United States EPA-600/R-97-()l8b
Environmental Protection
Agencv March 1997
SEPA Research and
Development
EVALUATION OF POLLUTION PREVENTION
TECHNIQUES TO REDUCE STYRENE EMISSIONS
FROM OPEN CONTACT MOLDING PROCESSES
Volume II, Appendixes
Prepared for
Office of Air Quality Planning and Standards
Prepared by
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
-------
FOREWORD
The U. S. Environmental Protection Agency is charged by Congress with pro-
tecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions lead-
ing to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet this mandate, EPA's research
program is providing data and technical support for solving environmental pro-
blems today and building a science knowledge base necessary to manage our eco-
logical resources wisely, understand how pollutants affect our health, and pre-
vent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks
from threats to human health and the environment/The focus of the Laboratory's
research program is on methods for the prevention and control of pollution to air,
land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites and groundwater; and prevention and
control of indoor air pollution. The goal of this research effort is to catalyze
development and implementation of innovative, cost-effective environmental
technologies; develop scientific and engineering information needed by EPA to
support regulatory and policy decisions; and provide technical support and infor-
mation transfer to ensure effective implementation of environmental regulations
and strategies.
This publication has been produced as part of the Laboratory's strategic long-
term research plan. It is published and made available by EPA's Office of Re-
search and Development to assist the user community and to link researchers
with their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/R-97-Ol8b
March 1997
Evaluation of Pollution Prevention
Techniques to Reduce Styrene Emissions
from Open Contact Molding Processes
V
Volume II, Appendixes
By
Emery J. Kong, Mark A. Bahner, Robert S. Wright, and C. Andrew Clayton
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709
EPA Cooperative Agreement CR 818419-03
EPA Project Officer: Geddes Ramsey
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Prepared for
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
Abstract
Pollution prevention options to reduce styrene emissions, such as new materials and
application equipment, are commercially available to the operators of open molding processes.
However, information is needed about the percent reduction in emissions that is achievable
with these options.
To meet this need, several of these pollution prevention options were examined.
Options examined were operator techniques, air flow velocities in the spraying area, gel coat
and resin formulations, and application equipment. Styrene emission factors calculated from
this test result were compared with the existing AP-42 emission factors for gel coat sprayup
and resin applications.
The study found that using controlled spraying (i.e., reducing overspray), low-styrene
and styrene-suppressed materials, and nonatomizing application equipment can reduce styrene
emissions from 11 to 52 percent. Facilities should investigate the applicability and feasibility
of these pollution prevention options to reduce their styrene emissions. The calculated
emission factors were from 1.6 to 2.5 times the mid-range AP-42 emission factors for the
corresponding gel coat and resin application. These results indicate that facilities using
existing AP-42 emission factors to estimate emissions in open molding processes are likely to
underestimate actual emissions.
n
-------
Contents
Appendixes
A. A Category III Quality Assurance Project Plan for the Evaluation of Pollution Prevention
Techniques to Reduce Styrene Emissions from Open Contact Molding Processes
B. Reichhold Standard Test Methods
C. Verification and Intercomparison of Compressed Gas Calibration Standards
D. Summary of Calibration Data, Calibration Error Tests, and Drift Checks
E. A Summary of Emission Measurements, Gravimetric Measurements, and Calculated
Emission Quantities and Emission Factors for the Test
F. Statistical Analyses of Test Results
G. THC Analyzer Evaluation: Sampling Line. Loss and Pressure Effect
H. An RTI Technical System Audit Report
1. EPA Performance Evaluation of Total Hydrocarbon Analyzer
in
-------
Appendix A
A Category IE Quality Assurance Project Plan for the Evaluation of Pollution Prevention
Techniques to Reduce Styrene Emissions from Open Contact Molding Processes (minus
Appendixes) and RTFs Responses to EPA's Comments on the Category IE Quality Assurance
Project Plan
A-i
-------
POLLUTION PREVENTION TECHNOLOGY DEMONSTRATION
EVALUATION OF POLLUTION PREVENTION TECHNIQUES
TO REDUCE STYRENE EMISSIONS FROM OPEN CONTACT
MOLDING PROCESSES
CATEGORY III QUALITY ASSURANCE PROJECT PLAN
EPA Cooperative Agreement CR 818419-03
RTI Project No. 96U-5171-016
Submitted to:
EPA Project Officer - Carlos Nunez
U.S. Environmental Protection Agency
Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
Prepared by:
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709
April 28, 1995
A-ii
-------
SIGNATURE PAGE
Project Officer r U.S. EPA, ORD
Air & Energy Engineering Research Laboratory
Carlos Nunez
tfate
Quality Assurance Officer - U.S. EPA, ORD
Air & Energy Engineering Research Laboratory
Nancy Aaams
Date
Project Manager
Research Triangle Institute
Jesse Baskir
Date
Project Leader
Research Triangle Institute
Quality Assurance Manager
Research Triangle Institute
Emery Kong
Cynthia Salmons
Date1
Date
A-iii
-------
Recipients of official copies of this Quality Assurance Project Plan:
Mr. Carlos Nunez
Mr. Geddes Ramsey
Dr. Nancy Adams
Dr. Jesse Baskir
Mr. Emery Kong
Ms. Cynthia Salmons
Mr. Mark Bahner
Mr. Andrew Clayton
Dr. R. K. M. Jayanty
Mr. Mark Callicutt
Mr. Federico Linares
Mr. Lorenzo Esposito
Mr. Mark Hollenbech
Mr. Tom Hedger
Mr. Charles Stard
Mr. Casey Herbert
Mr. Bob Lacovara
Staff
U.S. EPA, Air & Energy Engineering Research Laboratory
Project Officer
U.S. EPA, Air & Energy Engineering Research Laboratory
Quality Assurance Officer
Research Triangle Institute
Pollution Prevention Program Acting Manager
Research Triangle Institute
Project Leader
Research Triangle Institute
Quality Assurance Manager
Research Triangle Institute
Testing Crew Chief
Research Triangle Institute
Experimental Design and Data Analysis
Research Triangle Institute
Emission Measurement Team
Reichhold Chemicals, Inc.
Technical Service Supervisor
Reichhold Chemicals, Inc.
Manager of Physical Testing and Application
Reichhold Chemicals, Inc.
Senior Technical Service Representative
Cook Composites and Polymers
Technical Manager for Gelcoat
Magnum Industries
Research and Development
W.E.T. Inc.
Composites Fabricators Association
Director of Technical Service
Research Triangle Institute
Testing Crew Members
A-iv
-------
TABLE OF CONTENTS
1.0 PROJECT DESCRIPTION A-1-1
1.1 Background A-1-1
1.2 Data Quality Objectives A-1-2
1.3 Intended Use of Data A-1-2
1.4 Scope of Work A-1-2
1.4.1 Pilot Experiment A-1-3
1.4.2 Gelcoat Experiment A-1-4
1.4.3 Resin Experiment A-1-5
1.5 Schedule/Milestone Chart A-1 -7
1.6 Facility Description . A-1 -9
1.6.1 Total Enclosure System A-1-9
1.6.2 Resin Property Testing Laboratory A-1-11
1.6.3 Open Contact Mold A-1-11
1.7 Experimental/Test Matrix Design A-1-11
1.7.1 Critical and Noncritical Measurements A-1-11
1.7.2. Experimental Design A-1-13
1.7.2.1 Pilot Experiment A-1-13
1.7.2.2 Gelcoat Experiment A-1-16
1.7.2.3 Resin Experiment A-1-17
1.7.2.4 Combining the Gelcoat and Resin Experiments A-1-18
2.0 PROJECT ORGANIZATION AND RESPONSIBILITIES '. . A-2-1
3.0 DATA QUALITY INDICATOR GOALS FOR CRITICAL MEASUREMENTS A-3-1
3.1 Objectives for Quantitative Data Quality Indicators ..-.'.' A-3-2
3.1.1 Precision A-3-2
3.1.2 Accuracy A-3-3
3.1.3 Detection Limit A-3-3
3.1.4 Completeness A-3-3
3.2 Objectives for Qualitative Data Quality Indicators A-3-4
3.2.1 Representativeness A-3-4
3.2.2 Comparability A-3-4
4.0 SAMPLING PROCEDURES A-4-1
4.1 Total Enclosure and Capture Efficiency Test A-4-1
4.2 Sampling Location and Duration of Test Run A-4-1
4.3 Testing Procedures A-4-1
4.4 Emission Measurement A-4-3
4.5 Exhaust Air Flow Rate Measurement A-4-4
4.6 Mass Balance Determination A-4-4
4.7 Gelcoat and Resin Properties A-4-4
A-v
-------
4.8 Environmental Conditions A-4-5
4.8.1 Linear Air Velocity in Spray Booth A-4-5
4.8.2 Ambient Temperature A-4-6
4.8.3 Relative Humidity A-4-6
4.9 Equipment Type and Setup A-4-6
4.9.1 Equipment Type A-4-6
4.9.2 Equipment Setup A-4-6
4.10 Operating Parameters for Parts A-4-7
4.10.1 Catalyst Ratio A-4-7
4.10.2 Gelcoat/Resin Thickness A-4-7
4.10.3 Glass/Resin Ratio A-4-7
5.0 ANALYTICAL PROCEDURES A-5-1
5.1 Types of Gelcoat and Resin Materials A-5-1
5.2 Styrene Content A-5-1
5.3 Gel time, Time to Peak, and Peak Exotherm Characteristics of
Polyester Resins A-5-1
6.0 DATA REDUCTION, VALIDATION, AND REPORTING A-6-1
6.1 Data. Reduction A-6-1
6.2 Data Validation A-6-3
6.3 Data Reporting A-6-4
7.0 INTERNAL PERFORMANCE AND SYSTEM AUDITS A-7-1
7.1 Technical Systems Audits A-7-1
7.2 Performance Evaluation Audits A-7-1
7.3 Audits of Data Quality A-7-1
8.0 CALCULATION OF DATA QUALITY INDICATORS A-8-1
8.1 Precision A-8-1
8.2 Accuracy A-8-1
8.3 Method Detection Limit A-8-2
8.4 Completeness A-8-2
9.0 CORRECTIVE ACTION A-9-1
A-vi
-------
LIST OF FIGURES
Figure 1-1. Proposed schedule and milestones A-1-8
Figure 1 -2. Side view of the Reichhold Chemicals spray booth in a
permanent total enclosure A-1-10
Figure 1-3. Sketch of a male mold A-1-12
Figure 2-1. Project organization chart A-2-3
Figure 6-1. The data handling, reduction, validation, and reporting
procedures A-6-2
LIST OF. TABLES
Table 1-1. Anticipated Performance of Evaluation Test A-1-3
Table 1-2. Test Runs for Pilot Experiment .. . '. . A-1-4
Table 1 -3. Test Runs for Each of the Gelcoat Formulations and
Equipment Combinations A-1-5
Table 1-4. Test Runs for the Resin Formulations and Equipment Types ...... A-1-7
Table 1-5. Summary of Critical and Noncritical Measurements A-1-14
Table 1 -6. Randomly Ordered List of Trials for the Main Experiment . A-1 -20
Table 3-1. Summary of Styrene Emission Data from a Previous RTI Testing .. A-3-1
Table 3-2. Objectives for Quantitative Data Quality Indicators A-3-2
Table 4-1. Summary of Measurement Location and Frequency A-4-2
Table 4-2. Air Flow Pattern and Velocity in the Spray Booth A-4-5
Table 4-3. Data Recording Sheet for the Pollution Prevention Technique Evaluation
Test .; A-4-8
Table 5-1. Data Recording Sheet for Gelcoat and Resin Properties A-5-2
A-vii
-------
Section No.: 1.0
Revision: 0
Date: April 28, 1995
Page: A-1-1of20
1.0 PROJECT DESCRIPTION
1.1 Background
The Research Triangle Institute (RTI) is under a cooperative agreement with the
U.S. Environmental Protection Agency (EPA), Air and Energy Engineering Research
Laboratory (AEERL), to evaluate pollution prevention techniques to reduce styrene
emissions from open contact molding processes. The open contact molding process is
one of the most common production processes used by the reinforced plastics and
composites (RP/C) industry. This process is used to manufacture boats, bathtubs,
shower stalls, truck body parts, swimming pools, storage tanks, corrosion-resistant
equipment, furniture and accessories, electrical and equipment housings and
enclosures, duct and air handling equipment, etc. It is one of the RP/C processes that
consumes the most polyester resins. It also has the greatest potential of emitting
styrene due to the spraying equipment used and the openness of the process.
Styrene is emitted during the application stage when a catalyzed gelcoat or resin
is applied to the surface of an open contact mold. Styrene continues to emit from wet
gelcoat or resin during gelation and curing. The open contact molding process usually
is conducted in a facility with ample ventilation to maintain the ambient styrene
concentrations under current Occupational Safety and Health Association
-------
Section No.: 1.0
Revision: 0
Date: April 28.1995
Page: A-1-2of20
the most effective and practical pollution prevention techniques to reduce styrene
emissions.
This Quality Assurance Project Plan (QAPP) describes the scope of the
proposed pollution prevention techniques evaluation test and the procedures that will be
employed in the evaluation test to ensure that the data collected are of sufficient quality
to achieve project objectives.
1.2 Data Quality Objectives
The objective of this testing is to determine the styrene emission reduction for
several pollution prevention techniques from the baseline conditions. Pollution
prevention techniques will be evaluated for gelcoat and resin applications on open
contact molding processes. The baseline emissions will be determined for a regular
gelcoat and a regular resin formulation using an air-assisted airless spray gun under a
typical environmental condition. Comparison of emissions will be based on styrene
emission factors expressed as weight percent of available styrene (% AS) and as mass
per unit surface area (g/m2). The former unit is the unit used in the EPA Compilation of
Air Pollutant Emission Factors, AP-42 Document.
Table 1 -1 summarizes the anticipated performance of the evaluation in terms of
the widths of confidence intervals for differences m mean %AS associated with the
primary comparisons of interest. The widths of the intervals depend on the magnitude
of the measurement error standard deviation, which is denoted by o. Results are
presented for o = 1, 3, and 5 percentage points. The prior test results (shown in Table
3-1) suggest that standard deviations in this range should be achievable by mass
balance calculation method.
1.3 Intended Use of Data
The test results will be analyzed, summarized, and presented to the RP/C
industry in an EPA report. The report will provide quantitative emission reduction
potentials for the pollution prevention techniques evaluated, so that a facility owner or
operator can identify the most effective and practical pollution prevention techniques to
reduce styrene emissions from its operation.
1.4 Scope of Work
This testing will include a pilot experiment and a main experiment. From the pilot
experiment, the linear air flow velocity and the spraying technique will be determined for
the main experiment. The main experiment will include a gelcoat experiment, which will
examine two gelcoat formulations with three pieces of gelcoating equipment, and a
resin experiment, which will examine six resin formulations and four pieces of resin
A-1-2
-------
Section No.:
Revision: 0
Date:
Page:
1.0
April 28,1995
A-1-3o<20
application equipment. The formulations and equipment selected for the testing are
representative of current and evolving technologies available to the RP/C industry.
