxygenates and
Hydrocarbons

2-3: Comp Fuels Hydrocarbons that Are PAHs

2-4: WMPT Scoring and Ranking, and Categorization of Comp Fuels
Hydrocarbons and 

       Oxygenates

3-1: Properties of Fuel Oils

3-2: Gasoline Specifications 

3-3: CO Emissions by run (ppmv @7% O2) from Industrial and Commercial
Boilers

3-4: PCDD/F Emissions from Oil-Fired Industrial Boilers

5-1: Hazardous Waste Burning Watertube Steam Boilers with Risk Burn Data
 

5-2: Toxic Organic Emissions Data for Oil-Fired Industrial Boilers

5-3: Toxic Organic Emissions from Hazardous Waste Boilers for Compounds
for Which 

        Emissions Data for Oil-Fired Industrial Boiler Are Available

5-4:Comparison of Organic Emissions from HW Boilers and Oil-Fired
Boilers

5-5: Toxic Organic Emissions from Hazardous Waste Boilers for Compounds
for which 

       Emissions Data for Oil-Fired Industrial Boiler Are Not Available

5-6: PCDD/F Emissions from Hazardous Waste Boilers

5-7: Results of Abbreviated Comparative PCDD/F Risk Assessment

7-1: Classification of Wastestreams from ACC Survey.

7-2: ACC Survey based ECF Waste Quantities

7-3: Burdened Labor Rates for Economic Analysis.

7-4: Summary of ACC Waste Quantities and Savings for CFE Regulatory
Options

Introduction

The United States Environmental Protection Agency (EPA) is proposing to
revise the RCRA comparable fuel exclusion under 40 CFR 261.38 by
expanding the conditional exclusion from RCRA for fuels that are
produced from a hazardous waste but which generate emissions when burned
in an industrial boiler that are comparable to emissions from burning
fuel oil.  The revised rule would establish a new category of excluded
waste fuel called emission-comparable fuel (ECF).

ECF is a hazardous waste-derived fuel product that meets the
specifications for comparable fuel under §261.38(a) except for certain
hydrocarbons and oxygenates.  Generators who claim the exclusion and
burners must comply with various conditions for handling and storage. 
In addition, ECF must be burned in an industrial, watertube, steam
boiler under specified combustion conditions that ensure that emissions
from combustion of ECF are comparable to emissions from combustion of
fuel oil.

Background

Comparable Fuels Exclusion (1998): The comparable fuels exclusion was
promulgated on June 19 1998 and is codified in 40 CFR 261.38 (See 63 FR
33782).  The rule excludes waste fuels from the definition of solid
wastes if they have levels of toxic constituents and physical properties
similar to commercial (benchmark) fuels, in particular gasoline and fuel
oils.  Comparable fuels must meet certain specifications for physical
properties and constituents: 

Physical Properties:  The heating value of the fuel must exceed 5000
Btu/lb and kinematic viscosity (as fired) must be less than 50
centistokes.

Constituent Specification: The fuel must meet specifications for 14
metals and approximately 176 organic compounds listed in table 1 of 40
CFR 261.38. The organics include:

13 Hydrocarbons 

24 Oxygenates

9 Sulfonated organics 

61 Nitrogenated organics 

79 Halogenated organics.

  	

	Additionally there are specifications for total nitrogen, total
halogen, total PCBs and total cyanide.  The development of the
specifications is summarized below and is discussed in detail in the
technical support document that accompanied the final rule. 

	The comparable fuel rule also restricts burning of comparable fuels to
certain types of combustions units (including boilers, industrial
furnaces, incinerators and gas turbines) as defined in 40 CFR
261.38(c)(2)

Technical Amendments (1999): EPA issued technical amendments to the
final rule on September 30, 1999.  (See 64 FR 53027).  These amendments
include numerous changes to the specification tables for metals and
organics.  The changes were primarily due to errors in applying EPA’s
methodology for determining the specification level.  The amendments are
described in USEPA, “Final Technical Support Document of HWC MACT
Standards, Vol IV:  Compliance, July 1999.

Development of the Specification Table: In the 1998 rulemaking, the
individual constituent specifications for organics (and metals) were
developed based on four benchmark liquid fossil fuels (gasoline and
three grades of fuel oil).  Samples were collected from eight geographic
regions across the country.  A total of 27 samples were collected. 
These included:

Gasoline (8 samples), 

Fuel oil #2 (11 samples), 

Fuel oil #4 (1 sample)  and 

Fuel oil #6 (7 samples)

 

These samples were analyzed for all 40 CFR Part 261 Appendix VIII
hazardous constituents that were measurable.  The analyses for organics
included SW-846 method 8240 (Volatile Organics by GC/MS) and method
8270A (GC/MS for semivolatile organics).  

The comparable fuels specifications for organics were based on the
highest level observed in the benchmark fuels.  However, two approaches
were used depending on whether a constituent was detected in any of the
samples.  (Note that only benzene, naphthalene, and toluene were
detected in any of the samples.)

If a constituent were detected in any of the samples, the detect
level(s) was normalized to 10,000 Btu/lb based on the measured heating
value of the sample.  The specification was the higher of.

Highest observed laboratory quantization limit(QL) for nondetect samples


Highest normalized concentration limit for detected samples.

   

2.    If a constituent were not detected in any of the samples, the
highest laboratory quantization limit was used.  However, for VOC
constituents which were not detected in any gasoline samples, the
gasoline QLs were deemed unreliable and not used.  Therefore the
specification for nondetect VOCs were based on the highest QL among the
fuel oil samples. 

Note that for organic nondetects, the specification is nondetect at a
minimum detection limit—the quantitation limit for the benchmark fuels
analyses, except for hydrocarbons and oxygenates where the specification
is the quantitation limit because these compounds could be expected to
be found in benchmark fuels

Motivation

	Using excluded hazardous waste as fuel saves energy by reducing the
amount of hazardous wastes that would otherwise be treated and disposed,
promotes energy production from domestic, renewable sources, and reduces
the use of fossil fuel.  As part of the Resource Conservation Challenge,
EPA is examining the effectiveness of the current comparable fuels
program and is considering whether other industrial wastes could be used
as fuel. 

	EPA contacted the American Chemistry Council (ACC) in early 2003 to
determine to what extent the comparable fuels program was being utilized
and if additional industrial waste streams could be excluded.  ACC
distributed a survey to its members in Spring 2003 and provided the
results to EPA in late 2003.

	

	ACC provided EPA with responses from 52 surveys representing 14
companies. The survey provided information on both qualifying and
nonqualifying fuels on a waste stream by waste stream basis.  For
qualifying fuels (i.e. those that are meeting the comparable fuels
exclusion), information included annual quantity produced, heating value
of the waste, whether or not the exclusion was being used, and the
combustion device that was used or anticipated to be used.  For
nonqualifying fuels, that were candidates for an expanded definition of
comparable fuels, the data provided included:

Annual quantity of waste generated.

Heating value of waste.

Description of waste and primary constituents

Constituents failing CFE specifications and limits achievable.

Current disposal method

Burn location (onsite vs. offsite) and combustion device used

Availability of CO CEMS and ability to meet a CO limit of 100 ppmv.

Additionally, the survey respondents identified barriers to a larger
scale use of the exclusion and recommended changes to the current
specification.  ACC and EPA narrowed respondents’ suggestions down to
3 potential revisions:

Conditional Exclusion of fuels that are off-specification for
hydrocarbons and oxygenates where the conditions ensure emissions are
comparable to benchmark fuels or are otherwise at trace levels.

Change specifications for seven volatile oxygenates to base them on
gasoline quantization limits rather than fuel oil quantization limits

Allow blending to meet the specifications for hydrocarbons and
oxygenates 

Note however that only the first revision is being pursued at this time
because it appears that the other revisions would have limited utility
because few if any additional waste streams would be excluded. 

Summary of Proposed Expansion

EPA is proposing to exclude emission-comparable fuel (ECF) from the
definition of RCRA solid waste.  This exclusion would be conditioned
upon the following requirements being met:

ECF is burned in a boiler meeting the design requirements and under
combustion conditions summarized below.

ECF is stored in compliance with certain storage controls summarized
below.

Boiler Controls: Excluded ECF must be burned in boiler meeting certain
design criteria and under certain burner conditions that are associated
with good combustion practice. These include the requirements that ECF
be burned in a watertube steam boiler (excluding stokers) and specific
combustion controls:

CO emissions, monitored by a CEMS must be less than 100 ppmv @ 7% O2.  

The boiler must fire at least 50% fossil fuel, 

The boiler load must be greater than 40%

The ECF must be fired into the flame

The boiler must provide sufficient atomization for ECF streams

Units with dry air pollution control devices must maintain the gas inlet
temperature below 400°F unless coal is the primary fuel.  

The CO monitor and the gas temperature monitor (if applicable) must be
linked to the automatic ECF feed cutoff system. 

	

Storage Controls- ECF must be stored in above ground or underground
tanks, tank cars, or tank trucks.  See proposed §261.38(c)(1).  Above
ground tank systems would be subject to:  (1) most of the Spill
Prevention and Control Countermeasure (SPCC) requirements of 40 CFR
112.7, 112.8, 112.20, and 112.21 as though ECF met the definition of
oil; (2) engineered secondary containment requirements (i.e., a liner,
vault, or double-walled tank) and leak detection requirements, and (3)
the applicable air emission controls of 40 CFR 63 Subpart EEEE, Part 63,
for organic liquids distribution as though all RCRA oxygenates were
hazardous air pollutants listed in Table 1 to Subpart EEEE.  Underground
tanks storing ECF are subject to 40 CFR Part 280.  

In addition ECF would be subject to all of the management and
implementation conditions applicable to existing comparable fuel. 

	The rest of this document presents supporting information for the
proposed expansion of the comparable fuels rulemaking.  Section 2
discusses ECF constituents with a particular emphasis on the 37
oxygenates and hydrocarbons for which specifications are currently
provided in Table 1 to §261.38.  In addition, a relative hazard ranking
scheme for the 37 oxygenates and hydrocarbons is described in this
section. Section 3 provides information on the industrial boiler designs
and fuels used in these boilers.  Section 4 discusses the use of CO as
an indicator for incomplete combustion.  In Section 5 we describe a
qualitative risk assessment conducted for dioxins and furans and compare
organics data between hazardous waste burning boilers and industrial
boilers. The compliance conditions for ECF generators and burners are
described in Section 6, and, finally, the costs and savings for
generators and burners are presented in Section 7. 

Constituents of Emission-Comparable Fuel

	This section begins by defining emission-comparable fuels.  This is
followed by a discussion of the properties of the 37 oxygenates and
hydrocarbons for which specifications are provided in Table 1 to
§261.38 and a discussion of a relative hazard ranking scheme for these
constituents. 

Definition of Emission-Comparable Fuel 

Emission-comparable fuels must meet the following physical and
constituent specifications.

Physical specifications

Heating value.  The heating value must exceed 5,000 btu/lb (11,500 J/g)
as-generated. 

Viscosity.  The viscosity must not exceed 50 centistokes, as-fired.

Constituent specification

(A) The specification levels and, where nondetect is the specification,
minimum required detection limits as listed in the Table 1 to §261.38. 
Note the specifications would be waived for the compounds listed below.

(B) Waived specifications.  The specification levels in Table 1 would
not apply for the following hydrocarbons and oxygenates under the
special conditions provided for emission-comparable fuel:  

Benzene (CAS No. 71-43-2)

Toluene (CAS No. 108-88-3)

Acetophenone (CAS No. 98-86-2)

Acrolein (CAS No. 107-02-8)

Allyl alcohol (CAS No. 107-18-6)

Bis(2-ethylhexyl)phthalate [Di-2-ethylhexyl phthalate] (CAS No.117-81-7)

Butyl benzyl phthalate (CAS No. 85-68-7)

o-Cresol [2-Methyl phenol] (CAS No. 95-48-7)

m-Cresol [3-Methyl phenol] (CAS No. 108-39-4)

p-Cresol [4-Methyl phenol] (CAS No.106-44-5)

Di-n-butyl phthalate (CAS No. 84-74-2)

Diethyl phthalate (CAS No. 84-66-2)

2,4-Dimethylphenol (CAS No. 105-67-9)

Dimethyl phthalate (CAS No. 131-11-3)

Di-n-octyl phthalate (CAS No. 117-84-0)

Endothall (CAS No. 145-73-3)

Ethyl methacrylate (CAS No. 97-63-2)

2-Ethoxyethanol [Ethylene glycol monoethyl ether] (CAS No. 110-80-5)

Isobutyl alcohol (CAS No. 78-83-1)

Isosafrole (CAS No. 120-58-1)

Methyl ethyl ketone [2-Butanone] (CAS No. 78-93-3)

Methyl methacrylate (CAS No. 80-62-6)

1,4-Naphthoquinone (CAS No. 130-15-4)

Phenol (CAS No. 108-95-2)

Propargyl alcohol [2-Propyn-1-ol] (CAS No. 107-19-7)

Safrole (CAS No. 94-59-7); or

	The remaining parts of this section discuss the properties of a
oxygenates and hydrocarbons listed in Table 1 to §261.38 and the
rationale for not waiving the specification for particular hydrocarbons.

Properties of Oxygenates and Hydrocarbons.

Hydrocarbons occur naturally in virgin fossil fuels.  Oxygenates are a
class of compounds that are added to gasoline to improve combustion and
reduce carbon monoxide emissions.  

	

The 37 hydrocarbons and oxygenates that are part of the current
comparable fuel specification are shown in Table 2-1.  For each compound
the table displays:

The chemical abstract service (CAS) number, 

The composition of constituent found in benchmark fuels, and

The current concentration limit for comparable fuels.

	

All the hydrocarbons in the table are aromatics.  Ten of the 13
hydrocarbon compounds (benzene, toluene, and naphthalene are the
exceptions) belong to a group of compounds referred to as Polycyclic
Aromatic Hydrocarbons (PAH).  The oxygenates belong to various classes
of organic compounds including alcohols, aldehydes, ketones, and
phthalates. 

Table 2-1 also shows physical properties such as molecular weight,
boiling point, vapor pressure, and heating value for each of the
compounds.  Each of the physical properties is briefly discussed below:

	

Molecular Weight:  The molecular weights of the 37 hydrocarbons and
oxygenates range from 56 g/gmol to 390 g/gmol for Di-n octyl phthalate.

Table 2-1: Current Comparable Fuels Specification Limits and Properties
of Oxygenates and Hydrocarbons.

Volatility: The table shows the classification of the constituents
between VOC and SVOC.  As discussed below VOCs have lower boiling points
and higher vapor pressures when compared to SVOCs.  Two hydrocarbons
(benzene & toluene) as well as seven oxygenates (acrolein, allyl
alcohol, ethyl methacrylate, methyl methacrylate, isobutyl alcohol,
methyl ethyl ketone, and propargyl alcohol) are classified as VOCs.

Boiling Points:  The boiling points of the 37 compounds range from
53-117°C for VOCs and 136 – 536°C for SVOC.  Acrolein has the lowest
boiling point while indeno pyrene has the highest.

Vapor Pressures:  Values for vapor pressure were extracted from various
references as shown in the table.  The values range from 1.3 to 37 kPa
for VOC with acrolein having the highest vapor pressure.  For SVOC the
range is from 0.62 kPa to as low as 8.9 x 10-13 kPa.

Heating Values: Table 2-1 also shows the higher heating value (HHV) or
gross heat of combustion for the 37 constituents.  These values were
tabulated from several references as shown in the table.  For seven
compounds tabulated data were not readily available and heating values
were estimated using the Dulong formula which estimates heating value
from composition of the substance.

 

	With the exception of one compound, the heating values of the
hydrocarbons and oxygenates fall in the range 10,000 to18,500 Btu/lb.

Incinerability Index

	The incinerability index, a metric developed by researchers at
University of Dayton Research Institute (UDRI) and EPA’s Risk
Reduction Engineering Laboratory (RREL), categorizes all 40 CFR 261, App
VIII toxic organic compounds into seven thermal stability classes. 
Additionally each compound is ranked from 1 to 320 based on the
temperature required to achieve 99% destruction in two seconds under low
oxygen conditions.  The index was developed to aid selection of
principle organic hazardous constituents (POHC) for trial burn testing
to demonstrate compliance with the RCRA destruction and removal
efficiency (DRE) standard and was derived using a combination of
theoretical and experimental data.

	

	The destruction of an organic compound is a function of temperature,
residence time at the temperature, and compound specific properties.  A
simplified approach is to treat the destruction of a POHC using first
order kinetic considerations.  The rate of destruction of a POHC is
given by,

            

  ------- (2.3.1)

where, C is the concentration of the POHC, t is time, and k is the
reaction rate constant that is a function of temperature and is given by
the Arrhenius equation

 ---- (2.3.2)

where:

A= frequency factor (units of s-1), 

E= Activation energy (units of cal/gmol)

R = universal gas constant (1.987 cal/gmol.K) and

T = Temperature (K)

	Here A and E are a function of the POHC and also depends on if POHC
destruction occurs under fuel-lean or fuel rich conditions.

	Substituting for k in equation (2.3.1) and integrating one gets
concentration at time t

    ------------------------ (2.3.3)

Where Co is the inlet POHC concentration. 

Re-arranging equation (2.3.3) in terms of temperature and using
destruction removal efficiency (DRE) = (Co-C)/Co = 1-C/Co, we get,

  ---------------------- (2.3.4)

	Therefore, if A and E are known, and if elapsed time and DRE are
specified, the temperature required to achieve a given DRE within a
given time can be calculated.

Table 2-2 lists the incinerability ranking, thermal stability class, and
T99(2) values for oxygenates and hydrocarbons.  One can see from the
table that 10 of 13 hydrocarbons are in Class 1 in terms of thermal
stability which means they are the most difficult to destroy.  In fact,
benzene and napthalene have the third and fourth highest T99(2) values
respectively.  On the other hand, oxygenates are in lower thermal
stability classes and are in general easier to destroy in comparison to
hydrocarbons.

	

Table 2-2: Incinerability Rank and Thermal Stability Class for
Oxygenates and Hydrocarbons

Relative Hazard Categorization Scheme

We assessed the relative hazard of the 37 constituents in Table 1 to
§261.38 using a stepped approach, including:  (1)  ranking 37 chemicals
using the Waste Minimization Prioritization Tool (WMPT); (2) subdividing
ranked chemicals into categories; and (3) assigning “worst case”
subscores to constituents with insufficient data to generate WMPT
scores.

Step 1:  Ranking Chemicals Using MWPT

We used the Waste Minimization Prioritization Tool (WMPT) (U.S. EPA,
2000) to rank the 37 hydrocarbons and oxygenates according to their
relative hazard.  The WMPT was attractive and appropriate because it had
been so thoroughly reviewed, and used in previous RCRA decision-making
(see the side-bar regarding the PBT Chemical list development).  The
WMPT is a joint product of EPA’s Office of Solid Waste (OSW) and
EPA’s Office of Pollution Prevention and Toxics (OPPT).  It provides a
screening-level assessment of potential chronic (i.e., long-term) risks
to human health and the environment.  The relative chemical rankings
derived from the WMPT can complement other risk or cost information in
the decision-making process. 

