Boat Resin Emissions

BASELINE CHARACTERIZATION OF EMISSIONS
FROM FIBERGLASS BOAT MANUFACTURING

FOR
NATIONAL MARINE MANUFACTURERS ASSOCIATION
Prepared for
National Marine Manufacturers Association
3050 K Street, N.W., Suite 145
Washington, D.C. 20007
Prepared by
Stelling Engineering, P.A.
1319 Arnette Avenue
Durham, North Carolina 27707
Air-Tech Environmental LLC

P.O. Box 12353
Research Triangle Park, North Carolina 27709
Radian International LLC

P.O. Box 13000
Research Triangle Park, North Carolina 27709
August 1997
Baseline Emission Testing
Table of Contents
Page
1.0 INTRODUCTION......................................................................................................1-1
1.1 Background ....................................................................................................1-2

1.2 Purpose of Testing .........................................................................................1-3
1.3 Testing.............................................................................................................1-3
1.3.1 Test Enclosure.....................................................................................1-4

1.3.2 Process.................................................................................................1-5
1.3.3 Sampling Methods and Procedures..................................................1-7
1.3.4 Molds...................................................................................................1-8
1.3.5 Materials..............................................................................................1-9
1.3.6 Schedule ............................................................................................1-10
1.4 Organization of the Report .........................................................................1-10
2.0 RESULTS ....................................................................................................................2-1
2.1 Summary of Results.......................................................................................2-1

2.2 Evaluation of Emissions from Continuing Trace Concentrations .............2-6
2.3 CFA Mold Results........................................................................................2-10
2.4 Results of Gelcoat Application Testing......................................................2-13
2.5 Results of Resin Lamination Testing..........................................................2-16
2.5.1 Emissions as Percent of Available Styrene ....................................2-17

2.5.2 Flux Emissions..................................................................................2-20
2.5.3 Normalized Emissions.....................................................................2-21
3.0 CALIBRATION AND SAMPLING PROCEDURES...............................................3-1
3.1 Process Procedures and Measurements.......................................................3-2
3.1.1 Application Procedures .....................................................................3-3

3.1.2 Spray Equipment Calibration ...........................................................3-4
3.1.3 Gel Time ..............................................................................................3-5
3.1.4 Styrene and Methyl Methacrylate Content ......................................3-6
3.2 Emission Testing Equipment and Procedures ............................................3-7
iii Rev. 1
Table of Contents (Continued)
Page
3.2.1 Leak Check and Instrument Calibration ..........................................3-7
3.2.1.1 Total Hydrocarbon Analyzer ............................................3-8

3.2.1.2 Gas Chromatograph.........................................................3-10
3.2.2 Flow Rate Measurements ................................................................3-11

3.2.3 Temperature .....................................................................................3-12
3.2.4 Weight of Material Used..................................................................3-12
3.3 Air Flow Over the Mold..............................................................................3-13

3.4 Mass Balance Styrene Recovery Check......................................................3-15
3.5 Data Custody................................................................................................3-16
4.0 QUALITY CONTROL AND QUALITY ASSURANCE .........................................4-1
4.1 Concentration Measurements .......................................................................4-1
4.1.1 Total Hydrocarbon Measurements (Method 25A) ..........................4-2

4.1.2 Organic Compound Concentrations (Method 18)...........................4-6
4.1.3 Comparison of Methods in Determining Emissions.....................4-12
4.2 Gelcoat and Resin Flow Rate ......................................................................4-14

4.3 Air Velocity over Mold Surface ..................................................................4-15
4.4 Exhaust Flow Rate .......................................................................................4-15
4.5 Exhaust Air Temperature............................................................................4-17
4.6 Resin/Gelcoat Balance ................................................................................4-17
4.7 Technical Systems Audit .............................................................................4-18
4.7.1 Total Hydrocarbon Measurement ..................................................4-19

4.7.2 Weigh Cell Balance ..........................................................................4-20
4.7.3 Flow Measurement...........................................................................4-22
4.7.4 Styrene Evaporation Test.................................................................4-22
iv Rev. 1
Table of Contents (Continued)
APPENDICES
APPENDIX A Reference Methods
APPENDIX B Construction of a Temporary Total Enclosure for Volatile Organic

Compound Emission Assessment During Manufacture of Fiberglass
Boats
APPENDIX C Calculations
APPENDIX D Procedure for and Results of Gel Time Determination
APPENDIX E Material Safety Data Sheets
APPENDIX F Certifications for Gas Standards and Equipment Calibrations
APPENDIX G Bag Standards
APPENDIX H Location and Results of Air Flow Over the Mold Measurements
APPENDIX I Comparison of NMMA Results with Those from Other Studies
v Rev. 1
LIST OF TABLES
Page
1-1 Parameters Studied ...................................................................................................1-8
1-2 Surface Areas of Molds Used in Test Program.......................................................1-9
1-3 Properties of Gelcoat and Resin Materials ............................................................1-10
1-4 Test Schedule ...........................................................................................................1-11
2-1 Summary of Test Results ..........................................................................................2-2
2-2 Emissions Measured During Application of Resin to CFA Mold.......................2-11
2-3 Emissions from Gelcoating as Percentage of Available Volatiles.......................2-14
2-4 Total Emissions from Gelcoating Related to Mold Surface Area........................2-15
2-5 Emissions Measured During Resin Lamination ...................................................2-17
3-1 Measurements During Test Program ......................................................................3-1
3-2 Gel Time Results........................................................................................................3-6
3-3 Styrene and MMA Contents.....................................................................................3-6
4-1 Acceptance Criteria for Concentration Measurements ..........................................4-2
4-2 Calibration Error Measurements for Total Hydrocarbon Analyzer .....................4-3
4-3 Calibration Drift Measurements for Total Hydrocarbon Analyzer......................4-4
4-4 Specific Compound Response Factors for the Total Hydrocarbon Analyzer ......4-5
4-5 Method 18 Response Factors for Target Compounds ............................................4-7
4-6 Gas Chromatography Calibrations (after April 12) .............................................4-10
4-7 Comparison of Emission Results by Two Methods .............................................4-12
vi Rev. 1
LIST OF TABLES (continued)
Page
4-8 Spray Equipment Data Summary ..........................................................................4-14
4-9 Variability Measures for Flow Rate During Test Runs........................................4-16
4-10 Results of Balance Calibration................................................................................4-18
4-11 Audit Results for Method 25A Measurements .....................................................4-19
4-12 Results of Audit of Balance ...................................................................................4-21
4-13 Results of Flow Rate Audit.....................................................................................4-22
4-14 Results of Styrene Evaporation Tests ....................................................................4-23
vii Rev. 1
LIST OF FIGURES
Page
2-1 Summary Concentration Traces for Boat Lamination and Gelcoating.................2-3
2-2 Emissions as Percentage of Available Monomer....................................................2-7
2-3 Total Hydrocarbon Concentration Trace for NMMA-4-1......................................2-8
2-4 Concentration v. Cumulative Mass Emissions (NMMA-4-1)................................2-9
2-5 Emissions from Lamination and Gelcoating of the CFA Mold...........................2-12
2-6 Gelcoat Emissions as Percent of Available Volatiles............................................2-14
2-7 Emissions from Gelcoating Related to Mold Surface Area .................................2-16
2-8 Emissions from Lamination as Function of Available Styrene ...........................2-18
2-9 Comparison of Results by Test...............................................................................2-19
2-10 Emissions from Resin Lamination as Flux Measurement....................................2-20
2-11 Average Emission Flux Measurements for Two Resins ......................................2-21
2-12 Normalized Emissions from Resin Lamination ...................................................2-22
4-1 Method 18 Response Factors for Styrene ................................................................4-6
4-2 Method 18 Response Factors for Methyl Methacrylate..........................................4-7
4-3 Method 18 Response Factors for Selected Test Days .............................................4-8
4-4 Comparison of Total Hydrocarbon Traces by Two Methods..............................4-13
viii Rev. 1
ACKNOWLEDGEMENTS
The authors gratefully acknowledge U.S. Marine for the use of their facility in Arlington,
Washington, for this program. They provided space, an outstanding facility for the
program, an excellent enclosure (designed and constructed by U.S. Marine), and
outstanding support by their personnel. We wish to specifically recognize the following
people for their help in making the field work in Arlington such a success:
Dennis Pearson, who, as our U.S. Marine site coordinator, provided constant oversight and
help in scheduling space, personnel, molds, and materials; Dennis also spearheaded the
design and fabrication of the temporary total enclosure, which exceeded all expectations
for a test enclosure of its size;
Larry Dargitz, U.S. Marine’s lead lamination training specialist, who conducted the boat
lamination work during this program;
Ken Warren, U.S. Marine gelcoating specialist, who conducted all gelcoating runs during
the program; and
Don Barnhill, who extended the invitation and ensured daily that the program had
everything needed.
We wish to express appreciation to Research Triangle Institute and the U.S. Environmental
Protection Agency for funding the supplemental testing of flow chopper technology. We
also appreciate the support of this program by the EPA through its audit of our technical
systems.
Finally, we wish to thank John McKnight, the NMMA Project Manager, for his continued
support of this testing program before, during, and after the field work.
ix Rev. 1
1.0 INTRODUCTION
Under contract to the National Marine Manufacturers Association

(NMMA), Stelling Engineering, Air-Tech Environmental, and Radian
International conducted testing in April 1997 to characterize baseline emissions
from gelcoating and lamination of fiberglass boats. Testing was conducted in
accordance with the Quality Assurance Project Plan for the National Marine
Manufacturers Association Baseline Emissions Testing Project (QAPP) approved by
the U.S. Environmental Protection Agency (EPA) in March 1997. A quality
assurance (QA) audit was performed by the EPA during this program. Visits by
state regulatory authorities and other interested parties were also made during
testing.
The test program was designed to be as representative as possible

of actual fiberglass boat manufacturing. Testing was conducted at the
U.S. Marine (Bayliner) fiberglass boat manufacturing, research, and
development facility in Arlington, Washington, in an enclosure (inside a wooden
model fabrication area) made available for this purpose. The volatile organic
compounds (VOC) measured during this program were generated from
operations identical to those in actual production, that is, using actual boat part
molds, spray equipment, resins and gelcoat, and laminating procedures and
techniques typical of the industry. Experienced manufacturing personnel
conducted the lamination and gelcoating. Only a few runs incorporated a mold
or spray techniques not typical of the industry; such runs were designed to
produce data for comparison with previous emission tests.
During this sampling program, Research Triangle Institute (RTI)

conducted a research program to ascertain the viability of measuring emissions
1-1 Rev.1
from lamination and gelcoating operations using a mass balance approach. RTI’s
program was funded by the EPA’s Office of Research and Development (ORD)
and will be reported independently.
1.1 Background
Within the past two years, several important sampling and analysis
programs have sought to determine styrene emissions from open molding of
polyester resin fabricated parts. Most important among these studies were the
studies conducted by RTI for the EPA and by Dow Chemical USA for the
Composite Fabricators Association (CFA). The results from these programs raise
questions about the emission factors published in the EPA’s compendium of
emission factors, AP-42, especially for molding of large parts manufactured by
the marine industry.
Still, none of the tests to date adequately represent industry

practice or conclusively demonstrate the effect of mold size and styrene content
of resin on emissions. For example, the results of this test program indicate that
the relationship between emissions of styrene as percent of available styrene in
the resin and increasing styrene content does not appear to be constant. Further,
an increase in mold size resulted in increased flux emissions (pounds styrene per
square foot of surface area of mold) but decreased normalized emissions
(pounds styrene per 1000 lb resin per square foot of mold surface area). Also,
none of the previous tests included gelcoat containing methyl methacrylate
(MMA), which is common in the boat building industry.
1-2 Rev.1
1.2 Purpose of Testing
The primary project objective was to characterize total hydrocarbon

(THC), styrene, and MMA emissions from the open molding process that is
representative of fiberglass boat manufacturing, specifically spray up gelcoat
application and resin application. Hand lay-up, defined as the application of
glass and resin that are mixed manually and applied by brush, was not part of
this study. A technique comparable to hand lay-up found in the industry is the
use of a flow chopper, a glass chopper similar to the resin chopper gun but
equipped with a low-pressure, non-atomized resin delivery system resembling a
nozzle not unlike a shower head that exudes catalyzed resin. The principal
difference between this technique and conventional spray up application is that
resin atomization is avoided. Sampling during lamination using a flow chopper
was conducted to augment the test matrix planned for this program; this
supplemental sampling was funded by the EPA through RTI.
This study was designed to produce data for estimating emissions

from lamination and gelcoating in fiberglass boat manufacturing plants. The
data and analyses reported in this document are intended to be useful to plants
in compiling emission inventories and assessing permit needs or changes.
1.3 Testing
Sampling was completed by Darrell Doerle of Air-Tech

Environmental and John Stelling of Stelling Engineering. Testing was
coordinated with U.S. Marine personnel by John McKnight, NMMA’s project
manager. Larry Dargitz of U.S. Marine operated the resin application
1-3 Rev.1
equipment, and Ken Warren of U.S. Marine applied gelcoat for those
experiments.
Tests were also coordinated with Bob Wright, Emery Kong, and
Mark Bahner of RTI. RTI conducted a study in an adjacent enclosure under
contract to the EPA’s ORD to measure emissions from the same lamination
operations as conducted for this program but applied to a small panel. The goal
of their program was to develop a simplified material balance approach for
determining emissions from fiberglass lamination.
1.3.1 Test Enclosure
The testing was carried out at the U.S. Marine manufacturing and
research facility in Arlington, Washington. The test area was located in a
wooden model fabrication area where tool plugs are fabricated. A large test
enclosure (20 ft by 45 ft by 14 ft) meeting the total temporary enclosure (TTE)
requirements of EPA Method 204 (included in Appendix A) was erected in this
27-ft by 50-ft area to accommodate tools (i.e., molds) ranging from 18 to 28 ft in
length. Sufficient space was provided in the enclosure to allow technicians to
work around the tool and to move the tool as needed to apply gelcoat and resin.
A description of the TTE is contained in Appendix B. Natural draft openings
(NDOs) were arranged in accordance with EPA Method 204. Measurements
were made to ensure that the air velocity across the mold surface was
comparable to that measured in the manufacturing area.
The exhaust system for the entire model preparation room

comprised a dual pick-up system at one end of the room (near the constructed
enclosures) exhausting to a single induced draft fan. The fan had 20,000-cfm
1-4 Rev.1
capacity, far more than needed to exhaust the two enclosures. At the start of the
program, some of the pick-ups were closed to develop sufficient draw through
the enclosures. The flow rate was balanced with both enclosures in use so that
the turnover through the large enclosure used for NMMA testing was
comparable to that in the manufacturing areas on site (i.e., about 16 room
changes per hour). On-site checks in the production area verified this range.
The enclosure was equipped with two 10-in. plena for exhausting
the enclosure. Three-inch openings were installed every 3 ft in each plenum and
covered with spun fiberglass filter media. Air was supplied to the enclosure
through a 12-in. plenum in the ceiling of the enclosure. This plenum was open
on both ends, limiting the NDO associated with the inlet air plenum to 1.571 ft2.
The enclosure was equipped with a door (which remained closed during testing)
and a zippered end for changing application equipment and molds between
tests. Thus, the openings on both ends of the inlet air plenum were the only
NDOs. The openings in this plenum satisfied the requirements of Method 204 for
a TTE, representing only 0.043 percent of the total surface area of the enclosure.
Also, considering a flow rate of 2,600 to 3,300 cfm, the velocity at the NDO was
at least 1,655 ft/min, also satisfying Method 204 requirements. Air coming into
the enclosure was monitored continuously for hydrocarbon concentration during
each test run in accordance with Method 204; these data were recorded with
other continuous monitoring data.
1.3.2 Process
The polyester resin spray-up application process commonly used

