VIA E-MAIL

April 12, 2010

Mr. Jeffrey A. Telander

Sector Policies and Programs Division

Metals and Minerals Group (D243-02)

Office of Air Quality Planning and Standards

U.S. Environmental Protection Agency

109 TW Alexander Drive

Research Triangle Park, NC  27711

Mr. William Neuffer

Sector Policies and Programs Division

Metals and Minerals Group (D243-02)

Office of Air Quality Planning and Standards

U.S. Environmental Protection Agency

109 TW Alexander Drive 

Research Triangle Park, NC  27711

 

RE:	Follow-Up on Test Method Issues

Dear Messrs. Telander and Neuffer:

The North American Insulation Manufacturers Association (“NAIMA”)
appreciated the opportunity to discuss with the Environmental Protection
Agency (“EPA”) on Monday, March 29, 2010 the appropriate test
methods for emissions data that will be provided to EPA pursuant to the
Information Collection Requests (“ICRs”) for the wool fiberglass and
mineral wool MACT Standards.  In a letter to EPA dated March 26, 2010,
NAIMA requested EPA’s consideration of alternative test methods to
those identified in the draft ICRs.  NAIMA explained that use of
alternative test methods was important because of the extensive data
collected by these alternative methods and the fact that every company
is operating at reduced production, which means that many lines are not
operating and new data could not be collected for those non-operating
lines.

During the March 29, 2010 conference call, EPA agreed to accept data
collected by the following alternatives:

EPA Method 316 for formaldehyde

EPA Method 308 for methanol

SW-846 Method 8270D for phenol

EPA Method 10 for carbon monoxide

EPA Methods 15 and 320 for carbonyl sulfide

NAIMA also asked for EPA to consider other alternatives for which the
Agency indicated it would need additional support or explanation as to
the relevance and reliability of these test methods.  This letter sets
forth that additional support for the following alternative test
methods:

California Method ST 16/Gas Chromatography – Phenol

As previously noted, California Method ST 16 is as effective as Method
318 for testing of phenol and has the decided advantage of providing
reliable results and leveraging extensive available data.  NAIMA
specifically referenced side-by-side comparisons of Method 318 and ST
16.  This comparison is provided by Owens Corning:

During December 2008, Owens Corning responded to an EPA Section 114
request to quantify its VOC emissions from its Kansas City plant.  Part
of that request involved testing for formaldehyde, methanol, and phenol
at the plant’s #70 line in December 2008.  A protocol was submitted
and reviewed by technical experts from EPA headquarters and Region 7 and
approved by Adam Kushner at EPA headquarters.  The approval specifically
authorized California Method ST 16:

“At Owens Corning’s discretion EPA Method 8270C ‘Semi-volatile
Organic Compounds by Gas Chromatography/Mass Spectrometry’ may be used
in lieu of Analytical Procedure Lab-8, BAAQMD Method ST-16.”  (EPA
letter to Owens Corning dated October 27, 2008.)

The protocol identifies ST 16 for phenol and a method modification
substituting a standard Method 5 train for sample collection.  The
samples are collected in the impingers which are filled with a solution
of 0.1 normal (0.1N) sodium hydroxide which absorbs the phenolic
compounds. The phenolic compounds are then analyzed by gas
chromatography. This method was employed, and additional Method 318
sampling was run to identify VOC species previously undetected.  The
Method 318 results were evaluated for known and unknown VOC species. 
The 1 hour Method 318 test was evaluated for all VOC species including
phenol.  The Method ST 16 and the Method 318 phenol results are compared
in the chart below.

Phenol Comparison Results

70 line Testing December 2008

Location	Phenol Results  lbs/hr

Modified ST-16 	Phenol Results lbs/hr

FTIR (M-318)

Forming Zone A	.249	< .07

Forming Zone C	.436	< .09

Forming Zone D	.258	< .09

Oven Charge	.032	< .02

Oven Dis-Charge	.024	< .02

Cooling	ND	< .03



The testing results were obtained on the same product during subsequent
days because simultaneous testing was not possible.  From this data,
there appears to be a large amount of line loss for phenol using Method
318.  Phenol also comes off near the water band where for forming
sources with saturated levels of water, the phenol peak may be hard to
accurately quantify.  The stated range for ST 16 for phenolic compounds
is 200 ppb at the sample volume specified in this method; however, the
sample volume was approximately 3 times the standard volume.

EPA Method TO-8 – Phenol Testing

NAIMA also requested that EPA accept Method TO-8 as an alternative test
method for phenol.  Again, this Method was as effective as Method 318
and has the decided advantage of reliable results and leveraging
extensive available data.  CertainTeed has provided additional analysis
on the efficacy of EPA Method TO-8 for phenol.  Specifically, in
response to EPA Section 114 requests to conduct source testing for
specific VOC emissions at several wool fiberglass insulation plants, EPA
approved the use of EPA Method TO-8 for determining phenol and methyl
phenols emissions.  Testing associated with these Section 114 requests
was conducted at the following CertainTeed Plants:

	Mountain Top, PA		August 2007

	Athens, GA			January 2009

	Kansas City, KS		January 2009

	Chowchilla, CA		January 2009

The detailed source test reports were submitted to both the applicable
Regional (3, 4, 7 and 9) and headquarters offices of EPA.  Cary Secrest
at EPA headquarters was very involved in these Section 114 test
protocols.