The purpose of the WMPT scoring method is to develop chemical-specific
scores for a screening-level risk-based ranking of chemicals.  The
scoring method was designed to generate an overall chemical score that
reflects the potential of a chemical to pose risk to either human health
or ecological systems.  A measure of human health concern is derived,
consistent with the risk assessment paradigm, by jointly assessing the
chemical’s human toxicity and potential for exposure.  Similarly, a
measure of the ecological concern is derived by jointly assessing the
chemical’s ecological toxicity and potential for exposure.

 	The WMPT’s scoring method is modeled after the general risk
calculation equation used by U.S. EPA and others, where the risk of a
chemical is derived by combining estimates of the toxicity of the
chemical with estimates of the actual or potential exposure to the
chemical.  The WMPT uses a small number of relatively simple measures to
represent the toxicity (e.g. oral Cancer Slope Factor) and exposure
potential (e.g. Bioaccumulation Factor) of each chemical, consistent
with a screening-level approach and with other systems of this type. 
Figure 2-1 graphically represents examples of the various measures, and
how they culminate in WMPT scores.  A detailed list of the measures is
available in USEPA (2000).

Figure 2-1:  Overview of the WMPT Scoring Method



The Human Health Concern score is derived by adding two factor-level
scores, one reflecting the toxicity of the chemical to humans and the
other the potential for exposure to the chemical.  The Human Toxicity
score is derived by taking the higher of two “subfactor” scores: 
(1) Cancer Effects, and (2) Noncancer Effects.  The Human Exposure
Potential score is derived as the sum of two subfactor scores:  (1)
Persistence and (2) Bioaccumulation Potential.  Similar to the Human
Health Concern score, the Ecological Concern score is derived by adding
two factor-level scores, one reflecting the toxicity of the chemical to
aquatic ecosystems and the other the potential for exposure to the
chemical.  

Scores are first generated at the subfactor level.  A score for a given
subfactor is derived by evaluating certain chemical-specific parameters
that represent that subfactor.  Each chemical-specific parameter is
assigned a score depending on the level of concern associated with the
chemical’s tendency to demonstrate the characteristic (see the
side-bar regarding the fenceline approach to scoring).  These scores are
then “aggregated upward” to generate an overall chemical score.  

The parameters used to score the various subfactors are briefly
described below: 

Persistence – measured or estimated half-life in water, soil, or
sediment;

Bioaccumulation Potential – measured or estimated Bioaccumulation
Factors (BAFs) or Bioconcentration Factors (BCFs);

Human Toxicity – Cancer Slope Factors and non-cancer Reference Doses; 


Ecological Toxicity – a number of data elements representing measured
or estimated chronic and acute aquatic toxicity extracted from a number
of U.S. EPA sources (e.g., Final Chronic Values, measured and estimated
aquatic chronic values, EC50s, LC50s, and aquatic toxicity reportable
quantities).

U.S. EPA (2000) provides more detailed descriptions of the types of data
used, as well as the hierarchies of data sources.

Ranking Poly-Aromatic Hydrocarbons (PAHs) as a Category

For the Priority Chemical List PBT List effort, the WMPT collected
chemical-specific data element values, and assigned sub-factor and
factor scores for all constituents.  However, to be consistent with the
TRI, which combines individual PAH compounds into a single PAH category,
the WMPT also created a PAH category, and assigned a single final
ranking score to the entire category, rather than rank each PAH
individually.  The PAH category was assigned a WMPT score based on the
highest scoring PAH in the category.  To be consistent with the WMPT
methodology, we also collected chemical-specific data and scored each
constituent separately, but ranked the category as a whole rather than
ranking individual PAHs.  To make sure we used the most inclusive and
up-to-date definition of the PAH category for this effort, we used the
U.S. EPA (2001) list of PAHs.  The comparable fuel hydrocarbons which
qualify as PAHs according to USEPA (2001) include the following:

TABLE 2-3:  Comp Fuels Hydrocarbons that Are PAHs

Chemical	CAS#

3-Methylcholanthrene	56-49-5

7,12-Dimethylbenz(a)anthracene	57-97-6

Benzo(a)anthracene	56-55-3

Benzo(a)pyrene	50-32-8

Benzo(b)fluoranthene	205-99-2

Benzo(k)fluoranthene	207-08-9

Chrysene	218-01-9

Dibenzo(a,h)anthracene	53-70-3

Fluoranthene	206-44-0

Indeno(1,2,3-cd)pyrene	193-39-5



Updating/Collecting Constituent-Specific Data

Some, but not all of the 37 hydrocarbons and oxygenates were assessed
for the Priority Chemical List PBT List.  For those chemicals that had
been ranked for the PBT List, we updated the chemical-specific values
where appropriate (e.g. if the original source had been updated since
the PBT list was generated) and re-evaluated each chemical, to see if
their WMPT scores changed with more up-to-date data.  

For those hydrocarbons and oxygenates that hadn’t been assessed for
the PBT List, we collected data from the same hierarchy of sources as
the WMPT, and used the WMPT methodology to score each chemical.  Table
2-4 summarizes sub-factor and factor scores. A detailed list of
parameter values (including sources) which led to the scores is
available in Appendix I.

Due to time and budget restraints, we decided to collected data to
develop only the Human Concern Score as a first round.  We would collect
ecological toxicity data only for constituents which a high Toxicity
Score might elevate them to a different Category. 

Step 2:  Subdividing the Ranked Constituents into Categories

We applied a series a filter criteria to categorize the constituents
according to their relative hazard.

Phase 1 (Chemicals with a WMPT score of 8 or 9):  As the first step in
its Priority Chemical List methodology, the Agency identified for
further study only those chemicals which received a WMPT Overall
Chemical Score of 8 or 9.  Consistent with the approach the Agency used
to identify chemicals for the Priority Chemical List, we consider
chemicals that have a WMPT score of 8 or 9 to be high hazard chemicals. 


Any chemicals satisfying this criterion were assigned to Category A. 
All remaining chemicals continued to Phase 2.

Phase 2a (Chemicals with a WMPT Human Toxicity score based on inhalation
toxicity):  Because the current effort relates to combustor stack
emissions and, thus, inhalation as the initial exposure pathway, we
considered it important to consider if a chemical’s WMPT toxicity
score was driven by inhalation toxicity.  

All chemicals satisfying this criterion continued to Phase 2b.  All
other chemicals were assigned to Category C.

Phase 2b (Chemicals USEPA/IRIS has designated a “known human
carcinogen” OR that have a High WMPT Human Tox score):  There are some
chemicals with sufficiently robust toxicologic databases that the
EPA/IRIS program not only generates a carcinogenic slope factor, but
also designates them “known human carcinogens.”  The Agency
considers this pertinent information to consider when assessing the
potential hazard of a substance.  Likewise, the magnitude of non-cancer
toxicity is also pertinent to consider when assessing the potential
hazard of a substance.  We consider chemicals satisfying a Phase 2b
criterion to be moderate hazard chemicals.

All chemicals found to be either a known human carcinogen or to have a
high WMPT toxicity subscore were assigned to Category B.  All other
chemicals were assigned to Category C. 

We subdivided the 37 hydrocarbons and oxygenates into categories.  The
categories are defined as follows:

Category A:  WMPT score 8 or 9; (The first step in the “classic” PBT
methodology).  To qualify, a constituent would need to score High in two
of the three Factors (i.e. Persistence, Bioaccumulation, and Toxicity),
and at least Medium in the third.

Category B:  WMPT score 6 or 7.  A constituent would need to score High
for at least one Factor, score Low for bioaccumulation, and have a Human
Toxicity score based on the inhalation pathway.  Please see below for
more on this category;

Category C: WMPT score 5 – 7.  A constituent would need to score Low
for one Factor, but not satisfy the criteria for Category B;

Category D:  WMPT score 3 or 4.  A constituent would need to score Low
for at least two Factors (in effect, all remaining constituents with
sufficient data to evaluate, not qualifying for Categories A - C);

Unknown:  Insufficient data to evaluate all three Factors;

Step 3:  Chemicals with Insufficient Data to Generate WMPT Scores

There were five chemicals (1,4-napthoquinone, isosafrole, propargyl
alcohol, safrole, di methyl phthalate) with insufficient data to develop
a WMPT score.  The Agency assigned these substances a worst-case Human
Toxicity subscore of 3, to determine if any chemical might qualify for
either Category A or B.

Outcome

Table 2-4 lists the scores and final WMPT ranking of the 37 chemicals. 
Some sub-factor scores for re-evaluated chemicals did change (e.g.
Phenol’s Human Toxicity score changed from 1 to 2), but these changes
never led to a change in a final WMPT score.

To make reading the Table more intuitive, we sorted the list as follows:

Whether the constituent is a PAH according to the U.S. EPA (2001)
definition;

Category;

By Concern score (descending); and

Alphabetically.

Because the entire PAH category was assigned the final score of the
highest scored constituent (see “Ranking PAHs as a Category” above),
all PAHs qualified for Category A, regardless of their
constituent-specific Higher Concern score.  Naphthalene was the only
other constituent to qualify for Category A.

Benzene, acrolein, and phenol have WMPT toxicity scores based on
inhalation toxicity, satisfying the Phase 2a criterion.  USEPA/IRIS
classifies benzene as a “known human carcinogen” (USEPA 2006),
satisfying the Phase 2b criterion.  Acrolein’s very high inhalation
toxicity qualified it for a High human toxicity score in WMPT,
satisfying the Phase 2b criterion.  Phenol, however, is not classified a
known human carcinogen, and has only a Medium WMPT Toxicity Factor. 
Therefore, phenol did not satisfy either Phase 2b criterion, and was
assigned to Category C.

The remaining constituents were assigned to Category C.

As detailed in Table 2-4, the distribution of Persistence Factor and
Bioaccumulation Factor scores for the remaining constituents was such
that a high Ecological Toxicity score wouldn’t have elevated any of
the constituents to a different Category.  We therefore didn’t collect
any additional ecological toxicity data.

Also, the distribution of Persistence Factor and Bioaccumulation Factor
scores for the remaining constituents was such that none of the five
chemicals without WMPT scores (see Section 1.1.3 above) would have
qualified for either Category A or B.

More on Category B

Because of the structure of the WMPT ranking system, some constituents
might not qualify for Category A, yet demonstrate characteristics
suggesting that they are of greater concern than other constituents. 
Specifically, some constituents might have both a bioaccumulation score
of Low and inhalation as their driving exposure pathway. 
Bioaccumulation is only relevant for indirect exposure pathways.  With
the particular combination of bioaccumulation score and dominant
exposure pathway mentioned above, the dominance of the inhalation
pathway would tend to mitigate the relevance of the Low bioaccumulation
score which disqualified the constituent for Category A.  We decided to
assign these constituents their own category – Category B – to
highlight them for further study.

Scoring Outcome

Table 2-4 lists the scores and final ranking of the 37 constituents. 
Some sub-factor scores for re-evaluated chemicals did change (e.g.
Phenol’s Human Toxicity score changed from 1 to 2), but these changes
never led to a change in a final WMPT score.

To make reading the Table more intuitive, we sorted the list as follows:

Whether the constituent is a PAH according to the U.S. EPA (2001)
definition;

Category;

By Concern score (descending); and

Alphabetically.

Because the entire PAH category was assigned the final score of the
highest scored constituent (see “Ranking PAHs as a Category” above),
all PAHs qualified for Category A, regardless of their
constituent-specific Higher Concern score.  Naphthalene was the only
other constituent to qualify for Category A.

Benzene and acrolein both satisfy the criteria for Category B.. IRIS
classifies benzene as a “known human carcinogen” (USEPA 2006). 
Benzene is also difficult to destroy via combustion, very persistent in
the environment, and a PAH precursor (USEPA 2005).  Acrolein’s very
high inhalation toxicity further brings into question the relevance of
its low bioaccumulation score.  This information suggests that both
benzene and acrolein warrant special consideration in future
decision-making.

As detailed in Table 2-4, the distribution of Persistence Factor and
Bioaccumulation Factor scores for the remaining constituents was such
that a high Ecological Toxicity score wouldn’t have elevated any of
the constituents to a different Category.  We therefore didn’t collect
any additional ecological toxicity data.

There were five constituents for which no human toxicity data was
available (see Table 2-4).  This meant there was insufficient data to
assign a Human Toxicity score, and therefore no Human Concern score was
possible.  

REFERENCES for Section 2.4:

U.S.EPA 2000.  Waste Minimization Prioritization Tool Background
Document for the Tier III PBT Chemical List.

U.S. EPA 2001.  Emergency Planning and Community Right-to-Know Act
(EPCRA) - SECTION 313: Guidance for Reporting Toxic Chemicals:
Polycyclic Aromatic Compounds Category.  Office of Environmental
Information.  EPA 260-B-01-03. August.

U.S. EPA 2005.  Human Health Risk assessment Protocol for Hazardous
Waste Incineration Facilities.  Office of Solid Waste. EPA-R-05-006. 
September.

U.S. EPA 2006.  Integrated Risk Information System (IRIS):  Benzene.  
HYPERLINK "http://www.epa.gov/iris/"  http://www.epa.gov/iris/  .
Accessed September 2006.

TABLE 2-4:  WMPT Scoring and Ranking, and Categorization of Comparable
Fuels Hydrocarbons and Oxygenates

TABLE 2-4:  WMPT Scoring and Ranking, and Categorization of Comp Fuels
Hydrocarbons and Oxygenates

Chemical Name	CASRN	EPCRA 313 PAH?	Persistence Score	Bioaccumulation
Score	Human Toxicity Score	Ecological Toxicity Score	Human Concern Score
Ecological Concern Score	Higher Concern Score	CATEGORY

 	 	 	(P)	(B)	(HT)	(ET)	(P+B+HT)	(P+B+ET)	 	 

3-Methylcholanthrene	56495	X	3	3	3	3	9	9	9	A

7,12-Dimethylbenz(a)anthracene	57976	X	3	3	3	3	9	9	9	A

Benzo(b)fluoranthene	205992	X	3	3	2	3	8	9	9	A

Benzo(k)fluoranthene	207089	X	3	3	2	3	8	9	9	A

Dibenzo(a,h)anthracene	53703	X	3	3	3	3	9	9	9	A

Indeno(1,2,3-cd)pyrene	193395	X	3	3	1	3	7	9	9	A

Benzo(a)pyrene	50328	X	3	2	3	3	8	8	8	A

Fluoranthene	206440	X	3	2	2	3	7	8	8	A

Benzo(a)anthracene	56553	X	3	1	2	3	6	7	7	A

Chrysene	218019	X	3	1

3

7	7	A

Naphthalene	91203	 	3	2	2	3	7	8	8	A

Benzene	71432

3	1	2	3	6	7	7	B

Acrolein	107028	 	2	1	3	3	6	6	6	B

Bis(2-ethyhexyl)phthalate	117817

3	2	2	1	7	6	7	C

Dibutyl phthalate	84742

1	2	1	3	4	6	6	C

Diethyl phthalate	84662

3	1	1	2	5	6	6	C

2-Ethoxy ethanol	110-80-5	 	2	1	2	 	5	 	5	C

Allyl alcohol	107186

1	1	2	3	4	5	5	C

Butyl benzyl phthalate	85687

1	1	1	3	3	5	5	C

Di-n-octyl phthalate	117840

1	2	2	0	5	3	5	C

Endothall	145-73-3	 	2	1	2	 	5	 	5	C

Toluene	108-88-3	 	3	1	1	 	5	 	5	C

2,4-Dimethylphenol	105679

1	1	2	2	4	4	4	D

Acetophenone	98-86-2	 	2	1	1	 	4	 	4	D

Cresol, o-	95487

1	1	2	2	4	4	4	D

Ethyl Methacrylate	97-63-2	 	2	1	1	 	4	 	4	D

Isobutyl alcohol	78831

1	1	1	2	3	4	4	D

m-Cresol(3-methyl phenol)	108-39-4	 	1	1	2	 	4	 	4	D

Methyl methacrylate	80-62-6	 	2	1	1	 	4	 	4	D

p-Cresol(4-methyl phenol)	106-44-5	 	1	1	2	 	4	 	4	D

Phenol	108952

1	1	2	2	4	4	4	D

Methyl ethyl ketone	78-93-3	 	1	1	1	 	3	 	3	D

1,4 Naphthoquinone	130-15-4	 	3	1	ND	 	ID (7?)	 	 	Unknown

Isosafrole	120-58-1	 	2	1	ND	 	ID (6?)	 	 	Unknown

Propargyl Alcohol	107-19-7	 	2	1	ND	 	ID (6?)	 	 	Unknown

Safrole	94-59-7	 	2	1	ND	 	ID (6?)	 	 	Unknown

Di methyl phthalate	131-11-3	 	1	1	ND	 	ID (5?)	 	 	Unknown













SORT ORDER:  PAH?, Category, Higher concern score, alpha







	CATEGORY DEFINITIONS:











        A = WMPT score 8 or 9 = first step in classic PBT methodology







	        B = WMPT score 6 or 7, low (B), (HT) based on inhalation
pathway







	        C = WMPT score 5 - 7 (i.e. low score for at least one
criterion), except those that qualify for Category B





        D = WMPT score 3 or 4 (i.e. low score for at least 2 criteria)







	        Unknown = Insufficient data to categorize





















	ND = No Data











ID = Insufficient Data to classify (maximum possible)









gray = not in original PBT List  effort









	



Restrictions on High Hazard Compounds

Based on the hazard category rankings discussed in the previous sections
the following restrictions are proposed.

The current comparable fuels specifications are retained for hazard
category A compounds.  These include naphthalene and ten other
hydrocarbon compounds that are classified as PAHs.  Note that this
category includes all comparable fuels hydrocarbons (with the exception
of benzene) that are in class 1 of the thermal stability ranking.

For hazard category B compounds (benzene & acrolein) there are
additional firing rate restrictions (beyond what is discussed in section
1.3).  If the ECF contains more than 2% by mass of benzene or 2% by mass
of acrolein, the firing rate of ECF will be restricted to 25% on a
volume or heat input basis whichever results in a lower volume of ECF. 
The 2% cutoff level was selected so the limit on benzene and acrolein
would be no more stringent than their corresponding limits for current
comparable fuels.

          

 

Industrial Boilers

	This section summarizes industrial boiler designs and fuels used in
industrial, commercial, and institutional boilers.  Although various
types of boiler designs and fuels are discussed in this section it
should be noted that ECF combustion would only be allowed in industrial
(or utility), nonstoker, watertube steam boilers that fire fossil fuel
as the primary fuel.

	Boilers burning ECF must meet the RCRA definition of boilers codified
in 40 CFR 260.10.  A boiler is defined as an enclosed device using
controlled flame combustion with the following characteristics:

•	The combustion chamber and energy recovery section must be of
integral design.

•	Thermal recovery efficiency must be greater than 60%, determined as
the ratio of the recovered energy to the thermal value of the fuel.

•	Greater than 75% of the recovered energy must be exported and used
(i.e., this does not include internal boiler uses such as preheating
combustion air or driving combustion air fan or feedrate pumps).

	In September 2004, EPA promulgated national emissions standards for
hazardous air pollutants (NESHAP) for industrial, commercial and
institutional (ICI) boilers.  The inventory database for this rule
contains more 58,000 ICI boilers and process heaters. Industrial boilers
are used in all major industrial sectors but primarily by the paper
products, chemical, food and petroleum industries.  Industrial boilers
typically range in size from 10 to 250 MMBtu/hr although larger units do
exist.  Meanwhile commercial/institutional boilers supply steam or hot
water in hospitals, schools, hotels, restaurants etc. and are usually
smaller than 10 MMBtu/hr.  