Each of the experiments is described in the following subsections.
Table 1-1. Anticipated Performance of Evaluation Test
Difference in mean
%AS due to:
Alternative gelcoat formulation
vs. baseline formulation
(based on 12df)b
Alternative gelcoat spray
equipment vs. baseline
equipment
(basedon12df)b
Alternative resin formulation vs.
baseline formulation
(based on 15df)c
Alternative resin spray
equipment vs. baseline
equipment
(based on 9 df)c
Expected half width of 95% confidence interval on :
difference3
0=1
ฑ1.03
ฑ1.26
ฑ1.51
ฑ1.60
o = 3
ฑ3.08
ฑ3.77
ฑ4.52
ฑ4.80
0 = 5
ฑ5.14
ฑ6.29
ฑ7.54
ฑ8.00
Construction of the interval is not meaningful if there is an interaction of the gelcoat formulations
and equipment types.
c Half-widths are conservative in that larger numbers of degrees of freedom (df) may be available;
this will lead to narrower confidence intervals.
1.4.1 Pilot Experiment
Before these formulations and equipment are examined, we will conduct a pilot
experiment to determine the air flow velocity in the spray booth and the spraying
method that will be used throughout the entire test. A low and a high air flow velocity in
the spray area will be examined. A low air flow velocity in the spraying area will be
established by diverting the makeup air away from the spraying area to the sides of the
spray booth. A normal and a more careful spraying method will be examined for the
application technique. The pilot experiment will be conducted using a low-profile resin
or a regular gelcoat catalyzed with methyl ethyl ketone peroxide (MEKP) and by an air-
assisted airless (AAA) spray gun. Reichhold Chemicals or Cook Composites and
A-1-3
-------
Section No.: 1.0
Revision: 0
Date: April 28.1995
Page: : A-1-4of20
Polymers will provide the resin or gelcoat, respectively. Magnum Industries will provide
the AAA spray gun and an operator to operate it. The results will be analyzed to
determine whether there are any differences in styrene emissions resulting from
different air flow velocities and spraying methods. Following the pilot experiment, one
air flow velocity and one spraying method will be selected for the gelcoat and resin
experiments in the evaluation test.
The number of test runs for airflow velocity and spraying method are
summarized as follows and presented in Table 1-2.
A. Air flow velocity
A1. Low air flow velocity (40 to 100 fpm)
A2. High air flow velocity (100 to 200 fpm)
B. Spraying method
M1. Normal technique without conscious control of overspray from flanges
M2. Alternative spraying technique with more conscious control of overspray
Table 1-2. Test Runs for Pilot Experiment
M1 -Normal technique
M2-Altemative technique
A1 -Low air flow velocity
3
3
A2-High air flow velocity
3
3
1.4.2 Gelcoat Experiment
The geleoat formulations selected are one regular gelcoat with isophthalic
acid/neopentyl glycol (ISO/NPGฎ)-based resin and a low volatile organic compound
(VOC) gelcoat with the same base resin: Cook Composites and Polymers (CCP) will
provide these two gelcoats. For the purpose of this testing, both gelcoats will contain
straight styrene without any methyl methacrylate. MEKP catalyst will be used and the
catalyst ratios will follow those suggested by CCP.
The gelcoating equipment selected includes: one AAA external catalyst mixing
spray gun, one high-volume low-pressure (HVLP) external catalyst mixing spray gun,
and one HVLP internal catalyst mixing spray gun. The AAA external mixing spray gun
is considered the baseline condition of the industry. The AAA spray gun will be
compared with HVLP spray guns. The effects of internal and external catalyst mixing
will be evaluated for the HVLP spray guns. Magnum will provide the AAA and the other
two HVLP spray guns. A pump ratio of 20:1 will be selected for the gelcoat pump
systems. The spray guns will be compared at similar gelcoat delivery rates.
A-1-4
-------
Section No.: 1.0
Revision: 0
Date: April 28, 1995
Page: A-1-5of20
The gelcoat formulations and application equipment are denoted as follows:
a. Formulations
GF1. a regular (ISO/NPGฎ) gelcoat containing only styrene monomer
(baseline condition)
GF2. a low VOC styrene-suppressed (ISO/NPGฎ) gelcoat containing only
styrene monomer
b. Equipment
GE1. an AAA external catalyst mixing spray gun (baseline condition)
GE2. an HVLP internal catalyst mixing spray gun
GE3. an HVLP external catalyst mixing spray gun
Table 1-3 shows the number of test runs for each of the gelcoat formulation and
equipment combinations in the gelcoat experiment.
Table 1-3. Test Runs for Each of the Gelcoat Formulations
and Equipment Combinations
Formulation
GF1 -Regular gelcoat
GF2-Low VOC gelcoat
Equipment type
GEI-AAA(ext)
3
3
GE2-HVLP(int)
3
3
GE3-HVLP(ext)
3
3
ext=External catalyst mixing.
int=lnternal catalyst mixing.
1.4.3 Resin Experiment
The resin experiment will examine six resin formulations with an AAA spray gun
and four application equipment with one standard resin.
The resin formulations selected are one dicyclopentadiene (DCPD)-based low-
profile resin catalyzed with MEKP, one DCPD-based low-styrene resin, one
orthophthalic(ORTHO)-based styrene suppressed resin, one DCPD-based low-profile
resin catalyzed with benzoyl peroxide (BPO), a water-emulsified resin catalyzed with
MEKP, and the same ORTHO-based styrene-suppressed resin at a higher suppressant
concentration. All the resin formulations will be sprayed by an air-assisted airless spray
gun. Reichhold Chemicals, Inc., will provide all resin formulations and catalyst, except
the water-emulsified resin. The W.E.T., Inc. will provide the water emulsified resin.
The catalyst ratios will follow those suggested by Reichhold and W.E.T., Inc.
The resin application equipment selected are one AAA external catalyst mixing
spray gun; an internal catalyst mixing flow coater; a pressure-fed roller; and an AAA
A-1-5
-------
Section No.: 1.0
Revision: 0
Date: April 28,1995
Page: A-1-6 of 20
external catalyst mixing spray gun modified for the resin catalyzed with BPO. The AAA,
external mixing spray gun is consider the baseline condition of the industry. The AAA
spray gun is to be compared with other nonspraying equipment (i.e., the flow coater and
the pressure-fed roller). Magnum will provide all the equipment for evaluation. A pump
ratio of 11:1 will be selected for the resin pump systems. The equipment will be
compared at similar resin delivery rates.
Resin formulations and application equipment are denoted as follows:
a. Formulations
RF1. a DCPD-based low-profile resin catalyzed with MEKP (baseline
condition)
RF2. a DCPD-based low-styrene resin
RF3. an ORTHO-based styrene-suppressed resin
RF4. a DCPD-based low-profile resin catalyzed with BPO
RF5. a water-emulsified resin
RF6. the same ORTHO-based styrene suppressed resin at a higher
suppressant concentration
b. Equipment
RE1. an AAA external catalyst mixing spray gun (baseline condition)
RE2. an internal catalyst mixing flow coater
RES. an internal catalyst mixing pressure-fed roller
RE4. a modified AAA, external catalyst mixing spray gun for the resin
catalyzed with BPO
Table 1-4 show the number of test runs for the resin formulation and equipment
to be examined in the resin experiment.
A-1-6
-------
Section No.:
Revision: 0
Date:
Page:
1.0
April 28,1995
A-1-7of20
Table 1-4. Test Runs for the Resin Formulations
and Equipment Types
Formulation
RF1-DCPD-based low-
profile resin with MEKP
RF2-DCPD-based low-
styrene resin
RF3-ORTHO-based styrene-
suppressed resin
RF4-DCPD-based low-
profile resin with BPO
RF5-water-emulsified resin
RF6-the ORTHO-based
styrene-suppressed resin at
a higher suppressant
concentration
Equipment type
REI-AAA(ext)
6
3
3
NA
3
3
RE2-flow
coater(int)
3
NA
NA
NA
NA
NA
RE3-pressure-
fed roller(int)
3
NA
NA
NA
NA
NA
RE4-modified
AAA (ext)
NA
NA
NA
3
NA
NA
ext=Extemal catalyst mixing.
lnt=lnternal catalyst mixing.
NA = Not included in the experiment.
1.5 Schedule/Milestone Chart
The schedule and milestones for this testing are shown in Figure 1 -1. The
schedule and milestones are determined by the overall project completion date at the
end of September. The evaluation test is scheduled to start in the first week of June
1995 and will take 4 weeks to complete. The resin and gelcoat manufacturers and
equipment vendors will need to provide the materials and equipment to the test site at
Reichhold Chemicals Inc., c/o Mr. Mark Callicutt, 2400 Ellis Road, Durham, North
Carolina 27703-5543 by May 26,1995. RTI will prepare a temporary total enclosure
setup at Reichhold's spray booth in late May. RTI will conduct preliminary testing to
.ensure that the enclosure and emission measurement instrument meet EPA
requirements. The preliminary testing will also ensure that the emissions are within the
proper concentration ranges of the total hydrocarbon (THC) analyzer prior to the actual
testing.
The proposed schedule shows that the testing will take place during the
month of June. The testing schedule is based on the assumption that three test
runs can be conducted in each working day and 19 working days will be needed to
A-1-7
-------
Major activities
1. Prepare and submit a Category III QAPjP
to EPA
2. EPA reviews and approves the QAPjP
3. RTI revises the QAPjP as needed
4. Make arrangements to have materials
and equipment delivered to Reichhold
5. Set up the temporary total enclosure
system and instrumentation at Reichhold,
conduct preliminary testing to ensure that
the setup, instrument, and equipment
work property
6. Conduct the testing at Reichhold
7. Analyze test results
8, Prepare and submit a draft report to
involved organizations for comments
9. Technical review of the draft report
by involved organizations
10. Finalize the project report
Project schedule and milestones
March
April
May
June
July August
COT
Sept
-0 O 30 W
* II 8
s R *
Figure 1-1. Proposed schedule and milestones.
f > O -
dol b
(0
ซJ
Ol
-------
Section No.: 1.0
Revision: 0
Date: April 26, 1995
Page: A-1-9of20
complete the proposed 57 test runs. The final test period may be longer than the
proposed duration, if Reichhold needs the spray booth for its own testing or the test
team encounters technical difficulties that need to be resolved before the test can be
resumed.
1.6 Facility Description
The evaluation test will be conducted in an isolated spray booth in the Reichhold
Chemicals' physical testing laboratory, located in Research Triangle Park, North
Carolina. This type of spray booth is commonly used in a gelcoating area of an RP/C
facility. The Reichhold Chemicals' physical testing laboratory is used to conduct testing
for their resin users. It is not a production facility; therefore, the background styrene
concentration can be minimized.
1.6.1 Total Enclosure System
The spray booth is situated in an enclosed room with a double door leading to
the physical testing laboratory. Figure 1-2 shows the side view of the spray booth. The
room is 12 feet wide, 19 feet high, and 15 feet deep, which can be considered a
permanent total enclosure. The double door measures 6 feet wide by 7 feet high,
which can be consider the natural draft opening to the enclosure.
The spray booth is 7 feet high, 11.5 feet wide, and 7.5 feet deep from the front
edge to the filter bank. The filter bank is 6 feet high by 11 feet wide. The distance
between the front edge of the spray booth to the double door is 4 feet 10 inches. The
conditioned makeup air is provided through a duct (3 feet 9 inches by 4 feet) above the
open space between the spray booth and the double door. This duct is considered a
forced draft opening to the total enclosure system. The makeup air flows downward,
then turns horizontally through the spray booth. The exhaust air flows through the filter
bank at the end of the spray booth and is exhausted upward by a duct 34 inches in
diameter. The exhaust flow rate from the spray booth is estimated to be 8,000 cfm
when the double door is closed. The actual flow rate will be determined in preliminary
testing.
Emission measurements and exhaust air flow rate will be monitored from the
exhaust duct. The sampling location is 8 diameters downstream of the last bend as
shown in Figure 1-2. EPA Methods 1 and 2 will be used to determine the exhaust gas
velocity and volumetric flow rate. EPA Method 25A will be used to determine total
gaseous organic emissions. EPA Method 204 will be used to ensure that the enclosed
room meets the criteria for a total enclosure. The sampling procedures are outlined in
Section 4.0.
A-1-9
-------
Section No.:
Revision:
Data:
Page:
1,0
0
April 28, 1995
1-10 of 20
Emission and Air Row
Rate Monitoring Location
Makeup
Air Duct
3'9"x4'
Ground Floor
7'x6' (W)
Double Door
Application Equipment
(Could be Placed Inside
or Outside the Enclosure)
Figure 1-2. Side view of the Reichhold Chemicals spray booth in
a permanent total enclosure (19'H x 12'W x 15'L)
A-1-10
-------
Section No.: 1.0
Revision: 0
Date: April 28,1995
Page: A-1-11of20
1.6.2 Resin Property Testing Laboratory
Reichhold Chemicals has a resin property testing laboratory located in the same
building as the spray booth. The laboratory has all the instrument and equipment
necessary to determine the styrene contents and curing characteristics for the gelcoat
and resin formulations. Reichhold personnel will help the project team examine these
properties for every gelcoat and resin formulation; Available instruments include an
analytical balance, a forced-air oven, several Brookfield viscometers, a thermocouple,
and a temperature recorder.
1.6.3 Open Contact Mold
A male mold will be used for this evaluation test. Figure 1-3 shows a drawing of
this mold. The male mold will have five exposed smooth surfaces similar to a
rectangular box. The mold will measure 2 feet high, 2,5 feet long, and 2 feet wide. A 2-
inch wide flange surrounds the bottom of the mold for ease of part removal. The total
surface area, including flange, equals 24.5 ft2. The shape and surface area of this mold
are selected to simulate real conditions in the open contact molding process. The mold
is constructed of traditional reinforced plastics material to represent real molds used by
the industry. The mold will be placed on a cart with wheels so that the operator can
spray on all mold surfaces by turning the cart and without moving his position to the
down-wind location.
1.7 Experimental/Test Matrix Design
1.7.1 Critical and Noncritical Measurements
The critical and noncritical measurements, the frequency of measurement, the
locations where these measurements are taken, and methods of measurements are
shown in Table 1-5. The emission measurements, exhaust airflow rate, and mass
balance calculations are critical measurements for this study. The emission rates are to
be determined from styrene emissions measured as total hydrocarbon concentration
and the exhaust air flow rate monitored over the duration of the test run. Mass balance
calculations will be used to determine transfer efficiency of the application equipment
and to determine the weight losses from the mold and from the overspray. The styrene
contents of the gelcoat and resin formulations are critical measurements; because the
information will be used to express the styrene emission factors as the percent of
available styrene in the materials.
The linear air velocity over the mold and the ambient temperature are non-critical
measurements, but they will be recorded to document actual test conditions. The type
of gelcoat or resin materials and the equipment type will be noncritical parameters, but
the emission quantities will be compared for different materials and equipment.