Phenol Emissions – EPA Compendium Method TO-8

The concentration of phenol in the flue gas stream was collected using
sampling procedures outlined in Method TO-8.  The flue gas was withdrawn
at a constant rate through a Teflon probe and passed through a series of
glass midget impingers contained in an ice bath.  The first three
impingers contained 15 ml of 0.1 N NaOH followed by a dry impinger, and
then an impinger containing silica gel.  At the completion of each test,
the contents of the first two impingers were recovered together in a
glass vial and sealed with a Teflon lined screw cap.  The contents of
the third and fourth impinger were recovered separately in a glass vial
and analyzed separately to determine sample breakthrough.  The phenols
were trapped as phenolates.  All samples were subsequently analyzed by
high performance liquid chromatography (HPLC).

The Athens, Georgia, sample event, for example, had an ND blank value of
5.5 (g per sample for the TO-8 sample train.  This would equate to
in-stack detection limits of 0.0481 ppm, 0.0677 lb/hr, and 0.00857
lb/ton.  The actual samples had “catches” ranging from 241 to 309 (g
– well above the detection limits of the method at the ~ 2 ppm range
observed in the stack.

Section 1.3.2. of Appendix A to Part 63 – Test Method (Method 318 –
Extractive FTIR Method for the Measurement of Emissions from the Mineral
Wool and Wool Fiberglass Industries) provides: “In general, a 22 meter
pathlength cell in a suitable sampling system can achieve practical
detection limits of 1.5 ppm for three compounds (formaldehyde, phenol,
and methanol) at moisture levels up to 15 percent by volume.”

Based on CertainTeed’s experience, NAIMA asserts that the use of
Method TO-8 as an alternative to Method 318 is an appropriate source
test method.  The in-stack detection limit, using the Athens, Georgia,
test as an example, is significantly better than the published practical
detection limit for Method 318, 0.05 vs. 1.5 ppm, respectively. 
Similarly, the measured phenol concentration of approximately 2 ppm is
40 times higher than the TO-8 detection limit and only slightly above
the Method 318 detection limit.

EPA Method 5/202

NAIMA requested that EPA Method 5/202 be allowed as an alternative to
EPA Other Test Methods (“OTM”) 27 and 28.  NAIMA expressed concern
that some stacks might not be able to accommodate an OTM 27 probe
without significant blockage, which adds variability to emission test
procedures and results.  In addition, Method 5 is often required to be
used in place of Method 27 due to the presence of water droplets that
cause Method 27 to provide erroneous readings.  Method 5 has been a
reasonable substitute for Method 201A which is similar to Method 27
because all of the captured emissions were presumed to be PM10 or
finer.  In fact, Method 5/202 has been accepted as an alternative in
previous ICRs.  See EPRI, “Responding to the EPA Information
Collection Request for Electric Utility Steam Generating Units,”
January 2010.  In comparison testing on furnaces, this has proved to be
true as illustrated by the comparison below from Knauf Insulation
involving cold top furnaces with baghouses:



Knauf Shasta Lake, California - 641 Furnace

	Method 5/028

	08/06/09	08/06/09	08/06/09	Average

 	 	 	 

	Front Half Filter, lb/hr	0.00	0.00	0.00	0.00

 





Front Half Wash, lb/hr	0.06	0.01	0.00	0.02

 





Front Half, lb/hr	0.06	0.01	0.00	0.02

 





Back Half Aqeous, lb/hr	0.07	0.06	0.05	0.06

 





Back Half MeCl, lb/hr	0.03	0.09	0.05	0.006

 





Back Half, lb/hr	0.10	0.15	0.10	0.12

 	 	 	 

	Total, lb/hr	0.162	0.152	0.100	0.14



Knauf Shelbyville, Indiana - 602 Furnace

	Method 5/202

	11/09/07	11/09/07	11/09/07	Average

 	 	 	 

	Front Half, lb/hr	0.08	0.04	0.06	0.06

 





 





Back Half, lb/hr	0.82	0.51	0.56	0.63

 





 Total, lb/hr	0.990	0.55	0.62	0.69



The above tables provide test data from testing events at two TECO cold
top electric furnaces with baghouses for PM control: 641 Furnace at
Shasta Lake, California, and 602 Furnace at Shelbyville, Indiana,
facilities.  There is a third TECO cold top furnace with baghouse in
operation at the Shelbyville plant.  The data in the above tables are
from testing using EPA Method 5 Modified/202 (Shelbyville) and Method
5/028 (Shasta Lake).  The Shelbyville data was generated using the
Subpart NNN referenced modified Method 5 for front half PM.  The
modified Method 5 requires a filter temperature of no greater than
350°F +/- 25°F vs. the standard Method 5 filter temperature of 248°F
+/- 25°F.  The 202 vs. 028 data appears to show the bias between the
two methods and would support the move to the proposed revised/modified
202 (OTM 28).