in the boat building industry uses a “chopper gun” to apply a laminate. The
chopper gun dispenses polyester resin, catalyst, and chopped glass fibers.
1-5 Rev.1
Continuous strand fiberglass roving is fed to a chopper unit mounted on the

spray gun and is cut into chopped fiber lengths of approximately 1.5 in. long.
The chopped fiber is ejected from the chopper unit and is captured by the resin
fan pattern a short distance from the spray gun. The mixture of the catalyzed
resin and chopped fiber is deposited on the mold by the spraying action. Typical
of fiberglass boat manufacture, woven roving (a fabric material) was used with
resin application by spray gun and flow chopper for the addition of this
reinforcement. A Venus Gusmer GO3 low-pressure slave arm internal mix
airless chopper gun equipped with a 5003 tip, typical of guns used in the
industry, was used for spray application of resin and chopped glass during this
program.
Resin was supplied to the chopper gun and the flow chopper by a
Venus Gusmer system at a pump ratio of 11 to 1. The methyl ethyl ketone
peroxide (MEKP)/dimethyl phthalate (DMP) catalyst was internally mixed for
both chopper gun and flow chopper, with the catalyst delivered by a slave pump
system. The flow chopper was also a Venus Gusmer design. The resin is
delivered at a much lower pressure and no air is used to dispense the resin into
the chopped glass. The glass is chopped into lengths using the same method as
in the chopper gun, but because there is no atomizing air, the dispersion pattern
of glass and resin mixture is more narrow.
Gelcoat is applied in the boat building industry using a spray gun,

typically an airless air-assisted spray gun. For this program, gelcoat was
delivered to the spray gun using a Poly-Craft pump system at a pump ratio of 23
to 1. Catalyst, delivered using a metered feed pump, was mixed externally. A
Poly-Craft 755 airless air-assisted spray gun (equipped with a 0.026 tip size) was
used during this program.
1-6 Rev.1
Spray application was conducted using techniques common to the

industry. This study was not intended to compare spray techniques. Therefore,
skilled technicians from a manufacturing line operated the gelcoat and chopper
guns in all tests, except those included to produce data that could be more
directly compared to previous tests.
1.3.3 Sampling Methods and Parameters
Method 25A, continuous monitoring of THC, was used as the

primary technique to measure emissions. From the Method 25A results,
emissions were calculated using the measured THC concentration as propane,
the molecular weight of propane, and the exhaust gas flow rate measured in
accordance with Method 2. These results are presented in terms of THC
emissions (as propane) for lamination or gelcoating.
Speciation of styrene and MMA emissions (needed during

gelcoating) was effected through Method 18 analysis using a gas chromatograph
(GC) with flame ionization detector (FID). Method 18 analysis was conducted
concurrent with the Method 25A monitoring. The Method 18 results were used
primarily to establish the ratio of styrene emissions to MMA emissions during
gelcoating. This ratio allowed partitioning of the THC emission rate (as propane)
into styrene and MMA components. The Method 18 results were also used as a
check on the Method 25A results.
Other measurements made during this program included those

variables determined from other studies to have more significant influence on
emissions (Table 1-1).
1-7 Rev.1
1.3.4 Molds
Four molds were used in this program, three of which are boat part
molds in actual use by U.S Marine. Two hull molds represented the typical size
hulls common to the industry: an 18-ft runabout and a larger 28-ft cruiser. An
18-ft deck mold for a bow rider model was used to represent a more convex-
shaped mold. The mold used during Phase I testing by the CFA was included in
this program to provide data for comparison with the results of that previous
study. Surface areas of these molds (Table 1-2), determined from engineering
drawings and in previous studies, were used to calculate flux and normalized
emissions.
Table 1-1
Parameters Studied
Parameter Test Variable Measured/Fixed
Resin application method Yes (2) Fixed
Tool shape Yes (2) Fixed
Resin styrene content Yes (2) Measured (vendor)
Tool size Yes (2) Fixed
Gel time No Measured
Applied thickness No Measured
Resin densitya No Measured (vendor)
Resin percent non-volatilea No Measured (vendor)
Resin viscositya No Measured (vendor)
Resin peak exotherma No Measured (vendor)
Resin thixotropic indexa No Measured (vendor)
Resin flow ratea No Measured
Air flow rate No Measured
aThese parameters apply to resin and gelcoat.
1-8 Rev.1
Table 1-2
Surface Area of Molds Used in Test Program
Mold Surface Area (ft2)
18-ft Deck 171.09
18-ft Hull 220.5
28-ft Hull 454
CFA Phase I Mold (controlled spray) 28.06
CFA Phase I Mold (uncontrolled spray) 37.28a
aUncontrolled spray of this mold inevitably coated flange extensions, effectively increasing the
surface area of the final laminate structure.
1.3.5 Materials
Materials chosen for this program are typical of those used in

fiberglass boat manufacturing (Table 1-3). Two styrene polyester resins were
used, one with a nominal 35 percent styrene by weight content and the other
with a nominal 42 percent styrene by weight content. The gelcoat was a white on
white gelcoat containing nominally 32 percent styrene and 5 percent MMA. The
formulations used are considered most representative of the range of resin
styrene and gelcoat contents in the industry. Based on a survey of industry
representatives, the marine industry uses gelcoat that contains MMA as an
inhibitor to degradation by exposure to ultraviolet (UV) light. MEKP in a DMP
base is the most common catalyst used in the industry for both gelcoat and resin;
MEKP/DMP was used in this program.
1-9 Rev.1
Table 1-3
Properties of Gelcoat and Resin Materials
Gelcoat Resin 1 Resin 2
Product number 954WP53 80.654-NMMA 80.604-NMMA
Manufacturer Cook Composites Alpha/Owens Alpha/Owens
and Polymers Co. Corning Corning
Styrene content, % 32 35.1 42.2
MMA content, % 5 0 0
Specific gravity 1.32 1.082 1.09
Thixotropic index 5.5 4.9 5.71
Viscosity, cps RVF #2@ 20 RVF #2@ 20 rpm-
rpm-700 700
LVF #4@ 6 rpm- RVF #2@ 2 rpm- RVF #2@ 2 rpm-
14500 3400 4000
MEKP catalyst 2% 1.5% 2% 1.5% 2%
ratio
Gel time, min 16.5 23.22 17.98 23.48 18.70
1.3.6 Schedule
Testing was conducted from April 2 to April 19, 1997. Set up at the
site began March 30, continuing to April 1 (Table 1-4). One to three runs were
made each day. Sampling was coordinated with the U.S. Marine personnel
supporting the lamination and gelcoating operations and with RTI personnel
conducting sampling in an adjacent Method 204 enclosure.
1.4 Organization of the Report
The complete report including all appendices and separately

bound field data contains all data from the testing program. The text of the
report is intended to stand alone to provide the essence of the results from the
sampling and analysis program. This introduction (Section 1) provides a brief
1-10 Rev.1
Table 1-4
Test Schedule
Date Test Description Run
2-Apr NMMA-6-P 18-ft Deck Gelcoat 0402-01
3-Apr NMMA-8-1 28-ft Hull Gelcoat 0403-01
3-Apr NMMA-4-1 18-ft Deck 35 % Styrene Resin 0403-02
4-Apr NMMA-7-1 28-ft Hull 35 % Styrene Resin 0404-01
8-Apr NMMA-6-1 18-ft Deck Gelcoat 0408-01
9-Apr NMMA-11-1G CFA Mold Gelcoat 0409-02
9-Apr NMMA-11-1 CFA Mold 35 % Styrene Resin 0409-03
10-Apr NMMA-14-1 18-ft Deck 35 % Styrene Resin - Flow Chopper 0410-01
10-Apr NMMA-13-1 18-ft Hull 35 % Styrene Resin - Flow Chopper 0410-02
11-Apr NMMA-6-2 18-ft Deck Gelcoat 0411-01
11-Apr NMMA-14-2 18-ft Deck 35 % Styrene Resin - Flow Chopper 0411-02
11-Apr NMMA-13-2 18-ft Hull 35 % Styrene Resin - Flow Chopper 0411-03
15-Apr NMMA-16-1 18-ft Deck 42 % Styrene Resin-Flow Chopper 0415-01
15-Apr NMMA-15-1 18-ft Hull 42 % Styrene Resin-Flow Chopper 0415-02
16-Apr NMMA-16-2 18-ft Deck 42 % Styrene Resin-Flow Chopper 0416-01
16-Apr NMMA-15-2 18-ft Hull 42 % Styrene Resin-Flow Chopper 0416-02
1-11 Rev.1
introduction to the purpose of the testing and the schedule for completing the
on-site activities. The results of the sampling program are presented in Section 2.
Results are presented in tabular format and graphically to simplify review.
Section 3 describes the methods used in the sampling program, including those
used to quantify parameters not varied during the program, such as gel time and
styrene content. Quality control (QC) measures and QA data are presented in
Section 4.
Data and material supporting the test report are contained in the
appendices, as follows:
• Reference Methods
• Construction of a Temporary Total Enclosure for Volatile Organic Compound
Emission Assessment During Manufacture of Fiberglass Boats
• Calculations
• Procedure for and Results of Gel Time Determination
• Material Safety Data Sheets
• Certifications for Gas Standards and Equipment Calibrations
• Bag Standards
• Location and Results of Air Flow over the Mold Measurements
• Comparison of NMMA Results with Those from Other Studies
Copies of all raw data generated by the GC and the THC analyzer, organized by
day, and including pertinent process data (spray gun calibrations), have been
bound separately for archive at NMMA offices.
1-12 Rev.1
2.0 RESULTS
The results of emission testing at the U.S. Marine site in Arlington,

Washington, are presented in this section with comparison of the data between
tests and with the results of the CFA Phase I studies; further comparison of these
results with other studies is appended. The results are presented in terms of
emissions as a percent of available compound (i.e., styrene, MMA, and total
volatiles), as emissions per unit area of mold surface, and as normalized
emissions, that is, emissions per unit mass of resin per unit area of mold surface.
All calculations done for these analyses are appended (Appendix C), including a
sample calculation showing the complete analysis of data from test runs.
Results demonstrate that the enclosure constructed at the site

provided good capture of emissions from the molding process and that all
emissions were accounted for in the sampling. The styrene evaporation tests
demonstrated that the emission monitoring system provided a good measure of
the actual losses. Testing of emissions from lamination of the box mold from the
Phase I CFA studies demonstrated that the techniques used in the two studies
(i.e., CFA’s study and the portion of this study designed to reproduce some of
the testing done in the CFA study) were comparable. Results of duplicate testing
for each test case provided good closure, within the acceptance criteria specified
in the QAPP. In accordance with the plans outlined in the QAPP, all tests of boat
mold use were completed with only two runs. A third run was done for one of
the test conditions to incorporate the CFA study mold.
2.1 Summary of Results
Table 2-1 presents the average values from the sixteen emission
tests conducted at U.S. Marine’s Arlington, Washington, site. Three of the tests
2-1 Rev. 1
were conducted using the same box mold used by CFA/Dow during their
Phase I studies (to provide data for comparison with those studies); all other
tests were conducted using boat molds and procedures typical of U.S. Marine’s
boat manufacturing processes. The results of this study are taken as
representative of industry practice.
The amount of resin used per part during these studies was much
greater than the amounts used in previous studies (Table 2-1). This relative scale
of material use and measurement contributed to the consistency of results.
Table 2-1
Summary of Test Results
Emissions as Percent of Available
Emission Normalized
Resin Flux Emissions
Descriptiona Use (lb) Styrene MMA Volatiles (lb/SF) (lb/1000 lb/SF)
CFA Gel 3.2 41.1% 84.0% 46.9% 0.0203 6.36
18 Deck Gel 20.0 42.3% 75.6% 46.9% 0.0228 1.14
18 Hull Gel 25.7 46.4% 73.2% 50.0% 0.0261 1.03
28 Hull Gel 65.7 50.4% 78.6% 54.3% 0.0389 0.59
CFA 35R 8.1 41.8% 41.8% 0.0345 4.89
18 Deck 35R 124.6 12.9% 12.9% 0.0130 0.11
18 Hull 35R 144.5 14.8% 14.8% 0.0143 0.10
28 Hull 35 R 354.4 17.3% 17.3% 0.0234 0.07
b
18 Deck 35R - FC 114.7 11.9% 11.9% 0.0092 0.08
18 Hull 35R - FC 141.7 10.8% 10.8% 0.0072 0.05
CFA 42R 5.1 48.7% 48.7% 0.0324 6.28
18 Deck 42R 111.8 21.1% 21.1% 0.0284 0.25
18 Hull 42R 142.9 20.7% 20.7% 0.0272 0.19
28 Hull 42R 304.2 23.3% 23.3% 0.0357 0.12
18 Deck 42R-FC 122.2 13.4% 13.4% 0.0125 0.10
18 Hull 42R-FC 154.9 11.4% 11.4% 0.0089 0.06
aSee Table 1-4 for full description.
bApplication using flow chopper.
The THC measurements made during this program serve as the

principal basis for estimating emissions. Figure 2-1 presents summary
2-2 Rev. 1
Figure 2-1
Summary Concentration Traces for Boat Lamination and Gelcoating
2-1[a] - Spray v. Flow Chopper: 18 Hull and Deck - 35% Styrene Resin
300
THC Concentration (ppmv propane)
250
200
150
100
50
0
0:00 0:28 0:57 1:26 1:55 2:24 2:52
Elapsed Time
35R 35R FC
2-1[b] - Spray v. Flow Chopper: 18 Hull and Deck - 42% Styrene Resin
500
450
400
350
300
250
200
150
100
50
0
0:00 0:28 0:57 1:26 1:55 2:24 2:52 3:21
Elapsed Time
42R 42R FC
2-3 Rev. 1
2-1[c] - Spray: 28 Hull - Both Resins