A-1-11
-------
Section No.:
Revision:
Date:
Page:
1.0 '
0-
April 28, 1995
1-t2 of 20
2ฐ Flange surrounding
the bottom of the mold
to ease part removal
Figure 1-3. Sketch of a male mold.
A-1-12
-------
Section No.: 1.0
Revision: 0
Date: . April 28,1995
Page: A-1-13of20
The equipment setup (such as pump and air pressure, spray tip size and angle,
delivery rate, and catalyst ratio setting) are noncritical. The duration of application,
gelcoat/laminate thickness, glass/resin ratio, gel time, total time to peak, and peak
exotherm are noncritical measurements. However, these equipment conditions and
gelcoat/resin parameters are important information that should be documented.
1.7.2 Experimental Design
1.7.2.1 Pilot Experiment
Before executing the main experiment, several preliminary runs are proposed.,
The primary purpose is to help establish "standard" conditions under which the main
experiment will be conducted. A secondary purpose is to gain insight into the
magnitude of measurement error variability that might be anticipated in the main
experiment. If the pilot experiment indicates major difficulties with the planned
approach, this QAPP will be amended to indicate changes to the main experiment. -RTI
will acquire verbal approval from the EPA Project Office for changes that occur during
the test activities and submit a QAPP change as soon as possible. The proposed pilot
consists of 12 trials - namely, three replicates for each of the following conditions:
a. Normal spraying method (M1), low air flow velocity (A1)
b. Alternative spraying method (M2), low air flow velocity (A1)
c. Normal spraying method (M1), high airflow velocity (A2)
d. Alternative spraying method (M2), high air flow velocity (A2)
Resin RF1 or gelcoat GF1 will be used in all cases, the following random
ordering of the 12 trials will be used: b,c,d,c,b,a,d,a,d,b,c,a. Since one of the four
conditions wil* be chosen as the standard method for the subsequent trials, it may be .
possible to use those three pilot trials in the analysis of the.main resin experiment (e.g.,
to provide more degrees of freedom for error variability).
The pilot will provide only a limited amount of information that can be used for
statistical purposes. The data can be used, however, to give some idea of the impact
of these two parameters on emissions. The analysis of variance for the experiment is
as follows:
A-1-13
-------
Section No.:
Revision: 0
Date:
Page:
1.0
April 28, 1995
A-1-14of20
Table 1-5. Summary of Critical and Noncritical Measurements
Measurement
THC concentration
Exhaust air flow rate and velocity head Up)
Humidity
Mass balance calculations
1) gelcoat/resin materials used
2) gelcoat/resin applied on mold
3) cured material on mold
4) cured material on other ground cover
Types of gelcoat and resin materials
Styrene content
Gelcoat/resin properties
1) gel time
2) total time to peak
3) peak exotherm
Linear air velocity in the spray booth
Ambient temperature
Equipment type
Equipment setup
1) pump pressure
2) air pressure
3) spray tip size
4) spray tip angle
5) catalyst ratio setting
6) equipment delivery rate
Gelcoat/Resin data
1) catalyst ratio
2) gelcoat/laminate thickness
3) g[ass/resin ratio
Classificatio
n
Critical
Critical
Noncritical
Critical
Noncritical
Critical
Noncritical
Noncritical
Noncritical
Noncritical
Noncritical
Noncritical
Method
EPA Method 25A
EPA Methods 1 and 2
Sling psych rometer or relative
humidity detector
High precision scales with 1 50 kg
capacity and 1 g readability
Manufacturer data
Reichhold standard test method
No. 18-001
Reichhold standard test method
No. 18-050 and 18-051
Hot wire anemometer
Thermocouple
Vendor information
Vendor information and actual
setting on equipment
Mass calculation and mil gauges
A-1-14
-------
Section No.: 1.0
Revision: 0
Date: April 28,1995
Page: A-1-15o(20
Source of
variation
M (spraying method)
A (air flow velocity)
M x A (interaction)
Error
Degrees of
freedom
1
1
1
8
Total .11
The magnitude of differences in mean %AS that can be detected via the pilot
experiment can be expressed in terms of the half-width of confidence intervals (C.I.) on
the differences of interest:
Difference in mean %AS Expected half-width of 95% C.I.
M2vs. M1 1.3310
A2vs. A1 1.3310
For any two cells 2.306o
In this table, o denotes the standard deviation associated with measurement
variability of emissions in %AS. The width of the confidence intervals is based on an
assumption that the data for a given combination of M and A will be approximately
normally distributed with a common measurement error/variability (similar assumptions
apply to all other confidence intervals described herein). The 95 percent confidence
interval half-width on the difference can be related to a pairwise hypothesis test ..-.-
(conducted at a significance level of 0.05) in two ways. First, if the estimated
confidence interval does not include zero, then the corresponding test of no difference
in mean %AS will be rejected. Second, if the true difference between the means (for
A1 and A2, say) is equal to the expected half-width, then we will have a 50 percent
chance of detecting a difference in the means. (Of course, if the true difference is
larger than the half-width, then there will a higher likelihood that we will be able to
detect a difference.) Thus if the underlying error variability of emissions in %AS is 5
percentage points, then we should have about a 50 percent chance of finding a
difference in the two flow rates if the true difference between them is about 6.7 (i.e.,
1.331x5).
A similar statement can be made regarding the methods. Both of these
statements assume that there is not a method by air flow interaction (in which case the
overall comparisons would not generally be meaningful). It should be noted that the
above analysis of variance (ANOVA) and confidence interval statements rely on an
assumption of measurement-error variance homogeneity across the four combinations
of factors M and A.
A-1-15
-------
Section No.: 1.0
Revision: 0
Date: April 28,1995
Page: A-M6of20
1.7.2.2 Gelcoat Experiment
The gelcoat experiment is one of the two major components in the main
experiment. This experiment is aimed at evaluating how styrene emissions are affected
by type of gelcoat formulation (factor GF) and type of equipment (factor GE). The
factor combinations (two formulations and three equipment types) and proposed
sample sizes are given in Table 1-3. The 18 trials are to be performed in random order
(trials will be interspersed with the trials of the resin experiment, described below).
The ANOVA associated with the design (a completely random design with three
replications) has the following structure:
Source of Degrees of
variation freedom
GF (gelcoats) 1
GE (equipment) 2
GF x GE (interaction) 2
Error ' 12
Total 17 .
For the gelcoat experiment, the magnitude of differences that can be detected
can be expressed in terms of the half-width of confidence intervals on the differences of
interest:
Difference in mean %AS Expected half-width of 95% C.I.
GF2vs. GF1 1.0270
GE2vs. GE1,GE3vs. GE1,
orGE3vs. GE2 1.258o
For any two cells 1.779o
Again, o denotes the standard deviation associated with measurement error
variability. The 95 percent confidence interval half-width on the difference can be
related to a pain/vise hypothesis test (conducted at a significance level of 0.05) in two
ways. First, if the estimated confidence interval does not include zero, then the
corresponding test of no difference will be rejected. Second, if the true difference
between the means (for GF1 and GF2, say) is equal to the expected half-width, then we
will have a 50 percent chance of detecting a difference in the means. Hence if the
underlying error variability of emissions in percent available styrene is 5 percentage
points, then we should have (1) about a 50 percent chance of finding a difference Hi the
gelcoat formulations if the true difference between them is about 5.1, and (2) about a .
50 percent chance of detecting a difference in two pieces of equipment if the true
difference between them is about 6.3. Both of these statements assume that no
A-1-16
-------
Section No.: 1.0
Revision: 0
Date: April 28. 1995
Page: A-1-17of20
difference between them is about 6.3. Both of these statements assume that no
interactions are present. It should be noted that the above ANOVA and confidence
interval statements rely on an assumption of measurement error variance homogeneity
across equipment, gelcoat formulations, and levels of percent available styrene. If the
variability appears to change with level, then transformations such as logarithms will be
considered (for the logarithmic transformation, the o can then be interpreted as the true
underlying relative standard deviation).
1.7.2.3 Resin Experiment
The resin experiment is the other major component of the main experiment. This
experiment is aimed at evaluating how styrene emissions are affected by type of resin
formulation (factor RF) and type of equipment (factor RE). Table 1 -4 shows the
proposed combinations and associated test runs. This design consists of 27 trials,
which will be run in random order. Separate ANOVAs are used to test for different
equipment (REs) and for the different resin formulations (RFs).
The ANOVA for equipment comparisons is as follows:
Source of Degrees of
variation freedom
RE (equipments) 2
Error 9 or 12 or 19 or 22
The various choices for the error degrees of freedom (df) result from which set of
data is used for estimation of error variability: If just the 12 observations associated
with the different equipment are employed, then 9 df result; if the 3 trials from the pilot
are included, then 12 df result; if all trials in the resin experiment (Table 1-4) are used,
19 df are available; and if the 3 pilot trials are added, 22 df are available. If variances
across all cells in the experiment appear homogeneous and are also consistent with the
variance of the pilot trials, then use of the larger degrees of freedom is warranted.
The expected half-width of confidence intervals on the differences of equipment
are as follows:
Difference in mean %AS Expected half-width of 95% C.I.
9 df 19 df
RE2 vs. RE1. RES vs.REI 1.599o 1.48Q.O
RES vs. RE2 1.8470 1.709O
A-1-17
-------
Section No.: 1.0
Revision: 0
Date: April 28, 1995
Page: A-1-18of20
The width of the confidence intervals will be slightly narrower if the same
observations from the pilot experiment can be used for estimating measurement error
variability (i.e., there will be more df for the error component of the ANOVA). If the
underlying error variability of emissions in %AS is 5, then (assuming the 19 df situation)
we should have about a 50 percent chance of finding a difference between the baseline
spraying equipment and one of the alternatives if the true difference is 7.4 percentage
points (i.e., 1.480x5).
For resin formulation comparisons, the following ANOVA applies:
Source of Degrees of
variation freedom
RF (resins) 5
Error 15 or 18 or 19 or 22
The various choices for the error df result from which set of data one uses, as
previously discussed.
The expected half-width of confidence intervals on the differences of resin
formulations are as follows:
Difference in mean %AS Expected half-width of 95% C.I.
15 df 19 df
RF2, RF3, RF4, RF5,
or RF6 vs. RF1 1.5070 1.4800
All comparisons not
involving RF1 1.7400 1.7090-
It should be noted that any comparison involving RF4 is a comparison not only of
resin formulations but also of application methods. The width of the confidence
intervals will be slightly narrower if some of the observations from the pilot experiment
can be included in the data analysis. If the underlying error variability of emissions in
%AS is 5, then (assuming the 19 df situation) we should have about a 50 percent
chance of finding a difference between the RF1 and one of the alternative resin .
formulations if the true difference is 7.4 percentage points (i.e., 1.480 x 5).
1.7.2.4 Combining the Gelcoat and Resin Experiments
The gelcoat and resin experiments can be run effectively if trials of the two
experiments are interspersed. In particular, it is desirable if each gelcoat trial is
followed by at least one resin trial. To accomplish this and to randomize the ordering of
A-1-18
-------
Section No.: 1.0
Revision: 0
Date: April 28,1995
Page: A-1-19of20
trials within each separate experiment, a random ordering of the 27 trials in the resin
experiment was first determined. Then a list of 27 "pseudo-trials" was created for the
gelcoat experiment; the list contained the actual 18 gelcoat trials plus 9 dummy trials.
These 27 "trials" were also independently randomly ordered and then the two randomly
ordered lists were merged to produce the final composite set of trials, which is shown in
Table 1-6.
A-1-19
-------
Section No,:
Revision: 0
Date:
Page:
1.0
April 28,1995
A-1-20 of 20
Table 1 -6. Randomly Ordered List of Trials for the
Main Experiment
Trial
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
GF
1
1
t
1
.
2
,
^
1
t
1
t
2
f
1
t
1
1
B
2
f
2
.
t
.
2
<
2
ซ
1
,
2
,
2
2
GE
3
2
,
1
,
1
*
*
1
1
^
3
3
s
2
3
.
2.
2
t
4
2
.
1
I
2
1
3
3
RF
5
,
6
5
1
1
1
1
3
1
2
4
2
.
6
.
6
2
.
1
1
1
4
1
ซ
3
3
5
.
1
4
1
1
RE
1
f
1
.
1
2
3
,
3
1
1
1
t
1
4
,
1
f
1
1
1
1
1
3
4
2
1
1
1
1
t
4
*
1
,
2
Key:
GF1
GF2
GE1
GE2
GE3
RF1:
RF2 =
RF3 =
RF4 =
RF5 =
RF6 =
RE1
RE2
RE3
RE4
a regular (ISO/NPGฎ) gelcoat containing only styrene
monomer (baseline condition)
a low VOC styrene-suppressed (ISO/NPGฎ) gelcoat
containing only styrene monomer
an AAA external catalyst mixing spray gun (baseline
condition)
an HVLP internal catalyst mixing spray gun
an HVLP external catalyst mixing spray gun
a DCPD-based low-profile resin catalyzed with MEKP
(baseline condition)
a DCPD-based low-styrene resin
an ORTHO-based styrene-suppressed resin
a DCPD-based low-profile resin catalyzed with BPO
a water-emulsified resin
the same ORTHO-based styrene suppressed resin at a
higher suppressant concentration
an AAA external catalyst mixing spray gun (baseline
condition)
an internal catalyst mixing flow coater
an internal catalyst mixing pressure-fed roller
a modified AAA, external catalyst mixing spray gun for the
resin catalyzed with BPO
A-1-20
-------
Section No.: 2.0
Revision: 0
Date: April 28,1995
Page: A-2-1 of 3
2.0 PROJECT ORGANIZATION AND RESPONSIBILITIES
Figure 2-1 depicts the organizations and personnel involved in this pollution .
prevention technique evaluation test. Carlos Nunez is the EPA Project Officer for this
research project. Nancy Adams is the EPA Quality Assurance Officer.
Emery Kong, the RTI Project Leader, will coordinate the preparation and testing
activities. Emery Kong will be responsible for developing the QAPP, all data generated
under this study, all corrective action, and the overall technical quality of the evaluation
test. Andrew Clayton will assist in the experimental design and data analysis. Mark
Bahner is the RTI Testing Crew Chief. He and Keith Leese will conduct the actual
testing. Craig Whitaker and Robert Wright will operate the THC analyzer and measure
the exhaust air flow rate. Mark Bahner will also be responsible for the data reduction
activities, analyzing the test results, and preparing the final report. Cynthia Salmons, the
RTI QA Manager, William Yeager, and Shrikant Kulkami will provide assistance in the
QAPP preparation and ensure that the data collected adhere to the quality assurance
requirements specified herein. They are independent of the technical activities on this
project.
Reichhold will provide technical assistance, resin materials, and catalysts for the
testing. Mark Callicutt, Reichhold's Technical Service Supervisor, will be the primary
contact at Reichhold. Federico Linares, Manager of Physical Testing and Application,
will provide facility support, and Lorenzo Esposito, Fleichhold's Senior Technical Service
Representative, will provide technical support to the test. Reichhold's technical service
personnel will analyze some specific properties of the gelcoat and resin formulations in
their laboratory.