Fuels

	ICI boilers use a myriad of solid, liquid and gaseous fuel types from
fossil fuels (such as natural gas, fuel oil, coal) to non-fossil fuels
such as wood, tires, industrial sludge, baggasse, and municipal waste. 
Some of these fuels are discussed below.

Fuel Oil 

	Approximately 10% of all ICI boilers burn fuel oil.  The term fuel oil
can refer to any product derived from petroleum that has volatility
lower than that of gasoline.  The ASTM D396-2(a) specification divides
fuel oil into several classes, from fuel oil No.1 to fuel oil No. 6,
based on boiling range, composition, and other physical properties.  As
the fuel oil number increases, the number of carbon atoms increases from
about 9 to about 70, the boiling range and viscosity also increase. 
Additionally, the value of fuel oil decreases with from No.1 to No.6. 
Fuel oils are generally classified as either distillate or residual
based on whether or not they are vaporized in normal refining
operations. Usually, fuel oils No. 1 and 2 are distillate fuels, No. 5
and No. 6 fuel oils are residual fuels, and No.4 is a blend of
distillate and residual fuels.  

	Table 3-1 shows typical properties of various classes of fuel oil such
as flash point, specific gravity, API gravity, heating value, and
kinematic viscosity.  The flash point is the temperature to which the
liquid must be heated to produce vapors that flash but do not burn
continuously.  

 

Table 3-1: Properties of Fuel Oils

API gravity is calculated by:

 

	Fuel oil heating values range from 18,000 to 20,000 Btu/lb or 130,000
to 150,000 Btu/gal. As the table shows, the residual fuels are extremely
viscous and must be heated in order to be transported and atomized in a
burner.  

	Note that diesel has similar specifications to fuel oil No.2 and
contains hydrocarbons with boiling points in the range 175 – 380 °C. 


Gasoline: Although gasoline is not used in industrial boilers, gasoline
was used as a benchmark fuel to establish the comparable fuel
specification because it provides a reasonable upper boundary for
volatile organics which are fuel-worthy constituents.  Table 3-2
outlines the ASTM D4814 specifications for gasoline with regards to
vapor pressure (@38 °C) and distillation temperature for six volatility
classes.,  The table shows a distillation profile for each volatility
class.  (For example for class AA at least 10% of the fuel must
evaporate at 70°C, 90% at 190 °C and 100% at 225 °C). 

Table 3-2: Gasoline Specifications

These specifications generally indicate that gasoline constituents must
have a boiling point no higher than 225 °C and not much lower than
50°C and vapor pressures no higher than 776 mm Hg.

	The Agency for Toxic Substances and Disease Registry (ATSDR) has
developed a toxilogical profile for gasoline.  The ASTDR report says
that gasoline contains more than 150 chemicals and the actual content of
specific organics is a function of the source of crude, the manufacturer
and the time of year.  The primary components of gasoline are aliphatic
hydrocarbons (in particular straight chain and branched alkanes) and
aromatic hydrocarbons from carbon numbers C4 to C12.

Oxygenates: Oxygenates have been added to gasoline starting with ethanol
in the 1970’s. The introduction of CAA requirements to use
reformulated gasoline (RFG) in ozone non-attainment areas starting in
the 1990’s increased the use of oxygenates in gasoline.  RFG is
blended gasoline that is cleaner burning than conventional gasoline. 
One requirement for RFG is oxygen content.  A survey conducted by an
association of blenders, refiners, and importers of gasoline, and shown
on the EPA office of transportation and air quality’s (OTAQ) website,
indicates that in 2003 RFG had an average oxygen content of 2.15-2.3%. 
Alcohols and ethers are typically used as oxygenates with ethanol,
methyl tertiary butyl ether (MTBE), and tertiary-amyl methyl ether
(TAME) being the most widely used oxygenates in the US.  However, in
Europe the EN 228 specifications permits several oxygenates at various
concentrations including isobutyl alcohol at 10%. 

	Additionally, oxygenates have also been added to diesel fuel to improve
combustion and lower emissions.  E diesel, a blend of diesel with up to
15% ethanol is an experimental fuel being demonstrated in trucks, buses,
and farm machinery.  Many studies have been conducted on use of
oxygenates in diesel including one by Southwest Research Institute that
evaluated several alcohols, ethers and esters as oxygenates for diesel.

Natural Gas

	Natural gas is widely used in industrial boilers accounting for about
46,000 of the ICI boilers and process heaters.  Natural gas is a
desirable fuel for boilers due to its low content of ash and ease of
handling, and ability to combust with easily.  However, DOE’s energy
information administration notes that the price of natural gas has more
than doubled in the last decade.  The primary constituent of natural gas
is methane.  Other paraffinic hydrocarbons such as ethane, propane, and
butane as well as other gases such as nitrogen are also present.  The
heating value of natural gas, like its composition, varies by region but
usually is in the range from 900 to1000 btu per standard cubic feet.

Coal

	Coal is also widely used in industrial boilers partly due to its lower
price when compared to other fossil fuels.  About 2500 ICI units burn
coal.  The ASTM classification for coal has four primary categories
which rank coals based on volatile matter and fixed carbon content as
well as heating value.  The four categories (from low to high rank) are:
 lignite, submituminous, bituminous, and anthracite.  The higher rank
coals have higher carbon content and are “older” i.e., they have
undergone a comparatively longer coalification process.  Anthracite coal
has a carbon content of over 90% and is typically not used in industrial
boilers.  Most coal-fired industrial boilers burn either subbituminous
coal (heating values ranging from 8,300 to 11,500 Btu/lb) or bituminous
coal that (with heating values ranging from 10,500 to 14,000 Btu/lb). 
Lignite, a low rank coal that typically has a heating value below 8300
Btu/lb is not commonly burned in industrial boilers.  All coal types in
general have a higher content of ash and metals than fuel oil.

Non Fossil Fuel

	Industrial boilers also burn a range of non-fossil fuels such as wood,
tires, industrial sludge, baggasse, and municipal waste.  

Heat Transfer Configuration	

	Boilers can be subclassified into four different types based on their
heat transfer configuration set-up:  (a) watertube; (b) firetube; (c)
cast iron; and (d) tubeless.  The choice of design depends on factors
including the desired steam quality, thermal efficiency, size,
economics, fuel type, and responsiveness.  Watertube industrial boilers
have a large size range while the other configurations are typically
smaller than 10-30 MMBtu/hr.

Watertube

F); (b) achieve high thermal efficiency; (c) respond rapidly to
changes in steam demand; and (d) potentially burn a variety of fuel
types including coal, oil, gas, and other fuel types such as wood and
municipal wastes.  ICI watertube boilers range from 0.4 to 1500
MMBtu/hr. 

	

	A typical industrial watertube boiler is shown in Figure 3-1.  A
watertube boiler contains furnace and convective sections.  Fuel is
burned in the lower furnace section. Depending on the burner and fuel
feed design set-up, gas, liquid, and solid fuels can be burned.  The
furnace section is lined with small diameter tubes which carry flowing
water.  Radiative heat from the fuel combustion flame heats the water in
the tubes, creating steam.

	The combustion flue gases are routed from the furnace into the
“convective” section of the boiler.  This section typically contains
a superheater, reheater, economizer, and air preheater heat exchangers. 
The superheaters and reheaters are designed to increase the temperature
of the steam generated in the furnace section.  Following the
superheater and reheater, an economizer counterflow tube heat exchanger
is used to initally heat the boiler water before entering the furnace
tube wall.  The air heater is used to preheat the furnace combustion
air.  These separate operations all increase the boiler thermal
operating efficiency.

	Steam tubes are both imbedded in the furnace wall and mounted in the
convective heat exchanger bundles which are exposed to the hot flue
gases (such as in the superheater and reheater).  The steam tubes are
connected to one or more “steam drums” which collect the generated
steam.  Residues that collect and concentrate in the water/steam are
collected at the “mud drum” located at the bottom of the tubes.

Figure 3-1: Watertube Boiler

Soot, ash, and other solid deposits that are generated from combustion
tend to deposit and buildup on the boiler tube surfaces. 
“Sootblowers” are used periodically to clean the tubes of this
particle buildup.

Firetube

	Firetube design boilers are used for applications where smaller steam
production and lower steam quality is required, and steam load
requirements are relatively constant.  Most units are less than 30 MM
Btu/hr in size.  Firetube boilers are compact, modular, and have low
initial capital and installation cost.  Packaged units usually have the
capability of firing gas and liquids.  Solid fuel firing in firetube
boilers is rare due to clogging of tubes with ash and slag residue. 
Disadvantages to the use of firetube boilers include:  (a) inability to
superheat steam; (b) limit on steam pressure of 150 to 250 psi; (c) slow
response to changes due to larger thermal inertia; and (d) lower thermal
efficiency compared with watertube units.

	Firetube boiler design is similar to a shell-and-tube heat exchanger. 
Shown generally in Figure 3-2, firetube boilers consist of a
water-filled cylinder with immersed tubes passing through it, usually
making multiple passes back and forth through the cylinder.  Combustion
gases are routed through the inside of the tubes and transfer heat to
the pool of water to produce steam.  Depending on the tube and firing
arrangement, firetube boilers are generally classified as horizontal
return tube, scotch marine (or shell) or firebox.

Figure 3-2: Firetube Boiler

	

Cast Iron & Tubeless Boilers

	Cast iron boilers are typically smaller than 10 MM Btu/hr and operate
by passing hot combustion gases through sets of heat exchanger tubes. 
They are generally used for producing low quality stream or hot water
for commercial or institutional boiler applications.  Pressure limits
range from 15 to 100 psi for hot water and steam units.  

	The tubeless design is also limited to small applications.  A tubeless
boiler is typically vertically arranged with the burner located at the
bottom or side of the unit. Steam is collected over the water in large
jacket or U tube.

Burner Design

	In a liquid/gas burner, atomized liquid fuels are mixed with combustion
air in a swirling manner to provide a stable flame.  Liquids can be fed
and atomized in the main burner, or injected into the main flame through
auxiliary lances.  Liquid atomization is achieved through mechanical
methods such as rotary cup or pressure atomization systems, or by
twin-fluid nozzles with the assistance of high-pressure air or steam. 
With a high surface area, the atomized particles vaporize quickly,
forming a combustible mixture of fumes and combustion air that rapidly
ignite and burn.

	Liquid ECF with high solids would need to be filtered prior to feeding
given that the fuel must pass 200 mesh.  Additionally, wastes with
viscosities of greater than 50 cs would require pretreatment such as
heating to decrease viscosity or blending with lower viscosity fuels
prior to combustion.  See Part Two, Section II.B.7 of the preamble to
the proposed rule on restrictions on particle content and size.

	In firetube boiler designs, a single burner is usually used.  Watertube
boilers use one of the configurations described below.

Suspension Firing -- Gas, Oil, and Pulverized Coal

	Suspension firing designs are used in watertube boilers for gas (most
commonly natural gas), liquids (e.g., fuel oil) and pulverized solids
(e.g., pulverized coal).  

Suspension firing arrangements in watertube boilers include:

•	Wall (face) fired – Most wall fired boilers are larger than 100
MMBtu/hr. Horizontally mounted burners in either a single (front) wall
or opposed wall set-up.  

--Front wall fired -- Usually use a single burner, although there are
some older and larger units which may use multiple burner rows.  Newer
units use single burners which can provide required control and
turndown.

--Opposed wall -- Used mostly in larger utility applications. 

•	Tangential (corner) fired -- Horizontally mounted burners in the
four corners of a rectangular furnace, all firing toward the center to
produce a cyclonic fireball.

•	Cyclone -- Fuel (usually pulverized coal) and air is fed
circumferentially into a cylindrical combustion chamber.  This design is
not widely used for industrial purposes (mostly used for larger utility
applications).

	Pulverized coal units can be either a wet bottom or dry bottom design,
depending on if the ash is handled as a dry solid (dry bottom) or a
molten liquid slag tap (wet bottom).  Pulverized coal units are usually
large (greater than 100 million Btu/hr) due to the high cost of the coal
pulverizing and handling equipment.

Stoker Firing -- Solids

	Stoker fired boilers are designed to burn solid fuels (including coal,
wood, municipal wastes, etc.) on a bed.  Stoker systems are used on many
coal-burning (and other solid fuel) industrial, commercial, and
institutional boiler applications.  This is because fuel handling and
pretreatment procedures are not typically required.

	Stokers are mechanical or pneumatic devices that feed solid fuels onto
a grate at the bottom of the furnace and remove the ash residue after
combustion.  They consist of: (a) a fuel supply system; (b) stationary
or moving grate which supports the burning mass of fuel and admits most
of the combustion air to the fuel; (c) an overfire air system, provided
over the burning bed, to complete combustion; and (d) an ash or residual
discharge system.  In most stokers, fly ash collected downstream of the
furnace is reintroduced into the bed to ensure complete combustion of
the fuel.

	There are three main classes of stoker set-ups:  (a) underfeed, (b)
overfeed, and (c) spreader stoker:

Fluidized Bed Boilers - Solids

	Fluidized bed systems can be used to efficiently combust various types
of solid and liquid fuels although most ICI boilers fire coal and to a
lesser extent other solid fuels. Size reduced fuel (ground or shredded)
is fed into a bed of inert particles (sand and/or a sorbent such as
limestone).  The bed is kept suspended (“fluidized”) by an upward
flow of combustion air through the bed.  Fluidized beds operate at lower
temperatures than conventional suspension pulverized coal fired boilers.
   

 Emissions from Industrial Boilers

	The emissions database developed for the industrial boiler NESHAP
contains information from fewer than 2000 units.  This section discusses
emissions of CO, methane and PCDD/F.  Nondioxin organic HAP emissions
from industrial boilers are discussed in Section 5.1.

CO Emissions 

The industrial/commercial boiler database contains CO emissions data
from nearly 1000 runs.  Less than 25% of the CO measurements are from a
CEMS – many appeared to have performed a one time test.

Table 3-3 summarizes the CO data by fuel type (note that there were
dozens of fuels and fuel mixes and certain fuel types are combined for
ease of presentation). 

 	CO Emissions (ppmv @ 7% O2)	# of Runs	 

Fuel Type(s)	Min	Max	Avg	Median	Total	<100 ppmv	Percent runs

 	 	 	 	 	 	 	< 100 ppmv

Natural Gas only	0.0956	595	102	23	30	11	37%

Process Gas only	5.6	38200	18803	18403	6	3	50%

NG/Process Gas only	0.325	1.83	1	1.1	3	0	0%

Landfill Gas only	0.0187	131000	5654	1.2	46	44	96%

Fuel Oil only	0.956	70	21	10.6	17	17	100%

Coal and/or Coke only	0.46	94	22	14	43	43	100%

Coal/Coke w gas fuel	4.78	234	73	30.85	12	9	75%

Process Liq	73.4	384	214	217	6	1	17%

Wood only	0.148	7480	619	323	459	86	19%

Wood/RDF/MSW w any other fuel	7.09	6710	489	202	332	109	33%

Industrial Sludge w fossil fuel	0.115	8630	1290	56.6	19	10	53%



Table 3-3: CO Emissions by run (ppmv @7% O2) from Industrial and
Commercial Boilers

The table shows the range of CO emissions for each fuel type or
combination as well as showing the fraction of individual CO runs below
100 ppmv.  Rows in red indicate emissions from units that would likely
not belong to the universe of boilers that meet the criteria for
combusting ECF because:

They are not firing fossil fuel or

They are firing wood or MSW which would require either a stoker or a
fluidized bed unit. 

These data are also plotted in Figure 3-3.  There are 11 instances where
natural gas fired units exceed 100 ppmv CO.  These units with high CO do
not specify the boiler design type.  All units firing fuel oil only and
all those firing coal and coke only are below 100 ppmv for all runs. 
There are 3 runs with CO>100 from a circulating fluidized bed boiler
that burns coal with natural gas.

					(a)

							(b)

Fig 3-3: CO Emissions from industrial boilers (a) Gas/liquid fired
units, (b) Solid/sludge fired units.

CO vs. Methane

	Figure 3-4 shows methane emissions vs. CO emissions for all boilers for
which both CO and methane emissions are available.  The vertical and
horizontal lines on the plot show CO = 100 ppmv and Methane =30 ppmv
(which is equivalent to THC=10 ppmv on a propane basis if methane is the
only hydrocarbon present) 

All fuel oil fired units have low methane emissions with most around 1
ppmv.  No data are available from units that fire only coal.  A few
natural gas fired units have methane emissions near 10 ppmv.  This plot
also shows that there are no instances where methane emissions exceed 30
ppmv when CO is below 100 ppmv.

Fig 3-4: Methane vs. CO emissions for Industrial Boilers

PCDD/F

	For dioxins and furans, only 15 runs from five test conditions are
available. As seen in Table 3-4, emissions were generally low, with test
condition averages ranging from 0.006 to 0.042 ng TEQ/dscm and an
average value of 0.013 ngTEQ/dscm. 

Table 3-4: PCDD/F Emissions from Oil-Fired Industrial BoilersCO as an
Indicator of Incomplete Combustion

	Carbon monoxide is a universally accepted indicator of combustion
efficiency and limits on CO have been historically used to control
emissions of organic HAP.  For example in the recently promulgated
industrial boiler NESHAP CO is used as a surrogate for organic HAP.  CO
is a conservative indicator of deteriorating combustion conditions. 
Generally when CO is low organic emissions are low because destruction
(DRE) is high.  However, because CO oxidation to CO2 is the slowest and
last step in the combustion of organic waste high DRE and low organics
emissions levels can be achieved at high CO levels.

Combustion Failure Modes

	EPA has identified four combustion failure modes which would result in
poor DRE.

Total Ignition Failure:  Combustion does not occur and products of
incomplete combustion (PICs) are not formed.  This results in the
absence of CO, high HC if the unreacted fuel/waste has significant
organic content, and low DRE.

Partial Ignition Failure:  Part, but not all of the fuel/waste combusts
forming PICs.  This results in high CO, high HC if the unreacted
fuel/waste has significant organic content, and low DRE.

Combustion Air Failure:  Insufficient combustion air leads to incomplete
combustion.  Very high CO concentrations result when combustion air is
less than that required to stoichiometrically burn the waste and fuel. 
As the air to fuel/waste ratio decreases further and below
stoichiometric, high HC concentrations (initially dominated by methane)
result; and, at even lower air to fuel/waste ratios, low DRE may result
as non-methane organics begin to appear.

Rapid Quench Failure:  The fuel/waste is incompletely combusted because
it is not exposed to temperatures necessary to sustain oxidation for a
sufficient period of time.  The combustion gas cools so quickly that
combustion is interrupted.  The resulting CO emissions are high because
the oxidation of CO to CO2 is the last step in the combustion process,
and the reaction rate of CO oxidation is slow.  This failure mode may
also result in high emissions of HC and low DRE.

	Of the above failure modes, only total ignition failure would result in
high DRE without correspondingly high CO.  EPA also recognized that DRE
failure could occur if the POHC was injected at low concentrations,
especially if the POHC was also a likely PIC.   

	In addition, EPA examined data from hundreds of DRE tests with
concurrent measurements of CO and HC.  The rare instances where DRE
failure occurred without corresponding high emissions of CO were all
explainable.  EPA concluded that CO is a conservative indicator of DRE
failure--in other words DRE failures would usually be accompanied by
high CO.  Note that EPA has supplemented this analysis by including more
recent data, as presented in Appendix A.