700
600
500
400
300
200
100
0
0:00 1:12 2:24 3:36 4:48 6:00 7:12
Elapsed Time
35R 42R
2-1[d] - Gelcoating
900
800
700
600
500
400
300
200
100
0
0:00 0:28 0:57 1:26 1:55 2:24 2:52
Elapsed Time
18 Deck Gel 18 Hull Gel 28 Hull Gel
2-4 Rev. 1
concentration profiles for the testing conducted during the April 1997
deployment. Each of the four frames presents THC concentration traces derived
by averaging data from the individual test runs. This figure illustrates some of
the differences in the data summarized in Table 2-1.
For example, the data from sampling the exhaust during

lamination of the 18-ft mold illustrates the differences between spray up
application and application by flow chopper, as well as the increased THC
concentrations in the exhaust noted when laminating with a higher styrene
content resin (42.2 percent styrene). The data from the 28-ft mold lamination test
show that, although the peak concentrations using the two resins were
essentially the same, the concentration trace for the 42 percent styrene resin has
broader peaks, corresponding to greater mass emission rates.
The summary illustration of the gelcoat data shows the effect of

geometry on the concentrations measured. The greater THC concentrations in
the exhaust noted during lamination of the 28-ft mold are likely the result of the
broader strokes used in gelcoating the part and the distance between the
operator and the part during gelcoating. The atomized gelcoat remains airborne
for a longer period of time, resulting in greater volatilization and emissions of
VOC. The operator can stand closer to the 18-ft hull and make shorter strokes to
apply the gelcoat. Also, the operator can coat half of the 18-ft mold surface in a
single pass, rather than the minimum two passes required to coat the 28-ft hull
mold; this difference is noted in the differences in peak widths. Finally, the 18-ft
deck mold is relatively flat, with complex geometry requiring the operator to
work close to the mold surface and coat some portions of the mold by brush,
reducing the amount of gelcoat atomized.
2-5 Rev. 1
The results presented in the tables and figures in this report reflect
the exclusion from testing of emissions in the long, low-concentration “tail” in
each THC trace for each run. This phenomenon has been noted in previous
studies and was highlighted in the QAPP as a concern for completing this testing
in a timely fashion. Section 2.2 presents the results from some of the early test
runs that illustrate that elimination of the “tail” has little if any effect on the
results reported herein.
All the data related to emissions from boat gelcoating and

lamination derived from this NMMA testing program are summarized in
Figure 2-2. This figure presents the results for all tests in terms of emissions as a
percent of available monomer. As the results clearly indicate, emissions from
gelcoating differ from those associated with resin lamination. Sections 2.4 and
2.5 contain more detailed presentations of the results from gelcoating and resin
lamination emission testing.
2.2 Evaluation of Emissions from Continuing Trace Concentrations
The QAPP included the evaluation of a cutoff concentration in the

exhaust stream that would allow cessation of sampling before the concentration
of THC in the enclosure exhaust reached the background concentration,
provided at least 95 percent of total emissions were characterized before
cessation of sampling. The first test runs were used for this assessment. The
assessment determined that 95 percent of total emissions from lamination and
gelcoat operations would be characterized by the time the exhaust concentration
reached 75 ppmv as propane. Ceasing sampling at this concentration would
reduce run sampling by about 1 hour per run, allowing more sampling runs per
day.
2-6 Rev. 1
Figure 2-2
Emissions as Percentage of Available Monomer
Emi ssi ons as Percent of Available M onomer for Boat Tests
60%
Emissions (% of available monomer)
50%
40%
30%
20%
10%
0%
As it turned out, it was practical to continue the test runs until

concentrations measured about 7 ppmv as propane. Because the background
concentration was about 2-4 ppmv, the difference in concentration was about
4 ppmv, relating to less than 2 ppmv as styrene and representing only
0.2 percent of emissions.
The assessment of the concentration trace for NMMA-4-1

(lamination of an 18-ft deck with 35 percent styrene resin) illustrates this
analysis. The concentration trace for run NMMA-4-1 (Figure 2-3) shows the
typical pattern of concentration peaks and valleys as resin is applied
systematically to the sides of the mold. Resin and glass are applied to the first
side, and the mixture is rolled out to remove air bubbles. After the initial coat on
this half of the mold begins to gel, the mold is rotated and the first coat is
applied to the second side. After the coat on the second side begins to gel, the
mold is rotated again and the second layer of glass and resin is applied to the
2-7 Rev. 1
first side, with some of the glass being woven roving. This process continues
until the specified thickness is achieved. As part of the manufacturing
procedure, each layer of glass and resin are rolled out. After the final coat of
resin is applied and the peak exotherm of the polymerization reaction is
achieved, the concentration of styrene in the exhaust stream begins to decrease.
All the glass and resin were applied to this mold in about 1 hour. The styrene
concentration in the exhaust stream decreased to near background levels in
about 1 hour after that point.
Figure 2-3
Total Hydrocarbon Concentration Trace for NMMA-4-1
N M M A - 4 - 1 ( 1 8 D e c k 3 5 R)
400 100%
350 90%
80%
Cumulative Mass Emissions

300
THC Concentration (ppmv
70%
250
60%
propane)
(percent)
200 50%
40%
150
30%
100
20%
50
10%
0 0%
14:55:44
15:05:49
15:15:44
15:25:44
15:35:44
15:45:44
15:55:44
16:05:44
16:15:44
16:25:44
16:35:44
16:45:44
16:55:44
T i m e ( PST ) THC o (ppm)

Cumulative M ass Percent
Because the THC associated with the latter part of the

concentration trace (from 7 to 4 ppmv) represents about 0.2 percent of emissions,
the emissions associated with sampling completely to background levels from
the cutoff concentration were less than 0.015 lb. Even this amount represents a
worst case assessment (i.e., greatest emissions in the tail) because it assumes that
2-8 Rev. 1
emissions from all tests are the same, whereas the total emissions from
laminating the 18-ft deck mold were less than the emissions from laminating
either other part evaluated. The 0.015-lb total would represent a larger portion of
emissions for smaller parts, however, such as those included in the previous
CFA/Dow and EPA/RTI studies.
Another approach to assessing an appropriate cutoff concentration

is depicted in Figure 2-4, a plot of outlet concentration and cumulative
emissions. This figure, developed from the same NMMA-4-1 test run, shows that
95 percent of total emissions from lamination were quantified when the exhaust
concentration decreased to about 75 ppmv as propane.
Figure 2-4
Concentration v. Cumulative Mass Emissions (NMMA-4-1)
Exhaust Concentration v. Cumulative Mass Emissions
400
350
THC Concentation outlet (ppmv
300
250
propane)
200
150
100
50
0
70% 75% 80% 85% 90% 95% 100%
Cumulative Mass Emissions
2-9 Rev. 1
2.3 CFA Mold Results
Part of the test matrix for this program included lamination of the
mold used in the CFA Phase I study so that results of both studies could be
compared. Testing was conducted using the 35 percent styrene and 42 percent
styrene resins, nearly identical to the styrene contents in the resins used in the
CFA Phase I study. The initial test run using 35 percent resin (NMMA-11-1R)
was conducted using the techniques employed by experienced laminators
during actual boat manufacturing. This approach resulted in lower emissions
than those measured during the CFA study. To allow comparison to the CFA
study, subsequent tests were conducted using methods that resulted in
“uncontrolled” spray, including one run during lamination by a completely
inexperienced operator (NMMA-11-2R), as evinced by the 10.1 lb of resin used.
Although not included among the planned tests, one gelcoat run (NMMA-11-1G)
was also conducted by an experienced operator using good technique.
The results of testing during coating of the CFA mold (Table 2-2
and Figure 2-5) indicate an average loss of 38.0 percent of available styrene
during lamination using 35 percent styrene resin (not using Run 11-1R in the
average) and 48.7 percent of available styrene loss during lamination using
42 percent styrene resin. The results show a higher degree of variability than the
results for lamination on the boat molds, largely as a result of the high degree of
variability in application technique for inexperienced operators and the
influence of the resin weight measurement. The resin extraction equipment
disproportionately affected the measurement of small quantities of resin use
(Section 3.2.4). Although the results of the various test runs met acceptance
criteria, they were not as consistent as the results of emission testing during use
of boat molds. Emissions measured during CFA mold lamination in this study
were about twice those of comparable tests from the CFA study.
2-10 Rev. 1
Table 2-2
Emissions Measured During Application of Resin to CFA Mold
Percent
Resin of Flux Normalized
Run Material Content (lb) Available (lb/ft2) (lb/1000lb/ft2)
11-1G Gelcoat 37.0 3.2 46.9 0.0203 6.36
11-1R-Ctrl Resin 35.1 3.5 30.6 0.0114 3.25
11-2R-Unc Resin 35.1 10.1 29.9 0.0233 2.31
11-3R-Unc Resin 35.1 6.1 53.6 0.0456 7.48
12-1R-Unc Resin 42.2 5.4 55.7 0.0441 8.16
12-2R-Unc Resin 42.2 4.73 41.7 0.0208 4.39
Styrene losses were also considered in terms of emissions per unit

area and normalized emissions. Average losses per unit mold surface area (flux)
were 0.0268 lb/ft2 and 0.0325 lb/ft2 for the 35 and 42 percent styrene resins,
respectively. Normalized losses averaged 4.35 lb/1000 lb/ft2 for the 35 percent
styrene lamination and 6.28 lb/1000 lb/ft2 for the 42 percent styrene resin. Like
the results in terms of percent of available styrene, these results indicate greater
losses from use of greater styrene content resins than would be proportional to
the actual styrene content. The flux losses show a different trend, but
comparison of the emissions in this format is inconclusive because of the only
slight difference in averages and the difficulty in consistent resin weight
measurement.
In this test series of the program (i.e., using the CFA mold),
emission measurements were not always comparable to the amount of resin
used, i.e., the emissions as percent of available monomer was not always
consistent. The equipment was set up to measure large quantities of resin and
gelcoat continuously. The expected quantities were between 100 and 300 lb resin
for each test run. The scale selected, therefore, had an accuracy of 0.5 lb, which
was satisfactory for the boat lamination and gelcoating tests, but which did not
2-11 Rev. 1
Figure 2-5
Emissions from Lamination and Gelcoating of the CFA Mold
CFA M ol d - Em i ssi ons ( % A v a i l a b l e )
90%
80%
Emissions (% available
70%
60%
Styrene
monomer)
50%
MMA
40%
Volatiles
30%
20%
10%
0%
CFA Gel CFA 35R CFA 35R CFA 35R CFA 42R CFA 42R
CFA M ol d - Em i ssi ons ( l b / SF)
0.050
0.045
0.040
0.035
Emissions (lb/SF)
0.030
0.025
0.020
0.015
0.010
0.005
0.000
CFA M ol d - Em i ssi ons ( l b / 1 0 0 0 l b Resi n / SF)
9.0
8.0
Emissions (lb/1000 lb Resin/SF)
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
2-12 Rev. 1
prove to be adequate for measuring the small quantities of resin (on the order of
only 3.5 lb) used to laminate the CFA mold. Thus, the precision of the scale used
for resin and gelcoat led to some results from CFA mold testing that were more
variable than results from boat testing.
2.4 Results of Gelcoat Application Testing
All gelcoat application was done to a uniform thickness of 20 mils

with a gelcoat containing 32 percent styrene and 5 percent MMA. A preliminary
test run (NMMA-6-P) was made using a U.S. Marine production gelcoat
supplied by the same supplier for the test gelcoat. Although the production
gelcoat had different styrene content (approximately 31 percent), the results
generated were included in this report because they are very consistent with the
other results from gelcoat testing in this study (Figure 2-6).
Results for boat gelcoating are presented in terms of losses of

styrene, MMA, and total volatiles (Figure 2-6). The total emissions measured for
the deck and two hull sizes are consistent (especially for styrene and total
volatiles), with slightly greater emissions resulting from gelcoating the hulls,
and the greatest emissions resulting from gelcoating the larger hull (Table 2-3).
The most notable finding of this testing is the relative contribution to emissions
of the individual constituents in the gelcoat. From 60 to 80 percent of the MMA
in the gelcoat is lost as emissions in the gelcoat process. MMA is more volatile
than styrene and, thus, is lost at a much greater rate than the styrene. Only one
test run resulted in MMA losses of less than 60 percent of available MMA. That
run (NMMA-6-1) was discarded from the average values presented in Table 2-3
as unrepresentative because of variability of the GC response factors.
2-13 Rev. 1
Figure 2-6
Gelcoat Emissions as Percent of Available Volatiles
90%
80%
70%
Emissions (% available monomer)
60%
50% Styrene
MMA
40% Volatiles
30%
20%
10%
0%
CFA Gel
18 Deck Gel
18 Deck Gel
18 Deck Gel
18 Hull Gel
18 Hull Gel
28 Hull Gel
28 Hull Gel
(USM)
Table 2-3
Emissions from Gelcoating as Percentage of Available Volatiles
Percent of Percent of Percent of
Gelcoat Use Available Available Available
(lb) Styrene MMA Volatiles
18-ft Deck 20.0 42.3 75.6 46.9
18-ft Hull 25.7 46.4 73.2 50.0
28-ft Hull 65.7 50.4 78.6 54.3
The results from the first test run on the 18-ft deck after the
preliminary test using gelcoat (the second run for the 18-ft deck) yielded a loss of
available MMA that was substantially less than all other measurements made
during all other gelcoat test runs, only about 25 percent compared with 60 to 85
2-14 Rev. 1
percent of available MMA. Excluding that one run (NMMA-6-1) provides an

average MMA loss of 75.6 percent of available MMA, a value that is very
consistent with the MMA loss values for the 18-ft and 28-ft hulls. The average
styrene loss for this test was relatively unchanged as a result of excluding that
one run; likewise, the total volatile loss was relatively unchanged, owing to the
contribution of styrene losses to the total volatile loss.
Gelcoating results evaluated in terms of loss per unit mold surface

area and normalized emissions (i.e., emissions per unit mass of resin per unit
mold surface area) show some interesting trends (Table 2-4 and Figure 2-7). As
would be expected, the emissions from the 18-ft molds are essentially equal,
owing to nearly equal resin use and surface area. Normalized emissions appear
to be inversely related to mold area, assuming the same thickness of the gelcoat
is applied. Normalized emissions show the general trend of increasing emissions
with decreasing mold surface area. This trend was also evident from the
emission testing for resin lamination.
Table 2-4
Total Emissions from Gelcoating Related to Mold Surface Area
Gelcoat Use Mold Area Emissions Emissions
(lb) (ft2) (lb/ft2) (lb/1000 lb/ft2)
18-ft Deck 20.0 171.09 0.0228 1.14
18-ft Hull 25.7 220.5 0.0261 1.03
28-ft Hull 65.7 454 0.0389 0.59
2-15 Rev. 1
Figure 2-7
Emissions from Gelcoating Related to Mold Surface Area
0.045 7.0
0.040
6.0
0.035
5.0
0.030