Mark Hollenbech of Cook Composites and Polymers (CCP) will provide both the
ISO/NPGฎ-based regular gelcoat and low VOC gelcoat for testing. Casey Herbert of
W.E.T. Inc., will provide the water emulsified resin for testing. Tom Hedger of Magnum
Industries wili provide gelcoat and resin application equipment. Charles Stard from
Magnum Industries will be onsite during the test period to operate the equipment.
This testing will not have any off-site sample analysis. All measurements and
process data will be collected onsite or analyzed in Reichhold Chemicals' laboratory
during the test. Emery Kong will ensure that all testing procedures and quality
assurance requirements are correctly followed. Emery Kong will communicate with
other contributing organizations that will provide materials, equipment, and support for
the test. He will communicate with Reichhold personnel for necessary facility,
laboratory, and technical support for the test. If Emery Kong is absent, Mark Bahner
will assume all coordination responsibilities.
A-2-1
-------
Section No.: 2.0
Revision: 0
Date: April 28,1995
Page: A-2-2 of 3
As the Testing Crew Chief, Mark Bahner will coordinate all activities and on-site
personnel during the testing and report to the Project Leader. He will communicate with
the application equipment operator and the THC analyzer operator to address any
concerns or problems that they might encounter. Decisions to stop or continue testing
will be made by the Project Leader.
A-2-2
-------
Section No.:
Revision:
Date:
Page:
2.0
0
April 28, 1995
2-3 of 3
EPA Quality
Assurance Officer
Nancy Adams
919-541-5510
RTI Quality
Assurance Officer
Cynthia Salmons
919-541-6948
William Yeager
919-541-6287
Shrikant Kulkarni
919-541-5919
Provide QA support
and review QAPP
RTI Experimental
Design and
Data Analysis
Andrew Clayton
919-541-6392
r
Operates application
equipment on-site
Magnum Industries
Charles Stard
919-361-2095
EPA/AEERL
Project Officer
Carlos Nunez
919-541-1156
RTI Project Leader
Emery Kong
919-541-5964
Responsible for
overall planning,
coordination, and
report preparation
Testing Crew
Mark Bahner (Chief)
919-541-6016
Keith Leese
919-541-8020
Responsible for data
collection, analysis,
and report preparation
RTI Emission
Measurement Team
Craig Whitaker
919-541-5988
Robert Wright
919-541-6263
Operate the THC
analyzer and measure
air flow rate
Composites Fabricators
Association
Bob Lacovara
215-721-9246
Provides technical
review and suggestions
Reichhold
Chemicals, Inc.
Mark Callicutt
919-990-8013
Federico Linares
919-990-8083
Lorenzo Esposito
919-990-8022
Provide resins, catalysts,
spray booth, and
laboratory analysis
Cook Composites
and Polymers
Mark Hollenbeck
816-391-6000
Provides gel coat
formulations
Magnum Industries
Tom Meager
813-573-2955
Provides application
equipment and operator
W.E.T. Inc.
Casey Herbert
803-556-2506
Provides water emulsifier
resin
Figure 2-1. Project organization chart.
A-2-3
-------
Section No.:
Revision: 0
Date:
Page:
3.0
April 28, 1995
A-3-1 of 5
3.0 DATA QUALITY INDICATOR GOALS FOR CRITICAL MEASUREMENTS
This section presents the qualitative and quantitative descriptors that are used
to interpret the degree of acceptability of the test data. The principal data quality
indicators are precision, accuracy, detection limits, and completeness.
The primary objectives for this testing are to quantify the emissions from
selected pollution prevention techniques and to compare the emissions of these
techniques with the baseline conditions. Styrene emissions for baseline conditions and
each of the pollution prevention techniques will be expressed as percent of available
styrene in the gelcoat and resin formulations. The data quality indicator (DQI) goals
specified in this section are based on results obtained from previous tests. If these DQI
goals are attained in this testing, sufficient valid data of known quality will be collected
to evaluate different pollution prevention techniques.
The results of a recent test conducted by RTI in early March 1995 are
summarized in Table 3.1. This table presents the styrene emission data (expressed as
%AS) and preliminary statistical analysis of the results measured by mass balance and
THC emission measurement methods. In this table, a denotes the standard deviation
associated with measurement variability of emissions in %AS.
Table 3-1. Summary of Styrene Emission Data from a Previous RTI Testing
Formulation
Gelcoat
Resin
Resin
Equipment
Spray-up
Hand lay-
up
Chop and
spray-up
Method
MBa
THC
MB
THC
MBa
THC
No. of test
runs
3
3
3
3
4
4
Mean (% of
AS)
65.0
66.7
20.3
19.7
24.5
20.8"
o
(standard
deviation
in %AS)
2.6
4.6
1.2
0.7
1.3
5.3b
Relative
standard
deviation
(%)
4
7
6
4
5
26b
MB=mass balance method.
THC=THC emission measurement method.
8 Mass balance standard deviations for gelcoat and resin spray-up include corrected values.
b Includes one chop spray test run that had a known large error.
A-3-1
-------
Section No.:
Revision: 0
Date:
Page:
3.0
April 28,1995
A-3-2 of 5
3.1 Objectives for Quantitative Data Quality Indicators
Quantitative DQIs are typically defined in terms of measurement precision,
accuracy, detection limits, and completeness. The DQI objectives for the critical
measurements are summarized in Table 3-2. These DQI objectives are based on the
results of a recent RTI test and data in Table 3-1. Precision is typically determined from
duplicate measurements and is usually expressed as percent difference or standard
deviation in either absolute or relative terms. Accuracy is the degree of agreement
between an observed value and an accepted reference value. For the THC analyzer
and balance, the accuracy will be determined from standard reference styrene gases
and weights, respectively. Detection limits are the lowest concentration or amount of
weight that can be determined to be different from zero. Completeness is defined as
the ratio of the amount of valid data obtained compared to the planned amount.
Procedures for determining these quantitative DQIs are discussed in more detail below.
Table 3-2. Objectives for Quantitative Data Quality Indicators
Measurement
(unit)
THC cone.
(ppm)
Exhaust air
flow rate
(cfm)
Mass balance
(9)
Styrene
content (%)
Method
EPA Method
25A
EPA Method 2
Floor-type, high-
precision
balance
Reichhold
standard test
method No. 18-
001 and a high-
precision
analytical
balance
Precision
(RPD or
RSD)
10%
10%
5%
0.5 %
Accuracy
(%)
47-5
47-5
+/-1
+/-1
Detection limit
1 ppm
NA
19
0.0001 g
Completeness
(%)
90
90
90
90
RPD = Relative percent difference as calculated from duplicate measurements.
RSD = Relative standard deviation as calculated from three or more replicates
3.1.1 Precision
Precision objectives for all the listed measurements are presented as relative
percent difference (RPD) of duplicate measurements or as relative standard deviation
(RSD) of three or more replicates. The number of replicates for each test are shown in
Tables 1-3 and 1-4. Precision for THC measurement, air flow rate measurement, and
mass balance measurement is shown in Table 3-2. The styrene content for each
A-3-2
-------
Section No.: 3.0
Revision: 0
Date: . April 28,1995
Page: A-3-3 of 5
gelcoat or resin formulation is determination by duplicate samples having an RPD of +/-
0.5 percent. If duplicate samples are not within +/- 0.5 percent, the entire test to
determine styrene content will be repeated.
3.1.2 Accuracy
The accuracy of the THC analyzer will be determined following the calibration
error test procedures in Section 6.4 of EPA Method 25A. Immediately prior to the test
series, a zero gas and high-level calibration gas are introduced at the calibration valve
assemble. Then the analyzer output is adjusted to the appropriate levels, if necessary.
The predicted response for the low-level and mid-level gases based on a linear .
response line between the zero and high-level responses is calculated. Then low-level
and mid-level calibration gases are introduced successively to the measurement
system. The analyzer responses for low-level and mid-level calibration gases are
recorded and the differences between the predicted responses are determined. These
differences must be less than 5 percent of the respective calibration gas values. If not,
the measurement system is not acceptable and must be replaced or repaired prior to
testing.
The accuracy of the floor-type, high-precision balance (150,000 g capacity with
1 g readability) will be determined using standard reference weights. The balance has
an internal calibration weight that is used to calibrate the balance initially. Then the
accuracy of the floor-type balance will be checked by placing reference standard
weights from 1 g up to 1 kg with and without the cart and empty mold. The accuracy of
the analytical balance will be checked using standard weights suitable for its capacity
range. The accuracy of the floor-type balance and the analytical balance should be
less than 1 percent of the respective standard weights.
3.1.3 Detection Limit
The detection limit is defined as the lowest concentration or amount of the target
analyte that can be determined to be different from zero from a single measurement at
a stated level of probability. The detection limits for the instruments used for critical
measurements are presented in Table 3-2. These instruments include the THC
analyzer for emission measurement, the floor-type high-precision balance for material
balance determination, and the analytical balance for styrene content determination.
The detection limits specified in Table 3-2 provide adequate quantification for the
measurements of interest.
3.1.4 Completeness
Completeness is defined as the amount of valid data obtained compared to the
planned amount. The completeness objective of 90 percent was selected based on the
results of a recent RTI test that compared emission measurement and mass balance
A-3-3
-------
Section No.: 3.0
Revision: 0
Date: April 28,1995
Page: A-3-4of5
methods to quantify styrene emissions from open contact molding processes. A
completeness level of 90 percent ensures that sufficient valid data of known quality are
collected to evaluate different pollution prevention techniques for styrene emission
reduction. The results from emission measurements and mass balance methods will
complement each other, because each method has good precision. In the event that
both test methods failed for more than 10 percent of the planned test runs and the
completeness level of 90 percent is not met, then those invalid test runs will be
repeated.
3.2 Objectives for Qualitative Data Quality Indicators
Qualitative DQIs are typically defined in terms of representativeness and
comparability. The representativeness is the degree to which the collected data
accurately and precisely represent the population or the actual operations. The
comparability is the degree or confidence to which one data set can be compared to
another. These qualitative DQIs are described in more detail in the following sections.
3.2.1 Representativeness
The representativeness of this testing is best determined by the materials and
equipment used, the environmental conditions of the testing, and the operator
techniques. The gelcoat and resin materials, except the water-emulsified resin, and the
application selected for testing are currently available to, and used by, the industry.
Water-emulsified resin is currently used by the industry to a very limited extent;
however, this testing will determine whether this resin can significantly reduce styrene
emissions. The environmental conditions (i.e., air flow velocity and ambient
temperature) for the testing will be controlled so that they are representative of typical
conditions in an operating facility. The person who will operate the equipment is an
experienced technical support person from Magnum Industries. His experience will
ensure that the operating procedures are consistent between test runs and
representative of industry practice.
The emissions in the spray booth will be representative of actual industry
practice. As is discussed.in more detail in Section 4.1, EPA Method 204 for a total
enclosure is to be followed to ensure that 100 percent of the emissions is captured.
3.2.2 Comparability
The baseline emissions of this testing will be compared to the results RT1
collected from an early March testing at Dow Chemical that used similar spraying
techniques and the same gelcoat and resin materials. Another styrene emission study
is planned by Composites Fabricators Association (CFA) to be conducted at Dow
Chemicals in a time frame similar to this testing. The same regular gelcoat and low-
profile resin materials will be applied by similar spraying techniques in the CFA testing.
A-3-4
-------
Section No.: 3.0
Revision: 0
Date: April 28, 1995.
Page: A-3-5 of 5
The measurement procedures and methods will be similar for the CFA tests and this
testing; therefore, results from both testings should be comparable.
A-3-5
-------
Section No.: 4.0
Revision: 0
Date: April 28,1995
Page: A-4-1 of 9
4.0 SAMPLING PROCEDURES
4.1 Total Enclosure and Capture Efficiency Test
The open contact molding process will be conducted at a spray booth in a total
enclosure setup in Reichhold Chemicals' physical testing laboratory. The spray booth is
described in Section 1.6.1. The enclosure will be tested prior to the pilot and main
experiments to ensure that the enclosure meets EPA's total enclosure guidelines (EPA
Method 204), so that the emissions from the test can be assumed 100 percent
captured. The capture efficiency of the enclosure will be examined by (1) evaporating a
known quantity of styrene (determined by a high- precision scale) and measuring the
total styrene emissions at the exhaust stack and (2) making sure that all air flows at
natural draft openings flow inward and the velocity is at least 200 fpm.
4.2 Sampling Location and Duration of Test Run
Emissions from the open contact molding processes will be measured at the
exhaust stack of the enclosure. A test run will start when gelcoat or resin material is
applied to a mold and finish when the gelcoat or resin material is cured (as determined
by a negligible rate of emissions). An earlier RTI test at Dow Chemical Company
indicated that a test run may last from 1 -1/2 to 2 hours. At the end of a test run, the
mold will be removed from the enclosure and the enclosure will be flushed with fresh
makeup air for the next test run. The THC analyzer will measure the background
concentration before each test run. Any other VOC emission sources will be eliminated
from the immediate area to minimize the background VOC concentration.
Table 4-1 shows the locations and frequencies of the measurements for the
test.
4.3 Testing Procedures
Procedures for individual test runs are outlined in this section. More detailed
descriptions of the procedures are presented in the following sections.
A Before the Test Run
1. Measure and record the gelcoat/resin properties
2. Calibrate the THC analyzer
3. Measure the baseline concentration in the spray booth with the THC analyzer
4. Measure and record initial air temperature, velocity head, and relative humidity
A-4-1
-------
Section No.:
Revision: 0
Date: .