CO vs. DRE

	Early studies by EER and others in the 1980’s pointed to a potential
correlation between POHC emissions and CO.  Other subsequent work
primarily by Dellinger et al at UDRI generally found no correlation or a
poor correlation.  However, the use of a CO limit as a DRE failure
indicator does not require that CO and DRE be directly correlated;
rather, a DRE failure must simply be indicated by high CO. 

	In 1986, Hall et al tested a 12 component POHC mixture in a laboratory
scale thermal destruction unit.  The primary purpose of the work was to
identify thermally stable POHCs for future emissions testing during a
trial burn.  Many of the compounds were constituents (including
nitrogenated/halogenated organics and oxygenates).  An important
characteristic of this experiment was that while it was intended to
simulate a liquid injection incinerator the tests were conducted in a
flameless controlled temperature reactor.  The POHC mixture was injected
into a stream containing a nitrogen/air mixture (at either 2.5% O2 or
10% O2).  Water vapor was introduced to the stream such that the
moisture content was 5%.  The concentrations of most of the POHC
constituents were 1000 ppmv in air (acetone and methanol were at higher
concentrations).  Tests were conducted at two temperatures 650 °C(~1200
°F) and 775 °C (1430 °F) at residence times of 0.5 seconds and 2
seconds.  Tests showed very poor DREs in general.  In some cases poor
DREs were accompanied by low CO (as low as 15 ppmv).  Under the
conditions most conducive to POHC destruction-among those tested (10%
oxygen, 2 sec residence time and 775 °C), all constituents were below
detection limits with the exception of chlorobenzene  and acetonitrile
which had DREs of about 98%.  The authors conclude that “The apparent
lack of correlation between CO levels and changes in the decomposition
of the constituents of the mixture places doubt on the usefulness of CO
measurements as an indicator of destruction efficiency”

However, we believe that this study in no way invalidates the use of CO
as the sole indicator of good DRE because:

The study was conducted in a flameless decomposition reactor and
generally creates conditions similar to “total ignition failure”
which is one the four DRE failure modes identified by EER’s work in
the 1990’s.

The reactor temperatures were low (< 1430 °F), well below temperatures
one would expect to see in a combustion chamber of a typical boiler or
incinerator.

The CO analysis was not done using a CO CEMS.  Instead, gas samples from
the exhaust were collected and sent through a separation column to
separate the CO. This CO stream was then sent through a methanizer to be
converted to methane which was subsequently analyzed by a FID.  Moreover
the experimenters used a 10,000 ppmv CO in helium as a standard. 
Therefore, accuracy of the low CO measurements is questionable.

And most importantly, we believe that since the study was conducted
during a time when CO compliance limits were established based on levels
achieved during the trial burn, the authors’ intent was to point out
that basing CO limits on the trial burn was inappropriate.  The paper
does not say anything about the 100 ppmv CO level that was later
developed as a level that minimizes POHC emissions and PIC formation. 
See EPA handbook “Guidance on Setting Permit Conditions and Reporting
Trial Burn Results EPA/625/6-89/019, p.52” 

	In work published in the Journal of Hazardous Materials, Dellinger et
al studied the flameless decomposition of a 3-component POHC mixture
(toluene, ethyl cyanide, and 1,3,5 trichlorobenzene) in an isothermal
flow reactor.  The mixture was exposed to a given temperature for a 2
second residence time at an equivalence ratio of 3.0 (or 33% theoretical
air).  Under these fuel-rich conditions, up to 50 PICs were detected
from this 3-POHC mixture.  The authors arrived at the following
conclusions while acknowledging that the conclusions could be debatable.


PIC formation is a natural consequence of POHC destruction.

The formation and destruction of PICs are kinetically controlled and
levels predicted by thermodynamic equilibrium calculations are often
orders of magnitude too low.

Many of the PICs appeared more stable than the parent POHCs because they
are exposed to high temperatures for an effectively shorter residence
time.  Thus, they may require a higher destruction temperature.

	Another study was conducted in a pilot-scale 150,000 Btu/hr turbulent
flame reactor (TFR).   A mixture containing a selection of chlorinated
hydrocarbons in heptane (>90%) was introduced to the reactor via a
pressure atomizing nozzle.  Unlike the previous Dellinger experiments,
the overall stoichiometry was fuel-lean with exhaust oxygen
concentrations varying between 3% and 6.5%.  Exhaust CO and oxygen were
measured using a non-dispersive IR and paramagnetic analyzers
respectively.  Exhaust VOCs were analyzed using a VOST train.  The
approach was, however, limited to compounds with boiling points below
150 °C.  The authors show that no correlation was found between the
POHC fraction remaining in the stack and CO emissions.

	The authors state the predominant pathway of organics destruction is
formation of CO with subsequent oxidation to CO2 and the CO to CO2
oxidation is the slowest rate determining step.  The authors also
contend that the CO to CO2 conversion can only occur in an oxidative
environment (and most likely occurs by oxygen atom transfer involving
the hydroxyl (OH) radical).  The contention is that small amounts of the
feed will see localized regions that are fuel rich and/or low in
temperatures and residual POHC and PIC emissions are due to pyrolysis
pathways (oxygen deficient pockets).  Also low temperatures (<800°C)
are required before measurable POHC/PIC emissions occur in oxidative
pathways.  Since POHC degradation and PIC formation/degradation can
occur in different parts of the combustor than reactions involving CO
formation/destruction, CO emissions may not directly correlate to POHC
destruction efficiency.

Figure 4- 1: POHC Fraction Remaining vs. CO (reproduced from Staley et
al, 1989-lines for 100% CO, 99.99% DRE added to original plot)

	The authors used the above diagram to illustrate that POHC emissions do
not correlate with CO emissions.  The plot shows POHC fraction remaining
(1-DRE/100) as a function of measured CO emissions. 

However note the graph does show that:

There are only two cases where DRE is less than four nines and these
occur at high CO (525 ppm @ 0% O2 or 350 ppm @ 7% O2).

For the four cases where CO is less than 100ppmv, DRE is well above
99.99%.

Also, PIC emissions measured during these experiments were about an
order of magnitude lower than the highest POHC emissions.

Conclusions

	CO has historically been used as an indicator for good combustion and a
surrogate for organic HAP.  Although several studies have shown no
direct correlation between CO and DRE, the use of CO as good combustion
indicator does not require there to be a direct correlation--only that a
DRE failure would be indicated by high CO levels.  Also, the reason
there is no correlation is that, as explained earlier, high DREs can
occur even at relatively high CO levels (i.e., because CO is a
conservative indicator of combustion efficiency).  However, under
oxidative conditions seen in boilers, a CO limit coupled with the other
burner conditions required for ECF combustion would ensure that organic
emissions are minimized.  (Emissions are low and DREs are high when CO
is low, and high organic emissions are evidenced by high CO emissions).

Organic Emissions & Qualitative Risk Assessments

This section presents the results of several analyses that were
conducted to assess whether emissions of toxic organic compounds from
ECF-fired boilers could be expected to be comparable to emissions from
oil-fired industrial boilers, and whether dioxin/furan emissions from
burning ECF could be expected to be protective of human health and the
environment.  These include:

A comparison of speciated organic emissions data between fuel oil fired
industrial boilers and watertube steam boilers (that are not
stoker-fired) burning hazardous waste.

An abbreviated comparative risk assessment for dioxin/furan emissions
from hazardous waste-burning watertube steam boilers.

	

	In the absence of emissions data from boilers burning ECF, we evaluated
organic emissions data from watertube steam boilers (that were not
stoker-fired) burning hazardous waste and compared those emissions with
emissions from oil-fired industrial boilers.  Using hazardous waste
boiler emissions as a surrogate for ECF boiler emissions is reasonable
because the ECF exclusion would be conditioned on the ECF boiler
operating under conditions that ensure good combustion efficiency.  The
operating conditions would be more restrictive than those for
RCRA-permitted hazardous waste boilers.  (See discussion in Part II:
Section II.A of the preamble to the proposed rule.)

	

	To perform these analyses, we evaluated boiler designs for boilers with
risk burn test data in the hazardous waste combustor database.  Since
the HWC database only contains information prior to 2003-2004, more
recent test reports for risk burn testing were obtained.  The boilers
were screened to exclude firetube boilers, stoker-fired coal-boilers, as
well as process heaters.  In addition, data from three watertube boilers
whose combustion chamber was not of integral design with the boiler
section were also screened out.

 	The remaining 27 boilers are listed on Table 5-1.  The table shows
source ID number, EPA ID number, facility name and location as well as
information on burner design.  Note that all of these boilers are
classified as liquid fuel boilers under the HWC MACT standards because
there were no risk burn data available from pulverized coal-fired
boilers.

Table 5-1: Hazardous Waste Burning Watertube Steam Boilers with Risk
Burn Data  

Individual Organics

Industrial Boiler Benchmark

The Industrial Boiler emissions database discussed in Section 3.4 above
was surveyed for speciated organics data.  The following data were
screened out:

Nondetects

Emissions from firetube boilers, stokers, and process heaters.

Emissions from boilers other than those that fired only fuel oil
(Boilers that fired other fuels in combination with fuel oil were also
screened out.)

Emissions from distillate fuel oil boilers.

Emissions for nontoxic compounds—compounds not listed as RCRA toxic
organics in 40 CFR Part 261, Appendix VIII, or not listed as CAA
hazardous air pollutants (HAP).

	

The remaining emissions data were converted to standard units (ug/dscm)
and the results are shown in Table 5-2.  The table shows emissions
information for 26 specific organics, including number of total runs,
average emissions as well as the range of emissions.  The 95 %ile
emissions level, which is used as a benchmark for comparison with
hazardous waste boilers, (discussed below in Section 5.1.2) is shown in
bold font.  The toxic organics with the highest emission concentrations
among industrial boilers burning fuel oil are formaldehyde, benzene,
toluene, and xylene.

In addition to the information gathered from the emissions database, the
table also shows available AP-42 fuel oil combustion data for these
organics. 

Table 5-2: Toxic Organic Emissions Data for Oil-Fired Industrial Boilers



Comparative Data from Hazardous Waste Boilers

Organic emissions data were extracted for 28 risk-burn test conditions
from 26 hazardous waste boiler facilities shown in Table 5-1. 
(Note--Two units had two test conditions each and one unit did not
perform testing for individual organics).  In this section, we discuss
the procedure used to extract emissions data from the risk-burn test
reports and compare emissions from hazardous waste boilers to emissions
from oil-fired industrial boilers. 

The risk burns involved separate sampling trains for evaluating the
target analytes among VOCs and SVOCs.  In a few cases a PAH sampling
train was also used.  The test reports typically contained run by run
emissions for three runs for each test condition. However, each test run
was a composite measurement:

VOC- Each run is a composite of at least 3 measurements (Typically 3 or
more VOST tube pairs (tenax for the front half, and tenax/charcoal for
the back half) and a condensate measurement).  VOST tube pairs are
sometimes analyzed separately but the separate front and back half data
were rarely available.

 

SVOC/PAH- Typically there is only one pair of tenax tubes.  Separate
front and back half data were rarely available.

Several stack testing companies use the “<” qualifier to data that
are either fully nondetect or partially nondetect.  So, looking at
individual VOST tube pair data for VOCs and run level data for SVOCs we
made the following assumption:

If data had a "<" qualifier we assumed that this was a nondetetect
unless there was information available to distinguish nondetects from
partial nondetects.

The emissions from detect runs (or the detect component of partial
detect runs) were tabulated for each test condition for each of the 26
organic compounds for which we also have emissions data for oil-fired
industrial boilers.  The results are shown in Table 5-3 for all 28 test
conditions.  An “NM” in the table indicates that the constituent was
not measured by that source, while an “ND” indicates that the
constituent was measured but all three runs were fully nondetect.  For
the other cases, the condition averages of the detect runs (or detect
component of partial detect runs) are shown in units of ug/dscm
corrected to 7% oxygen. The last row on the table shows the number of
conditions with detect (or partial detect) runs for each compound. 
There are total of 173 such conditions when considering all 26
compounds.

Table 5-3: Toxic Organic Emissions from Hazardous Waste Boilers for
Compounds for Which Emissions Data for Oil-Fired Industrial Boiler Are
AvailableExceedances of Industrial Boiler Benchmark

The 173 hazardous waste boiler emissions conditions were compared to an
industrial boiler benchmark for each compound--the 95%ile emissions
level from the industrial boiler database.  There are 24 exceedances of
the benchmark level from 15 boilers.  Nine of these exceedances involved
dichloromethane (methylene chloride) and six involved benzene.

The results are summarized in Table 5-4.  For many exceedances, the test
report indicates that the sample may have been contaminated.  This was a
particular problem for dichloromethane.  Lab contamination was known or
thought to be a problem for 12 of the exceedances.  For nine cases (six
with dichloromethane, two with benzene, and one with toluene) the
constituent being measured was found in the blank.  There were three
additional exceedances for dichloromethane, a common lab contaminant
that is frequently found in laboratory samples and in the environment.  

Seven other exceedances were at trace emission levels.  Hazardous waste
boiler emissions were below 8 ug/dscm for benzo[a]pyrene, ethylbenzene
(accounting for two exceedances), 2-methylnaphthalene, anthracene,
fluorine, and phenathrene.  

Moreover, we note that only five of the 15 boilers had exceedances that
were not suspect because of known or suspected lab contamination and
that were at significant concentration levels.  Four of the exceedances
were for benzene and one was for acetaldehyde.  None of those five
boilers were operating under the conditions that would be required for
ECF boilers, however.  Although this is not unexpected because these
boilers were not required to operate under those conditions, operating
under combustion conditions less stringent than would required for ECF
boilers could result in higher organic emissions.  Three of these
boilers burned hazardous waste fuel with a heating value of 2,000 Btu/lb
or below while ECF must have an as-fired heating value of 8,000 Btu/lb. 
One boiler fired less than 20% primary fuel (natural gas) while ECF must
be fired with at least 50% primary fuel.  And, the hazardous waste fired
in one boiler had virtually no heat content and had a viscosity of 165
cs, while ECF must have an as-fired viscosity of 50 cs.



Nondetect Analysis—Limitations of the Database

	As Table 5-3 indicated, toxic organic emissions were nondetect for a
majority of compounds for several of the hazardous waste boilers. 
Appendix B summarizes these nondetect data and compares detection limits
to the industrial boiler benchmark.  For several compounds, detection
limits were higher than the industrial boiler benchmark for a
significant fraction of the boilers.  In instances where the detection
limits exceed the benchmark level, the actual emissions level from the
hazardous waste boiler may or may not be higher than the benchmark
level.

Emissions of Other Toxic Organics from Hazardous Waste Boilers

	In addition to the 26 organic compounds with available emissions data
from oil-fired industrial boilers, we also extracted detect data from 40
other toxic organics emitted by the hazardous waste boilers.  Table 5-5
shows the average emissions from hazardous waste boilers for these toxic
organics.  Test condition average emissions are at trace levels—below
10 ug/dscm—for all compounds except acetophenone, phenol,
bis(2-ethylhexyl)phthalate, and chloroform.

	All four of these compounds had a single test condition which appeared
to be a high outlier.  Therefore we conducted a statistical outlier
analysis for the data sets for these compounds.  The analysis involved: 
(a) identifying the suspected outlier, (b) assessing if the remaining
data fit either a normal or lognormal distribution using the
Shapiro-Wilk test procedure, and (c) using the Dixon and Grubbs
statistical tests to assess if the suspected data point is an outlier.  

	This process, which is detailed in Appendix C, determined that the test
condition with the highest emissions average was an outlier for
acetophenone and phenol.  As seen in the far right column of Table 5-5,
the average emissions for these two compounds are below 5 ug/dscm
without the statistical outlier.  As explained below, the highest test
condition average was not a statistical outlier for the other two
compounds:  bis(2-ethylhexyl)phthalate and chloroform.     

 	We have bis(2-ethylhexyl)phthalate emissions data for 15 test
conditions (generally comprised of three runs) representing 15 different
boilers.   Test condition average emissions ranged from 0.34 ug/dscm to
600 ug/dscm for the boilers, with an average of 69 ug/dscm.  Although
the highest test condition average—600 ug/dscm—appeared to be an
outlier given that the second highest average was 130 ug.dscm and 12
test conditions were below 42 ug/dscm, we determined that it is not a
statistical outlier.  Nonetheless, we note that:  (1) the boiler with
the highest emissions—600 ug/dscm—was not operating under the
conditions that would be required for an ECF boiler (which could result
in higher emissions)—the primary fuel firing rata was approximately
30% rather than a minimum of 50%, and boiler load was approximately 30%
rather than a minimum of 40%; and (2) bis(2-ethylhexyl)phthalate is
known to be a common lab contaminant, and thus the reported emissions
levels may be suspect.

For chloroform, we have emissions data for 9 test conditions (generally
comprised of three runs) representing 9 different boilers.  Test
condition average emissions ranged from 0.28 ug/dscm to 270 ug/dscm for
the boilers, with an average of 45 ug/dscm.  Although the highest test
condition average—270 ug/dscm—appeared to be an outlier given that
the second highest average was 85 ug/dscm and the remaining test
conditions did not exceed 16 ug/dscm, we determined that it is not a
statistical outlier.  We note, however, that the boiler with the highest
emissions—270 ug/dscm—was not operating under the conditions that
would be required for an ECF boiler—it burned a waste fuel with a
heating value below 8,000 Btu/lb and it is not clear whether the boiler
burned process vent gas or natural gas as primary fuel.

	

Table 5-5:  Toxic Organic Emissions from Hazardous Waste Boilers for
Compounds for which Emissions Data for Oil-Fired Industrial Boiler Are
Not Available



	Dioxins and Furan (PCDD/F)

	Given that polychlorinated dibenzo dioxins and furans (PCDD/F), like
other persistant organic pollutants (POPS), bioaccumulate in human and
animal fatty tissue and remain in the environment for long periods, they
merit special consideration when expansion of the comparable fuels
exclusion is being considered.  This section contains a summary of
PCDD/F/furan formation mechanisms and discusses available PCDD/F
emissions data from boilers that burn hazardous waste.  Additionally,
the results of an abbreviated risk assessment for PCDD/F is also
presented.

PCDD/F Formation Mechanisms

	This section provides a brief summary of formation mechanisms for
PCDD/F.  Note that several extensive reviews of PCDD/F formation
mechanisms are available.,   The primary formation mechanisms can be
described as:

Homogeneous (gas-gas) formation from organic precursors (such as
chlorinated aromatics) in combustion zone (500-800 °C).

Heterogeneous (gas-solid) condensation reactions between gas phase
precursors and a catalytic particle surface. 

De novo synthesis- Heterogeneous surface-catalyzed reactions between
carbon containing particles and organic or inorganic chlorine.

	Mechanism (1) usually plays a minor role in PCCD/F formation in
combustion facilities, although it has been theorized that gas-phase
formation of Cl2 leads to PCDD/F formation when a highly chlorinated
waste is burned and the combustion gases undergo slow quench (e.g., in a
waste heat boiler).   