Emissions (lb/SF)
4.0
0.025
0.020
3.0
0.015
2.0
0.010
1.0
0.005
0.000 0.0
CFA 18 18 18 18 18 28 28
Gel D eck D eck D eck Hull Hull Hull Hull
Gel Gel Gel Gel Gel Gel Gel
(USM )
Flux Normalized
2.5 Results of Resin Lamination Testing
Testing was conducted during lamination of three mold size and

shape combinations selected to be representative of the majority of resin use in
the industry. Because the quantities of resin used for each test are much greater
than those used in previous testing, the results were very reproducible
(Table 2-5).
Figures are provided to illustrate the tabulated data. All tests are
shown in the figures and are grouped by resin (styrene content) to allow
comparison of the results across the different mold sizes and shapes.
2-16 Rev. 1
Table 2-5
Emissions Measured During Resin Lamination
Emissions
Percent of Normalized
Resin Use Available (lb/1000
Test (lb) Styrene Flux (lb/ft2) lb/ft2)
18 Deck 35 R 124.6 12.9 0.0130 0.11
18 Deck 35 R FC 114.7 11.9 0.0092 0.08
18 Deck 42 R 111.8 21.1 0.0284 0.25
18 Deck 42 R FC 122.2 13.4 0.0125 0.10
18 Hull 35 R 144.5 14.8 0.0143 0.10
18 Hull 35 R FC 141.7 10.8 0.0072 0.05
18 Hull 42 R 142.9 20.7 0.0272 0.19
18 Hull 42 R FC 154.9 11.4 0.0089 0.06
28 Hull 35 R 354.4 17.3 0.0234 0.07
28 Hull 42 R 304.2 23.3 0.0357 0.12
2.5.1 Emissions as Percent of Available Styrene
Figure 2-8 presents the data for emissions as a function of available

styrene. In each set of test runs for a given mold size and shape, the same pattern
of emissions is seen: the emissions from the 42 percent styrene resin lamination
are greater than the emissions from lamination using 35 percent styrene resin,
and the use of a flow chopper to apply the resin does result in less emissions.
The latter effect is likely the result of not atomizing styrene in the spray up
process. Based on these measurements, the effect of flow chopper use on
emissions is greater for the higher styrene content resins.
The greatest emissions in terms of percent available styrene were

measured for the larger mold size and the greater styrene content resin. An
average loss of 23.3 percent of available styrene was measured from laminating
the 28-ft hull with 42 percent styrene resin. Only 17.3 percent loss was measured
for the same hull using 35 percent resin. Because this format of presenting
2-17 Rev. 1
Figure 2-8
Emissions from Lamination as Function of Available Styrene
25%
20%
15%
10%
5%
0%
18 Hull 42R-FC
18 Hull 42R-FC
18 Hull 35 R
28 Hull 35 R
28 Hull 35 R
18 Deck 42R-FC
18 Deck 42R-FC
18 Hull 35R
18 Hull 42R
18 Hull 42R
28 Hull 42R
28 Hull 42R
18 Hull 35R - FC
18 Hull 35R - FC
18 Deck 35 R
18 Deck 35 R
18 Deck 42R
18 Deck 42R
18 Deck 35R - FC
18 Deck 35R - FC
emissions should account for differences in the resin styrene content, the results
indicate that emissions are not directly proportional to resin styrene content, as
might have been inferred from an emission factor based on percent available
styrene.
The comparisons become clearer when the same data are reviewed
in terms of the various molds and tests for the different resins. Figure 2-9
presents emissions as a percentage of available styrene in the resin for the two
resins tested. One observation from these results is that emissions as a
percentage of available styrene for a given styrene content in the resin increase
with surface area of the mold for surface areas substantially greater (i.e., the 28-ft
2-18 Rev. 1
hull). This trend is apparent for the two styrene content resins tested. Also, the
use of a flow chopper to apply resin reduces the percentage of available styrene
emitted for a given mold size or shape. Greater reductions are noted for the
greater percentage styrene content resin, but in general, reductions of 8 to
45 percent are noted: from about 12.9 percent available to 11.9 percent available
for the 35 percent styrene resin (18-ft deck) and from 20.7 percent available to
11.4 percent available for the 42 percent styrene resin (18-ft hull).
Figure 2-9
Comparison of Results by Test
25%
20%
15%
10%
5%
0%
2-19 Rev. 1
2.5.2 Flux Emissions
Emissions per unit area of mold surface (Figure 2-10) show the
consistency of the runs made for testing and indicate general trends in the data.
For example, emissions per unit surface area increase with the increase in resin
styrene content. Also, the results are fairly consistent for the two 18-ft mold
types, despite the difference in configuration. The hull is a concave mold and the
deck (a bow rider design) has convex character.
Figure 2-10
Emissions from Resin Lamination as Flux Measurement
0.040
0.035
0.030
0.025
Emissions (lb/SF)
0.020
0.015
0.010
0.005
0.000
18 Deck 35R - FC
18 Deck 35R - FC
18 Deck 42R-FC
18 Deck 42R-FC
18 Hull 35R - FC
18 Hull 35R - FC
18 Hull 42R-FC
18 Hull 42R-FC
18 Deck 35 R
18 Deck 35 R
18 Deck 42R
18 Deck 42R
18 Hull 35 R
18 Hull 35R
18 Hull 42R
18 Hull 42R
28 Hull 35 R
28 Hull 35 R
28 Hull 42R
28 Hull 42R
The same data, reduced to averages for each test set, are presented
in Figure 2-11 for the two different styrene resins tested. These results again
illustrate that emissions from the 42 percent styrene resin are greater than the
2-20 Rev. 1
emissions from the 35 percent styrene resin lamination. In this case, because the
mold surface areas are more nearly equal, flux emissions from the surface of the
two 18-ft molds are almost equal. Figure 2-11 also shows that emissions from
lamination using a flow chopper are less than the baseline cases.
Figure 2-11
Average Emission Flux Measurements for Two Resins
0.040
0.035
0.030
0.025
Emissions (lb/SF)
0.020
0.015
0.010
0.005
0.000
2.5.3 Normalized Emissions
Figure 2-12 compares results from all resin lamination test runs
presented as normalized emissions (mass of emissions per mass of resin per
surface area). For both resins, the normalized emissions decrease with increasing
surface area, indicated by the 18-ft deck to the 18-ft hull to the 28-ft hull.
Normalized emissions were about 80 percent greater from lamination with 42
percent styrene resin. Similar trends are noted for the emissions from lamination
using a flow chopper; normalized emissions decrease with increasing mold
2-21 Rev. 1
surface area and are always less than the normalized emissions for the
corresponding spray up operation.
The differences between emissions generated using spray devices

and those generated using flow choppers can be seen in Figure 2-1, which
compares the concentration traces for 18-ft hull and deck mold lamination using
two different styrene content resins and these two application methods. The
difference in the concentration profile is due to the application devices. The
utility of the flow chopper in reducing the amount of styrene atomized in
application is much greater for the higher styrene content resin. As Figure 2-1
illustrates, the concentration tail after resin is applied is nearly identical for these
two application methods.
Figure 2-12
Normalized Emissions from Resin Lamination
0.30
0.25
0.20
0.15
0.10
0.05
0.00
18 Deck 42R-FC
18 Hull 42R-FC
18 Hull 35R
18 Deck 42R
18 Hull 42R
28 Hull 42R
18 Deck 35R - FC
18 Hull 35R - FC
18 Deck 35 R
28 Hull 35 R
2-22 Rev. 1
3.0 CALIBRATION AND SAMPLING PROCEDURES
The extensive battery of tests conducted and measurements made

for this program includes those that were repeated routinely, either daily or
during every test run, and those that were repeated only once or twice over the
course of the whole program (Table 3-1). This section describes the procedures
for all the tests conducted and measurements made during this program,
including QC procedures (calibrations) and QA procedures followed to produce
data for assessing the quality of the other measurements. The procedures are
presented in terms of those associated with the process of manufacturing the
part (i.e., hull or deck) and those associated with the measurement of emissions.
Additional measurements (i.e., of air flow over the mold and styrene recovery)
and data handling are also addressed in this section.
Table 3-1
Measurements During Test Program
Measurement Designation Frequency
THC Concentration Method 25A Continuous
Styrene (MMA) Concentration Method 18 Intermittent (6-15 min
intervals)
Wet Laminate Thickness Wet Mil Gauge Each Test Run
Wet Film Thickness Wet Mil Gauge Each Test Run
Styrene (MMA) Content Manufacturer’s Method Per Batch Delivered
Gel Time U.S. Marine Method Per Batch Delivered
Resin Flow Rate Spray Gun Calibration Pre/Post Test Daily
Air Flow Hot Wire Anemometer For Each Mold
Air Temperature Exhaust Gas Temperature Continuous
Exhaust Flow Rate Vent Stack Air Flow Rate One traverse set per test;
continuous point
measurement
Styrene Recovery Material Balance/Method 25A Pre/Post Program
Resin/Gelcoat Weight Precision Balance Pre/Post Test; continuous
measurement
Two test runs were made most test days; the number of tests
completed in a day was dictated by the duration of the test and the availability
3-1 Rev. 1
of material, molds, and personnel to construct the parts. On some days, only a
single test run was completed, while on other days, as many as three test runs
were completed. The procedures implemented most test days include the
following:
• Morning spray equipment calibration;

• Morning leak check and calibration of the THC analyzer and the GC;
• Measurements made during the test runs, including THC, styrene and MMA
(using the GC), flow rate, weight of material (resin or gelcoat) used during
the run, and temperature;
• Post-test quality control measurements (calibrations);
• Measurements for second test run (THC, styrene and MMA, flow rate,
weight, and temperature);
• Post-test quality control measurements; and
• Evening spray gun flow rate calibration.
Additional tests conducted include measuring the rate of air flow

over the mold and mass balance styrene recovery checks (QA).
3.1 Process Procedures and Measurements
The procedures used to apply the resin were those used by the
skilled U.S. Marine staff in applying gelcoat or resin/glass in the manufacturing
operation. Resins and gelcoat selected for this program had characteristics
typical of industry use and meeting the goals of this program (Section 1.3.5).
Staff used to manufacture molds and to train lamination staff performed all
gelcoating and lamination during this program. Additional personnel assisted
when glass was rolled out or woven roving added to the mold.
3-2 Rev. 1
3.1.1 Application Procedures
After the morning calibrations were completed, monitoring of air

flow was established, materials to be applied were weighed, and lamination (by
spray up or flow chopper) or gelcoating of a part in the TTE commenced. Skilled
operators applied resin and gelcoat as they would in normal production. The
operators measured laminate and gelcoat thickness periodically as the material
was applied, using a chop mil gauge for laminate thickness and a gelcoat mil
gauge for gelcoat thickness. During the application of both materials, multiple
readings were taken at three locations along the length of the mold as the
operators fabricated the part. The operators did not record gauge readings but
applied material to the desired thickness, in accordance with normal operating
procedures.
Gelcoat was applied using spray guns with a distance of 18-24 in.
from spray tip to mold surface for smaller (18-ft) parts. The distance was greater
for gelcoating the 28-ft hull mold, reaching 4-5 ft at the greatest distance from the
application equipment to the keel of the hull. When gelcoating, a fog coat was
applied first to one side of the mold and then a thicker layer was applied to
bring the coating to a nominal thickness of 20 mils. Using typical gelcoating
procedure, the operator checked the gelcoat thickness several times in at least
three locations on each side of the mold to ensure uniform thickness. After one
side was coated, the mold was rotated, and the process was repeated to coat the
other half of the mold surface. Complex portions of some molds, such as
encountered on the 18-ft deck mold, were coated using a cup and brush
technique. These portions were typically inaccessible with the spray equipment.
A similar procedure was used to apply resin and glass to the 18-ft
hull and deck and the 28-ft hull molds. In this procedure, though, a skin coat of
3-3 Rev. 1
resin and glass of about 90 mil thickness was applied first and allowed to cure
almost entirely before applying additional layers of the laminate. Additional
layers were then added by alternating resin/glass chop and resin/woven
roving. The chopped glass was rolled out at every layer. The laminate
thicknesses for the hulls and decks were much greater than the laminate
thicknesses used in previous studies. For example, a laminate thickness of
414 mils was used in the bottom of the 28-ft hull (including three layers of
woven roving, each about 48 mil thickness), and a thickness of 198 mils was used
for the sides and transom of the same hull (including one roving). A 0.010-in.
thickness gauge was used to make the measurements.
3.1.2 Spray Equipment Calibration
The spray gun resin or gelcoat flow rate was determined at the
beginning of every test day and confirmed at the end of every test day, that is,
before the first test and after the last test. The spray equipment flow rate was
adjusted (resin pump pressure was adjusted and spray tip sizes changed) by the
operator to provide the normal resin and gelcoat output for manufacturing
operations, and the flow rate was recorded. The flow rate calibration consisted of
spraying resin (or gelcoat) into a pre-weighed container for 15 seconds,
recording the material weight, and calculating the rate in pounds per minute.
For spray-up application, the glass content was adjusted to the required resin
flow rate; after the resin flow rate was established, the chopper speed was
adjusted to provide the proper glass ratio. The spray pattern was evaluated by
the experienced operators by spraying resin and glass onto a cardboard panel.
Adjustments were made based on the visual pattern to achieve the desired
result.
3-4 Rev. 1
The resin/gelcoat-to-catalyst ratio was calibrated by comparing the

catalyst volume measurement to flow rate and adjusting the catalyst ratio by the
slave pump setting. Both resin and catalyst are delivered to the applicator head
by positive displacement pumps. Because the air-actuated piston pumps deliver
a fixed amount of material with every stroke, the amount of catalyst delivered to
the applicator head can be accurately set based on the stroke count. The amount
of catalyst delivered to the applicator is very small compared with the amount of
resin delivered, so the amount of resin is set first and calibration completed.
Then the stroke count of the resin pump is set to the desired catalyst ratio. The
catalyst ratio is adjusted after application of the skin coat and then these stroke
count settings were maintained throughout each day.
3.1.3 Gel Time
Gel time was measured on site for each batch of resin used
(Table 3-2). The procedure used to measure gel time (Appendix D) is the same
procedure used by U.S. Marine in determining gel time for production resin and
gelcoat. Gel time for gelcoat was determined using a 2 percent 925 clear catalyst
(MEKP); a gel time of about 16.5 minutes is typical for the gelcoat. Two gel times
are common for lamination, one for the skin coat and another for the bulk of the
lamination. The skin coat gel time for the 35 percent styrene resin was about
18 minutes using a 2 percent Red 925 MEKP catalyst; the remainder of the
lamination was done using a gel time of about 23.2 minutes with a 1.5 percent
Red 925 MEKP catalyst. The latter materials achieved a peak gel temperature in
26 (for the 1.5 percent catalyst) to 34 minutes (for the 2 percent catalyst).
Comparable gel measurements were made for the 42 percent styrene resin.
3-5 Rev. 1
Table 3-2
Gel Time Results
1.5% Catalyst 2% Catalyst 2.5% Catalyst
Gelcoat
Gel time @ 77°F 16:30
35% Resin
Gel time 23:13 17:59 15:38
Gel time to 150°F 6:05 5:06 3:09
Time to Peak Temperature 33:57 26:18 22:11
Peak Temperature (°F) 350 360 376
42% Resin
Gel time 23:29 18:42 16:38
Gel time to 150°F 6:17 5:05 5:04
Time to Peak Temperature 35:55 28:47 27:08
Peak Temperature (°F) 320 352 367
3.1.4 Styrene and Methyl Methacrylate Content
Styrene and MMA content of the resins and gelcoat used during
this study were determined by the resin suppliers: Alpha/Owens-Corning for
the resin and Cook Composites and Polymers for the gelcoat. Industry
procedures were followed and certifications of styrene and MMA content
provided (Table 3-3). Material safety data sheets with material certifications are
in Appendix E.
Table 3-3
Styrene and MMA Contents
Material Styrene Content (wt. %) MMA Content (wt. %)
Laminating Resin (35 R) 35.1 0.0
Laminating Resin (42 R) 42.2 0.0
Test Gelcoat 32.0 5.0
Production Gelcoat (USM) 30.95 4.95
3-6 Rev. 1
3.2 Emission Testing Equipment and Procedures
This section describes the equipment used to sample emissions

from the enclosure and to measure other critical test parameters such as flow rate
from the enclosure, temperature of the exhaust stream, and amount of resin
applied. QC checks made on the sampling system are also presented in this
section.
Samples were collected from the enclosure’s exhaust duct through

Teflon® sample tubing at ambient temperature using a sample pump. This
sample was fed continuously to a THC analyzer and to the GC used to speciate
the organic compounds in the exhaust stream. The THC data were collected
continuously and recorded in a computer data acquisition system (DAS) through
a Dianachart data logger. GC data were collected separately.
The following data were collected continuously:
• TTE outlet THC concentration,