Page:
4.0
April 28,1995
A-4-2 Of 9
Table 4-1. Summary of Measurement Location and Frequency
Measurement
THC concentration
Exhaust air flow rate and velocity head (Ap)
Humidity
Mass balance calculations
1) gelcoat/resin materials used
2) gelcoat/resin applied on mold
3) cured material on mold
4) cured material on other ground cover
Types of gelcoat and resin materials
Styrene content
Gelcoat/resin properties
1) gel time
2) total time to peak
3) peak exotherm
Linear air velocity in the spray booth
Ambient temperature
Equipment type
Equipment setup
1) pump pressure
2) air pressure
3) spray tip size
4) spray tip angle
5) catalyst ratio setting
6) equipment delivery rate
Gelcoat/Resin data
1) catalyst ratio
2) gelcoat/laminate thickness
3) glass/resin ratio
Classification
Critical
Critical
Noncritical
Critical
Noncritical
Critical
Noncritical
Noncritical
Noncritical
Nonqritical
Noncritical
Noncritical
Location
Exhaust stack
Exhaust stack
Spray booth
Spray booth
Manufacturer
Laboratory
Laboratory
Spray booth
Spray booth
Equipment
vendor
Spray booth
Spray booth
Frequency
Continuous
Flow rate (weekly),
Ap (every 15
minutes)
Before each test run
Each test run
Each gelcoat and
resin formulation
Each gelcoat and
resin formulation
Each gelcoat and
resin formulation
Every week
Every test run
Each equipment
Each equipment
Each test run
A-4-2
-------
Section No.: 4.0
Revision: 0
Date: April 28,1995
Page: A-4-3 of 9
5. Calibrate the balance(s) and check the accuracy with standard weights
6. Prepare and adjust the application equipment
7. Measure equipment delivery rate at the setup conditions in a remote location
8. Record the equipment setup conditions
9. Record the initial weights for mold, gelcoat/resin, catalyst, fiberglass
reinforcement, protective skirt for cart, and ground cover
B. During the Test Run
1. Initiate the timer as soon as the application starts
2. Record styrene emissions with the THC analyzer continuously
3. Record velocity head (*p) and air temperature every 15 minutes
4. Record the time and the weight of gelcoat/resin reading as soon as the
application is completed (the application equipment can be removed for
cleaning)
5, Measure and record wet gelcoat/laminate thickness at the center of each mold
surface
6, Remove the protective skirt from the cart and record the weight of the wet mold
at the end of application
7. Reattach the protective skirt to the cart
8. Stop the test run when the concentration returns to the baseline concentration or
when incremental emissions are negligible
9. Record the time at the end of the test run
C. After the Test Run
1. Record the final weights for mold, catalyst, fiberglass reinforcement, protective
skirt for cart, and ground cover immediately
2. Conduct the baseline drift determination for the THC analyzer
3. Remove the mold from the enclosure and flush the enclosure with fresh makeup
air until the baseline concentration stabilizes
4.4 Emission Measurement
Styrene emissions will be measured from the exhaust stack of the enclosure
using a total hydrocarbon analyzer following the EPA Method 25A. The THC analyzer
will provide real-time measurements of the emission concentrations. Any other VOC
emission sources will be eliminated from the enclosure, so that the total VOC emissions
measured can be assumed to be from styrene emissions. Two to three concentration
ranges (0-11 ppm, 0-110ppm, and 0-1,100 ppm) on the THC analyzer will be used for
the emission measurement. If the highest concentration range (0-1,100 ppm) is used,
the maximum concentration is not expected to exceed 300 ppm. The following styrene
standard reference gases are available to establish the calibration curve for each of the
concentration ranges; 5.1,10.7, 49.5, 81.5, and 237.1 ppm. Exceptions to EPA Method
25A are that the calibration error check for lowest concentration range (0-11 ppm) will
be done with one instead of two standard reference gases.
A-4-3
-------
Section No.: 4.0
Revision: 0
Date: April 28,1995
Page: . A-4-4 of 9
4.5 Exhaust Air Flow Rate Measurement
The exhaust flow rate will be measured using EPA Methods 1 and 2 at least
once a week. EPA Method 2 specifies the use of either a standard or S-type pitot tube
to traverse the duct. Because the particulate loading of the exhaust air stream is
expected to be low for this test series, a standard pilot tube will be used. The exhaust
flow rate will be measured at traverse points using a standard pitot tube with a
differential pressure gauge that meets the specifications described in EPA Method 2,
Section 2.2. The exhaust flow rate will be correlated to the velocity head measured at
the center point. The center point velocity head (Ap) will be monitored periodically
(every 15 minutes) to ensure that air flow rate is consistent during the test run. The
relative humidity of the air in the spray booth will be measured by a sling psychrometer.
4.6 Mass Balance Determination
Weight losses due to styrene emissions will be determined using a floor-type,
high-precision balance (Sartorius Corporation, Model F160S) that has a 150,000-g
capacity and 1 g readability. The initial and final weights of mold, gelcoat/resin
materials, catalyst, fiberglass reinforcement, protective skirt for the cart, and ground
cover will be measured by two balances. (A protective skirt will encircle the cart to
prevent contamination during the application.) Total emission quantity will be
determined from the difference of total materials used and the final weights after curing
(see data reduction in Section 6.0). From the weights of materials used and materials
applied on the mold, the transfer efficiency can be calculated for each test run.
Weight loss due to emissions will be recorded for each test run and the results
will be compared to the emission measurement to determine whether these two
methods are comparable.
4.7 Gelcoat and Resin Properties
Roughly a liter of sample will be taken from each of the gelcoat and resin
container for analysis. Before sampling, the content in the container will be thoroughly
mixed by a hand-held mixer for 2 minutes. The content will be scooped out to a
nonreactive container and delivered to the laboratory. The type of material, lot/batch
number, and container number will be recorded on the sample container. This
information will be recorded on the data sheet in Section 5.0. Analytical procedures
used to determine gelcoat and resin properties are presented in Section 5.0. These
gelcoat and resin properties will be documented in the report to show the reader what
kind of gelcoat materials were examined in the testing.
A-4-4
-------
Section No.:
Revision: 0
Date:
Page:
4.0
April 28,1995
A-4-5 of 9
4.8 Environmental Conditions
Environmental conditions include linear air velocity and the ambient temperature
in the spray booth. These conditions will be maintained at constant levels when
feasible. The linear air velocity in the spray booth will be maintained at the level
selected from the pilot testing. The temperature in the spray booth is affected by the
temperature of the makeup air, which is heated or cooled according to the difference in
indoor and outdoor temperatures. The actual temperature will be recorded for every
test run on the data sheet shown at the end of Section 4.
4.8.1 Linear Air Velocity in Spray Booth
Linear air velocity in the spray booth will be measured with a hot wire
anemometer at various traverse points in the spray booth. The spray booth will be
divided into three sections (i.e., at front edge, in the middle, and at filter face), two
layers (in the middle of top and bottom filter banks), and four divisions from left to right.
These (3x2x4) volumetric traverse points will provide 24 readings to characterize the air
flow pattern and velocities in the spray booth. Linear air velocities will be verified once
every week or whenever the physical setup of the enclosure is changed. The air flow
pattern and velocities will be recorded in Table 4-2.
Table 4-2. Air Flow Pattern and Velocity in the Spray Booth
Filter bank (top)
Filter bank
(bottom)
Middle (top)
Middle (bottom)
Front edge (top)
Front edge
Jpottom)
1/5 width from left
wall
Dir.
Vel.
2/5 width from
left wall
Dir.
Vel.
3/5 width from
left wall
Dir.
Vel.
4/5 width from left
wall
Dir.
Vel.
Dir. = Air flow direction expressed by an arrow head.
Vel.= air flow velocity in fpm.
A-4-5
-------
Section No.: 4.0
Revision: 0
Date: April 28,1995
Page: A-4-6 of 9
4.8.2 Ambient Temperature
Ambient temperature in the exhaust air and in the spray booth will be monitored
with thermocouples. The temperature will be recorded every 15 minutes during the test
run.
4.8.3 Relative Humidity
Relative humidity in the spray booth will be measured before each test run by a
sling psychrometer or a relative humidity detector. The relative humidity reading will be
recorded on the data sheet shown later.
4.9 Equipment Type and Setup
The equipment type and setup will be documented in the report to show the
reader what kinds of equipment are examined in the testing.
4.9.1 Equipment Type
The types of equipment to be examined are described in Section 1.4.
4.9.2 Equipment Setup
The equipment setup affects the operation of the equipment and the emission
generated. These conditions include pump pressure, air pressure, spray tip size, spray
tip angle, catalyst ratio, and equipment delivery rate. These set-up conditions will be
adjusted according to the vendors' recommendations by an experienced operator so
that the equipment will be operated under their optimum conditions. The setup will be
recorded for each equipment in each test run. The same setup will be used for the
same equipment and material in the replicate test runs.
The equipment delivery rate will be determined during the standard calibration of
the equipment under the same conditions of the testing. The standard calibration
procedures consist of spraying the gelcoat or resin materials into a plastic bag for 30
seconds and weighing the amount of output materials. The weight is then multiplied by
2 to convert to a flow rate in pounds per minute.
A-4-6
-------
Section No.: 4.0
Revision: 0
Date: April 28,1995
Page: A-4-7 of 9
4.10 Operating Parameters for Parts
The operating parameters for parts include catalyst ratio, gelcoat/laminate
thickness, and glass/resin ratio. These parameters will be measured and documented
in the report.
4.10.1 Catalyst Ratio
The actual catalyst ratio will follow the catalyst ratio suggested by the gelcoat
and resin manufacturers.
4.10.2 Gelcoat/Resin Thickness
In the gelcoating and spray-up laminating tests, the operator will use a gelcoat
and chop mil gauge during the application to determine the thickness. The operator will
check and build the thickness of the gelcoat or laminate at various locations on the
mold surface to the specified thickness. The wet gelcoat thickness is expected to be 18
to 24 mils and is to be achieved by multiple passes of spraying. The spray-up laminate
thickness is expected to be 80 to 100 mils and is to be achieved by two passes of 40- to
50-mil thick laminate. Each pass of the spray-up laminate is equivalent to 1.5 oz/ft2 of
glass.
In the flow coating and pressure-fed rolling laminate tests, the primary thickness
control will be the thickness of the chopped strand mat reinforcement. Two ply of 1.5-
oz/ft2 mat will be used to build the.laminate so that the thickness of the laminates
fabricated by the flow coater and pressure-fed roller will be similar to that of the '
laminates fabricated by spraying equipment.
4.10.3 Glass/Resin Ratio
The glass/resin ratio will be determined from the weights of glass roving or
chopped strand mat and the amount of resin used to build the laminate. The
glass/resin ratio will be documented for resin laminate only.
The data recording sheet for the measurements described in Section 4.0 is as
follows:
A-4-7
-------
Section No.: 4.0
Revision: 0
Date: April 28,1995
Page: A-4-8of9
Table 4-3. Data Recording Sheet for the Pollution Prevention Technique Evaluation
Test
Date/Time span:
Test run No.:
Formulation type (gelcoat/resin):
Equipment type:
Recorded by:
1. Temporary Total Enclosure System
a. Air flow velocities at natural draft openings (fpm)
b. Air flow velocity over application area (fpm)
2. Air Temperature, Exhaust Air Flow Rate, and Relative Humidity
a. Ambient air temperature (ฐF)
0' 15'(min) 30' 45' 60'
75'. 90' 105' 120' . 135'
b. Velocity head (Ap) on the pitot tube (inches)
0' 15' 30' 45' 60'
75' 90' 105' 120' 135'
c. Exhaust air flow rate (cfm)
d. Relative humidity (%)
3. Emission Measurement
a. Readings of THC analyzer calibration check
Range 1 Low Med High
Range 2 Low Med High
Range 3 Low Med High
b. Reading of zero gas before the test run
c. Reading of zero gas after the test run
d. Zero and span potentiometer setting before calibration
e. Zero and span potentiometer setting after calibration
4. Mass Balance Calculation
Readings of balance accuracy check with standard weights
1g 5g 10g 50g 100g
200g 500g 1,000g
a. Weight of empty mold, g
b. Weight of empty ground cover, skirt for cart, gloves, and tapes, g
(cont.)
A-4-8
-------
Section No.: 4.0
Revision: 0
Date: April 28, 1995
Page: A-4-9 of 9
Table 4-3. Data Recording Sheet for the Pollution Prevention Technique Evaluation
Test (continued)
c. Initial weight of gelcoat/resin used, g (container No: )
Initial reading, g Final reading, g
d. Weight of catalyst used, g
Initial reading, g Final reading, g
e. Weight of fiberglass reinforcement used, g
Initial reading, g Final reading, g
f. Weight of wet mold at the end of application, g
g. Weight of mold with cured part, g
h. Weight of ground cover, skirt for cart, gloves, and tapes with cured
overspray, g
5. Open Contact Molding Processes
a. Type of spray gun
b. Spray gun brand name/model No.
c. Pump brand name/model No.
d. Pump ratio
e. Air supply pressure, psi
f. Pressure at pump assembly, psi
g. Spray tip number/size
h. Catalyst ratio setting
I. Delivery rate, gpm
j. Gelcoat/laminate thickness, mil .
top front left right back
k. Glass/resin ratio, %
A-4-9
-------
Section No.: 5.0
Revision: 0
Date: . April 28, 1995
Page: A-5-1 of 2
5.0 ANALYTICAL PROCEDURES
This section describes the analytical procedures that will be used to analyze the
gelcoat and resin materials tested. Gelcoat and resin properties include types of
materials, styrene contents, and curing characteristics of the materials. Laboratory
procedures commonly used by the industry will be followed to measure these
properties. The other measurements, such as emission and environmental
measurements, are described in Section 4 in conjunction with sampling procedures.
5.1 Types of Gelcoat and Resin Materials
The types of gelcoat and resin materials to be examined are listed in Section 1.4.
5.2 Styrene Content
Styrene content for each of the gelcoat and resin materials will be measured
using Reichhold Standard Test Method No. 18-001 {Appendix A) in the Reichhold
laboratory. The gelcoat and resin manufacturers will be asked to provide materials that
contain only styrene as the monomer. The Reichhold test method determines the
nonvolatile content of the materials, and the remainder is considered the styrene
content. The styrene content of the water-emulsified resin will be determined for the
base resin material before water is added.
5.3 Gel time, Time to Peak, and Peak Exotherm Characteristics of Polyester
Resins
The gel time, time to peak, and peak exotherm characteristics of polyester resins
will be measured following the Reichhold Standard Test Method No. 18-050 (in
Appendix A) for gelcoats and resins catalyzed with MEKP. The Reichhold. Standard
Test Method No. 18-051 (in Appendix A) will be used for the resin catalyzed with BPO.
The catalyst ratio suggested by the gelcoat and resin manufacturers will be used in the
curing characteristics determination and in the actual testing.
The data recording sheet for the measurements described in Section 5.0 is
presented on the next page.
A-5-1
-------
Section No.: 5.0
Revision: 0
Date: April 28,1995
Page: A-5-2 of 2
Table 5-1. Data Recording Sheet for Gelcoat and Resin Properties
Date/Time:
Test run No.:
Formulation type:(gelcoat/resin)
Container No.:
Recorded by:
a. Gelcoat/Resin type
b. Manufacturer, Lot/Batch No.
c. Styrene content (%)
d. Weight percent of water (water-emulsified resin only)
e. Catalyst type
f. Catalyst ratio (wt. %)
g. Geltime (minute/second)
h. Time to peak (minute/second)
I. Peak exotherm (ฐF)
A-5-2
-------
Section No.: 6.0
Revision: 0
Date: April 28, 1995
Page: A-6-1 of 4
6.0 DATA REDUCTION, VALIDATION, AND REPORTING
This section describes how data will be reduced, validated, and reported. The
data handling, reduction, validation, and reporting procedures are shown in Figure 6-1.
6.1 Data Reduction
After daily sampling is completed, the recording sheets containing sampling and
analytical results will be collected, verified, and analyzed by the Testing Crew Chief. If
THC analyzer readings are not available on a recorder, they will be entered into a
computer spreadsheet. The styrene concentrations will be calculated from the THC
analyzer readings and the styrene standard calibration curve for each corresponding
concentration range used. An average styrene emission concentration will be
calculated for the duration of the test run.