	Mechanisms (2) and (3) are generally more dominant.  The key
requirements of PCDD/F formation by these two pathways are particulate
holdup in the temperature window 200-400 °C (400-750 °F), and the
presence of chlorine or chlorinated organics.  Although it is not always
easy to distinguish between mechanisms (2) and (3), according to work
cited in the EPA risk burn guidance document mechanism (2) involves fast
reactions and may predominate in post-combustion and heat exchanger
sections where residence time at the critical temperature window may on
the order of 1 second. Conversely, the de novo process may dominate in
dry APCDs where particle residence times may be much longer.

Other factors – The bulk of the chlorine in a combustor would be
present as either HCl or Cl2.  HCl is converted to Cl2 via the Deacon
reaction and Cl2 is known to chlorinate aromatic organic PCDD/F
precursors.  The presence of metal catalysts (such as copper or nickel)
are usually required to overcome kinetic limitations of the Deacon
reaction.  Additionally, metals support the condensation reactions that
form PCDD/Fs from organic precursors.  Also, the presence of sulfur
inhibits formation of PCDD/F by depleting Cl2 and poisoning copper
catalysts.

	As shown in a comprehensive review by Stanmore, numerous PCCD/F
formation pathways have been studied and documented.  Many of these
pathways involve the formation or presence of chlorinated benzene or
phenol.  However, if an alternative chlorine source is present,
precursors can be nonchlorinated aromatics (such as benzene) or even
aliphatics such as propene or acetaldehyde.  Gullet and Seeker have
shown that heterogeneous PCCD/F formation pathways from aliphatics
involve PAHs as an intermediate step.  Proccccini et al showed by
injection of benzene into a cooldown section of a combustor (at
500-800°C) chlorobenzenes and chlorophenols may be rapidly formed.  
(Note that for this scenario to occur in a real boiler significant
amounts of benzene would have to escape the flame zone).

Factors influencing PCDD/F Emissions from ECF Fired Boilers

As noted in the previous section, PAH are intermediate compounds under
certain PCDD/F formation pathways.   ECF will not contain detect levels
of PAH because as discussed earlier the specifications for PAH
hydrocarbons would not be waived under the proposed ECF regulation.   

Many of the 26 hydrocarbons and oxygenates for which the specifications
would not apply are aromatics (e.g., benzene, toluene, phenol).  In the
presence of chlorine, these can be converted to known PCDD/F precursors
such as chlorobenzenes and chlorophenols.  (Note that the current 540
mg/kg specification limit on total chlorine is more than sufficient
chlorine for heterogeneous PCDD/F formation.)

All the mechanisms mentioned above require chlorine in the feed.  The
heterogeneous mechanisms are largely impacted by metals.  Chlorine
limits (total chlorine as well as individual halogenated organic
specifications) in Table 1 to §261.38 would apply to ECF.  Also, the
metal specification limits would apply to ECF as well.  Moreover, wastes
eligible for the ECF exclusion are not likely to contain significant
amounts of metals.  Thus, two factors that have a significant bearing on
PCDD/F emissions will likely not come into play.

PCCD/F formed in the combustion zone via homogeneous gas phase reactions
will likely be destroyed in the combustion zone.  The Dellinger thermal
stability index ranks 2,3,7,8 TCDD (the PCDD/F congener with the highest
TEQ) a Class 2 compound and thus easier to destroy than benzene. 
Similarly, aromatic hydrocarbons in the feed will be mostly destroyed in
the combustion zone.

The additional requirement of the 100 ppmv CO limit would serve to
establish good combustion conditions and minimize the formation of
chlorinated aromatic PCDD/F precursors as PICs.  Also, continuous CO
monitoring would warn of flame quenching or other process upsets that
could cause soot deposition in downstream boiler tubes and contribute 
to increased PCDD/F emissions.

Units with Dry Air Pollution Control Devices- For units with dry APCDs,
formation of surface-catalyzed PCDD/F can increase exponentially for
APCD temperatures above 400 °F.  Additionally, the next section will
show that hazardous waste burning boilers with dry APCD temperatures
below 400 °F have comparatively lower emissions of PDD/F.  Therefore a
gas temperature limit of 400°F is proposed for boilers that burn ECF,
unless the boiler’s primary fuel is coal.  Boilers that burn coal as
the primary fuel are exempt from this requirement because sulfur in coal
is known to inhibit PCDD/F formation. 

 PCDD/F Emissions from Boilers burning Hazardous Waste

	Table 5-6 shows PCDD/F emissions data from hazardous waste burning
boilers separated into three categories:  (a) units with dry APCS; (b)
units with wet APCS; and (c) units with no APCS or unknown APCS.  The
table shows condition average PCDD/F emissions in ng TEQ/dscm.

	

	The table includes data from the 2005 Hazardous Waste Combustor MACT
database. (which includes a few boilers that are no longer burning
hazardous waste) and data from additional test reports obtained since
the close of the HWC MACT database.

	Data are available from 11 test conditions for dry APCD-equipped units.
 Two dry APCD boilers have PCDD/F emissions above 0.4 ng TEQ/dscm, the
generic MACT standard for most HWCs, and a level below which D/F
emissions are generally considered de minimis.  One watertube boiler
which had PCDD/F emissions of 2.4 ng TEQ/dscm is not of integral design
and had very high levels of nickel in the feed.  The second unit, a
firetube boiler feeding mixed waste had emissions of 0.66 ng TEQ/dscm.  


    

Table 5-6: PCDD/F Emissions from Hazardous Waste Boilers

	We have nine test conditions for PCCD/F from units with wet APCDs.  Two
units measured PCDD/Fs greater than 0.4 ng TEQ/dscm.  Both of these are
fire tube units and one of them burns waste fuel containing 60%
chlorine.

Comparative Risk Assessment

Background:  In 1999, EPA conducted a comprehensive, multi-pathway risk
assessment for Phase I HWC (incinerators and kilns).  For each facility,
site-specific air modeling and fate and transport modeling was performed
to quantify multi-pathway exposures. Risk distributions were developed
for each source category and each pollutant.  Since the 2005 HWC MACT
replacement rule brought in additional source categories this analysis
was updated for the 2005 rulemaking.  A comparative analysis was
performed in lieu of a comprehensive risk assessment.  The comparative
analysis relied on predictions from the Phase I risk assessment and
comparisons between the Phase I and Phase II universes. Two primary
methodologies were utilized in the comparative analysis:

Weight-of-Evidence (WOE) Scoring - WOE scoring relied on a large array
of statistical comparisons involving all four megavariables (emissions
rates, stack parameters, population and meteorology) to evaluate whether
a Phase II source category would be expected to have risks either less
than, equal to, or greater than a Phase I category’s high-end risks
for each pollutant.

Margin of Exposure (MOE) Analysis - For situations where the WOE scoring
predicted in the direction of Phase II risk greater than Phase I, a
simple Phase II/Phase I emissions ratio was calculated (using upper
confidence limit mass emission rates for a given pollutant).  The
emission ratio was then evaluated against the risk “safety margin”
(i.e., MOE) determined from the Phase I risk assessment.

Comparative Risk Analysis:  An abbreviated comparative risk evaluation
was performed to assess the impact of the proposed comparable fuels
expansion on PCDD/F emissions.  The abbreviated evaluation utilized one
component of the Phase II hazardous waste combustor MACT comparative
risk evaluation, specifically, the Margin of Exposure (MOE) analysis. 
The emission-adjusted MOE analysis utilizes the risk “safety
margins” (i.e., modeled MOEs) determined from the MACT Phase I
comprehensive risk assessment to see whether, considering emissions
alone, risks for a second universe (here, the ECF boilers) could rise to
a level of concern.  Smaller MOEs correspond to a greater potential for
risk beyond the level of concern (i.e., 1E-05 lifetime cancer risk).  To
predict an emission-adjusted MOE for the ECF boilers, the Phase I
modeled MOEs are multiplied by the PCDD/F emission ratios for the Phase
I universe versus the ECF boiler universe.  The Phase I incinerator
category was used for predicting the boiler-adjusted MOEs.  Since there
are little or no PCDD/F emissions data specifically for the universe
under consideration (i.e., industrial boilers burning
emissions-comparable fuel), the PCDD/F emissions database from boilers
burning hazardous waste fuels was considered as a surrogate

The technical approach involved the following steps:

	 

Update/Revise boiler PCDD/Fs emissions data.

Calculate point estimates and confidence levels for revised boiler
emissions data.

Combine old phase I and revised Phase II data sets and conduct test for
common generalized percentile.

Adjust MOE if appropriate.

Update Boiler Database

	

PCDD/F emissions information available in the newly acquired test
reports was added:

Mallinkrodt, Raleigh, NC (1000)

Rohm & Haas, Louisville, KY (741)

Air Products, Wichita, KS (2007)

Kalama Chemical, Kalama, WA (771)

Sunoco, Philadelphia, PA (2008)

Merck, Rahway, NJ

Since there are no PCDD/F data available from nonstoker solid fuel
boilers, there were no revisions to the list based on inclusion of
nonstoker solid fuel boilers.

	Since firetube boilers and process heaters would not be permitted to
burn ECF, these units were removed from the “Phase II boiler” list 
that was used for the MACT 2005 comparative risk assessment.  Units 763
(SC), 776/777 (GA), 814/815 (LA), 2001/2/3 (TX), 1016(TX), 2020, 746(TX)
were removed).

	Once the revisions were made to the database, the facility emissions
estimates (in g TEQ/yr) were calculated.  Note that, consistent with the
approach used for the 2005 assessment, imputed data were used for
sources only if another source at the same facility had measured data. 
We also assigned identical sister units identical emissions rates.  For
the revised boiler databases there were 25 distinct facility
observations (compared to 23 for the 2005 analysis).  

Calculate Revised Point Estimates and Percentile levels for boilers

	The point estimates and upper and lower confidence levels for the
revised boiler category were calculated at the 90th, 75th, and 50th
percentile levels and are shown in Appendix D.”  The revised point
source estimates in g TEQ/yr are:  

Common Generalized Percentile Test for Combined Data

	The revised boiler database was combined with the unchanged old Phase I
incinerator database.  There are 47 total observations (note that number
of Phase I incinerator observations (n) =22 and number of revised boiler
emissions (mnew) = 25). These emissions are ranked from lowest to
highest and chi square analyses were performed to determine the
appropriate percentile level for the MOE analysis.  The procedure
followed involve the following steps.

Select a percentile level (typically 90% is chosen).

Calculate the point source estimate (for the combined data sets) at the
percentile level selected.

Calculate the number of observations above and below the value
calculated in step 2 for each data set. (n1 is the number of Phase I
incinerator observations less than or equal to the point source estimate
and n2 is the number of observations higher than this value.  The
corresponding number of observations for boilers is m1 and m2)

Calculate the chi square statistic (Χ2) which is a function of n1, n2,
m1 and m2.

If Χ2 is less than the critical level of 2.707 and one of the four
quandrants (n1, n2, m1 and m2) is less than 5 we would move on to
another percentile level closer to the median.  However, if Χ2 greater
than 2.707 an adjustment of MOEs is warranted.

		

Test at 90th Percentile

The 90th percentile pt, is the 43rd ranked observation = 0.083 gTEQ/yr

 n1 = # of Ph I INC observations that are less than or equal to 0.083 =
19

 and n2=  # of Ph I INC observations that are greater than 0.083 =
22-19=3

Similarly for the revised boiler set, m1= 24 and m2=1

Using the equation above we get  Χ2 = 1.395

And “p value” = Chidist(Χ2  , 1) = 0.237

Since “pvalue” > 0.1  (equivalent to Χ2  < 2.707) and two of the
four quadrants (m1,m2,n1,n2) are less than 5 we move to a different
percentile level closer to the median.

Test at 75th Percentile

Thus 75th percentile pt, is 36th ranked observation = 0.024 gTEQ/yr

We also have n1 =13 and n2= 9 as well as m1= 23 and m2=2

We calculate the chi square statistic

Χ2 = 7.069

And “p value” = Chidist(Χ2  , 1) = 0.0078

We now have the chi square statistic greater than the critical value of
2.707 so an 

adjustment to the MOEs warranted.

Adjusted Margin of Exposure

             

	The original MOE for Phase I incinerators (using the 1985 D/F slope
factor) are 50, 20 and 10 at the 90th, 95th and 99th percentile risk
distributions respectively.  These MOEs are adjusted by the upper
confidence level emissions ratio at 75th percentile (i.e. Ph I UCL/Ph.
II UCL= 0.166/0.062 =2.67).

The results are shown in Table 5-7.  Appendix E presents detailed
calculations.

	Phase I Modeled MOEs 

for 90th to 99th Percentile Risk Distributions

Phase I - All Incinerators

Complying w/ MACT	

90th Percentile	

95th Percentile	

99th Percentile





	- 1985 Dioxin Slope Factor	50	20	10

- 2000 Dioxin Slope Factor	8	4	1.7

	PCDD/F Emissions Rates (g TEQ/year)

	75th Percentile

Point Estimate	95th Percent Confidence Bounds	

Maximum



Upper	Lower

	Phase I All Incinerators	0.063	0.166	0.021	0.174

ECF Boilers	0.016	0.062	0.008	0.109

	ECF Predicted MOEs

for 90th to 99th Percentile Risk Distributions

ECF Boilers 

- 1985 Dioxin Slope Factor	90th Percentile	95th Percentile	99th
Percentile

	130	50	30

Table 5-7: Results of Abbreviated Comparative PCDD/F Risk Assessment

Conclusions

The emissions-adjusted MOEs representing the ECF boilers are 130, 50 and
30 at the 90th, 95th, and 99th percentile risk distributions,
respectively.  This suggests a lower potential for risk for the ECF
boiler category compared to the Phase I incinerator category.

The emission-adjusted MOE analysis should be considered a rough gauge of
protectiveness.  It is important to be aware of the potential error and
uncertainty associated with this approach.  Based on the
cross-validation analysis conducted for the Phase II MACT evaluation
(see Attachment 1), it would not be unusual (on the order of 33% of the
time) to have predictive errors that are greater than an order of
magnitude.

The emissions database itself is subject to substantial uncertainty. 
There are little or no PCDD/F emissions data specifically for the
universe under consideration (i.e., industrial boilers burning ECF under
the proposed design and operating conditions).  Therefore, the PCDD/F
emissions database from boilers burning hazardous waste fuels was
considered as a surrogate.  The surrogate emissions database may not be
representative for some situations.  For example, if the industrial
boilers were not required to control inlet temperature to dry air
pollution control devices, then PCDD/F emissions could be much higher
than measured for the hazardous waste boilers.  Similarly, PCDD/Fs could
be higher if the industrial boilers were permitted to burn ECF
containing PAHs.  Other differences might include combustion quality and
stack characteristics.

Use of the MOE analysis alone introduces greater uncertainty than for
the MACT Phase II comparative risk evaluation.  An important aspect of
the Phase II comparative evaluation was that the statistical analyses,
hypothesis testing, and WOE scoring preceded the MOE analysis.  Only
when the hypothesis testing/WOE scoring indicated the potential for
increased risks (relative to the Phase I source category) did the MOE
analysis come into play.  Ideally, differences in stack parameters,
location, nearby land use and meteorology should be taken into account
in any evaluation of protectiveness.  Nevertheless, in the absence of
additional information, an argument in favor of using the MOE component
is that the cross-validation analysis showed MOE to err on the side of
wrongly predicting greater risk.

EPA has been conducting a reassessment of the human health risk
associated with dioxin and dioxin-like compounds (U.S. EPA , Exposure
and Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin
(TCDD) and Related Compounds, NAS Review Draft, December 2003).  This
draft reassessment was reviewed by the National Academy of Sciences
(NAS) and EPA is currently evaluating the NAS report to determine next
steps.  Because this is still a draft report, the toxicity risk factors
presented in this document should not be considered EPA’s official
estimates of dioxin toxicity but rather reflect EPA’s ongoing effort
to reevaluate dioxin toxicity.  Evidence compiled from this draft
reassessment indicates that the carcinogenic effects of dioxin and
dioxin-like compounds may be six times as great as believed in 1985. 
However, given the emission-adjusted MOEs representing the ECF boilers,
risks above the 1E-05 level of concern would not be predicted were the
2000 dioxin slope factor utilized for the risk management decision. An
assessment using this alternative value should not be considered Agency
policy. 

If the 2000 dioxin slope factor is considered, the risk safety margin
would be reduced.  However, given the emission-adjusted MOEs
representing the ECF boilers, risks above the 1E-05 level of concern
would not be predicted were the 2000 dioxin slope factor utilized for
the risk management decision.



Special Conditions for Emission-Comparable Fuel

	This section summarizes the conditions applicable to generators that
claim the ECF exclusion and to ECF burners.  For more information, see
the preamble to the proposed rule, Part Two:  Rationale for the Proposed
Rule

	ECF would be subject to all of the conditions that apply to existing
comparable fuel, including 

Constituent specifications in Table 1 to §261.38 with the exception of
26 hydrocarbons& oxygenates for which the specifications would not
apply.

Minimum heating value of 5,000 Btu/lb, as-generated, and maximum
viscosity of 50 cs, as-fired.

Prohibition of blending to meet the specifications.

Notifications to state RCRA and CAA directors and public notification
(see proposed §261.38(b)(2)

Waste analysis plans for generators and burners (see proposed §261.38
(b) (4) and (5)

Sampling and analysis conditions per proposed § 261.38(b)(6)

Prohibition on speculative accumulation (proposed § 261.38 b(7))

Recordkeeping (proposed § 261.38 (b)(8) and (9)

Burner Certification to generator per proposed §261.38 b(10), and 

Ineligible waste codes.

However, ECF must also meet additional conditions to ensure that:

It is stored and handled in a manner protective of human health and the
environment given the higher concentrations of particular toxic,
volatile hydrocarbons and oxygenates it may contain compared to fuel
oil; and

It is burned under good combustion conditions that would ensure that
emissions are comparable to emissions from fuel-oil combustion. 

Storage of Emissions Comparable-Fuels

Storage conditions for ECF are stipulated in §261.38 (c) (1) of the
proposed rule.  ECF must be stored in tanks (above ground or
underground), tank cars or tank trucks. 

Above-Ground Storage Tanks

ECF stored in above ground tanks is subject to:  (a) certain SPCC
requirements of 40 CFR Part 112; (b) additional secondary containment
requirements, and (c) air emissions regulations under the Organic
Liquids Distribution NESHAP.

 

(A) SPCC requirements.  Emission-comparable fuel tank systems are
subject to the following requirements under 40 CFR Part 112 as though
emission-comparable fuel meets the definition of oil under §112.2 

Section 112.7, General Requirements for Spill Prevention, Control, and
Countermeasure Plans, except for paragraph (c) (secondary containment)
and paragraph (d) (waiver of secondary containment). 

Section 112.8, Spill Prevention, Control, and Countermeasure Plan
Requirements for Onshore Facilities, except for paragraph (b) (facility
drainage), paragraph (c)(2) (secondary containment for bulk storage
containers), and paragraph (c)(11) (secondary containment for mobile
containers).

Section 112.20, Facility Response Plans.

Section 112.21, Facility Response Training and Drills/Exercises

These regulations describe requirements including development and
submittal of a SPCC plan including-

A facility description (describing location and contents of each
container).

Discharge prevention measures (such as high liquid level alarms) and
countermeasures for discharge recovery & methods for disposal of
recovered materials.

Inspections, tests and recordkeeping, 

Personnel training and 

Security. 