• TTE inlet THC concentration,
• Velocity pressure head in the exhaust duct,
• Temperature of the exhaust stream, and
• Weight of the gelcoat or resin used.
3.2.1 Leak Check and Instrument Calibration
Leak checks and bias checks were incorporated into the calibration
and response factor procedures, respectively. Calibration gases were introduced
at the probe inlet, and calibration responses were obtained using the same
procedures as those used for sample responses. Certifications for all gas
3-7 Rev. 1
standards are included in Appendix F. No vacuum leak checks were performed

because any leaks, if present, would be small enough to be “calibrated out” and
stable enough to have a negligible effect on post-test drift measurements. Any
bias introduced by the system was similarly subsumed in the response factor
determinations because they were performed with the entire sampling system.
Instrument calibration checks were also made at the instrument

and the results compared with the overall calibration checks made at the sample
probe tip. The two sets of results were consistent, indicating that there was little
or no bias in the sampling system. Checks by the EPA during their technical
system audit verified the findings of the NMMA checks.
3.2.1.1 Total Hydrocarbon Analyzer
THC was measured at the outlet and inlet of the enclosure using an
FID in accordance with EPA Method 25A (Appendix A). Analysis of the outlet
gas stream was made using a Thermo Electron Model 51H THC analyzer. A
Ratfisch Model 52RF analyzer was used to monitor THC content in the inlet air
to the enclosure. A second Ratfisch was held on stand-by in case one of the other
instruments failed. Both instruments were calibrated using the same procedure.
Before testing each day, the calibration of the FID was verified
using zero, low span (30 percent), mid-span (60 percent), and 90 percent span
gases. Only one detection range, 1000 ppmv as propane, was used, which
covered concentrations of styrene from 0 to approximately 416 ppmv.
Calibration was simplified by using the DAS instead of recalibrating the two
instruments. That is, instead of using adjustment potentiometers on the
instruments themselves, the response signal to the DAS computer was adjusted
in the computer program. This approach, which basically acknowledges that the
3-8 Rev. 1
DAS is an integral part of the entire sampling, analysis, and data acquisition
system, allowed more rapid calibration of the system each morning. An
advantage, beyond mere simplification of the process, was that calibration
incorporating the DAS allowed calibration of the entire system, including
sampling lines, instrument, and data systems, incorporating potential biases
created by sample extraction equipment, instrument response, and the DAS into
a single system response. The acceptable calibration error of less than 10 percent
was met for both instruments for all tests (Section 4.1.1).
The FID was calibrated with propane. Instrument response factors

were periodically developed for styrene and MMA by analyzing known
concentrations of the target compounds in a manner identical to that employed
for propane calibration. This procedure generated a correlation of relative
response of the compounds on the THC instrument, a factor needed to compare
the results of Method 18 sampling with Method 25A sampling.
After the first test of the day, the instrument drift of the analyzer
was checked using zero, 30 percent, and 60 percent span gases. When the drift
was within the acceptable range, testing proceeded; otherwise, the instrument
was recalibrated. The post-test calibration drift was recorded before any
recalibration took place. The results of these measurements are reported in
Section 4.1.1.
Printouts of raw continuous emission monitoring data are included

by run in the field data bound separately. Monitoring data printouts with
reduced data are included with other calculations (Appendix C).
3-9 Rev. 1
3.2.1.2 Gas Chromatograph
Concentrations of the two specific organic compounds expected in

the exhaust from the enclosure (styrene and MMA) were measured to provide
speciation of their emissions during gelcoat operations and to provide a check on
the Method 25A emission measurements using Method 18. Specific
concentrations of styrene and MMA were determined using periodic
measurements by GC using a Hewlett Packard Model 5890 with a Model 3396
Series II integrator. A Supelco GP 10% 1,2,3-tris(2-cyanoethoxy)propene on
80/100 Chromosorb PAW column (1/8-in diameter X 6 ft long stainless steel)
was used for these analyses. Hydrogen at 25 psi was provided to the instrument
for the flame; air (at 33 psi) and nitrogen (at 50 psi) were also provided to the
instrument to support the flame and as a carrier gas (35 cc/min) for the sample.
The instrument was maintained at 130°C for all GC analyses.
Measurements and calibration were conducted in accordance with

EPA Method 18 specifications (Appendix A). Sample gas was continuously fed
to the GC using the sampling system from the enclosure exhaust duct. Part of the
sample stream was sent to the THC analyzer, with another portion of the stream
flowing to the GC. At least GC four measurements per hour were to be made in
accordance with the QAPP; more often, data were collected every 6 to 8 minutes.
The GC was calibrated according to Method 18 procedures using

styrene and MMA to indicate the retention time and the area counts. Styrene gas
standards in cylinders were used to provide styrene for these calibrations. MMA
standards were prepared in Tedlar® bags (Appendix G) using the procedure
detailed in Method 18. After establishing the retention time for styrene and
MMA, the integrator was programmed to identify the two compounds (by
retention time) and used the response factor determined daily to calculate the
3-10 Rev. 1
concentrations of each compound. Area counts for the instrument responses

were recorded and concentrations calculated off-line in spreadsheets. A linear
relationship between concentration and area count was generated for each
calibration to calculate concentrations. (These calibration “curves” are included
with the calculations in Appendix C.) All chromatograms generated during
daily tests have been bound separately with the other field data. Records and
reductions for every run are included in Appendix C.
Day to day calibration checks met the acceptance criteria or the

instrument was recalibrated. Results of instrument calibration are presented in
Section 4.1.2. The audit of the GC system that the EPA conducted is discussed in
Section 4.7.
3.2.2 Flow Rate Measurements
Air flow rate from the enclosure was measured using EPA
Method 2 protocols once per test. Flow rate was also monitored continuously at a
single point in the TTE exhaust duct using a standard (Type L) pitot tube by
recording the velocity pressure head at the centroid of the duct. The pressure
head was logged through the Dianachart data logger to the computer DAS with
other parameters, such as temperature, resin weight, and THC concentrations at
the inlet and outlet to the enclosure. The Method 2 velocity traverse was
performed for each test run using the same standard pitot tube to monitor air
flow through the duct.
A velocity for each recorded THC measurement was calculated

based on a ratio of the average pressure head readings. The constant velocity
monitor reading at the time of the traverse was noted. This reading was scaled to
the instantaneous reading (recorded in the DAS) and to the flow rate measured
3-11 Rev. 1
using a complete traverse. The average of each calculated 1-min flow rate
reading based on 1-min velocity head readings was used in calculating an
average volumetric flow rate over the entire sampling period of the test run. The
change in flow rate over the course of any test run was less than 3 percent.
3.2.3 Temperature
The type K thermocouple used for DAS and flow measurements

was calibrated on site at two points: ice bath (32°F) and boiling water (212°F).
Air temperature within the enclosure was monitored continuously at the location
of the fixed pitot. Therefore, this temperature reflects the temperature at the
sampling location rather than the temperature within the enclosure. Air flow
velocities within the exhaust duct were approximately 30-35 ft/sec. The TTE
exhaust temperature was therefore measured within two seconds of leaving the
enclosure. The temperature of the well-mixed TTE exhaust provided a better
measurement of the average TTE temperature than a single fixed point within
the TTE because of the short retention time in the exhaust duct and the small
temperature differential between ambient and exhaust air. The temperature was
monitored continuously using the DAS calibrated at two points.
3.2.4 Weight of Material Used
The weight of the resin or gelcoat used during a test run was
determined by weighing the container of material on a platform balance
immediately before and after application to the mold. The balance was equipped
with an analog output that allowed changes in weight to be tracked on the DAS.
A small degree of variability in the weight measurement was noted

during the testing on the CFA mold, where only 3.5 to 10 lb of resin were used.
3-12 Rev. 1
The variability resulted from the movement of the ancillary equipment

associated with the resin and catalyst delivery systems. As indicated by the
weigh scale calibration checks (Section 4.6) and audit (Section 4.7.2), the scale
met the desired performance criteria and provided accurate measurements of the
resin and gelcoat use during testing on boat parts.
The weight of the catalyst was determined from the resin (or
gelcoat) weight and the established catalyst-to-resin ratio. These ratios were set
based on stroke count of the positive displacement piston pumps used for both
materials. The amount of catalyst was insignificant compared to resin or gelcoat
use.
The balance used to determine weight of resin and gelcoat used

during lamination was calibrated by Weigh Tronics in Seattle before it was
delivered to the site. All adjustments were made at the signal output and DAS;
no adjustment was made to the scale that would have affected its calibration.
Calibration certification is included with other certifications in Appendix F.
Calibration of the scale was checked periodically during the testing

program. These results are presented in Section 4.6. The precision and accuracy
of the balance was checked during the on-site technical system audit conducted
by the EPA. The results of this audit are presented in Section 4.7.2.
3.3 Air Flow Over the Mold
Air flow over the mold in the TTE was measured to enhance the
documentation of the experiments conducted, but no adjustment was made to
the rate of air flow through the enclosure based on the measurement. The more
important adjustment was the flow rate through the enclosure to simulate the
3-13 Rev. 1
number of room volume exchanges found in the U.S. Marine production area (an
integral part of the TTE design). The velocities over the parts were measured
merely to provide a record.
To support the validity of the measured velocities over the parts,

measurements were made (with a hot wire anemometer) early in the test
program to determine the range of velocities over parts encountered in the
production environment. Air flow velocities in a production lamination area
used for laminating large boats and large parts for those boats ranged from 20–
25 ft/min near the air pick-ups at the exhaust plenum to 34–40 ft/min in the
center of the lamination area. (In this production area, air was supplied through
air handlers at one side of the room and exhausted from pick-up plena on the
opposite side of the room.) Velocities near 200 ft/min were measured on covered
molds (convex architecture) near the exhaust from the air handlers servicing the
lamination area. This higher velocity was not typical in the lamination area.
Measurements made in the testing enclosure showed that velocities

of 20-24 ft/min were typical for the enclosure. These velocities were typical of
measurements for induced draft environments, such as those found nearer the
exhaust plenum in the lamination room. Measurements on the part (near the
door to the enclosure) showed that the bulk velocity into the enclosure was
approximately 40 ft/min. (Velocities at the NDOs to the TTE were much greater
than the required 200 ft/min, as noted in Section 1.3.1.)
Air flow over the mold was first evaluated for the 18-ft deck mold.
Measurements at five locations on the mold surface ranged from 20-24 ft/min on
the lower portion of the mold to 30-35 ft/min at the upper region of the mold,
nearer the air intakes for the enclosure. Velocities were checked on the 28-ft hull
mold at seven locations; velocities ranged from 20 to 36 ft/min. Measurements
3-14 Rev. 1
were made parallel to the surface of the mold, where the maximum velocity was
noted on the readout from the hot wire anemometer. Sketches from the log book
showing locations of measurements and results are found in Appendix H.
During the site audit by the EPA and its contractor Midwest
Research Institute (MRI), air velocity measurements at the mold surface were
made to verify that the velocity was in the range typical of the industry (and
plant) practice. Measurements made by MRI (Appendix H) corroborate the
measurements made on the molds.
Velocities were not measured during each test or run. These

velocities were not key measurements of the program; further, the velocities
would not be expected to vary during the program because the flow rate
through the enclosure was not a variable during this program and the geometry
of the molds did not vary.
3.4 Mass Balance Styrene Recovery Check
In accordance with the QAPP, styrene evaporation tests (the towel

tests) were conducted to evaluate the capture and recovery of styrene by the
sampling system. Two tests were completed, one at the initiation of the testing
program and one at its conclusion.
A covered container was loaded with a large pan, a collapsible

clothes drying rack, towels, and a gallon of styrene and was weighed on a
precision scale outside the TTE. (The scale used for this test was used by RTI in
their testing to measure smaller quantities of materials.) The container was then
placed inside the TTE, the rack erected in the pan, towels draped over the bars,
and, with the flow rate and emissions being measured, the styrene was poured
3-15 Rev. 1
over the towels and into the pan. A fan was used to induce flow over the towels
and accelerate styrene evaporation. After about 30 minutes, the equipment was
repacked into the container and the TTE allowed to clear of styrene (while
emission monitoring continued). When emissions reached background
concentrations, the container was removed from the TTE and reweighed.
Emissions were measured using the THC analyzer and the GC

during the first styrene recovery check, but using only the THC analyzer during
the second. Recoveries are based on THC results, and the sampling system’s
capture and recovery of styrene was assessed by comparing the quantity of
styrene evaporated (determined from weight loss of the container) with the
quantity of styrene measured in the exhaust (calculated from the average styrene
concentration and exhaust flow rate). The results of these tests, reported in
Section 4.7.4, indicate good agreement between the measured evaporative loss
and the measured mass emission quantity.
3.5 Data Custody
Because all emission measurements were made using continuous

monitors that vent emission streams after analysis, no sample retention was
necessary, or possible. Data collected in the field include gel time, spray gun
calibration, quantity of resin material used, temperature of the exhaust, static
pressure of the exhaust stream, pressure head of the exhaust stream, THC
concentration of the inlet stream to the enclosure, THC concentration on the
outlet stream from the enclosure, and concentrations of specific compounds in
the exhaust from the enclosure. Test runs are designated by day, for example,
0408-01 indicates the first run on April 8, 1997. Test numbers are designated with
“NMMA” prefix, for example, NMMA-1-1 indicates test condition 1,
assessment 1 (Table 1-1). Thus, a test can be identified by date or by type using
3-16 Rev. 1
the total description: 0405-01, NMMA-8-2 indicates a second replicate of a 28-ft

hull gelcoating and was the first run on April 5, 1997.
3-17 Rev. 1
4.0 QUALITY CONTROL AND QUALITY ASSURANCE
Quality control/quality assurance (QC/QA) guidelines outline

pertinent steps to be followed during the production of gas composition data to
ensure the reliability and acceptance of the data generated. QC, in the context of
this report, refers to a system of activities designed to achieve a level of precision
and accuracy as specified in the project plan. The QC procedures are targeted at
maintaining a level of quality but in themselves are not a measure of the degree
of quality achieved. QA, as used in this report, refers to the system of activities
implemented to measure the effectiveness of the QC system.
The QC/QA procedures implemented during this test program

were presented in the Quality Assurance Project Plan for the National Marine
Manufacturers Association Baseline Emission Testing Project (QAPP), which was
approved by the EPA before the on-site sampling program. This section presents
the results of the measurements taken to demonstrate adherence to the data
quality objectives for critical measurements, as presented in Section 3 of the
QAPP.
4.1 Concentration Measurements
The measurements taken for the characterization of THC content of

the TTE exhaust gas stream were made according to the procedures of EPA
Method 25A; this approach was the same as that used in previous studies,
ensuring comparablity with those previous tests. Styrene and MMA
concentration measurements were made according to Method 18 procedures.
Previous studies have not evaluated resins containing MMA. Method 18 results
are critical to speciation of THC from the gelcoat testing where these two species
are measured. No other compound was identified in the Method 18
4-1 Rev. 1
chromatograms. The QA measures and the acceptance criteria used for this test
program, which were approved by the EPA in the QAPP, are presented in
Table 4-1.
Table 4-1
Acceptance Criteria for Concentration Measurements
Method Measurement Acceptance Criteria
25A Calibration Error ≤ 10 percent of calibration gas value
25A Calibration Drift ≤ 10 percent of calibration gas value
18 Calibration Error ≤ 10 percent of calibration gas value
18 Calibration Drift ≤ 5 percent of calibration gas value
Note: Calibration error measurements were substituted for performance audits because no audit
materials were provided for routine assessment.
4.1.1 Total Hydrocarbon Measurements (Method 25A)
Tables 4-2 and 4-3 present the results of the Method 25A calibration
error and drift tests for all the test runs. Detailed results are presented in
Appendix C. Both assessments of instrument performance were made before
and after each test. The calibration errors presented in Table 4-2 are within the
acceptance criteria presented in the QAPP. The results of instrument drift
assessments demonstrate similar results; all post-test assessments are well within
10 percent of pre-test values. Based on review of these Method 25A calibration
assessments, all the data presented in this report are accurate to within
10 percent of the values reported, with a variability in precision of no greater
than 10 percent.
Because Shirley Wasson of the ORD during her review of the