Both sets of data support the contention that front half Method 5 data
is representative of PM/PM10/PM2.5 emissions from a cold top electric
furnace with a baghouse.  The filterable PM emissions are extremely low.
 It is reasonable to stipulate that all of the measured filterable PM
from this source category is PM2.5 or less due to the low mass collected
and the fact that the sampling occurred after a high efficiency
baghouse.  Therefore, the existing Shasta Lake Method 5/028 data should
be accepted as representative for the cold top electric furnace with
baghouse source category for purposes of responding to the draft ICRs. 
Both of the testing events referenced were witnessed by State and/or
local agency representatives and submitted to the applicable agency as
required in Knauf’s operating permits.  Copies of the full test
reports are available and will be submitted in the response to the
applicable draft ICR questions.

OTM 28 is a new method that has not been used extensively in the fiber
glass industry.  Limited use of this method indicates that on dry
melter sources with little SO2, there is virtually no difference in
measurement results when compared to Method 202.  The data comparison
between Method 202 and OTM 28 indicate that the use of OTM 28 is more
important on wet sources.  Therefore, data captured by Method 5/Method
202 is virtually equivalent to data captured by OTM 27/OTM 28 for
electric melters and furnaces.

EPA Method 30B

NAIMA requested that EPA Method 29 be used to test for mercury.  NAIMA
pointed out that mercury was not used in the manufacturing process, so
there was not a need to separately test for mercury.  EPA indicated that
it would consider NAIMA’s request if it could be demonstrated that the
threshold for mercury emissions was sufficiently low.  Member companies
provided any mercury emissions test data they had to NAIMA’s outside
counsel.  Only two companies had mercury emissions test data.  One fiber
glass facility’s four total tests of three different furnaces yielded
emissions ranging from 0.17 to 6.5 pounds of mercury per year. 
Similarly, one mineral wool facility reported mercury emissions from
multiple tests that range from 0.18 to 0.97 pounds per year.  Based on
the low mercury emissions, separate testing for mercury pursuant to EPA
Method 30B is not warranted for the wool fiberglass and mineral wool
insulation industries.

EPA Method 306 and CARB 425

NAIMA requested that EPA consider EPA Method 306 and CARB 425 as
alternatives to EPA Method 0061 for hexavalent chromium because Method
0061 cannot always be successfully used in high temperature
environments.  EPA was willing to consider these alternatives if a third
party could validate the problems described by NAIMA and its members. 
NAIMA acquired the opinion of Bureau Veritas, attached hereto, which
confirmed NAIMA’s characterization of the possible complications with
using EPA Method 0061 in high temperature environments.

EPA Method 26A

EPA designated EPA Method 26A for testing of hydrochloric acid and
hydrogen fluoride.  Initially NAIMA did not believe there was a basis
for testing for either of these emissions.  EPA requested that NAIMA
survey its members to determine if fluorspars or similar minerals are
used in the manufacturing process that might trigger hydrogen fluoride
emissions or other substances that would trigger hydrochloric acid
emissions.  Based on member company responses provided to NAIMA’s
outside counsel, there is a basis for wool fiberglass manufacturers to
test for hydrogen fluoride.  This data will be included in the industry
response.  There is not, however, a basis for such a test in the mineral
wool industry as mineral wool facilities do not utilize fluorspars or
similar minerals which would provide a basis for testing of hydrogen
fluoride.  There was a fraction of hydrogen fluoride in a mineral wool
facility in a report from the early 1990s that was based on an emissions
factor.  The survey indicated that neither wool fiberglass nor mineral
wool have a basis to test for hydrochloric acid.  NAIMA requests that
EPA therefore not require the mineral wool industry to test for either
hydrogen fluoride or hydrochloric acid and that EPA excuse the wool
fiberglass industry from testing for hydrochloric acid.

If EPA desires further discussions on these issues, NAIMA would gladly
schedule a conference call to bring a final resolution to any of these
remaining issues.  Based on this letter, NAIMA requests that EPA allow
the use of alternative test methods as described herein in order to
expedite completion of the ICR.

Sincerely,

Angus E. Crane

Angus E. Crane

Executive Vice President, General Counsel

%

*

7

G

P

v

ˆ

‰

ø

ù

␃愃̤摧笿®

␃愃̤摧䋀?ᜀv

w

x

y

}

~



·

C

D

‡

ˆ

‰

Š

ù

há

6

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

혈F鐃᳿ꐋⰖ"蠆

蠆

蠆

kdÙ

摧䄟-Ê欀ɤ

*▪  Suite 310  ▪  Alexandria, Virginia  22314  ▪  Tel: (703)
684-0084  ▪  Fax: (703) 684-0427