An average exhaust air flow rate will be calculated from the average velocity
head monitored during the test run using the equations in EPA Method 2. Styrene
emission quantity (Em) for each test run will be calculated by the following equation:
Em, Ib = 2.6x10'9 x Q x MW x C x T
Em, g = 1.18x10'6 x Q x MW x C x T
where
2.6x10'9, 1.18x10"6 = conversion factors to standard conditions (68 ฐF.and 29.92
inches mercury) in English and metric units, respectively
Q = average exhaust air flow rate (dry standard cubic feet per minute)
MW = molecular weight of styrene (104)
C = average styrene emission concentration (ppmv dry)
T = duration of test run (minute).
The weight loss due to emissions (Wloss) and the transfer efficiency (TReff) will
be calculated by the following equations for each test run:
Wloss, g = (Wa+Wb+Wc+Wd+We) - (Wg+Wh)
TReff, % = (Wf-Wa) / (Wc+Wd+We) x 100%
where
Wa = Weight of empty mold (g)
Wb = Weight of empty ground cover, skirt for the cart, gloves, tapes (g)
A-6-1
-------
Section No.:
Revision:
Date:
Page:
6.0
0
April 28. 1995
6-2 of 4
Testing Crew Chief collects
data recording sheets from
testing crew members and
verifies raw data
Testing Crew Chief
performs data reduction
Project Leader reviews
raw and reduced data
for correctness
Project Leader and Testing
Crew Chief validate emission
measurement and weight
loss data for each test run
if A < 30%
Forward the reduced
data to statistician
for analysis
Project Leader and Testing
Crew Chief summarize and
interpret the data and prepare
a final report which also includes
a QA/QC evaluation section
If DQI for test
completeness is not met.
The test team repeats
the flagged test run.
if A > 30%
Flag the test run. Project Leader
and Testing Crew Chief identify
and resolve the problem.
If DQI for test
completeness is met.
Note: A is the difference between the emission measurement and weight loss results for each test run.
Figure 6-1. The data handling, reduction, validation, and reporting procedures.
A-6-2
-------
Section No.: 6.0
Revision: 0
Date: April 28.1995
Page: A-6-3 of 4
We = Weight of gelcoat/resin used (g)
Wd = Weight of catalyst used (g)
We = Weight of fiberglass reinforcement used (g)
Wf = Weight of wet mold at the end of application (without the skirt on the
cart) (g)
Wg = Weight of cured part with mold (g)
Wh = Weight of ground cover, skirt for the cart, gloves, and tapes with
cured overspray (g).
The emission quantity and weight loss data will be expressed as emission factor
(Ef) in percent available styrene according to the following formula.
Ef, %AS = (Em or Wloss, g) / (We, g x styrene content) x 100%
where styrene content is the weight fraction of styrene in the gelcoat or resin
formulation.
The same emission quantity and weight loss data can also be expressed as
emissions per unit mold area (Ea) in gram per square meter according to the following
formula:
Ea, g/m2 = (Em or Wloss, g) / (surface area of the mold, m2).
The field data will be reduced by the Testing Crew Chief as they are generated.
At the end of the pilot experiment, these field data and reduced data will be analyzed to
establish "standard" conditions under which the main experiment will be conducted.
The standard conditions will be selected that best represent the actual operating
conditions in the industry.
At the end of the main experiment, the field data and reduced data will be
reviewed by the Project Leader for correctness. The reduced data are then passed to
an RTI statistician for the analysis of variance as outlined in Section 1.7. The reduced
data and the results of the statistical analysis will be included in the final report.
6.2 Data Validation
A key element in assessing data quality and validity is the comparison of
emission measurement and the weight loss data for the pure styrene emissions test
(described in Section 4.1) and for each test run. The Testing Crew Chief will perform
the basic review and audit of the field data sheet for completeness and accuracy. The
Testing Crew Chief will also compare the reduced emission measurement with the
weight loss data to ensure that these two results are comparable within ฑ 30 percent. If
these two results are not within ฑ 30 percent, the test run will be flagged. The test team
will investigate the possible cause of the difference and correct the problem
A-6-3
-------
Section No.: 6.0
Revision: 0
Date: . April 28,1995
Page: A-6-4 of 4
immediately. At the end of the main experiment, if the data quality indicator for
completeness is not met for emission measurement and weight loss determination,
then these flagged test runs will be repeated.
6.3 Data Reporting
An RTI statistician will analyze reduced data and prepare the results of the
statistical analysis. The Project Leader and the Test Crew Chief will be responsible for
data summary, interpretation, and final report preparation. The report will present the
effects of airflow velocity, spraying methods, gelcpat formulations, gelcoat application
equipment, resin formulations, and resin application equipment on styrene emissions.
The emissions from different formulations or equipment will be compared with the
emissions from the baseline conditions. The ability to differentiate any emission
reduction potential will be determined by the variability of the measurement method and
the actual difference between the two compared conditions. The final report will also
contain a Quality Assurance/Quality Control (QA/QC) evaluation report to document the
QA/QC activities and results. The final report will include a statement indicating
whether the data quality objectives were met or not. If the QA objectives were not met,
an explanation of the impact of not meeting the project's QA objectives will be included.
A-6-4
-------
Section No.: 7.0
Revision: 0
Date: April 28.1995
Page: A-7-1 of 1
7.0 INTERNAL PERFORMANCE AND SYSTEM AUDITS
7.1 Technical Systems Audits
A technical systems audit (ISA) is a qualitative on-site evaluation. A ISA
evaluates compliance with the QAPP and any standard operating procedures (SOPs).
One internal ISA is planned for this project. The TSA will be conducted by the project
QA Manager, Ms. Salmons, or her designee and will cover sampling, analysis, and data
handling steps. A written report will be prepared, summarizing the results of the audit
and noting any deviations from the QAPP, within one month of completion of the audit.
In addition, RTI will cooperate fully with any external audits performed by EPA.
7.2 Performance Evaluation Audits
A performance evaluation audit (PEA) is a quantitative evaluation of a
measurement system. No internal performance audits are planned for this project. If
EPA provides performance evaluation samples, RTI will analyze them.
7.3 Audits of Data Quality
Audits of data quality (ADQs) involve assessments of the methods used to
collect, interpret, and report the information required to characterize data quality. While
no formal ADQ is planned for the project, the project QA Manager or her designee will
review the data at the end of the project, before the report is finalized. This review will
check that reduction and validation, as described in Section 6, have been performed,
and that data can be tracked from data forms and notebooks to the summary tables in
the report.
A-7-1
-------
Section No.: 8.0
Revision: 0
Date: April 28, 1995
Page: A-8-1 of 2
8.0 CALCULATION OF DATA QUALITY INDICATORS
The exhaust air flow rate, the THC analyzer readings, and weight loss data are
measured and recorded at the test site. These field data are immediately available for
data quality review. The field data will be reduced during the testing to the extent
possible to provide a means of immediately assessing the field data quality. If it is
found during the testing that accuracy, precision, method detection limit, and
completeness measurements deviate from the DQI goals indicated in Section 3, the
source of error will be identified and the problem corrected as soon as possible. A pure
styrene emission test as described in the capture efficiency test for the total enclosure
may be used to identify the source of error.
The following calculations will be used for this study.
8.1 Precision
For precision, relative standard deviation will be reported:
RSD = (s/y)x100%
where
RSD = relative standard deviation
s = standard deviation
y = mean of replicate analyses.
Standard deviation is defined as follows:
(v - \A*
s= ^(y> w
- 1
where.
y; = measured value of the with replicate
n = number of replicates.
8.2 Accuracy
For accuracy, percent recovery will be reported.
A-8-1
-------
Section No.: 8.0
Revision: 0 ' -
Date: April 28,1995
Page: A-8-2 of 2
When a standard reference material (SRM) is used:
%/? = 100% x
where
%R = percent recovery
Cm = measured concentration of SRM
Csrm = actual concentration of SRM.
8.3 Method Detection Limit
MDL is defined as follows for all measurements:
MDL = ^.
where
MDL = method detection limit
s = standard deviation of the replicate analyses
V-u-a=o.99) = students' t-value for a one-sided 99% confidence level and a
standard deviation estimate for n-1 degrees of freedom.
8.4 Completeness
Completeness is defined as follows for all measurements:
%C = 100% x V
n
where
%C = percent completeness
V = number of measurements judged valid
n = total number of measurements planned
A-8-2
-------
Section No.: 9.0
Revision: 0
Date: - April 28,1995
Page: ' A-9-1 of 1
9.0 CORRECTIVE ACTION
The need for corrective action may be identified through internal performance
and system audits (described in Section 7); whenever measurement precision,
accuracy, detection limit, or completeness deviates from the objectives established in
Section 3; or whenever the comparison of emission measurement and weight loss
calculation for individual test runs differs by more than 10 percent.
Corrective action begins with identifying the source of the problem. Potential
sources of problems include failure to adhere to prescribed test procedures or methods,
equipment malfunction, or error in data reduction. Pure styrene emission tests .
described in Section 4.1 may be used to identify problems related to the THC analyzer:
If an instrument calibration check does not meet the specified acceptance criteria,
recalibration will be required.
The Testing Crew Chief has the primary responsibility for initiating and
completing corrective action required to resolve measurement problems encountered
during the testing. The Project Leader and the Testing Crew Chief will determine
whether the corrective action has resolved the problem or not and when to resume the
testing. The Quality Assurance Manager will be notified of all corrective actions
undertaken at the test site. If necessary, the Project Leader will work with the Quality
Assurance Manager to resolve major problems such as THC analyzer malfunction and
to. obtain concurrence from the EPA Project Officer and QA Officer. All corrective
actions and the nature of problem will be documented in the QA/QC evaluation in the
final report.
A-9-1
-------
RESEARCH TRIANGLE INSTITUTE
/Ml
Center for Environmental Analysis
MEMORANDUM
DATE: May 25, 1995
TO: Carlos Nunez, EPA Project Officer
FROM: Emery J. Kong
SUBJECT: Responses to EPA's Comments on the Category HI Quality Assurance Project
Plan (QAPP)
RE: Evaluation of Pollution Prevention Techniques to Reduce Styrene Emissions
from Open Contact Molding Processes
EPA Cooperative Agreement No. CR 818419-03
RTI Project No. 96U-5171-016
Our responses to EPA's comments on the QAPP are presented in the attachment. As
I have discussed with Dr. Nancy Adams on May 5, RTI will make the following changes to
the QAPP: (1) RTI will calibrate the THC analyzer with propane standards and establish a
response factor relationship between the propane and styrene standards, and this relationship
will be used to determine the styrene concentrations monitored, and (2) RTI will use
hydrogen gas as the fuel for the Ratfisch THC analyzer (as called for in the instrument
manual) instead of a hydrogen/nitrogen mixture.
In addition to the above changes, we will separate the gelcoat experiment from the
resin experiment because we are able to resolve some technical problems in gelcoat and resin
application. This change will not affect the validity of measurements for either gelcoat
experiment or resin experiment because data analysis for each experiment is done separately.
We will perform the pilot experiment first using the regular gelcoat, then the gelcoat
experiment, and finally the resin experiment.
Please forward our responses to Dr. Nancy Adams. If you believe our responses
have adequately addressed EPA's concerns in the QAPP, please sign and date the attached
signature page and return it to me as soon as possible. Please call me at 541-5964
immediately, if you think the responses are not adequate. Thank you very much.
Attachments
cc: Cynthia Salmons, RTI
Mark Bahner, RTI
Bob Wright, RTI
Andy Clayton, RTI
Jesse Baskir, RTI
3040 Cornwallis Road Post Office Box 12194 . Research Triangle Park, North Carolina 27709-2194 USA
' Telephone 919 541-5816 Fax 919 541-7155
A-9-2 \\<\
-------
Attachment
Section Comment
1. Table 1-4 TEST RUNS FOR THE RESIN FORMULATIONS AND EQUIPMENT
TYPES
How will RE-3 (pressure-fed rollers) and RE-4 (modified AAA) be
compared when different resin formulations are being used in the testing of
these two types of equipment? RE-1, RE-2, and RE-3 are all being tested
with the same resin formulation, but RE-4 testing is proposed using a
different formulation. A discussion of the way in which the four
equipment types (REs) will be compared would add to the plan.
Response:
The resin catalyzed with benzoyl peroxide (BPO) catalyst requires the
modification of the air-assisted airless (AAA) spray gun and pump system,
so it is not possible to test the BPO system using the conventional AAA
spray gun. This means that it will not be possible within the present test
design to separate the effects of the BPO catalyst from the effects of the
modified spray gun. Therefore, the test report will acknowledge that no
separation of these effects can be made. As noted on page 1-18, any
comparison of RF-4 with other resin formulations will be confounded with
the equipment difference.
2. Table 1-5 SUMMARY OF CRITICAL AND NONCRTTICAL MEASUREMENTS
Reichhold Method No. 18-001 is proposed for the measurement of styrene
content in the formulations tested. This method (Appendix A) involves
weight loss with heating of a small sample on foil. Method 18-001 seems
to be a measurement of volatiles and not styrene. This matter is discussed
in Section 5.2; the manufacturers will be asked to "provide materials that
contain only styrene as the monomer." However, throughout the
document, styrene measurement is listed as a critical measurement, and
the proposed method is not measuring styrene. Is there any additional data
that could be supplied to verify that the proposed method really is an
accurate measure of styrene?
Response:
Reichhold Method No. 18-001 is commonly used by the industry to
determine the non-volatile (NV) content of polyester resins. Section VI of
1
A-9-3
-------
the method shows that the % monomer content is calculated using the
following formula:
% monomer = 100 - % NV.
Since the materials used in the testing will contain only styrene monomer,
we can use the method as an indirect measurement of styrene content.
3. Section 3.1.2 ACCURACY
These accuracy goals pertain to calibration error only. What about zero
and calibration drift (post test checks)?
There is a minor typo noted in the text. The word is "assembly".
Response:
The zero and mid-level calibration drift will be checked according to the
procedures outlined in Method 25A section 7.2, at the end of each run (but
not hourly during the run). The acceptable drift will be taken to be ฑ3
percent of span value. If the drift exceeds the acceptable level, the. THC
response will be checked for all calibration gases within the range(s) used
in the run. The test results will be reported using both sets of calibration
data (before and after the run). The THC analyzer will then be
recalibrated as described in Method 25A section 6.4, prior to the following
run.
4. Table 3-2 OBJECTIVES FOR QUANTITATIVE DATA QUALITY INDICATORS
The accuracy objectives should be clearly defined. Is % bias from full
scale or from a known standard? Is the THC detection limit of 1 ppm
realistic in the 0 - 1, 100 ppm range? How was the detection limit
derived? Does the equation in Section 8.3 apply for non-discrete
measurements such as those from CEMs?