(B) Secondary Containment—Because ECF contains higher concentrations
of certain hazardous hydrocarbons and oxygenates than fuel oil and can
pose a higher risk to human health and the environment, ECF tanks must
be equipped with “engineered” secondary containment which is more
protective than the SPCC requirements of dikes, berms or retaining
walls.  Secondary containment must be provided by a liner, a vault, a
double-walled tank, or an equivalent device approved by the regulatory
authority that meets the specifications of proposed §261.38 (c) (1)(ii)
(B).  Additionally, ancillary equipment such as pumps, valves and piping
must be equipped with secondary containment unless they are visually
inspected on a daily basis.

(C) Air emissions.  ECF fuel storage tank systems would be subject to
the applicable air emission controls under the organic liquids
distribution (OLD) NESHAP codified in 40 CFR 63 Subpart EEEE except that
the following compounds must be considered in addition to the organic
hazardous air pollutants in Table 1 to Subpart EEEE when considering the
applicability of that subpart and the organic compounds that must be
controlled:

Allyl alcohol (CAS No. 107-18-6)

Bis(2-ethylhexyl)phthalate [Di-2-e thylhexyl phthalate] (CAS
No.117-81-7)

Butyl benzyl phthalate (CAS No. 85-68-7)

Diethyl phthalate (CAS No. 84-66-2)

2,4-Dimethylphenol (CAS No. 105-67-9)

Dimethyl phthalate (CAS No. 131-11-3)

Di-n-octyl phthalate (CAS No. 117-84-0)

Endothall (CAS No. 145-73-3)

Ethyl methacrylate (CAS No. 97-63-2)

(2-Ethoxyethanol [Ethylene glycol monoethyl ether] (CAS No. 110-80-5)

Isobutyl alcohol (CAS No. 78-83-1)

Isosafrole (CAS No. 120-58-1)

Methyl ethyl ketone [2-Butanone] (CAS No. 78-93-3)

1,4-Naphthoquinone (CAS No. 130-15-4)

Propargyl alcohol [2-Propyn-1-ol] (CAS No. 107-19-7)

Safrole (CAS No. 94-59-7); 

The OLD NESHAP requires tanks that are larger than 5000 gallons, and
that contain organic liquids that exceed specified vapor pressure limits
(that are a function of tank size), to meet certain emissions limits or
work practice standards.  This can be accomplished by certain types of
air emissions controls that are analogous to the “level 2” air
emissions controls required for hazardous waste storage tanks under 40
CFR 264 subpart CC.  These controls may include:

An internal floating roof (as described in Part 63, Subpart WW)

External floating roof (Part 63 subpart WW) 

Tank vented through a closed vent system to a control device (Part 63,
Subpart SS)

	Alternatively, tanks emissions can be either routed back to a fuel gas
system or the process (per Subpart SS), or controlled with a vapor
balancing system that complies with §63.2346 (a)(4).

Underground Storage Tanks

Underground storage tank systems managing ECF are subject to the
requirements under 40 CFR Part 280.  These regulations include
requirements for leak detection, spill and overfill protection and
corrosion protection.  Additionally, new USTs must have secondary
containment.  

Combustion of Emissions-Comparable Fuel

Combustor Type

As prescribed in the proposed rule at §261.38(c)(2)(i), ECF must be
burned in a industrial or utility boiler.  Additionally, the boiler must
be a watertube, steam boiler that is not a stoker. 

	Combustion in firetube boilers or stokers would not be allowed. 
Firetube boilers are smaller and have a relatively higher surface to
volume ratio which makes for colder burner temperatures due to higher
heat loss.  Therefore, there may be a greater potential for localized
coal spots and poorly mixed zones which can result in poor combustion
conditions.  This is also true for stokers because they burn fuel with
large particle size. Also, stokers in general have relatively higher
emissions of CO and organics, evidence of less than optimum combustion
conditions. 

	Burning of ECF would not be allowed in process heaters either because
process heaters often have operating practices  (such as quenching
combustion gases to avoid overheating a process liquid) that could
reduce combustion efficiency and result in higher emissions of PICs.

	Also, ECF boilers must be industrial or utility boilers as defined in
proposed §261.38(b)(2)(i)(B).  ECF could not be burned in commercial or
institutional boilers (e.g., boilers at hospitals, schools) for the same
reasons that existing comparable fuel cannot be burned in those units. 
See 63 FR at 33798.  Burning in industrial or utility boilers would
ensure that ECF was burned in a unit subject to Federal/State/local air
emission regulations and that was capable of handling excluded fuel
(and, for ECF, was capable of complying with the conditions on burning
ECF)..

Operating Conditions

The burning of ECF must be conducted under boiler operating conditions
that ensure a hot stable flame and are consistent with good combustion
practices.  These conditions include:

Fossil fuel as primary fuel.  A minimum of 50% percent of fuel fired to
the device shall be fossil fuel, fuels derived from fossil fuel or tall
oil.  The 50%  primary fuel firing rate shall be determined on a total
heat or volume input basis, whichever results in the greater volume of
primary fuel fired;

Fuel heating value.  Primary fuels and emission-comparable fuel shall
have a minimum as-fired heating value of 8,000 Btu/lb, and each material
fired in a firing nozzle where ECF is fired must have a heating value of
at least 8,000 Btu/lb, as-fired;

CO CEMS.  When burning emission-comparable fuel, carbon monoxide
emissions must not exceed 100 parts per million by volume, over an
hourly rolling average (monitored continuously with a continuous
emissions monitoring system (CEMS)), dry basis and corrected to 7
percent oxygen.  

Dioxin/furan control.  Boilers equipped with a dry air pollution control
device must monitor the combustion gas temperature at the inlet to the
initial dry particulate matter control device, and the gas temperature
must not exceed 400°F on an hourly rolling average.  As discussed in
Section 5.3.1 above, heterogeneous surface-catalyzed formation of PCDD/F
is increased at temperatures above 400 °F.  Note that the temperature
limit is not required for boilers firing coal as primary fuel because
sulfur in coal is known to inhibit formation of dioxins and furans.  

Automatic fuel cutoff system. The boiler must be equipped with a
functioning system that immediately and automatically cuts off the
emission-comparable fuel feed when:  (1) the emissions limit for CO or
the APCD inlet temperature limit is exceeded; (2) if the CO CEMS or the
gas temperature detector malfunctions; and (3) when any component of the
automatic fuel cutoff system malfunctions. Occurrences of automatic feed
cutoffs must be investigated, corrective measures taken, and findings
recorded in the operating record.  Excessive exceedances must be
reported to the Administrator as required under § 261.38(c)(2)(F)

Boiler load.  Boiler load shall not be less than 40 percent.  Boiler
load is the ratio at any time of the total heat input to the maximum
design heat input; 

ECF must be fired into the primary fuel flame.  This requirement is to
ensure that ECF does not completely bypass the flame.  If this were to
occur, potential exists for a combustion failure that would not be
detected by an increase in CO emissions.

Fuel atomization.  The emission-comparable fuel shall be fired directly
into the primary fuel flame zone of the combustion chamber with an air
or steam atomization firing system, mechanical atomization system, or a
rotary cup atomization system under the following conditions:

(a) Particle size.  The emission-comparable fuel must pass through a 200
mesh (74 micron) screen;

(b) Mechanical atomization systems.  Fuel pressure within a mechanical
atomization system and fuel flow rate shall be maintained within the
design range taking into account the viscosity and volatility of the
fuel;

(c) Rotary cup atomization systems.  Fuel flow rate through a rotary cup
atomization system must be maintained within the design range taking
into account the viscosity and volatility of the fuel.

Restrictions on benzene and acrolein.  If the as-fired concentration of
benzene or acrolein in the emission-comparable fuel exceeds 2 percent by
mass, the firing rate of emission-comparable fuel cannot exceed 25% of
the total fuel input to the boiler on heat or volume input basis,
whichever results in a lower volume input of ECF.

More on Atomization

	Atomization, the process of breaking up a liquid stream into small
droplets, ensures rapid evaporation and thereby fast and efficient
combustion of the stream by increasing the surface area to volume ratio
of the fluid.  As a condition of the exclusion, the ECF firing system
must provide proper atomization to ensure that ECF droplets are
sufficiently small.  The acceptable atomization systems for ECF burners
are those that are commonly used for firing liquid fuels.  These
include:

Air or steam atomization (twin fluid atomization)

Mechanical or pressure atomization.

Rotary cup atomization

Each of these atomization systems are briefly discussed below.

	Twin Fluid Atomization. In twin fluid atomization systems, either air
or steam is used to break up the liquid fuel into a fine mist.  When
steam is used for atomization, it can also serve to heat the liquid fuel
and thereby reduce its viscosity.  In addition to subdividing by type of
atomizing fluid used, these nozzles may also be categorized as
high-pressure versus low pressure.  In high pressure nozzles, steam or
air (at pressures of 30-150 psig) impinges on the liquid stream at high
velocity.  The atomizing fluid requirements are 20-200 ft3 air or 2 to 5
lbs of steam per gallon of liquid fuel.  These atomizers can handle
fluids with viscosities as high as 150-5000 SSU (33 -1100 cS). 

	In low pressure, air atomized systems, compressed air is provided at
1-5 psig.  The quantify of atomizing air required is higher than that
for high pressure nozzles, and is in the range 400 to 1000 ft3 per
gallon of fuel.  The higher steam/air flowrates used in these nozzles
cause shorter flames and thereby they require smaller combustion
chambers when compared with the high pressure nozzles. 

	Two other kinds of twin fluid nozzles that are occasionally used are
internal mix nozzles, where the atomizing medium is introduced within
the nozzle and impinges on the liquid stream prior to discharge, and
sonic nozzles, which use a compressed gas to create high-frequency sound
waves that break up the liquid stream.

	Mechanical Atomization. Mechanical atomizers do not require an
atomizing fluid such as steam or air but rather direct the fuel through
small discharge orifice to create small droplets.  The liquid is pumped
at a relatively high pressure and is given a strong “cyclonic”
velocity (from internal tangential guide slots) before it is sent
through the orifice.  The nozzle typically provides a conical spray
pattern and combustion air is provided on the periphery of the cone. 
The combination of combustion air and the action of the swirling fuel
produces effective atomization.  The turndown ratios are usually on the
order of 3:1 but higher turndown ratios can be obtained by using a
return flow for the liquid fuel.  Liquids with moderate viscosities may
be atomized to sufficiently small droplets with pressures in the range
75-150 psig but higher viscosity liquids will require higher pressures. 
Additionally, since low volatile fuels would need to be atomized to
comparatively smaller droplets to ensure efficient combustion, these
fuels would require a comparatively higher atomization pressure.

	

	In general mechanical atomizers are used with fuels with viscosity
below 100 SSU (~22 cS).  In comparison to other types of atomizers,
mechanical atomizers are more susceptible to erosion and plugging from
solids in the liquid stream.

   

	Rotary-Cup Atomization: In this configuration, the liquid stream is
atomized by discharging it centrifugally from a rotating cup.  The
rotary cup is mounted on a hollow shaft and rotates at speeds up to
several thousand revolutions per minute.  Liquid is torn from the lip of
the cup in the form of thin conical sheets.  Air is introduced through
an annular space around the rotating cup.  Since rotation rate and
combustion air flow are the main factors impacting atomization not much
liquid pressurization is required.  Rotary cup atomizers can handle
fuels with relatively high viscosities up to 170 - 300 SSU (40 to 72
cS), and they are relatively insensitive to solid impurities in the
fuel, and can handle waste fuels with solids with maximum particle size
below 35-100 mesh.

	Figure 6-1 shows a schematic of commonly used atomization systems.

Other Conditions

ECF may not be managed by any entity other than the generator,
transporter, and designated burner to facilitate compliance assurance. 
Also, additional requirements apply to ECF generators and burners
(beyond what is required for generators and burners of comparable fuels)
regarding notifications, reporting, and recordkeeping.  These are listed
in the proposed rule at §261.38(c)(4)  

	

Figure 6-1: Commonly-Used Atomization Systems for Liquid Fuels (Adapted
from Brunner (1989)

               

Engineering Costs and Savings

This section discusses waste quantities qualifying as emissions
comparable fuel and savings and costs incurred by generators and burners
of ECF under various regulatory options EPA considered for tracking and
storage of ECF.  Note that this analysis is limited to waste streams
identified by respondents to a survey by the American Chemistry Council.
 EPA also conducted an independent analysis of costs and savings.  See
USEPA, “Assessment of the Potential Costs, Benefits, and Other Impacts
of the Expansion of the RCRA Comparable Fuel Exclusion—Proposed
Rule,” May 2007. 

Regulatory Options

	Three primary regulatory options were considered for analysis.  Under
all options, generators and burners must provide a one time notification
to the Regional Administrator.  Offsite burners must also provide a one
time certification to the generator that they will comply with the
storage and burner conditions.

Option A (Currently Applicable Product Controls):  No specific controls
for storage; any existing controls for commercial products would apply.
Records must be kept of each shipment to an off-site burner.  DOT
shipping papers may be required for some shipments (i.e., for ECF that
meets the DOT definition of a hazardous material (HAZMAT)). 

Option B (Tailored Management Conditions):  Storage requirements per
Spill Prevention and Control Countermeasure (SPCC) regulations in 40 CFR
112 that apply to fuel tank systems, other than secondary containment. 
These include tank integrity assessment and tank inspections, tests and
recordkeeping.  Generators and burners must also conduct personnel
training and develop an SPCC plan including a response plan. 
Additionally, ECF storage tanks would need “engineered” secondary
containment:  an external liner, vault, or use of double walled tank. 
Further, certain storage tanks would need Level 2 air emissions controls
based on requirements of the organics liquid distribution (OLD) NESHAP. 
Similar to Option A, records must be kept of each shipment to an
off-site burner, and DOT shipping papers may be required for some
shipments.  Offsite burners must provide a one time certification to the
generator that they will comply with the storage and burner conditions. 

Option C (Subtitle C Storage/tracking) – ECF would be stored, tracked,
and transported as hazardous waste subject to RCRA Subtitle C storage
and tracking requirements.  This requirement means that the point of
exclusion would be moved to the ECF burner nozzle.  Thus, 40 CFR 265,
Subpart J tank standards would apply to storage of ECF by generators who
accumulate ECF for less than 90 days, and the permit standards under
Part 264 (including the Subpart J tank standards) would apply to
generators who accumulate for more than 90 days and to burner storage
units.  The Part 265 and Part 264 Subpart J tank standards require
engineered secondary containment virtually identical to Option B. 
Additionally, tank systems would be subject to the RCRA air emissions
controls under Subparts BB and CC of Parts 265 or 264.  The Part 264
standards would also require off-site burner storage units (and
generators who accumulate ECF for more than 90 days) to comply with
closure and financial assurance requirements and groundwater monitoring
requirements and to obtain a RCRA storage permit.  As discussed in
Section 7.5.3 below, we analyzed savings for options C under two
different scenarios.

Analysis of ACC Survey

As discussed in Section 1, the ACC survey provided detailed information
on 95 hazardous waste streams that are currently not qualifying for the
comparable fuels exclusion.  However, only 76 of these streams were
subject to further analysis because:

17 streams did not contain sufficient information on heating values
and/or annual quantities generated.

Two wastestreams, N-01-13 and N-01-14 are currently not handled as a
hazardous waste.

Many of these 76 streams were failing the current comparable fuel
specification for only one or two § 261.38, Table 1 constituents. 
However, the survey did include many streams that failed for
constituents other than oxygenates and hydrocarbons and would not
qualify as ECF.  Table 7-1 provides a summary of failing constituents
for each of the 76 streams.  Forty three of these streams are failing
the specifications for metals or halogenated, nitrogenated, or
sulfonated organics or Hazard Category A hydrocarbons.  See discussion
of the Relative Hazard Categorization Scheme in Section 2.4 above. 
These wastestreams would not qualify as ECF.

	

	As for wastestreams failing the specification for oxygenates or
hydrocarbons, the table shows which specific constituents are failing
the specifications.  Fifteen streams fail the current specifications for
toluene, ten fail for isobutyl alcohol, nine fail for methyl ethyl
ketone (MEK), and six fail for benzene.  

Table 7-1: Classification of Wastestreams from ACC Survey.

Estimate of Qualifying ECF Quantities

Based on information provided in the previous section, we estimated the
quantity of hazardous waste that would qualify as emissions-comparable
fuels by screening out streams that would not meet the definition of
ECF.

Forty three streams would not be eligible as ECF because they fail the
specifications in 40 CFR 261.38, Table 1 for compounds other than the
hydrocarbons and oxygenates for which the specifications would not
apply.  These streams fail the specification for compounds such as
metals, halogenated organics, or Relative Hazard Category A
hydrocarbons.  These streams amount to 121.1 million lbs per year (or
60,600 tons/yr) with a weighted average heating value of 10,500 Btu/lb. 
These will continue to be handled as hazardous waste.  See Table 7-2. 
Note that only 6 waste streams totaling 5200 tons/yr would be ineligible
as ECF solely because they fail the specification for a Hazard Category
A hydrocarbon. 

Thirty three streams meet the definition of emissions-comparable fuel. 
These 33 ECF streams amount to 253.4 million lbs/yr (or 126,700
tons/yr), are generated at 16 sites representing eight different
companies.  These streams have a combined thermal value of 2.97 x 106 MM
Btu/yr or a weighted average heating value of 11,700 Btu/lb.  Table 7-2
shows a stream-by-stream breakdown of these results.  The amount of
qualifying comparable fuel is dominated by two streams (one at 118
million lb/year and the other at 56 million lb/yr) together accounting
for more than 60% of the qualifying waste.

Preliminary Information Used in Costs/Savings Estimate

Value of Fuel Oil and ECF

The Department of Energy’s Energy Information Administration gathers
information on fuel oil pricing.  March 2006 prices for #2 fuel oil and
residual fuel oil were $1.79/gallon and $1.19/gallon respectively which
was a 33% increase (on average) from March 2005 levels.   The March 2005
price for fuel oil #4 was listed as $1.16/gallon (this is the latest
information available for fuel oil # 4).  Based on the average increase
for other types of fuel oil we can estimate the March 2006 price for
fuel oil #4 at $1.54/gallon.  This cost is equivalent to $426/ton or
$11/ MM Btu assuming a heating value of 139,400 Btu/gal.  

Table 7-2: ACC Survey based ECF Waste Quantities.

We assumed that the value of ECF as a fuel would be similar to that of
fuel oil # 4 on a thermal input basis.  However, given that ECF is a
conditionally-excluded hazardous waste subject to substantive storage
and burner conditions (and liability if those conditions are not met),
and may contain an increased level of some hazardous constituents when
compared with fuel oil, we estimate that the value of ECF would be 75%
of that of fuel oil and would be equivalent to $11* 75% or $8.25
$/MMBtu.

Hazardous Waste Disposal Costs

The hazardous waste disposal costs currently being incurred by some
generators for their ECF is estimated to be $0.96/gal.  This is the 2004
average tipping fee for bulk, liquid, nonhalogenated waste charged by
commercial incinerators.  Assuming ECF has a density equivalent to that
of water, we can estimate the disposal costs to be $231/ton.  This
disposal cost is not inconsistent with the limited information available
in the ACC survey which shows disposal cost information for five
streams.  The weighted average disposal cost is $231/ton, although a
range from $130/ton to $270/ton was seen.