QAPP expressed concerns about the stability of styrene stored in cylinders, THC
instrument calibrations were performed using propane. Response factors were
then determined for instrument response to styrene and MMA in terms of
4-2 Rev. 1
Table 4-2
Calibration Error Measurements for Total Hydrocarbon Analyzer
(Percent Actual Gas Value)
Actual Gas Concentration (ppmv)
Inlet Instrument Outlet Instrument
Date Test Run 0 15 30.4 45.5 0 297 600 914 1604
10-Apr NMMA-14-1 0410-01 N.A. 0.0 0.0 -0.7 N.A. 1.3 2.3 0.0
11-Apr NMMA-14-2 0411-02 N.A. 0.7 0.0 -0.4 N.A. -0.4 1.4 -2.1
15-Apr NMMA-16-1 0415-01 N.A. 2.7 2.3 0.4 N.A. -0.8 -0.1 -3.4
16-Apr NMMA-16-2 0416-01 N.A. 2.0 2.3 0.2 N.A. -0.9 0.1 0.0
2-Apr NMMA-6-P 0402-01 N.A. 0.7 1.3 -1.5 N.A. -0.3 -0.7 -0.3
8-Apr NMMA-6-1 0408-01 N.A. -1.3 0.0 -0.4 N.A. 0.7 2.3 0.0
11-Apr NMMA-6-2 0411-01 N.A. 3.3 4.3 0.4 N.A. -0.5 0.3 -0.1
3-Apr NMMA-4-1 0403-02 N.A. -0.7 -0.3 0.0 N.A. 1.8 3.0 0.2
8-Apr NMMA-4-2 0408-03 N.A. -1.3 0.0 -0.4 N.A. 0.7 2.3 0.0
12-Apr NMMA-5-1 0412-02 N.A. 2.0 1.3 0.2 N.A. -0.6 0.3 -0.1
14-Apr NMMA-5-2 0414-01 N.A. 2.0 1.3 0.2 N.A. 0.7 1.9 -1.2
10-Apr NMMA-13-1 0410-02 N.A. 0.0 -0.7 0.2 N.A. 2.0 2.8 -0.1
11-Apr NMMA-13-2 0411-03 N.A. 0.7 0.0 -0.4 N.A. -0.4 1.4 -2.1
15-Apr NMMA-15-1 0415-02 N.A. 2.7 2.3 0.4 N.A. -0.8 -0.1 -3.4
16-Apr NMMA-15-2 0416-02 N.A. 2.0 2.3 0.2 N.A. -0.9 0.1 0.0
4-Apr NMMA-3-1 0404-02 N.A. -1.3 -0.7 0.0 N.A. -0.3 -0.7 -0.3
8-Apr NMMA-3-2 0408-02 N.A. -1.3 0.0 -0.4 N.A. 0.7 2.3 0.0
5-Apr NMMA-1-1 0405-02 N.A. -4.7 -1.3 0.4 N.A. 2.5 4.8 2.8
9-Apr NMMA-1-2 0409-01 N.A. 0.7 0.3 0.2 N.A. -0.8 0.2 0.0
12-Apr NMMA-2-1 0412-02 N.A. 2.0 1.3 0.2 N.A. -0.6 0.3 -0.1
14-Apr NMMA-2-2 0414-03 N.A. 2.0 1.3 0.2 N.A. 0.7 1.9 -1.2
10-Apr NMMA-13-1 0410-02 N.A. 0.0 -0.7 0.2 N.A. 2.0 2.8 -0.1
11-Apr NMMA-13-2 0411-03 N.A. 0.7 0.0 -0.4 N.A. -0.4 1.4 -2.1
18-Apr NMMA-9-1 0418-01 N.A. 0.7 1.0 0.0 N.A. 0.0 0.0 -0.1 -1.7
19-Apr NMMA-9-2 0419-01 N.A. -1.3 0.7 -0.4 N.A. -0.7 0.8 0.1
3-Apr NMMA-8-1 0403-01 N.A. -0.7 -0.3 0.0 N.A. 1.8 3.0 0.2
5-Apr NMMA-8-2 0405-01 N.A. -1.3 4.6 -4.8 N.A. 4.7 4.8 1.1
4-Apr NMMA-7-1 0404-01 N.A. -1.3 -0.7 0.0 N.A. 0.5 2.6 -0.4
7-Apr NMMA-7-2 0407-01 N.A. -4.7 -3.9 0.7 N.A. 3.0 5.2 0.0
9-Apr NMMA-11-1 0409-03 N.A. 0.7 0.3 0.2 N.A. -0.8 0.2 0.0
10-Apr NMMA-11-2 0410-03 N.A. 0.0 -0.7 0.2 N.A. 2.0 2.8 -0.1
12-Apr NMMA-11-3 0412-01 N.A. -1.3 -1.3 -0.9 N.A. 2.8 3.8 0.6
14-Apr NMMA-12-1 0417-01 N.A. 3.3 2.6 0.2 N.A. -0.1 0.7 -0.1 -0.1
18-Apr NMMA-12-2 0418-02 N.A. 0.7 1.0 0.0 N.A. -0.9 0.0 -0.1 -1.7
4-3 Rev. 1
Table 4-3
Calibration Drift Measurements for Total Hydrocarbon Analyzer
(Percent Span)
Actual Gas Concentration (ppmv)
Inlet Instrument Outlet Instrument
Date Test Run 0 15 30.4 45.5 0 297 600 914 1604
10-Apr NMMA-14-1 0410-01 0.0 -0.4 -1.0 -4.2 0.0 -0.2 0.5 1.9
11-Apr NMMA-14-2 0411-02 0.0 -0.2 0.2 -0.2 0.0 0.1 -0.3 -0.1
15-Apr NMMA-16-1 0415-01 0.0 -0.6 -1.2 N.A. 0.0 -0.6 -1.1 N.A.
16-Apr NMMA-16-2 0416-01 0.0 0.0 -1.6 N.A. 0.0 -0.4 -0.2 N.A.
2-Apr NMMA-6-P 0402-01 -0.6 -0.8 N.A. 2.4 0.0 0.1 N.A. 0.6
8-Apr NMMA-6-1 0408-01 0.2 0.4 -2.2 N.A. 0.0 0.2 0.8 N.A.
11-Apr NMMA-6-2 0411-01 0.2 -0.8 -2.6 -0.8 0.0 0.0 0.6 -1.9
3-Apr NMMA-4-1 0403-02 -1.0 -1.2 -1.6 -3.8 0.0 -0.2 -0.3 1.5
8-Apr NMMA-4-2 0408-03 0.0 -0.4 -1.8 -2.6 0.1 -0.2 0.0 2.1
12-Apr NMMA-5-1 0412-02 0.0 0.8 1.6 N.A. 0.0 -1.1 -1.6 N.A.
14-Apr NMMA-5-2 0414-01 -0.8 -1.4 -3.2 N.A. 0.0 -0.1 0.1 N.A.
10-Apr NMMA-13-1 0410-02 -0.4 -1.0 -1.2 N.A. 0.0 -0.8 -1.1 N.A.
11-Apr NMMA-13-2 0411-03 -0.2 -0.4 -0.6 N.A. 0.0 -0.2 -0.3 N.A.
15-Apr NMMA-15-1 0415-02 0.0 -0.2 -1.2 N.A. 0.0 -0.6 -0.9 N.A.
16-Apr NMMA-15-2 0416-02 -0.2 -0.2 -1.4 N.A. 0.0 0.0 0.6 N.A.
4-Apr NMMA-3-1 0404-02 -1.6 -1.2 -1.0 N.A. 0.0 0.1 N.A. 0.6
8-Apr NMMA-3-2 0408-02 0.0 -0.2 -1.4 N.A. 0.0 -0.4 -0.3 N.A.
5-Apr NMMA-1-1 0405-02 -1.2 1.0 1.4 N.A. 0.0 0.3 0.4 N.A.
9-Apr NMMA-1-2 0409-01 0.0 -0.6 -0.6 N.A. 0.0 -0.2 0.0 N.A.
12-Apr NMMA-2-1 0412-02 -0.2 -0.6 -0.6 N.A. 0.0 -0.9 -1.5 N.A.
14-Apr NMMA-2-2 0414-03 -0.2 -2.0 -2.6 N.A. 0.0 -0.1 0.1 N.A.
10-Apr NMMA-13-1 0410-02 -0.4 -1.0 -1.2 N.A. 0.0 -0.8 -1.1 N.A.
11-Apr NMMA-13-2 0411-03 -0.2 -0.4 -0.6 N.A. 0.0 -0.2 -0.3 N.A.
18-Apr NMMA-9-1 0418-01 -0.2 -0.2 -1.2 N.A. 0.0 N.A. 0.0 0.5 N.A.
19-Apr NMMA-9-2 0419-01 -1.0 -0.4 -2.0 N.A. 0.0 0.1 0.1 N.A.
3-Apr NMMA-8-1 0403-01 -1.2 0.8 N.A. N.A. 0.0 N.A. 0.5 N.A.
5-Apr NMMA-8-2 0405-01 -3.8 -0.8 0.8 N.A. 0.0 -0.6 0.0 1.6
4-Apr NMMA-7-1 0404-01 -0.2 -0.6 N.A. -1.8 0.0 -0.2 N.A. 2.3
7-Apr NMMA-7-2 0407-01 -2.6 0.8 -2.8 N.A. 0.0 0.4 0.9 N.A.
9-Apr NMMA-11-1 0409-03 -0.2 -1.2 -0.8 -3.0 0.0 0.0 0.4 -0.2
10-Apr NMMA-11-2 0410-03 -0.4 -0.6 -0.8 N.A. 0.0 -0.4 -0.8 N.A.
12-Apr NMMA-11-3 0412-01 -0.4 -2.2 -0.6 N.A. 0.0 -0.1 0.5 N.A.
14-Apr NMMA-12-1 0417-01 -0.2 0.4 3.2 N.A. 0.0 0.1 0.5 1.3 N.A.
18-Apr NMMA-12-2 0418-02 0.0 -0.4 -1.0 N.A. 0.0 0.2 0.4 0.8 N.A.
4-4 Rev. 1
response to propane. The response factors, determined for the outlet THC
analyzer only, are listed in Table 4-4. The variation in instrument response
factors serves as a measure of the day to day precision of the measurements as
well as a correlation to styrene concentrations. The average relative response
factor determined as the average of the daily average relative response factors
for seven days of measurements, including values determined for all styrene
calibration standards brought on site, is 2.40, with a standard deviation of 0.053
(±2.2 percent). The issue of daily variation in response factors was not addressed
in the QAPP and no criterion for data acceptance has been suggested. The
standard deviation of the response factors is within the precision requirement for
the THC calibration and drift criteria.
Table 4-4
Specific Compound Relative Response Factors
for the Total Hydrocarbon Analyzera
Styrene THC Relative MMA THC Relative
Concentration Response Response Concentration Response Response
Date (ppmv) (ppmv) Factor (ppmv) (ppmv) Factor
1-Apr 204 487.7 2.39
3-Apr 45.6 62.7 1.38
7-Apr 204 506.8 2.48
9-Apr 122 300.1 2.46
9-Apr 204 490.3 2.40
12-Apr 204 468.5 2.30
12-Apr 59.2 148.4 2.51
15-Apr 59.2 138.5 2.34
16-Apr 59.2 137.6 2.32
19-Apr 59.2 141.5 2.39 139.6 166.8 1.19
19-Apr 122 304.3 2.49
Average 2.40 1.28
aRelativeresponse factor is the THC instrument measurement divided by the calibration gas
concentration, e.g., ppm propane/ppm styrene.
4-5 Rev. 1
4.1.2 Organic Compound Concentrations (Method 18)
Method 18 calibrations were conducted every morning before

starting a test run. Calibrations were checked after each run to ensure that the
instrument remained in calibration for the test series. A measure of calibration is
provided by the response factors (Figures 4-1 and 4-2), which are simply the
calibration gas concentration divided by the area counts from the chromatogram
(i.e., the GC response). As the figures indicate, there was more variability in the
response factors early in the sampling program. This variability resulted from
leaking of the gas injection valve during these early runs. On April 12, the valve
was replaced, and the response factor variation was reduced (after Analysis
No. 714).
Figure 4-1
Method 18 Response Factors for Styrene
Response Factors for Styrene - Method 18
0.00008
0.00007
0.00006
Response Factor
0.00005
0.00004
0.00003
0.00002
0.00001
0
29
17
63
104
180
270
363
438
487
515
534
588
633
693
714
762
819
876
911
960
1029
1074
Analysis Number
Method 18 data for this program have been used primarily to

determine the relative quantities of styrene and MMA emitted during
4-6 Rev. 1
gelcoating. The Method 18 results are also useful for comparison with the results
of Method 25A testing. The relative concentrations of these two compounds in
the TTE exhaust is important. An assessment of the ratio of the average response
factors for styrene and MMA before and after April 12 (Analysis No. 712) shows
that the leaking valve had little or no effect on the determination of the relative
concentrations of these two compounds (Table 4-5) on average.
Figure 4-2
Method 18 Response Factors for Methyl Methacrylate
Response Factors for MMA - Method 18
0.00016
0.00014
0.00012
Response Factor
0.0001
0.00008
0.00006
0.00004
0.00002
0
11
26
106
153
184
198
226
346
350
354
410
443
450
497
600
604
628
679
683
687
Analysis Number
Table 4-5
Method 18 Response Factors for Target Compounds
Before April 12 After April 12
Styrene 2.78E-05 1.17E-05
MMA 5.87E-05 2.27E-05
Styrene:MMA 0.473 0.516
4-7 Rev. 1
Daily response factor ratios were also assessed (Figure 4-3). Only
those days when styrene and MMA results were needed, and the final check
conducted on April 19, are presented in the assessment. The figure illustrates the
variability seen in the individual measurements, although the degree of
variability is lessened through the averages. The ratio of the response factors
(important in apportioning emissions to styrene and MMA for gelcoating test
runs) was relatively constant for all gelcoat test days except April 3 and April 8.
The results for April 3 appear to be consistent with other days; so the difference
in response factor ratio did not appear to affect results for that test. The results
from April 8, however, indicate much lower concentrations of MMA than noted
for all other gelcoat test runs. Styrene results for that test run were consistent
with other measurements and Method 25A results were typical. The abnormally
low reading for MMA is attributed to poor response for that calibration.
Figure 4-3
Method 18 Response Factors for Selected Test Days
8.0E-05 0.7
7.0E-05 0.6
Response Factor (ppm/area count)
6.0E-05
0.5
Reponse Factor Ratio
5.0E-05
0.4
4.0E-05
0.3
3.0E-05
0.2
2.0E-05
1.0E-05 0.1
0.0E+ 0 0 0.0
2-Apr 3-Apr 4-Apr 5-Apr 8-Apr 9-Apr 11-Apr 19-Apr
Styrene MMA Styrene:M M A
4-8 Rev. 1
The relative concentrations of styrene and MMA were determined

using the following approach. Area counts based on chromatographic analyses
were determined for each compound at varying levels of gas concentrations. An
average response factor was then determined for various gas standards. A point
to point calibration curve was generated plotting area count against gas
concentration assuming a linear function from point to point. These calibration
curves (Appendix C) were then used to determine the relative concentrations of
styrene and MMA in each sample analyzed.
Table 4-6 lists the QC measurements and their results. These

measurements pertain only to styrene calibrations, because all gelcoating runs
(i.e., the only runs involving MMA) had been completed by the time of
equipment repair (April 12). All the pretest measurements of precision met the
criteria defined for this program (within 10 percent of mean value). Drift
exceeded the acceptance criteria of less than 5 percent of pretest values on four
days; all these events occurred at the 59.2 ppmv styrene level. Data comparisons
made between GC and THC concentration measurements show a higher level of
styrene measured by the GC than by the THC analyzer, largely because the THC
results are average results and the GC reports the concentration of a grab
sample, generally selected by the GC operator to analyze temporal peaks in
concentration. Because the reported emissions were calculated from the
Method 25A results, drift in Method 18 measurements has no effect on the
reported emissions.
4-9 Rev. 1
Table 4-6
Gas Chromatography Calibrations (after April 12)
Analysis Concn Precision
Date No. (ppmv) Area Count RF (%Mean) Drift Comment
12-Apr 713 59.2 5971597 9.91E-06 Pre 5-2