Response:
RTI now plans to operate the total hydrocarbon analyzer on instrumental
ranges that correspond to 0 to 20 ppm styrene (C8) and 0 to 200 ppm
styrene (C8). These ranges are also equivalent to 0 to 53 ppm propane
(C3) and 0 to 533 ppm propane (C3). The calibration gases will be 16,
27, 45, 160, 267, and 453 ppm propane. These calibration gases will
correspond to 30 percent, 50 percent, and 85 percent of the two full-scale
ranges, as called for in EPA Method 25A.
2
A-9-4
-------
The accuracy data quality objective is being changed from +/- 5 percent to
+/- 3 ppm on the 0 to 53 ppm propane range and to +/- 31 ppm on the 0
to 533 ppm propane range. The values are the root-sum-squares of the
calibration gas accuracy, the calibration error, and the calibration drift. .
The following example is for the 0 to 53 ppm propane range:
calibration gas accuracy = 5 % of 45 ppm =2.25 ppm
calibration error = 5 % of 27 ppm =1.35 ppm
calibration drift = 3 % of 53 ppm = 1.59 ppm
root-sum-square =3.07 ppm
Please note that each of these accuracy components is measurable. The
calibration gas accuracy will be determined by comparing the specialty gas
producer's certified value with RTTs verification value for the high-level
calibration gases. The calibration error and calibration drift will be
measured according to the procedures outlined in Method 25A section 7.2,
at the end of each run.
The THC detection limit of 1 ppm (as styrene) is quite realistic for the 0-
200 ppm (as styrene) range. As shown in the attached figure, the zero gas
analysis in a March 1995 testing at Dow (using the same THC analyzer)
showed that a detection limit of less than 1 ppm was achieved in a 0-350
ppm (as styrene) range. The equation in Section 8.3 can be used if THC
readings are taken at fixed intervals (such as every 15 seconds).
5. Section 4.4 EMISSION MEASUREMENT
How many points will go into, the calibration? Based on the gases
discussed in the QAPP, only the 1,100 ppm scale can be calibrated per
Method 25A. What is the sample line made of? Is it heated? Are the
other components of the sample delivery system heated? Are any system
bias checks planned? Styrene is very reactive. What will be done to
evaluate the bias of the sample delivery system? Where will the probe be
located in the stack? Will an emissions profile be performed to assess
stratification in the duct? A probe can be built to sample representatively
across the duct (see 40 CFR Part 86.310-79).
Response:
In accordance with EPA Method 25, each of the two concentration ranges
used in the total hydrocarbon analyzer will be calibrated with zero gas and
low-level, mid-level, and high-level calibration gases. The concentrations
of these calibration gases are shown in the response to Comment No. 4.
3
A-9-5
-------
RTI anticipates using PFA Teflon for the sample line. The sampling
location will be 5 or 6 stack diameters downstream of the last bend. The
sample line will be capped at its end and a number of holes will be drilled
along its in-stack length to obtain a sample stream that is representative of
the entire stack. In general, RTI will follow the specifications for the
sample probe given in 40 CFR, Part 86.310.79 (b) except for the use of
PFA Teflon rather than stainless steel.
Method 25 states that the sample line should be heated, if necessary, to
prevent condensation in the line. The total hydrocarbon analyzer will be
sampling essentially room air at room.temperature. As such, there is no
need to heat the sample line to prevent condensation. Further, heated
sample line may cause styrene polymerization within the sample line.
During a previous gas chromatographic verification study of styrene
calibration gases, RTI checked for sample line losses using a calibration
gas containing approximately 5 ppm styrene. No differences among
unheated stainless steel, heated stainless steel, and unheated Teflon sample
lines were found. Any sample line losses would likely be less at the much
higher flow rate associated with the total hydrocarbon analyzer.
The bias of the sample delivery system will be tested, before the actual
testing, by comparing the instrumental responses to the styrene and
propane calibration gases when they are delivered through the analyzer's
calibration gas port and through the unheated sample line. If the responses
for both styrene and propane gases are the same, the sample line will not
be heated. If the response for styrene gas is different and the response for
propane gas is the same, then the need for heating the sample line will be
evaluated. If responses for both styrene and propane gases are different,
then the THC analyzer will be checked and repaired. A THC analyzer
rental unit can be arranged if the Ratfisch THC analyzer is not functional.
In addition to the direct sample delivery system bias test, an indirect
sample system bias check will be conducted in the preliminary testing using
pure styrene evaporation. The pure styrene evaporation test will identify
the bias in exhaust flow rate and THC measurements when the emission
quantity is compared to the known quantity of styrene evaporated. The
pure styrene evaporation test will release styrene at a constant rate;
therefore, it can be used to determine the emissions profile and to assess
stratification in the exhaust stack. The emission concentrations profile
could not be performed during an actual test due to the changing styrene
concentrations. Additionally, the multi-point sample probe described above
should eliminate bias due to potential stratification.
4
A-9-6
-------
6. Section 4.5 EXHAUST AIR FLOW MEASUREMENTS
Is the exhaust flow rate stable enough to only measure flows weekly? Why
was the center of the duct selected to monitor Ap? Wouldn't the average
square root of Aps across the duct be more appropriate? Has an Annubar
been considered to determine total flow with a single Ap?
Response:
Prior to the actual testing, several velocity traverse/centerline
measurements will be made on different days to determine whether the
exhaust flow rate changes over time and whether the centerline Ap
measurements accurately correlate with velocity traverse measurements. If
the preliminary exhaust flow rate measurements indicate that centerline Ap
measurements do no correlate to within ฑ 5 percent of the exhaust flow
rate as determined by traverse measurements, an Annubar probe will be
used for Ap monitoring in the actual testing.
If the exhaust flow rate is relatively stable over time, a velocity traverse
will be performed at the beginning of each week to determine the exhaust
flow rate, and the Ap (either at centerline or by Annubar) will be
monitored every IS minutes during the test run. The exhaust flow rate
during the test run will be calculated according to the following formula:
Q run = [avg (Ap run)M/(Ap weekly)0^] x Q weekly
where
Q run = exhaust flow rate during a test run (scfm)
Q weekly = exhaust flow rate determined by the weekly velocity traverse
(scfm) .
avg (Ap run)0-5 = average square root of the 15-minute Aps recorded
during the test run, either at the centerline or by Annubar
(Ap weekly)0-5 = square root of the Aps recorded during the weekly
velocity traverse
7. Section 4.7 GELCOAT AND RESIN PROPERTIES
Is the hand-held mixer electric? If so, please be sure the motor is
explosion proof since the flash point of styrene is 31ฐC.
Response:
The hand-held mixer will be powered by compressed air.
5
A-9-7
-------
8. Section 8.3 METHOD DETECTION LIMIT
Does the definition of s mean the (estimated) standard deviation of a single
observation, or the average of the n? Presumably the "MDL" refers to a
single observation but is based on a variance estimated from multiple
observations.
Response:
The method detection limit will be calculated using the equation shown
with s being determined from n measurements of zero gas taken at 15-
second intervals. RTI anticipates that n will equal 20.
9. GENERAL COMMENTS FROM ROSS LEADBETTER
Response:
The entries in Table 1-1 are approximated half widths.
The QAPP was arranged according to the format given in the QA manual;
therefore, some of the texts may be disconnected to the reader.
Nevertheless, the reader should be able to find the essential information by
referring to the table of contents.
6
A-9-8
-------
Zero gas addition during 030795AC
o
O
Elapsed time (minutes)
030785ACJCLC
-------
Appendix B
Reichhold Standard Test Methods
B-i
-------
Reichhold Standard Test Methods
Test Method Page
18-001. Determination of non-volatile content of polyester resins B-l
18-021. Determination of Brookfield viscosity and thixotropic index of polyester
resins : B-4
18-050. Determination of room temperature gel, time to peak and peak exotherm
characteristics of polyester resins B-9
18-152. Determination of static styrene emissions for compliance with SCAQMD
Rule 1162 B-12
B-ii
-------
REICHHOLD
REACTIVE
POLYMERS
DIVISION
TIEBT PROCEDURE
DETERMINATION OF NON-VOLATILE CONTENT
OF POLYESTER RESINS
METHOD; 18-001
ISSUED: 11/17/8?
REVISED: 02/27/89
PAGE: 1 OF 3
I. SCOPE
This method describes a procedure for the rapid determination of the
non-volatile content of polyester resin solutions.
II. SAFETY
Safety glasses are recommended for this procedure.
HI. EQUIPMENT
1. Aluminum foil - 6" x 12", Thomas Scientific 1086-F27-F32 or
equivalent. (Reynolds Wrap Heavy Duty)
2. Paper clips, #1 gem clips-or equivalent.
3. Cardboard sheet - 8" x 12" x 0.025".
4. Analytical balance, capable of accurately weighing to +/- 0.0001
grams.
5. Disposable syringe, 3cc capacity, B-D 15586 or equivalent.
6. Oven, forced air, of suitable capacity maintained at 120 +/- 2ฐC,
7. Thermometer, (1.0ฐC divisions).
8. Clean glass plates (2) approximately 8" x 8" x 1/4".
9. Stopwatch or timer capable of measuring to one second intervals.
The information herein is to assist customer* in determining whether our products are suitable tor their applications. Our products are Intended for
sale to industrial and commerciaJ customers. Wb request that customers inspect and test our products before use and satisfy themselves as to contents
and suitability. Ws warrant that our products wU meet our written specifications. Nothing herein shall constitute any other warranty express or implied,
inctufinQ any warranty of merchantability or fitness, nor is protection from any law or patent to be inferred. Al patent rights are reserved. The exclusive
remedy tor all proven claims is replacement of our materials and in no event shal we be HaHe for special, incidental or consequential damages.
REICHHOLD CHEMICALS, INC.
REACTIVE POLYMERS DIVISION
B-1
JACKSONVILLE, FL 32245 (904)739-2170
-------
METHOD: 18-001
ISSUED: 11/17/87
REVISED: 02/27/89
PAGE: 2 OF 3
IV. PROCEDURE
1. Fill disposable syringe with polyester resin solution to be tested,
cleaning away all excess resin from the exterior of the syringe and
removing all excess air from the interior. Immediately replace
syringe cap to insure minimal monomer loss due to evaporation. '
2. Fold a 6" x 12" aluminum foil sheet in half, shiny surface facing
in, and measure its dry weight (without resin) to the nearest 0.0001
gm. Record this weight as "A".
3. Unfold the aluminum foil sheet and rest it shiny side up on one of
the 8" x 8" x 1/4" glass plates. Place approximately 0.5 grams
(O.Scc) of polyester resin solution in the center of either of the
6" x 6" halves of the aluminum foil sheet. Replace the syringe cap.
4. Carefully fold foil sheet halves together and gently place second
glass plate over the folded foil sheet. Press carefully to ensure
even distribution of the resin sample into a thin film without
exuding from the edges of the aluminum foil.
5. Quickly reweigh the aluminum foil containing the resin sample, to
the nearest 0.0001 gram. Record this weight as "B".
6. Unfold the foil sheet and place (resin side up) onto the cardboard
sheet. Use the paper clips to carefully secure the foil to opposite
corners of the cardboard.
NOTE: Use care not to tear the aluminum foil sheet or allow the
paper clips to come in contact with the resin.
7. Place entire apparatus in 120 +/- 2ฐC oven, begin timer, and leave
in oven for 10 minutes.
8. Repeat steps two (2) through seven (7) for duplicate sample.
9. After 10 minutes, remove the sample apparatus from the oven.
Carefully remove the aluminum foil from the cardboard surface and
fold several times to avoid the loss of dried resin sample. This
will also help to minimize added moisture from condensation.
Quickly reweigh the aluminum foil to the nearest 0.0001 gram.
Record this weight as "C".
B-2
-------
METHOD: 18-001
ISSUED: 11/17/87
REVISED: 02/27/89
PAGE: _3_ OF
V. CALCULATIONS
Percent non-volatile can then be calculated using the following formula:
% NV = Weight of Resin Solids X 100
Weight of Resin Solution
Where: Weight of resin solids = C - A
Weight of resin solution = B - A
Duplicate samples should agree to within +/- 0.5%. If duplicate samples
are not within +/- 0.5%, rerun the entire test.
NOTE: Occasionally, high boiling monomers are used in manufacturing
polyester resins. Some may not evaporate as quickly as styrene. If
samples don't agree on the second run, this may be the cause.
VI. REPORT
The percent non-volatile is reported as an average of the duplicate
samples. Round the value to the nearest 0.1%. Once the % NV is known, %
Monomer Content may be calculated using the following formula:
% Monomer = 100 - %NV
Author:
Approved by:
B-3
-------
REICHHOLD
REACTIVE
POLYMERS
DIVISION
TIESJT PROCEDURE
DETERMINATION OF BROOKFIELD VISCOSITY METHOD: 18-021
AND THIXOTROPIC INDEX OF POLYESTER RESINS ISSUED: 02/17/84
REVISED: 02/27/89
PAGE 1 OF 5
I.
This method describes a procedure for determining the Brookfield
viscosity and/or thixotropy of polyester resins.
II. SAFETY
Safety glasses and protective gloves are recommended for this
procedure.
III. EQUIPMENT
1. Brookfield Viscometer, Model LVF, (6,12,30 and 60 RPM) with
spindles #1.through 14 and without guard.
2. Brookfield Viscometer, Model RVF (2,4,10 and 20 RPM) with
spindles #1 through 17 and without guard.
3. Brookfield Viscometer, Model RVT (0.5,1,2.5,5,10,50 and 100 RPM)
with spindles II through 17 and without guard.
4. Brookfield Laboratory Stand. Model A.
5. Circulating water bath controlled at 25ฐC +/- 0.2ฐC.
6. Thermometer, ASTM-17C (19-27ฐC)<
7. Stopwatch or timer capable of measuring to one second intervals.
8. Quart can and lid.
9. Brookfield factor finder or note page 5 of 5 .
The information herein is to assist customers in determining whether our products are suitable tor their applications. Our products are intended tor
sale to industrial and commercial customers. VAte request that customers inspect and test our products before use and satisfy themselves as to contents
and suitability. We warrant that our products win meet our written specifications. Nothing herein shall constitute any other warranty express or implied,
including any warranty of merchantability or fitness, nor is protection from any law or patent to be inferred. AH patent rights are reserved. The exclusive
remedy for all proven claims is replacement of our materials and In no event shall we be liable tor special, incidental or consequential damages.
REICHHOLD CHEMICALS, INC. REACTIVE POLYMERS DIVISION JACKSONVILLE, FL 32245 (904)739-2170
B-4
-------
METHOD: 18-021
ISSUED: 02/17/84
REVISED: 02/27/89
PAGE 2 OF 5
10. Laboratory monitored at 25+/-TC or 77+/-2ฐF.
IV. CALIBRATION
1. Brookfield 500 cps and 2500 cps oil standards are to be used. Each
viscometer calibration should be checked once per month using the .
Brookfield Standardized oils. Select a spindle and RPM that will
give a reading in the mid range of the viscometer scale; 40-60 on
the dial.
2. Mount the viscometer in air and level it. Deflect the needle
lightly from its zero position and let it swing back under its own
power. If the needle swings freely and returns to zero it is
acceptable.