Labor Costs

Labor costs are taken from the BIF rule ICR.  The rates are adjusted to
March 2006 levels by increasing the ICR rates based on the increase in
consumer price index.  As Table 7-3 shows, the rates vary from about
$34/hour for clerical support to $119/hour for legal/consultant support.

Estimation of the Number of ECF Shipments

For generators who would ship their ECF off-site for burning, we are
assuming that shipments occur when the volume of each stream reaches
9000 gallons.  However, we are also assuming that each stream is shipped
at least once every 90 days to satisfy accumulation time requirements
for 90-day HW generator tanks under §262.34.

So for a given generator site,

 

	Appendix E shows the estimated number of shipments by generator site
for each qualifying waste stream.  

Estimation of Tank Size for Generators and Burners

In order to estimate storage costs, it was necessary to estimate the
tank size for burners and generators.

Generators:  For generators, we assume that each ECF stream will be
stored in a separate tank.  Therefore:

Number of ECF tanks = Number of ECF streams

There is a large range of production rates among individual ECF streams.
 For example, the 90-day production rate varies from 500 gallons to 3.5
million gallons. Although facilities producing very large quantities can
ship and or transfer their waste to the burner frequently (even multiple
times a day) the generator must have storage capacity to account for
burner downtime and other contingencies (such as shipment delays). 

If combusting at onsite burner:  We assume the size of the generator
tank should be at least sufficient to store 2 weeks of ECF.  So:

 	

Tank Size = Biweekly ECF stream generation rate (gallons per 2 weeks) *
2

If combusting at offsite burner:  The 2 week storage minimum applies but
also consider that generators must ship at least once every 90 days.

If generation rate is greater than 9000 gallons/ 2 weeks, 

Tank Size (gallons) = 2* Biweekly generation rate (gallons)

If generation rate is less than 9000 gallons every 3 months:       

                                                       

Tanks Size (gallons) = 2* 90 day generation rate (gallon)

For intermediate generation rates:

Tank Size (gallons) = 2*9000 

Tank size is rounded up to the nearest 1000 gallons.  Sizes vary from
2000 to 1.1 million gallons.

Burners:  We estimate that all the ECF streams from a single generating
site will go to a single boiler.  And, we assume that this burner will
store all the ECF streams together in a single tank.  Assuming that the
burner tank needs to be large enough to hold a couple of days of the
aggregate ECF:

Burner Tank Size (gallons) = 2* 2* Total ECF sent to burner in a day
(gallons)

	Based on these estimates burner tank size varies from 2000 to 160,000
gallons

The results of these analyses are shown in Appendix E.

Generator Types and ECF burn Scenarios

	Based on the ACC survey, we ascertained that 16 sites representing
eight companies had qualifying ECF. These sites could be categorized
into two groups. 

Group 1:  There are nine sites with qualifying ECF currently incurring
disposal costs, i.e., these sites are currently paying a fee (to
incinerators, cement kilns or fuel blenders) to dispose of ECF as
hazardous waste.  These sites are D-01, D-02, D-03, D-04, F-01, J-05,
L-02, L-06, and M-03.  Group 1 sites taking the ECF exclusion will save
on hazardous waste disposal costs.  Additionally, since ECF will
displace fossil fuels burners will have fuel savings.  These generators
and burners will incur costs for recordkeeping and storing ECF,
analyzing ECF for toxic constituents, and for installing burner
controls. These costs, as discussed in the next subsection, will vary by
the regulatory option being considered. 

Group 2:  There are seven sites with qualifying ECF that we determined
are not currently incurring disposal costs.  These sites will not
realize any savings for displaced fuel or hazardous waste disposal but
they will have other cost savings such as storage.  Based on the
information provided, we assumed three different scenarios for these
sites. 

Scenario 1:  ECF is burned onsite in a boiler that remains a hazardous
waste boiler. (K-01, K-04)

Scenario 2:  ECF is burned onsite in a boiler that exits the HW/MACT
regulatory system. (B-01)

Scenario 3:  ECF is burned offsite at an affiliated boiler (i.e., a
boiler owned by the parent company) that remains a HW boiler. (B-02,
D-05, D-06, G-01)

Costs/Savings Estimation

	The costs and net savings to ECF generators/burners depend on the
regulatory option as well as the classification discussed in Section
7.4.  Primary assumptions used are:

All qualifying ECF generated at a single site will be burned at a single
steam watertube boiler that would be modified to meet ECF combustion
control requirements.  We assume that for each generating site such a
boiler is located close enough to the generator site that makes
transportation of ECF to the boiler economically feasible.

The boiler may be located onsite or offsite.  Based on the ACC survey it
was determined that the following facilities will send their ECF off
site for combustion:  B02, D05, D06, G01, D01, D04, F01, J05, L06, and
D02.

	Group 1 and Group 2 sites are handled differently for purposes of
estimating savings.  The itemized cost estimates were derived from
information collection request (ICF) supporting statements for several
different rules, review of information available from EPA, DOE and other
agencies, and vendor quotes.  Reasonable assumptions were made in cases
where the information was not readily available.  Appendix F lists unit
cost/savings for each item below as well as assumptions and information
sources used.  Appendix G contains the cost model that was developed
based on these assumptions.  

Details of the cost and savings analysis are provided below.

Group 1 Sites: 

      

Savings:

Fossil fuel savings assuming ECF is valued at $8.25/MMBtu.

Savings for not managing ECF as hazardous waste - $ 275/ton (includes
$231/ton for disposal, $14/ton for manifesting, and $30/ton for
generator storage.)

Costs

Costs vary by regulatory option and include costs to comply with the
regulatory requirements as well as costs deemed to arise from standard
operating procedures (SOP).  The cost model groups the costs into five
primary components.  (1) boiler costs associated with retrofitting a
typical fuel oil boiler to meet the requirements of an ECF boiler; (2)
analytical costs incurred to test candidate ECF streams for the
specification constituents and properties; (3) tracking costs for
shipping papers/manifests and maintaining records; (4) generator storage
costs to replace, inspect, and maintain storage tanks, provide secondary
containment and air emissions controls; and (5) burner storage costs. 
These components are discussed in detail below.

Boiler-  

Costs apply equally to all regulatory options.

Carbon Monoxide Monitor – Assume one time installed cost of $5800 per
boiler annualized over 10 years.

Automatic Feed Cutoff - Assume one time installed cost of $3800 per
boiler annualized over 10 years.

Firing Nozzle – Assume 50% of boilers would require a new firing
nozzle. One time cost (annualized over ten years) of $350 for an ECF
feedrate of 3.3 gallons per minute.  Costs scale with ECF feedrate.

Costs for maintenance, calibration, and recordkeeping related to CO CEMS
and ECF feed systems and automatic feed cutoff – 1.75 hours per day or
$30,400 annually per boiler.

 

Analytical 

Analytical costs apply equally to all options

Testing – Assume $8100 per test per ECF stream for analyzing all
261.38 constituents (including volatile and semi-volatile organics,
metals, PCBs, total nitrogen etc.) and for evaluating other properties
such as sulfur and ash content, heating value, and viscosity.  The cost
takes into account that iterative sampling may be required to achieve
low detection limits. Assume 50% waste streams are tested annually,
while the rest are tested semi-annually. 

Recordkeeping – Assume $1200 per test for maintaining lab
documentation of testing (test methods, quantization limits, QA/QC
records etc) and for certifying certain compounds are not in ECF based
on process knowledge.

Tracking – 

Tracking costs vary by regulatory option and include:

Shipping papers - Every shipment under Option C will require hazardous
waste manifests at a cost of $50 per manifest.  Assume all shipments
will be accompanied by DOT shipping paper (at half the cost of a
hazardous waste manifest) under Options A & B.  

One-time notification to the Regional Administrator by the generator and
off-site burner at $160 per notification annualized over 20 years
(Options A& B)

Transporters and offsite burners obtain EPA ID number at a one time cost
of $470 annualized over 20 years(Option C)

One-time certification by an off-site burner to the generator that
burner is in compliance with storage and burner controls at a one time
cost of $264 annualized over 20 years. (all options)

Recordkeeping for each shipment at $30 per shipment. (all regulatory
options)

Generator Storage 

A key assumption for the cost estimate for both generator and burner
storage is that ECF will be stored in above ground tanks only. 
Generator storage costs also vary by regulatory option and include: 

Replacement of storage tanks at an installed cost of $69,000 for a
20,000 gallon tank.  Cost for other tanks sizes are calculated using a
0.6 scaling factor.  The costs are annualized over 15 years.  (all
regulatory options)

Secondary containment for storage tanks.  Earthen berm for Option A (as
standard operating procedures) at $1300 for a 1000 gal tank.  Engineered
secondary containment with leak detection for Options B/C at cost of
$30,700 for 20,000 gallon tank.  Cost for other tanks sizes are
calculated using a 0.6 scaling factor and the costs are annualized over
15 years.

Piping to from generator tanks to burn tank (assume 500 ft) - Applies
for all options but only for onsite combustion at a cost of $4100 per
tank annualized over 10 years.

Daily visual inspection of tanks at a cost of $8 per inspection per tank
(all options)

Weekly visual inspection of  pumps, valves and piping at a cost of $17
per inspection per tank (all options)

Monthly monitoring for VOC leaks at cost of $150  per test per tank (all
options)

Professional engineer’s certification of major repairs – Assume $490
per repair and assume each tank has one major repair every 5 years. (all
options)- 

Cessation/containment of leaks- Leaks must be cleaned up as hazardous
waste Assume three leaks per generator site per year as HW and labor and
disposal costs total $1570 per leak.(all options)

Notification and reporting of leaks to Regional administrator at $1020
per leak (Option C)

SPCC Plan – For review of SPCC plan and certification by PE and
submitting response plan to RA (Option B).  Plans are reviewed every
five years. Equivalent costs are $289 per year per generator site.

Air emissions controls - Cost of Level 2 air emissions control are
estimated to be $20,100 per 20 gallon tank annualized over 15 years. 
Assume a fraction of tanks will install Level 2 controls for air
emissions for their replacement storage tanks under the various options.


50% under Option A (currently existing air emissions controls
requirements),

75% under Option B (requirements are per organics liquids distribution
(OLD) NESHAP, in 40 CFR 63 Subpart EEEE), and 

75% under Option C (40 CFR 265 Subpart CC)

Recordkeeping related to air emissions-$1400 per tank per year for
maintaining inspection records, tagging defective equipment etc. (all
options)

Operator Training – Assume 40 man hours of training annually at a
total cost of $3600 per generator site (all options)

Burner Storage – 

The following burner storage costs are identical to those of generators:

Cost of new storage tank for each burner. (all options)

Secondary containment (earth berm for Option A, engineered materials for
Options B/C)

Daily visual inspection of tanks (all options)

Weekly visual inspection of  pumps, valves and piping (all options)

Monthly monitoring for VOC leaks (all options)

PE certification of major repairs (all options)

Cessation/containment of leaks as HW (all options)

Notification and reporting of leaks to RA (Option C)

SPCC Plan – Review SPCC plan and certify by PE.  Submit response plan
to RA (Option B)

Air emissions controls (Assume 50% will install Level 2 air emissions
controls under Option A, and 75% will install controls under Option B/C
on replacement storage tanks)

Recordkeeping related to air emissions (all options)

Operator training (offsite burners- all options)

The following costs will differ from those for generator storage tanks:

Piping from burn tank to burner- $410 per burner annualized over 10
years. (all options)

Closure /financial assurance (Option C) – Assume facilities will
estimate cost of closure and make annual payments to a closure trust
fund.  Assume total estimated closure cost for site with 1 million
gallon ECF capacity is $1.48 million.  Smaller capacity sites’ costs
are estimated using a 0.6 scaling factor.  Costs include those for
flushing the tanks, decontamination, demolition, and removal of
containment systems, decontamination and removal of soil.  Annual
payment to a closure trust fund is estimated by annualizing total
closure costs over 20 years.  Also, assume $1800 per year for annual
amendments to closure plan.

Groundwater monitoring (offsite burners, Option C) – Assume detection
monitoring only,i.e., compliance monitoring and corrective action not
required.  Assume capital costs for groundwater monitoring system of
$143,000 (annualized over 15 years) and $40,300 per year for labor and
O&M costs.

Permitting - (offsite burner, Option C) - Requirements for part B permit
application including waste analysis plan, closure/financial assurance
plan, certification of tanks and equipment.  One time cost $171,500 for
labor and O&M annualized over 10 years.  Assume permit renewal costs in
10 years are 50% of that of the original permit.  This is equivalent to
$18,300 per year for each offsite burner.

Group 2 Sites:

Costs/Savings depend on scenario and regulatory option.

Scenario 1- ECF is burned onsite in a boiler that remains a hazardous
waste boiler. 

                    

Savings- 

Secondary containment on replacement storage and burn tanks (Option A)

Partial air emissions control costs (Option A)

Permit renewal costs for burn tanks (Option A, B)

                    

Costs- 

SPCC Plan renewal, certification, and submittal (Option B)

Scenario 2 – ECF is burned on site in a boiler that that would cease
being a part of the RCRA/MACT regulatory system

                      

Savings-   

BIF/MACT compliance related – Assume $65,300 per year savings due to
elimination of BIF/MACT compliance costs.   (All regulatory options) 

Secondary containment on replacement storage and burn tanks (Option A)

Partial air emissions control costs (Option A)

Permit renewal costs (Option A, B)

Costs- 

SPCC Plan renewal, certification, and submittal (Option B)

Scenario 3- ECF is burned offsite in an affiliated boiler that remains a
hazardous waste boiler. 

Savings- 

Tracking (difference between costs of manifests and shipping papers,
Option A-B)

Secondary containment on replacement storage and burn tanks.(Opt A)

Partial air emissions control costs (Option A)

Permit renewal costs for burn tanks (Opt A-B)

Costs- 

SPCC Plan renewal/certification and submittal (Opt B)

Additional Assumptions for Option C for Offsite Combustion of ECF

Given that Regulatory Option C would require offsite boiler to obtain a
storage permit for ECF, we calculated savings estimates under two
scenarios:

Low End Savings Estimate – Assume no offsite shipment of ECF. 
Facilities that would otherwise ship their ECF to offsite facilities
will not claim the exclusion.

High End Savings Estimate – Assume limited offsite shipments of ECF. 
Some generators will claim the exclusion even if they do not have
quantifiable savings for the intangible benefits associated with dealing
with less hazardous waste.  Estimate also assumes commercial HWCs
(incinerators and kilns) will reduce their fees rather than lose revenue
from these streams.

The following sites’ ECF streams are currently burned in an affiliated
boiler and would continue to do so:  B02, D05, D06, and G01 – These
sites claim the exclusion but incur no quantifiable savings.

The following sites’ streams are currently burned in an offsite
commercial HWC and generators would claim the exclusion and continue to
send ECF to the HWC under reduced fees:  D01, D04, F01, J05, and L06 –
These sites will incur savings of disposal costs. 

The following site will not claim the exclusion: D02

Results and Discussion

The results of savings estimates are shown in Table 7-4.  The results
are shown for all three regulatory options and are also broken down for
Group 1 and Group 2 sites. For each option, the number of companies,
sites, and ECF streams that are estimated to take the exclusion are
shown.

	The quantity of waste that is excluded would be 126,700 tons/year under
Regulatory Options A and B.  Under Option C the excluded waste quantity
will vary between 25,600 to 125,200 tons/yr.  The table also shows fuel
and hazardous waste management savings for Group 1 sites and other costs
or savings (note-table shows costs in parenthesis) for both Group 1 and
2 sites.  A site-by-site breakdown of these figures can be found in
Appendix H. 

 	Net savings to generators and burners are also shown in the table. 
Under Option A, net annual savings are $7.73 million with Group 1 sites
seeing 96% of these savings. Under Option B, slightly higher costs for
storage will result in slightly lower net savings of $7.58 million per
year or 98% of Option A savings.  Annual savings under Option C range
from $2.2 to $3.9 million (or 29-51% of option A savings).  The
reduction in savings is due to:–

Reduction in amount of waste excluded

Increase in number of generators that incur no savings despite taking
the exclusion.

Limitations of the Analysis

As noted earlier, the analysis in this section is limited to information
provided in the ACC survey.  The survey contained information on
generators that produce 187,000 tons/year of hazardous waste.  However,
according to EPA’s Biennial Reporting System, over 30 million tons of
hazardous waste was generated in 2003. A national estimate based on the
2003 Biennial Report and the 1996 hazardous waste constituent survey is
presented elsewhere.

Additionally, the analysis in this section does not account for
transportation costs/savings.  Note that generators who currently burn
ECF onsite and would ship ECF offsite under the exclusion will incur
transportation costs while others that are currently burning or
disposing of the waste offsite may see some savings on transportation
expenses. 

Appendices

DRE versus CO/HC: Supplement to 1997 TSD

Analysis of Nondetect Organic Emissions

Outlier Analysis for Hazardous Waste Boiler Organic Emissions

Calculations for DF Comparative Risk Assessment

ECF Tank size and Number of Shipments

Assumptions used for Savings Estimate

Cost Model for Economic Analysis.

Generator Savings by Option

 See section 2.1 for a definition of ECF.

 Each of these five groups included Volatile Organic Compounds (VOC) and
Semi-Volatile Organic Compounds (SVOC)

 See USEPA, “Final Technical Support Document for HWC MACT Standards
Volume VI: Development of Comparable Fuels Specifications”, May 1998.

 Method 8240 has been discontinued and replaced by Method 8260. The
newer method uses a capillary GC instead of a column GC.

 Gasoline has a higher fraction of volatiles than fuel oil and matrix
interferences required larger dilution factors for the samples.
Additionally, more dilution was required for gasoline to quantify one
the hydrocarbon VOCs (toluene) within the GC/MS calibration range.  The
higher dilution factors result in high detection limits for nondetect
substances.  EPA believed that waste fuels are likely to have a
composition more similar to fuel oil than gasoline, and thus the matrix
interferences for waste fuels would be more similar to fuel oil than
gasoline.  Therefore, EPA believed it was appropriate to base the
specifications for VOCs that were not detected in gasoline on fuel oil
QLs rather than gasoline QLs.

 USEPA, “Spring 2006 Regulatory Agenda”, Sequence # 3245- Expanding
the Comparable Fuels Exclusion under RCRA.

 Letter from American Chemistry Council (Carter Lee Kelly, Leader, Waste
Issues Team, and Robert A. Elam, Director, Regulatory Affairs, Waste
Issues Team) to Robert Springer and Matt Hale, USEPA, dated November 24,
2003.

 Not all information was provided for every waste stream. In fact about
20% of these waste streams had insufficient information for further
analysis. 

 Only a summary of the combustion control requirements are presented
here.  See the preamble to the proposed rule Part 2; Section IIB for a
complete description. Also see Section 6 of this technical support
document.

 	Note that hazardous waste may be treated by bona fide means (e.g.,
other than blending) to achieve a heating value of 5,000 Btu/lb.  Note
also that ECF must have an as-fired heating value of 8,000 Btu/lb. 
Consequently, ECF may be blended with other fuels to achieve the 8,000
Btu/lb as-fired heating value.

 This list of substances for which specification would be waived
includes all 22 oxygenates and two of the thirteen hydrocarbons (benzene
and toluene) that are in Table 1 of §261.38.  Note that the
specifications are unchanged for the remaining 11 hydrocarbons, other
organics (nitrogenated, sulfonated, and halogenated organics), metals
and other compounds listed in the table. 