714 59.2 6009331 9.85E-06 -0.31 Pre 5-2
12-Apr 729 59.2 5958909 9.93E-06 Post 5-2
730 59.2 6291424 9.41E-06 Post 5-2
731 59.2 6429882 9.21E-06 -1.09 Post 5-2
732 122 10977608 1.11E-05 Post 5-2
733 122 11537000 1.06E-05 -2.48 Post 5-2
734 204 16097608 1.27E-05 Post 5-2
735 204 14421208 1.41E-05 Post 5-2
736 204 21264480 9.59E-06 Post 5-2
737 204 19247328 1.06E-05 Post 5-2
738 204 19026992 1.07E-05 0.58 Post 5-2
12-Apr 759 59.2 5525818 1.07E-05 Post 2-1
760 59.2 5794490 1.02E-05 -2.37 Post 2-1
761 122 11881808 1.03E-05 Post 2-1
762 122 10850096 1.12E-05 4.54 Post 2-1
14-Apr 768 0 352953 pre 5-2
769 0 203009 pre 5-2
470 59.2 4755664 1.24E-05 pre 5-2
471 59.2 4167498 1.42E-05 pre 5-2
472 59.2 4384538 1.35E-05 -2.54 pre 5-2
473 122 9803558 1.24E-05 pre 5-2
474 122 8547834 1.43E-05 pre 5-2
475 122 8905798 1.37E-05 -2.05 pre 5-2
476 204 16823296 1.21E-05 pre 5-2
477 204 16883072 1.21E-05 pre 5-2
478 204 17030400 1.20E-05 -0.43 pre 5-2
14-Apr 796 204 17182512 1.19E-05 post 5-2
797 204 16327544 1.25E-05 post 5-2
798 204 16258032 1.25E-05 3.18 post 5-2
14-Apr 818 204 17429680 1.17E-05 post 2-2
819 204 16991360 1.20E-05 -0.59 post 2-2
820 59.2 5021901 1.18E-05 post 2-2
821 59.2 5041053 1.17E-05 -15.07 post 2-2
823 0 0
824 0 0
15-Apr 826 0 176395 0.00E+00 pre 16-1
827 0 213703 0.00E+00 pre 16-1
830 59.2 4179248 1.42E-05 pre 16-1
831 59.2 5030979 1.18E-05 -9.25 pre 16-1
832 122 9615066 1.27E-05 pre 16-1
833 122 9712975 1.26E-05 -0.51 pre 16-1
834 204 17192032 1.19E-05 pre 16-1
835 204 16329160 1.25E-05 2.57 pre 16-1
854 204 17244432 1.18E-05 -2.87 post 16-1
855 59.2 4923560 1.20E-05 -7.27 post 16-1
875 59.2 4928013 1.20E-05 post 15-1
876 59.2 4053978 1.46E-05 0.54 post 15-1
877 59.2 4739386 1.25E-05 post 15-1
878 204 17330048 1.18E-05 -3.35 post 15-1
4-10 Rev. 1
Table 4-6 (continued)

Analysis Concn Precision
Date No. (ppmv) Area Count RF (%Mean) Drift Comment
16-Apr 879 0 0 pre 16-2

880 0 0 pre 16-2
881 59.2 4915139 1.20E-05 pre 16-2
882 59.2 4863680 1.22E-05 0.53 pre 16-2
883 122 10776440 1.13E-05 pre 16-2
884 122 9171008 1.33E-05 pre 16-2
885 122 10341320 1.18E-05 pre 16-2
886 122 10572206 1.15E-05 -1.10 pre 16-2
16-Apr 905 59.2 4956816 1.19E-05 post 16-2
906 59.2 4863680 1.22E-05 -0.42 post 16-2
910 0 660 0.00E+00 post 16-2
911 122 11027728 1.11E-05 post 16-2
912 122 10367112 1.18E-05 -2.17 post 16-2
16-Apr 934 0 0 post 15-2
935 122 10505456 1.16E-05 post 15-2
936 122 10076096 1.21E-05 1.64 post 15-2
937 59.2 4985315 1.19E-05 post 15-2
938 59.2 5178906 1.14E-05 -3.76 post 15-2
17-Apr 939 0 0 pre 12-1
940 0 2044 pre 12-1
941 59.2 4324157 1.37E-05 pre 12-1
942 59.2 4535176 1.31E-05 pre 12-1
943 59.2 4768189 1.24E-05 -2.50 pre 12-1
944 122 10597432 1.15E-05 pre 12-1
945 122 9772608 1.25E-05 4.05 pre 12-1
17-Apr 960 204 17712032 1.15E-05 post 12-2
961 204 18269232 1.12E-05 -1.55 post 12-2
962 122 10231520 1.19E-05 -0.62 post 12-2
963 59.2 5499240 1.08E-05 post 12-2
964 59.2 5894784 1.00E-05 -18.30 post 12-2
18-Apr 978 0 1915 pre 9-1c
979 0 3089 pre 9-1c
980 59.2 5035434 1.18E-05 pre 9-1c
982 59.2 5181651 1.14E-05 -1.43 pre 9-1c
983 204 19345344 1.05E-05 pre 9-1c
984 204 18797024 1.09E-05 1.44 pre 9-1c
18-Apr 1018 204 18110944 1.13E-05 post 9-1
1020 204 19007808 1.07E-05 2.80 post 9-1
1021 59.2 5687642 1.04E-05 -10.20 post 9-1
19-Apr 1035 59.2 4908355 1.21E-05 pre 9-2
1036 59.2 4736861 1.25E-05 1.78 pre 9-2
1037 204 18883424 1.08E-05 pre 9-2
1038 204 19033904 1.07E-05 -0.40 pre 9-2
19-Apr 1070 204 18327536 1.11E-05 post 9-2
1071 204 18467392 1.10E-05 3.05 post 9-2
1072 122 11046080 1.10E-05 post 9-2
1073 122 10864040 1.12E-05 post 9-2
1074 122 11025512 1.11E-05 3.28 post 9-2
1075 59.2 5985747 9.89E-06 post 9-2
1076 59.2 5700285 1.04E-05 0.00 -17.44 post 9-2
19-Apr 1077 59.2 5780963 1.02E-05 0.01 Bias test
19-Apr 1078 6178867 MMA Bag 7
1079 6176253 MMA Bag 7
4-11 Rev. 1
4.1.3 Comparison of Methods in Determining Emissions
There is relatively good agreement between the results of the two

methods, with most of the measurements within 6-8 percent. The runs that had
more divergent results were typically those involving smaller molds. Table 4-7
presents a comparison of the results in terms of percent of available volatile
material, which for most of the testing was styrene only.
Table 4-7
Comparison of Emission Results by Two Methods
Emissions (% Emissions (%
Available) - Available) -
Date Test Description Method 25A Method 18
2-Apr NMMA-6-P 18 Deck Gel 47.0% 54.1%

4-Apr NMMA-7-1 28 Hull 35 R 15.7% 18.2%
7-Apr NMMA-7-2 28 Hull 35 R 18.8% 18.4%
8-Apr NMMA-6-1 18 Deck Gel 42.3% 43.1%
8-Apr NMMA-3-2 18 Hull Gel 54.2% 53.5%
8-Apr NMMA-4-2 18 Deck 35 R 16.1% 16.1%
9-Apr NMMA-1-2 18 Hull 35R 13.9% 13.2%
9-Apr NMMA-11-1G CFA Gel 43.4% 39.0%
9-Apr NMMA-11-1R CFA 35R 29.8% 37.1%
10-Apr NMMA-14-1 18 Deck 35R - FC 11.1% 10.4%
10-Apr NMMA-13-1 18 Hull 35R - FC 10.5% 11.0%
10-Apr NMMA-11-2R CFA 35R 28.8% 26.8%
11-Apr NMMA-6-2 18 Deck Gel 45.2% 46.4%
11-Apr NMMA-14-2 18 Deck 35R - FC 12.1% 11.1%
11-Apr NMMA-13-2 18 Hull 35R - FC 10.5% 10.9%
12-Apr NMMA-11-3R CFA 35R 52.0% 44.8%
12-Apr NMMA-5-1 18 Deck 42R 20.4% 14.1%
12-Apr NMMA-2-1 18 Hull 42R 22.3% 24.1%
14-Apr NMMA-5-2 18 Deck 42R 20.8% 22.4%
14-Apr NMMA-2-2 18 Hull 42R 18.2% 14.2%
15-Apr NMMA-16-1 18 Deck 42R-FC 12.2% 12.0%
15-Apr NMMA-15-1 18 Hull 42R-FC 11.6% 11.7%
16-Apr NMMA-16-2 18 Deck 42R-FC 14.0% 14.4%
16-Apr NMMA-15-2 18 Hull 42R-FC 10.6% 10.8%
17-Apr NMMA-12-1 CFA 42R 54.6% 44.8%
18-Apr NMMA-9-1 28 Hull 42R 22.8% 25.5%
19-Apr NMMA-9-2 28 Hull 42R 23.0% 23.5%
18-Apr NMMA-12-2 CFA 42R 39.2% 38.2%
4-12 Rev. 1
Certainly, the concentration measurements made using both

approaches yielded consistent results. The THC concentration trace for
Run NMMA-7-1 and an equivalent THC concentration trace generated using
measured styrene and MMA concentrations and the response factors developed
during this program (2.40 for styrene and 1.39 for MMA), although not identical,
show fair agreement for the different FIDs (Figure 4-4).
Figure 4-4
Comparison of Total Hydrocarbon Traces by Two Methods
NMMA-7-1
800
700
600
500
400
300
200
100
0
8:38 9:50 11:02 12:14 13:26 14:38 15:50
Time (PST)
M25A M18Eq
4-13 Rev. 1
4.2 Gelcoat and Resin Flow Rate
Spray gun resin and gelcoat flow rates were determined at the start
and end of each test day (Table 4-8). A single assessment was made during each
flow rate check. Resin-to-glass ratios were assessed during the initial resin
calibrations (Table 4-8). Catalyst-to-resin/gelcoat ratios were set according to the
catalyst pump stroke rate. These ratios remained constant during testing.
Table 4-8
Spray Equipment Data Summary
Flow Rate (g/min) Percent Resin Glass Percent
Equipment Date Pre test Post test Change (g) (g) Glass
Spray gun/ 3-Apr 327 180 35.5
Resin 4-Apr 324 128 28.3
5-Apr 2888 2916 1.0 148 77 34.2
7-Apr 2828 3092 9.3 277 149 35.0
8-Apr 2792 2896 3.7 301 176 36.9
9-Apr 2836 2776 -2.1 242 127 34.4
10-Apr 2956 2784 -5.8 222 114 33.9
12-Apr 2892 3100 7.2 218 117 34.9
14-Apr 3192 3096 -3.0 378 201 34.7
17-Apr 3100 317 152 32.4
18-Apr 2944 2992 1.6 300 160 34.8
19-Apr 3036 2904 -4.3 256 148 36.6
Flow 10-Apr 2956 2784 -5.8 222 114 33.9
Chopper 11-Apr 2556 2668 4.4 297 158 34.7
15-Apr 3164 3388 7.1 406 192 32.1
16-Apr 3084 3072 -0.4 317 162 33.8
Spray gun/ 3-Apr 1264 1256 -0.6
Gelcoat 4-Apr 1256 1260 0.3
5-Apr 1264 1264 0.0
8-Apr 1260 1268 0.6
9-Apr 1272 1276 0.3
11-Apr 1256 1260 0.3
4-14 Rev. 1
4.3 Air Velocity over Mold Surface
Air flow over the mold surface was assessed twice during the test
program. These measurements reflect a best assessment approach and no
presumption of accuracy has been made. In addition to being difficult to assess
accurately, air flow over the mold (i.e., velocity) was dependent on mold
position in the TTE, a factor that changed every time a tool was brought into the
enclosure or moved within the enclosure in the process of lamination or
gelcoating. The velocities measured at the mold surface were consistent with
measurements made over mold surfaces in the manufacturing area, taking into
account that the TTE was designed for induced draft flow. Greater emphasis was
placed on controlling air flow through the enclosure to reflect conditions found
in the facility than on attempting to duplicate the velocities measured during
previous studies using parts not typical of boat building.
Measurements made during a technical system audit performed by