 See discussion in Section 3.1.1

 Note that Naphthalene is also sometimes classified as a PAH. See also
Section 2.4

 This classification is based on the test methods used to measure the
constituent during the development of the specifications.  

 In a few instances the vapor pressure values were available
temperatures other than 25 C and were corrected to 25 °C assuming that
vapor pressure varies linearly with temperature.

 The one exception was endothall which had an estimated heating value of
7500 Btu/lb.

 Dellinger B., & Talor P.H., “ Designating Principal Organic Hazardous
Constituents,” App D in USEPA “Guidance on Setting Permit Conditions
and Reporting Trial Burn Results, 1989, EPA/625/6-89/019

 Brunner, C.R., 1988 “Handbook of Hazardous Waste Incineration”, TAB
Books 

 T99(2)This is the temperature required to achieve 99% DRE in two
seconds. Note-The compound with the highest T99(2) values is ranked #1
and so on.

 This Section was written by Tim Taylor, Environmental Scientist at EPA
OSW’s Economics Methods and Risk Analysis Division.

 The specific use of the current version of the WMPT rankings in
developing the RCRA Priority Chemical List is documented in the Tier III
Priority Chemical List Docket.

 Benzene also has a High WMPT Eco-toxicity score, is a PAH precursor,
and very difficult to destroy thermally .  This information suggests
that benzene warrants special consideration in current decision-making.

 See 69 FR 55218. 

USEPA, 2004,  “Regulatory Impact Analysis for the Industrial Boiler &
Process Heater NESHAP,” EPA-452/R-04-002

 Six thousand out of 57,000 ICI boilers and process heaters are
oil-fired units according to, USEPA, “The Upcoming Industrial Boiler &
Process Heater MACT Standard., “A&WMA MACT Web Confierence, May 2002.

 USEPA, “AP 42:Compilation of Air Pollution Emissions Factors, Vol I:
Stationary Point and Area Sources,” , 5th Edition, 1995, Ch.1.3 Fuel
Oil Combustion.

 Property values from, Stultz & Kitto, eds, “Steam: It’s Generation
and Use,” 40th Edition, 1992, Ch.8.

 Chevron Products Company, “Diesel Fuels Technical Review,” 1998

 See preamble to the 1998 comparable fuels rulemaking at 63FR 33785.

 Various geographic regions in the are assigned single volatility class
each month based on altitude and ambient temperature range

 From http://www.faqs.org/faqs/autos/gasoline-faq/

 ASTDR, 1995, “Toxicological Profile for Automotive Gasoline,”
http://www.atsdr.cdc.gov/toxprofiles/tp72.html

 This requirement is currently being removed.
http://www.epa.gov/OMS/rfg.htm

 http://www.epa.gov/OMS/regs/fuels/rfg/properf/rfg-params.htm

 The use of MTBE is being phased out.

 Chevron Products Co., “Motor Gasoline Technical Review”, 1996

 http://www1.eere.energy.gov/biomass/renewable_diesel.html

 Owens, E., “Oxygenates for Diesel Emissions Reduction” 7th Diesel
Engines Emissions Reductions workshop, Portsmouth, VA.

 Avallone, E.A., & Baumeister, T.(eds), “Mark’s Standard Handbook
for Mechanical Engineers,” 9th Edition, 1987

 Note there are subcategories within these main categories.

 See Section 6.2.3 for a discussion on atomization systems.

 See 63 Fr 55218 (Sept. 13, 2004) and 40 CFR Part 63, Subpart DDDDD.

 USEPA, “National Emission Standards for Hazardous Air Pollutants for
Industrial, Commercial, and Institutional Boilers and Process Heaters;
Final Rule,” September 2004, See 69 FR 55223

 USEPA, Draft Technical Support Document for HWC MACT Standards(NODA),
Volume II: Evaluation of CO/HC and DRE Database,” April 1997

 La Fond, et al, 1985, “Evaluation of Continuous Performance
Monitoring Techniques for Hazardous Waste Incinerators.

 Hall, D.L., Dellinger, B., Graham, J.L., and Rubey, W.A. “Thermal
Decomposition Properties of a Twelve Component Organic Mixture.” 
Hazardous Waste & Hazardous Materials, Volume 3, November 4 1986, pg
441-449.  Lieber, Inc. Publishers.

 Note that the low CO measurements are somewhat deceptive.  The POHCs
were injected at low concentration with no auxiliary fuel.  There was
very little carbon available to form CO.  The CO concentrations were not
corrected to 7% oxygen and this correction would be complicated because
the oxidant was a mixture much higher in nitrogen than standard
combustion air.  A preliminary estimate suggests that the 15 ppm CO
would be over 100 ppm at 7% O2  if also corrected for nitrogen dilution.

 Dellinger, B. Taylor, P., and Tirey, D., “Pathways of Formation of
Chlorinated PICs from the Thermal Degradation of Simple Chlorinated
Hydrocarbons.” Journal of Hazardous Materials, Elsevier Publishers,
1989. 

 Staley, Richards, Huffman, Olexsey, and Dellinger. “On the
Relationship Between CO, POHC and PIC Emissions from a Simulated
Hazardous Waste Incinerator.” JAPCA 39: 321-327, 1989.

 This database is available from the OAQPS website on the ICI NESHAP at
http://www.epa.gov/epaoswer/hazwaste/combust/finalmact/index.htm#finaloc
t

 These are two Rubicon units in Geismar, LA (Source IDs 812 & 813) and a
BASF unit (834) in LA.  These units have a boiler section that is
separated by ductwork from the primary combustion chamber and, thus, do
not meet the definition of a boiler in 40 CFR 260.10.  

 Information on boiler type was limited. 

 A majority of the units with emissions data burn residual fuel oil.

 See   HYPERLINK "http://www.epa.gov/ttn/atw/orig189.html" 
http://www.epa.gov/ttn/atw/orig189.html  for the list of HAP.

 USEPA, “AP 42:Compilation of Air Pollution Emissions Factors, Vol I:
Stationary Point and Area Sources,”  5th Edition, 1995, Ch.1.3 Fuel
Oil Combustion.

   HYPERLINK "http://www.atsdr.cdc.gov/toxprofiles/tp14-c6.pdf" 
http://www.atsdr.cdc.gov/toxprofiles/tp14-c6.pdf 

 For these compounds we did not extract partial detects.

  See Appendix C

 USEPA “Risk Burn Guidance for Hazardous Waste Combustion
Factilities”, EPA OSW/Region 4,  EPA 530-3-01-001, 2001

 USEPA, “Technical Support Document for HWC MACT Standards,”
September 2005, Vol; IV Compliance with the HWC MACT Standards, Section
3.

 Stanmore B.R., 2004,“The Formation of Dioxins in Combustion
Systems,” Combustion and Flame, v 36, pp398-427.

 Gullet B., and Seeker, R., 1997, “Chlorinated Dioxin and Furan
Formation, Control and Monitoring”, ICCR Meeting Presentation

 Porcaccini et al, 2003 “Formation of Chlorinated Aromatics by
Reactions of Cl., Cl2, and HCl with Benzene in the Cool-Down Zone of a
Combustor” Engineering Science and Technology, v .37 p, pp 1684-1689.

 http://www.atsdr.cdc.gov/toxprofiles/tp69.html

 Note that there is no specification limit for Copper.

 USEPA, “Technical Support Document for HWC MACT Standards, Vol IV,
Compliance with HWC MACT Standards,” September 2005, Section 3.2

 Ni is known to increase PCDD/F formation.

 Beth Antley, USEPA, Region 4, and David Layland, USEPA, OSW, made
significant contributions to this section.

 RTI International, “Inferential Risk Analysis in Support of Standards
for Emissions of Hazardous Air Pollutants from Hazardous Waste
Combustors,” Final Report, July 2005

 It must be emphasized that emission-adjusted MOEs should not be
construed as predictions of the level of risk.  Instead, they are only
intended to provide an indication of whether risks could exceed a level
of concern based on simplifying assumptions and as such, are subject to
a considerable degree of uncertainty

 Boilers that were not integrally designed were not excluded.

 Industrial boilers with dry air pollution control devices that elect to
burn ECF would be required to control their air pollution control device
inlet temperatures to 400F unless their primary fuel is coal

 The specification would not be waived for PAHs.

 We note, however, that there is no reason to believe that these
parameters (e.g., stack parameters, location, nearby land use) would be
any different for the universe of ECF boilers than for the universe of
MACT HW boilers.  

	Please note that, although the 2000 slope factor is consistent with
EPA’s most recent upper bound slope factor for human cancer risk,
based on human data, there are a variety of sources of uncertainty in
this estimate.  The 1985 estimate was derived from animal studies, which
EPA has reanalyzed as part of its dioxin reassessment.  These estimates
are not final and are subject to change once EPA finalizes the
assessment.

 Generators may blend hazardous waste fuels to meet the viscosity
specification under conditions stipulated in proposed §216.38 (a)(4). 
Note that the waste is not excluded until after blending to achieve a
maximum viscosity of 50 cs, and the blending is hazardous waste
treatment.

 EPA is seeking comment if storage in other containers (such as 55
gallon drums and other portable containers) is likely to occur.

 USEPA, “A Must for USTs” July 1995, EPA 510-K-95-002

 See proposed rule at §261.38(c) (2)(ii) for complete specifications
and see Preamble to the proposed rule Part Two, Section II.B for a
discussion of the rationale for these conditions.

 Note that oxygen must also be measured with a CEMS.

 For a spherical droplets the surface area to volume ratio is equal to
6/d, where d is the diameter of the droplet.  Therefore the smaller the
droplet the higher the surface area per given volume of fuel that is
available to absorb heat from the flame.

 Engineering Science, “Background Information Document for the
Development of Regulations to Control the Burning of Hazardous Wastes in
Boilers and Industrial Furnaces, Volume I: Industrial Boilers,”
January 1987.

 Brunner,C.R., 1989 “Handbook of Hazardous Waste Incineration,” Tab
Books Inc.,  

 Note that ECF cannot have a viscosity exceeding 50 cs.

 	Note that 35 and 100 mesh corresponds to a particle size less than 500
microns and 150 microns respectively.  Note, however, that ECF must be
able to pass through a 200 mesh (74 micron) sieve to ensure that
particles are small enough to ensure volatilization and destruction of
organic compounds.

 Letter from American Chemistry Council (Carter Lee Kelly, Leader, Waste
Issues Team, and Robert A. Elam, Director, Regulatory Affairs, Waste
Issues Team) to Robert Springer and Matt Hale, USEPA, dated November 24,
2003.

 This is similar to secondary containment required by the states of
Florida and Minnesota for above ground fuel oil tanks.

 Although the ACC survey table shows 96 streams, one stream L-02-12 was
duplicated on a 2nd row.

 Six streams totaling 9.4 million lbs/yr are failing the specifications
for naphthalene, a hazard category A hydrocarbon.

 The waste stream IDs in the ACC survey had the form x-XX-XX (for e.g.
A-01-12). We assumed that the letter in the code represented a single
company and the first 2 digit number represented a unique generating
site. This approach may have underestimated the number of generating
sites.

 http://tonto.eia.doe.gov/dnav/pet/hist/a803700002m.htm

 http://www.etc.org/costsurvey8.cfm

 USEPA, “Supporting Statement for EPA ICR 1361.10,” October 2005. 

 ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.tx

 This is based the size of a typical tanker truck.

An additional factor of 2 to provide sufficient headspace.

 The savings for HW manifests and generator storage are equal to the
costs for these items under regulatory option C.

 Absent a condition of the exclusion, we nonetheless assumed that
facilities would incur certain costs to implement standard operating
procedures (SOP), such as keeping records of each shipment of waste fuel
under Regulatory Option A.

 Assumes every boiler burning ECF will purchase a new CO CEMS and every
boiler is already equipped with an oxygen CEMS.

 A 0.6 scaling factor was used in all economic estimates where the size
of the item had an impact on the cost.

 This overestimates costs because ECF will require DOT shipping papers
only if ECF meets the DOT definition of a hazardous material.

 Assume separate tank for each qualifying ECF stream.  Tanks replacement
schedule:  33% replaced immediately, 33% after 5 years and 33% after 10
years.

 For example the installed cost for a 10,000 gallon tank would be =
$69,000* (10,000/20,000)0.6= $45,000.

 The visual inspections of tanks/equipment as well as monthly monitoring
are based on requirements for hazardous waste tanks in Subparts J and CC
of 40 CFR 265 and requirements for fuel oil tanks under SPCC
requirements.  It is assumed that these inspections are SOP and would be
performed under Option A as well.

 Assume spills are less than 42 gallons which is below  the SPCC
reporting threshold per 40 CFR §112.4

  Costs are for renewing existing plans.  The cost of a response plan is
included although very few generator sites would have a total storage
capacity for fuel oil and ECF of >1,000,000 gal, the threshold for which
a response plan is required under §112.20(f)(1).  

 Assume that application of the CAA OLD NESHAP and RCRA Part 265/264,
Subpart CC, will result in a similar number of tanks being subject to
air emissions controls under Options B and C.  However, assume that only
50% of tanks will be subject to Level 2 controls under existing product
controls (Option A). 

 All ECF streams combined in single new storage tank for each burner.
Tank life is 15 years.

 This estimate is based on a closure cost tool developed by Washington
State Dept of Ecology, Hazardous Waste and Toxics Reduction
Program,Publication # 05-04-009; May 2005

 These costs are based on, “Supporting Statement for EPA ICR 959.12,
Facility Groundwater Monitoring Requirements,” January 2005.

 From “Supporting Statement for EPA ICR # 1573.10, Part B Permit
Application, Permit Modification, Special Permits.***

 Annualized savings based on elimination of one time SSRA costs of
$300,000,  one time PCDD/F test cost of $7700, Comprehensive performance
test costs ($144,000 every 5 years), SSMP, notification of compliance,
documentation of compliance etc.

 Difference between engineered secondary containment and cost of earth
berm.

 For example, reduced taxes for managing hazardous waste.

 We assume commercial incinerators/kilns will reduce their disposal fees
by 75%.

 The qualifying waste quantity is the same for all options and is equal
to the excluded waste quantity under options A and B.

 See 7.5.3 for assumptions used to obtain high and low end estimates for
option C.

 Note that respondents to the survey reported waste that they believed
were candidate waste fuels that could potentially be excluded from RCRA
regulation.  The wastes were liquid hazardous wastes with a heating
value > 5000 Btu/lb and they failed the comparable fuel specification
for only a few constituents or for viscosity.

 USEPA, “The National Biennial RCRA Hazardous Waste Report (Based on
2003 Data), ***

 USEPA, “Assessment of the Potential Costs, Benefits, and Other
Impacts of the Expansion of the RCRA Comparable Fuel
Exclusion—Proposed Rule,” May 2007.

Draft Technical Support Document for the Expansion of the Comparable
Fuels Exclusion 

________________________________________________________________________
______________

 PAGE   

 PAGE   i 

Nov 17, 2006	

 PAGE   

 PAGE   14 

Nov 17, 2006

 PAGE   

 PAGE   15 

Nov 17, 2006	

The Fenceline Approach

Most of the subfactors are scored using a “fenceline” approach.  The
fenceline scoring approach involves comparing the value for a given
chemical data element against predefined “high” and “low”
threshold values for that data element, termed “fencelines.”  

S

Ä

„

	

1

2

3

M

N

O

P

Q

R

S

T

U

U

q

r

s

t

w

x

¢

£

¤

¾

¿

À

Á

Â

Ã

Ä

Å

Æ

â

ã

ä

å

è

é

ý

þ

ÿ

	



+

,

-

.

3

4

a

b

c

}

~



‚

ƒ

„

…

†

¢

£

#£

¤

¥

ª

«

Þ

ß

à

ú

û

ü

þ

ÿ

j¸

j;

j¬

j/

j

gd

I

gd

I

„Ð`„Ðgd

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

਀&䘋%摧䤋

摧䤋

萑Ũ葠Ũ摧䤋

萑ː葠ː摧䤋

h

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

萑ː葠ː摧䤋

摧䤋

h

I

h

I

h

I

h

I

h

I

਀&䘋&摧䤋

਀&䘋;摧䤋

摧䤋

萑Ũ葠Ũ摧䤋

h

I

h

I

h

I

摧䤋

萑ː葠ː摧䤋

h

I

h

I

h

I

h

I

h

I

h

I

h

I

h

I

摧䤋

h

I

h

I

h

I

h

I

h

I

h

I

h

I

摧䤋

hù

hÕ

'(

ë

ë

ë

ë

ë

ë

ë

ë

ë

ë

ë

ë

ë

S

T

w

Ñ

Ò

Y

[

d

e

g

r

s

Û

Ü

ë

R

S

T

Ñ

Ò

Ø

ñ

ò

÷

ø

û

ý

þ

X

Y

Z

[

c

d

e

f

g

q

r

s

w

Ù

Ú

Û

Ü

ã

ê

-ê

ë

ì

ô

ö

ý

þ

h

 h

h

h

h

 h

H*  h

h

h

h

h

 

਀&䘋

萑ː葠ː摧‛

옍ć̘栁؁摧‛

 

 hþ

萏Ũ葞Ũ摧岎)

 h‰

hO

h

hˆ~

h

 hO

hO

 hO

 hO

hO

h

h¤

 h¤

h

 h¤

h¤

h¤

h^{«f h¤

h¤

 h¤

h

摧纈

hˆ~

hˆ~

hˆ~

h

 h¤

摧纈

摧纈

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

hˆ~

摧纈

gdˆ~

摧纈

&

&

&

&

&

&

&

&

&

&

&

	㜀$␸䠀$摧ᗻ_

$

h)

h¦

h2



$

&

4

h)

h‡

J4

5

=

N

Z

o

p

q

Î

Ï

Ð

ù

 

:

A

B

J

K

L

M

T

_

`

a

b

c

ñ

ó

hd

 hd

h;

h;

hd

 h;

h;

 h;

h;

h;

਀&䘋

摧殉íᘀ

 hò

 hò

摧䤋

᐀mical-specific value is less than the “high” fenceline, the
subfactor is assigned a score of 3 (high concern).  If the chemical’s
value for that data element is between the “high” and the “low”
fencelines, the subfactor is assigned a score of 2 (medium concern). 
For other data elements (e.g., the cancer potency slope factor, BCFs),
lower numeric values denote lower concern; in such cases, the fenceline
logic is reversed.  All the subfactors are scored as low concern (score
= 1), medium concern (score = 2), or high concern (score = 3).

EPA’s PBT Priority Chemical List

After several rounds of internal expert and public comments, EPA used
the current version of the WMPT as the initial step in the process of
identifying the initial pool of persistent, bioaccumulative, and toxic
(PBT) chemicals that are national priorities for voluntary pollution
prevention activities across the agency.  EPA determined the initial
pool of (PBT) priority chemical candidates based on their rank.  The
rankings are based on the higher of available scores for human health
concern (i.e., the sum of the scores for persistence, bioaccumulation,
and human toxicity) and ecological concern (i.e., the sum of the scores
for persistence, bioaccumulation, and ecological toxicity).  The (PBT)
priority chemical candidate pool was limited to those chemicals with
WMPT scores of 8 or 9 (on a scale of 3 to 9).  

See USEPA (2000) for a more detailed description of the WMPT development
process.  The specific use of the current version of the WMPT rankings
in developing the RCRA(PBT) priority chemical list is documented in the
Tier III (PBT) Priority Chemical List Docket.