MRI and the EPA were consistent with the measurements made during this
program. Their findings are reported in MRI’s Technical Systems Audit of a
Laboratory Spray Booth at the U.S. Marine Facility in Arlington, Washington (May 29,
1997).
4.4 Exhaust Flow Rate
Air flow rate through the enclosure was measured and monitored
as described in Section 3.2.2. Accuracy of the velocity measurement can be
assessed by observing the variability of the velocity head measurements (which
were accurate to within 0.005 inches of water) during the test. This yields an
assessment in terms of percent which can be compared with the acceptable
criteria of the method. Variability measurements for all test runs are shown with
4-15 Rev. 1
the test averages (Table 4-9). The overall average variability in flow rate for the
program was 2.9 percent, and no measure for an entire day was greater than
6.9 percent. The two instances of greatest variability were affected by the two
measurements on April 4, early in the program, when TTE exhaust filters had
been allowed to become partially blocked and were first being changed during
the test runs.
Table 4-9
Variability Measures for Flow Rate During Test Runs
Flow Rate Variability (%)
Date Run1 Run 2 Run 3 Average
2-Apr 1.0 1.0
3-Apr 2.5 3.2 2.9
4-Apr 4.9 8.9 6.9
5-Apr 2.2 1.2 1.7
7-Apr 2.5 2.5
8-Apr 1.4 3.3 1.8 2.2
9-Apr 3.2 3.5 0.9 2.6
10-Apr 2.3 2.8 3.0 2.7
11-Apr 2.9 4.5 3.7 3.7
12-Apr 4.3 4.2 3.7 4.1
14-Apr 3.2 2.4 2.8
15-Apr 3.3 3.5 3.4
16-Apr 2.1 1.7 1.9
17-Apr 2.5 2.5
18-Apr 2.7 2.4 2.5
19-Apr 2.3 2.3
Average 2.9
Flow rates during the test program were observed (based on the
continuous, fixed point velocity head measurements) to vary based on filter
collection and other operations. The exhaust plena on both longitudinal sides of
the enclosure were equipped with filter media (spun fiberglass air filters). Flow
rate decreased with time as these filters collected air-borne resin containing
reactive styrene and MMA, which polymerized on the filter media. Periodically,
these filters were replaced to ensure more consistent flow rates through the
4-16 Rev. 1
enclosure for all test runs. The build-up of material on the filters was one cause
of the variability in flow rate measured during the overall program. To some
smaller extent, this affected flow rate over the course of the test run, more so for
the larger articles which required more resin. A second cause of the fluctuation
in flow rate was activity in the building where the enclosure was located.
Temperature was affected by the opening and closing of doors into the building.
Changes in the system operated by RTI also had some effect on the measurement
of flow because both systems were connected to the same exhaust system at the
U.S. Marine facility. Changes made to accommodate their operation had some
effect on exhaust system pressure, thus affecting flow rate through the larger
enclosure.
4.5 Exhaust Air Temperature
As described in Section 3.2.3, enclosure air temperature was

monitored continuously using a type K thermocouple at the location of the fixed
pitot and the DAS, calibrated at two points. The accuracy of this measurement
was 0.005°F (or 0.0001 mV). Measurements of precision were not made during
the test program; however, literature suggests the precision of type K
thermocouples, assessed in a laboratory environment, will yield a precision of
0.5°F, or within 1 percent of the measured values.
4.6 Resin/Gelcoat Balance
Balance calibration checks were often performed with barrels of

resin still on the platform and always performed without zeroing the balance.
Therefore, the calibration error was calculated in terms of percent of weight
change rather than the displayed weight value (Table 4-10). All calibration
results display a level of accuracy, measured as percent error, meeting the
4-17 Rev. 1
criteria of the QAPP (within 2 lb). Based on the maximum error measure of
0.2 percent and the high end of the scale measurements (i.e., 500 lb), the accuracy
was within approximately 1 lb.
Table 4-10
Results of Balance Calibration Checks
Tare Weight Initial Scale Scale Change in Error (%
Date (lb) Reading (lb) Response (lb) Weight (lb) Tare)
6-Apr 0 0.0 0 0.0 NA
500 0.0 500.1 500.1 0.02
300 0.0 300.1 300.1 0.03
100 0.0 100.1 100.1 0.10
8-Apr 0 0.0 -1.9 0.0 NA
100 0.0 98.1 100.0 0.00
200 0.0 197.9 199.8 -0.10
300 0.0 297.8 299.7 -0.10
400 0.0 397.6 399.5 -0.12
500 0.0 497.5 499.4 -0.12
10-Apr 0 201.7 201.7 0.0 NA
50 201.7 251.7 50.0 0.00
100 201.7 301.7 100.0 0.00
200 201.7 401.7 200.0 0.00
15-Apr 0 231.7 231.7 0.0 NA
100 231.7 331.7 100.0 0.00
200 231.7 431.9 200.2 0.10
250 231.7 481.8 250.1 0.04
19-Apr 0 0.0 -1.33 0.0 NA
100 0.0 98.5 99.8 -0.17
200 0.0 198.7 200.0 0.02
300 0.0 298.7 300.0 0.01
400 0.0 398.8 400.1 0.03
500 0.0 498.7 500.0 0.01
100 0.0 98.9 100.2 0.23
4.7 Technical Systems Audit
The performance of the sampling system was assessed through an

audit and styrene recovery tests. Audits of the THC sampling, balance, and flow
measurement systems were performed. The system audit was conducted by
4-18 Rev. 1
EPA/OAQPS and its contractor MRI. The EPA/OAQPS audit will report on this
project, but the final report has not been released. Also, styrene recovery tests
(styrene evaporation tests) were performed in accordance with the QAPP.
4.7.1 Total Hydrocarbon Measurement
The results of the Method 25A system audit are presented in

Table 4-11. The audit gas was introduced in two manners during the
performance of the audit. One method involved filling a Tedlar® bag with the
audit gas and attaching it to the sampling probe. In the second method, the audit
cylinder was connected to the calibration system and the gas was introduced to
the sampling system in the same manner as all on-site calibrations were
performed; the propane audit was performed only with this latter method. Both
methods yield results representative of the system performance.
Table 4-11
Audit Results for Method 25A Measurements
Analyzer Result Gas
Measurement (ppmv as Standard Result
Date (ppmv C3H8) styrene) (ppmv) Compound (%) Notes
15-Apr 84.3 36.8 39.7a Styrene -7.3 b, c
16-Apr 85.3 36.8 39.7 Styrene -7.3 b, d
1000 999 Propane +0.1 d
a. Gas standard reported at 41.7 ppmv; on site analysis by GC reported 39.7 ppmv.
b. Response factor to styrene is 2.289 on 4/15 and 2.318 on 4/16.
c. Test performed with audit gas in Tedlar® bag.
d. Test performed with audit gas in vented cylinder.
Conversion of THC concentration to equivalent concentrations of

styrene was done using daily relative response factors determined from the THC
analyzer (Method 25A) on the day of the audit. Using the pre-test analysis by
Scott Specialty Gases (41.7 ppmv styrene), the analysis by THC analyzer was
about 12 percent low. However, on-site analysis of the standard by GC indicated
4-19 Rev. 1
that the concentration in the audit sample was only 39.7 ppmv. Although this
assessment does not meet the requirements of a protocol gas certification, the
results indicate that the certified value of the audit cylinder may be slightly low.
Comparing to concentration from this reanalysis of the standard shows that the
THC analysis was only 7.3 percent low, within the criterion established for the
program. This low recovery does not appear to be a system malfunction. The
good results of the propane audit indicate the low results of the styrene audit
sample are not a result of sampling system bias (leaks). Further, procedures in
determining response factors included introduction of the gas at the probe, thus
incorporating any system characteristics into the results. The results of the
styrene audit indicate the sampling system, as used in the performance of the
audit, results in an under reporting of styrene concentrations by less than
10 percent at a 40-ppmv concentration, or about 3 ppmv styrene.
4.7.2 Weigh Cell Balance
The results of the weigh cell balance audit performed on April 16

indicate that greater than 5 percent error results when weighing less than 3000 g,
or approximately 6.6 lb, on an empty scale (Table 4-12). However, testing was
conducted with a drum of resin or gelcoat or resin in place, meaning that the
scale was used during the test program in its normal operating range. The
results of the audit and the calibration checks (Table 4-10) demonstrate that even
small weight changes can be measured accurately through the weight range
experienced during this test program. Although the audit was conducted with
an empty balance, the calibration checks were conducted with both empty and
loaded scale (i.e., with a drum of resin in place). Data collected from the
calibration check using certified weights on April 8 indicate that a less than 1 lb
change in weight at an elevated balance response (512 lb) was accurately
quantified. These data indicate the balance was capable of accurately measuring
4-20 Rev. 1
500-g (1.1-lb) changes in weight in the range of operation, well within

specifications of the QAPP (i.e., 2 lb). Two data points from the audit can be
used to assess the precision of the weight measurements. The replicate weighing
performed at 57 and 157 lb indicate a precision of greater than 0.2 lb, also within
QAPP specifications.
Table 4-12
Results of Audit on Balance
Scale Net Scale

Audit Response Scale Initial Response Difference
Date Weight (g) (lb) (lb) (lb) (lb) % Error
8-Apra 3000 504.3 497.6 6.7 0.086 1.30
5000 508.7 497.6 11.1 0.077 0.70
6000 510.9 497.6 13.3 0.072 0.55
6500 512 497.6 14.4 0.070 0.49
Scale Net Scale
Audit Response Scale Initial Response Difference
Weight (lb) (lb) (lb) (lb) (lb) % Error
16-Aprb 0 -1.92 -1.92 0 0.00 NA
0.002 -1.94 -1.92 -0.02 -0.022 -1007.18
0.01 -1.96 -1.92 -0.04 -0.053 -402.39
0.04 -1.99 -1.92 -0.07 -0.105 -298.45
0.1 -1.86 -1.92 0.06 -0.086 -58.76
0.4 -1.65 -1.92 0.27 -0.096 -26.22
2.6 0.55 -1.92 2.47 -0.101 -3.91
7 5.06 -1.92 6.98 0.000 0.00
57 54.9 -1.92 56.82 -0.180 -0.32
157 154.74 -1.92 156.66 -0.340 -0.22
457 454.9 -1.92 456.82 -0.180 -0.04
357 355 -1.92 356.92 -0.080 -0.02
157 154.9 -1.92 156.82 -0.180 -0.11
57 55 -1.92 56.92 -0.080 -0.14
7 5 -1.92 6.92 -0.060 -0.86
2.6 0.5 -1.92 2.42 -0.15 -5.86
aCalibration check using certified weights.
bAudit using certified weights.
4-21 Rev. 1
4.7.3 Flow Measurement
The results of the flow rate audits were reported in the draft audit
report. Comparison of the audit results to the measurements made for the test
runs of the same days indicate good agreement for velocities and flow rates
measured using Method 2 (Table 4-13). The preliminary velocity measurement
by hot wire anemometer made by the auditor on April 14 indicated a greater
velocity than calculated from pitot tube measurements. The differences are
attributed to the hot wire anemometer velocity being a point measurement
compared with an average velocity resulting from the pitot tube measurements
and the presence of hydrocarbon in the stream, which can cause hot wire
anemometer velocity readings greater than actual velocities.
Table 4-13
Results of Flow Rate Audit
Date Test Velocity (ft/min) Flow Rate (cfm)
14-Apr Audit-hot wire 1930
NMMA-5-2 1738 3071
NMMA-2-2 1728 3053
15-Apr Audit-pitot 1787 3157
NMMA-16-1 1866 3297
NMMA-15-1 1793 3168
4.7.4 Styrene Evaporation Test
In accordance with the QAPP, styrene evaporation tests (the towel

tests) were conducted to evaluate the capture and recovery of styrene by the
sampling system. Two tests were completed, one at the initiation of the testing
program and one at the conclusion of the program. In both cases, the recovered
quantity of styrene met the acceptance criteria for a good system (±10 percent)
(Table 4-14).
4-22 Rev. 1
Table 4-14
Results of Styrene Evaporation Tests
Towel-1 Towel-2
Date 4/1/97 4/19/97
Material Losses (g) 1199 1095
Temperature (°F) 65.6 66
Flow rate (acfm) 3017 3306
Average THC Concentration (ppmv) 117.8 116.7
Duration (min) 66 59
Measured Emissions (g) 1202 1165
Percent Difference 0.21 6.4
The first test represents the best measure of the system and its
operation because no testing had been conducted in the building and, therefore,
no background styrene in building air or booth air interfered with the results.
The results from the first test showed excellent agreement between the measured
quantity of styrene evaporated from the apparatus (provided by RTI) and the
emissions determined by measuring stack gas concentrations and flow rate. The
closure between these two measurements is not as close for the final test,
however. Although it is true that the flow rate through the enclosure was greater
during the second test, the greater emissions measured are attributed more to
the mold left in the enclosure after previous lamination testing. Although the
measured styrene concentrations were adjusted for the concentration in the
incoming air and attempts were made to account for the background
concentration in the enclosure, it is likely that some of the styrene included in the
measured emissions may have been generated by the mold and part left in the
enclosure (i.e., residual curing emissions).
4-23 Rev. 1

Boat Resin Emissions

Uploaded by

Copyright:

Available Formats

Boat Resin Emissions

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Boat Resin Emissions

Uploaded by

Copyright:

Available Formats

BASELINE CHARACTERIZATION OF EMISSIONS

FROM FIBERGLASS BOAT MANUFACTURING

Air-Tech Environmental LLC

Radian International LLC

1.1 Background ....................................................................................................1-2

1.3.1 Test Enclosure.....................................................................................1-4

1.4 Organization of the Report .........................................................................1-10

2.0 RESULTS ....................................................................................................................2-1

2.1 Summary of Results.......................................................................................2-1

2.5.1 Emissions as Percent of Available Styrene ....................................2-17

3.0 CALIBRATION AND SAMPLING PROCEDURES...............................................3-1

3.1 Process Procedures and Measurements.......................................................3-2

3.1.1 Application Procedures .....................................................................3-3

3.2 Emission Testing Equipment and Procedures ............................................3-7

Table of Contents (Continued)

3.2.1 Leak Check and Instrument Calibration ..........................................3-7

3.2.1.1 Total Hydrocarbon Analyzer ............................................3-8

3.2.2 Flow Rate Measurements ................................................................3-11

3.3 Air Flow Over the Mold..............................................................................3-13

4.0 QUALITY CONTROL AND QUALITY ASSURANCE .........................................4-1

4.1 Concentration Measurements .......................................................................4-1

4.1.1 Total Hydrocarbon Measurements (Method 25A) ..........................4-2

4.2 Gelcoat and Resin Flow Rate ......................................................................4-14

4.7.1 Total Hydrocarbon Measurement ..................................................4-19

Table of Contents (Continued)

APPENDIX A Reference Methods

APPENDIX B Construction of a Temporary Total Enclosure for Volatile Organic

APPENDIX D Procedure for and Results of Gel Time Determination

APPENDIX E Material Safety Data Sheets

APPENDIX F Certifications for Gas Standards and Equipment Calibrations

APPENDIX G Bag Standards

APPENDIX I Comparison of NMMA Results with Those from Other Studies

1-1 Parameters Studied ...................................................................................................1-8

1-2 Surface Areas of Molds Used in Test Program.......................................................1-9

1-3 Properties of Gelcoat and Resin Materials ............................................................1-10

1-4 Test Schedule ...........................................................................................................1-11

2-1 Summary of Test Results ..........................................................................................2-2

2-2 Emissions Measured During Application of Resin to CFA Mold.......................2-11

2-3 Emissions from Gelcoating as Percentage of Available Volatiles.......................2-14

2-4 Total Emissions from Gelcoating Related to Mold Surface Area........................2-15

2-5 Emissions Measured During Resin Lamination ...................................................2-17

3-1 Measurements During Test Program ......................................................................3-1

3-2 Gel Time Results........................................................................................................3-6

3-3 Styrene and MMA Contents.....................................................................................3-6

4-1 Acceptance Criteria for Concentration Measurements ..........................................4-2

4-2 Calibration Error Measurements for Total Hydrocarbon Analyzer .....................4-3

4-3 Calibration Drift Measurements for Total Hydrocarbon Analyzer......................4-4

4-5 Method 18 Response Factors for Target Compounds ............................................4-7

4-6 Gas Chromatography Calibrations (after April 12) .............................................4-10

4-7 Comparison of Emission Results by Two Methods .............................................4-12

LIST OF TABLES (continued)

4-8 Spray Equipment Data Summary ..........................................................................4-14

4-9 Variability Measures for Flow Rate During Test Runs........................................4-16

4-10 Results of Balance Calibration................................................................................4-18

4-11 Audit Results for Method 25A Measurements .....................................................4-19

4-12 Results of Audit of Balance ...................................................................................4-21

4-13 Results of Flow Rate Audit.....................................................................................4-22

4-14 Results of Styrene Evaporation Tests ....................................................................4-23