V. PROCEDURE (NON-THIXOTROPIC RESINS)
1. Pour approximately 800 ml of resin sample into a quart can and
adjust to 25 +/- 0.2ฐC using a thermometer (avoid air entrapment).
2. Place the resin sample into the constant temperature water bath at
25 +/- 0.2ฐC. Allow sufficient time for the sample to deaerate
completely since air bubbles will affect viscosity readings.
3. Select the appropriate viscometer, spindle and spindle speed
according to the Master Formula. If unspecified, select spindle and
speed which will give a reading in the mid range of viscometer dial.
4. Remove the lid and place the resin sample under the leveled
viscometer. Lower the viscometer to a point where the spindle
coupling is approximately two (2) inches from the resin surface.
Insert the clean, dry spindle into the sample at an angle to avoid
air entrapment under the spindle. Lift up on the spindle coupling
and attach the spindle. (NOTE: left-hand thread). Avoid putting
side or down thrust on the shaft when attaching spindle.
5. Center the spindle, then raise or lower the viscometer housing until
the upper surface of the sample is in the middle of the spindle
shaft indentation.
B-5
-------
METHOD: 18-021
ISSUED: 02/17/84
REVISED: 02/27/89
PAGE 3 OF 5
6. After selecting the desired viscometer RPM, turn on viscometer motor
and simultaneously start the stopwatch.
7. After one (1) minute has elapsed, depress the viscometer clutch,
stop the motor and take a dial reading.
Va. CALCULATION
1. Multiply the dial reading by the factor obtained from the factor
finder (see page 5 of 5 ) to obtain a centipoise value. Report
results in centipoises (CPS) showing temperature, Brookfield Model,
spindle number and RPM.
VI. PROCEDURE (THIXOTROPIC RESINS)
1. Pour approximately 800 ml of resin sample into a quart can and
adjust to 25 +/- 0.2ฐC using a thermometer (avoid air entrapment).
2. Place the resin sample into the constant temperature water bath at
25 +/- 0.2ฐC, undisturbed for exactly fifteen (15) minutes prior to
viscosity determination.
3. Select the appropriate viscometer, spindle and spindle speed.
Unless specified, all thixotropic resins will be evaluated using the
Brookfield Model LVF viscometer, #3 spindle at 6 RPM and 60 RPM.
4. Follow steps (4) and (5) for non-thixotropic resins. Handle the
sample with care to minimize disturbance of the resin.
5. Set the speed to 6 rpm, then start the viscometer simultaneously
with the timer. After 60 seconds, increase the speed to 60 rpm.
After three (3) minutes depress the viscometer clutch and take a
reading. Reduce the speed to 6 RPM and start the viscometer again.
Take a final reading at third (3) minutes.
Via. CALCULATIONS
Multiply the dial reading taken at each speed by the respective factor
obtained from the factor finder (see Page 5 of 5 ) to obtain
centipoise values. Report results in centipoises (CPS) showing
temperature, Brookfield Model, spindle number and RPM. (NOTE: In most
cases the viscosity is reported using the higher RPM value).
B-6
-------
METHOD: 18-021
ISSUED: 2/17/84
REVISED: 11/4/88
PAGE 4 OF 5
VII. THIXOTROPIC INDEX
CALCULATIONS
Divide the viscosity obtained at the slower spindle speed by the
viscosity obtained at the higher spindle speed to obtain the
thixotropic index. Thixotropic index has no units of measure and is
reported to the nearest 0.1.
T j = Viscosity @ 6rom
Viscosity @ 60rpm
Author:
Approved by:
B-7
-------
BROOKFIELD FACTOR FINDER
Viscometer
Model
Spindle
Number
.5
w 1
* 2
2.5
4
Speed q
RPM
10
20
50
100
RV
1
200
100
50
40
25
20
10
5
2
1
RV
2
800
400
200
160
100
80
40
20
8
4
RV
3
2M
IN
500
400
250
200
100
50
20
10
RV
4
4M
2M
1M
800
500
400
200
100
40
20
RV
5
8M
4M
2H
1.6M
1M
800
400
200
80
40
RV
6
20M
10M
5M
4M
26M
2M
1M
500
200
100
RV
7
80M .3
40M .6
20M 1.5
16M 3
10M 6
oM Speed . ,
8M RPM 12
4M 30
2M 60
800
400
METHOD: 18-021
ISSUED: 2/17/84
REVISED: 2/27/89
PAGE 5 OF 5
LV
1
200
100
40
20
10
5
2
1
i
LV
2
1M
500
200
100
50
25
10
5
Factor
LV
3
4M
2M
800
400
200
100
40
20
LV
4
20M
10M
4M
2M
1M
500
200
100
Factor
To convert viscometer dial reading to centipoise, locate the column which identifies
model and spindle number used. Locate viscometer speed, RPM's. The number found at
of viscometer, spindle and speed is the factor. Multiply viscometer dial reading by
determine viscosity in centipoise. M=1000
the viscometer
the intersection
the factor to
-------
REICHHOLD
REACTIVE
POLYMERS
DIVISION
TEST PRCXXDURE
DETERMINATION OF ROOM TEMPERATURE GEL, METHOD: 18-050
TIME TO PEAK AND PEAK EXOTHERM CHARACTERISTICS ISSUED: 2/17/84
OF POLYESTER RESINS REVISED: 2/27/89
PAGE 1 OF 3
I- SCOPE
This method describes a procedure for determining the gel, total-time to
peak and peak exotherm of promoted or unpromoted resins when catalyzed
with methyl ethyl ketone peroxide.
II. SAFETY
.Safety glasses and protective gloves are recommended for this procedure.
Use care when handling methyl ethyl ketone peroxide. Read and -thoroughly
understand the Material Safety Data Sheet provided with this material
before its use.
III. EQUIPMENT
1. Temperature recorder with 0. to 500ฐF range. Type J thermocouple
interface. Capable of speeds of 30"/hour and 60"/hour.
2. Type J thermocouple, iron-constantan, 6" sheathed in stainless
steel.
3. Constant temperature water bath maintained at 25 +/- 0.2ฐC with
suitable rack.
4. Laboratory balance, 400 gm minimum capacity, capable of weighing to
0.01 grams.
5. ASTM-17C thermometer,-(19-27ฐC).
6. 6" wooden handle stainless steel spatula or wooden tongue depressor
(6" X 3/4").
The information herein is to assist customers in determining whether our products an suitable for their applications. Our products are intended tor
sale to industrial and commercial customers. We request that customers inspect and tost our products before use and satisfy themselves as to contents
and suitability. We warrant that our products win meet our written specifications. Nothing herein shall constitute any other warranty express or Implied,
including any warranty of merchantability or fitness, nor is protection horn any law or patent to be inferred. All patent rights are reserved. The exclusive
remedy tor aK proven claims is replacement of our materials and in no event snail we be liable tor special, incidental or consequential damages.
REICHHOLD CHEMICALS, INC. REACTIVE POLYMERS DIVISION JACKSONVILLE, FL 32245 * (904)739-2170
B-9
-------
METHOD: 18-050
ISSUED: 2/17/84
REVISED: 2/27/89
PAGE: 2 of 3
7. Disposable 150 ml polypropylene beaker (VWR Scientific Cat. no.
13915-544 or equivalent).
8. Stopwatch or timer capable of measuring to one second intervals.
9. Laboratory - maintained @ 25 +/- 1ฐC or 77 +/- 2ฐF.
10. Repipet dispenser. 5 ml capacity with 0.05 ml graduations (fisher
Scientific-Cat, no. 13-687-54 or equivalent) or Tuberculin Syringe,
1.0 CC capacity with 1/100 graduations (Fisher Scientific-Cat, no
14-820-15 or equivalent.
IV. REAGENTS
1. Methyl Ethyl Ketone Peroxide (Type specified by Master Formula).
2. 12% (metal) cobalt octoate solution (optional).
V. PROCEDURE
1. Weigh 100 +/- 0.1 grams of resin into a 150 ml polypropylene beaker.
2. Insert metal spatula, wooden tongue depressor, or thermometer into
beaker. If the wooden tongue depressor is used, it must be coated
1/2 inch above the resin level with previously weighed resin to
prevent absorption.of cobalt solution, MEKP or any additional
additives.
3. Promote resin if necessary for room temperature cure. Unless
specifically stated by Master Formula, unpromoted resins are
promoted with 0.21 grams of 12% cobalt octoate. Any promoter must
be thoroughly mixed into the resin before proceeding.
4. Place the beaker containing resin into a constant temperature water
bath at 25 */- 0.2ฐC. Allow sufficient time for the resin sample to
equilibrate to 25 +/- 0.2ฐC. If a thermometer is used to facilitate
resin temperature adjustment it must remain in the sample until
after the MEKP has been added and thoroughly dispersed.
5. Add the type and amount of MEKP specifically stated in the Master
Formula into the sample resin; simultaneously start the stopwatch
and mix thoroughly for one minute in the water bath. Avoid air
entrapment while mixing. Allow the stopwatch to run for the entire
test.
B-10
-------
METHOD: 18-050
ISSUED: 2/17/84
REVISED: 2/27/89
PAGE: 3 of 3
6. Check the sample periodically by lifting the spatula or tongue blade
to.observe the resin flow rate watching for signs of gellation. Do
not stir the sample when checking it, but simply lift the spatula or
tongue blade straight up and replace it. The point at which the
resin ceases to flow and "snaps" off the stick back into the beaker
is called the gel point and the elapsed time from catalyst addition
to the gel point is called "gel time". Record the gel time. Oo not
stop the stopwatch.
7. Upon reaching the gel time, immediately remove the beaker from the
water bath, place on a non-heat-conductive surface (i;e., wood) and
insert the thermoccuple. The tip of the thermocouple is to be
located 3/16 inch from the beaker's bot;om and within the center of
the resin sample surface.
8. Observe the recorder and stopwatch. Record the time elapsed from
catalyst addition to the peak temperature. This is called, "Total
Time to Peak".
9. The maximum temperature reached is reported as the "Peak Exotherm".
NOTE: Some customers require an interval time rather than "Total Time to
Peak". The interval (also called Gel-to-Peak) is the Total Time to Peak
minus the Gel Time.
Example:
GEL TIME: 13'
TOTAL TIME TO PEAK: 28'
PEAK EXOTHERM: 325ฐF
Here, the interval (or Gel-to-Peak) is: 28' - 13' = 15'
AUTHOR:
APPROVED BY:
B-11
-------
REICHHOLD
REACTIVE
POLYMERS
DIVISION
TEST PROCIEDURiE
DETERMINATION OF STATIC STYRENE EMISSIONS METHOD: 18-152
FOR COMPLIANCE WITH SCAQMD RULE 1162 ISSUED: 06/07/88
REVISED: 11/15/88
PAGE: 1 OF 3
I. SCOPE
This method describes the procedure for determining the weight loss of
styrene from a polyester resin during its gel and cure under static
conditions. Results are reported in g/m for compliance with Rule 1162's
limit of 60 g/m maximum emissions.
II. SAFETY
Safety glasses and protective gloves are recommended for this procedure.
Use care when handling methyl ethyl ketone peroxide. Read and thoroughly
understand the Material Safety Data Sheet provided with this material
before its use.
III. EQUIPMENT
1. Constant temperature water bath maintained at 25 +/- 0.2 deg C.
2. ASTM-17C Thermometer, (19-27 deg C).
3. Laboratory balance, 400 gm minimum capacity, accurate to +/- 0.01 gm.
4. Stopwatch or timer capable of one second intervals.
5. Disposable 400 m-1 polypropylene beaker.
6. 6" wooden handle stainless steel spatula or wooden tongue depressor
(6" x 3/4").
7. Gallon can lid, 14.5 cm diameter, deep form to hold 100 gm resin.
The information herein is to assist customers In determining whether our products are suitable tor their applications. Our products are intended tor
sale to industrial and commercial customers. Wa request that customers inspect and test our products before use and satisfy themselves as to contents
and suitability. We warrant that our products wiD meet our written specifications. Nothing herein shall constitute any other warranty express or implied,
including any warranty of merchantability or fitness, nor is protection from any law or patent to be inferred. All patent rights are reserved. The exclusive
remedy tor aH proven claims is replacement of our materials and in no event shall we be liable tor special, incidental or consequential damages.
REICHHOLD CHEMICALS, INC. REACTIVE POLYMERS DIVISION JACKSONVILLE, PL 32245 (904)739-2170
B-12
-------
METHOD: 18-152
ISSUED: 06/07/88
REVISED: 11/15/88
PAGE: 2 OF 3
8. Paper clip; bent to 90 deg angle.
9. Relative Humidity Meter, +/- 3% accuracy.
10. Repipet dispenser 5 ml capacity with 0.05 ml graduations (Fisher
Scientific Cat. No. 13-687-54 or equivalent). A Tuberculin syringe,
1.0 cc capacity with 1/100 graduations (Fisher Scientific Cat. No.
14-820-15 or equivalent) may also be used.
IV. REAGENTS
1. MEKP (Type specified on Master Formula).
V. PROCEDURE
1. Place bent paper clip on can lid.
2. Weigh can lid.to +/- O.Olgm. Record weight.
3. Record relative humidity and ambient temperature.
4. Weigh 200 +/- O.lgm of resin into polypropylene beaker.
5. Adjust temperature of resin to 25 +/- 0.2 deg C.
6. Insert metal spatula, wooden tongue depressor or thermometer. If the
wooden tongue is used the depressor much be coated 1/2 inch above the
resin level with presiously weighed resin to prevent absorption of
cobalt solution, MEKP or any additional additives. Add the type and
amount of MEKP specified in the Master Formula, then start the
stopwatch. Mix thoroughly for one (1) minute.
7. Place the gallon lid on the balance then tare. Then pour 100 +/-
0.5gm of catalyzed resin into the lid. Record the weight of resin to
+/- O.Olgm.
8. Note the gel time of resin remaining in the beaker. Determine the gel
time of the resin in the can lid by cautiously lifting paper clip.
Record gel time of resin in the can lid. Avoid excessive movement of
the paper clip as this will interfere with the results.
B-13
-------
METHOD: 18-152
ISSUED: 06/07/88
REVISED: 11/15/88
PAGE: 3 OF 3
9. Allow the resin to cure in the lid for one (1) hour, or until the
resin has cooled to room temperature.
NOTE I: If balance use is required, remove the can lid from the
balance after gel and place the lid on a heat conductive surface until
weight back is required.
NOTE II: A draft-free area is required for the test.
10. Record the final weight of resin in the can lid to +/- O.Olgm, or
record the final weight of resin plus the can lid to +/- O.Olp.
VI. CALCULATION
1. Styrene loss in g/m (E)
a. Can lid left on balance for duration of test
E ป (W -H ) A .
Where: W = initial weight of resin
W = final weight of resin
A ซ 60.56 for 14.5cm lid
b. Can lid removed from balance during test for other balance use
E = (W + W-W)A
Where: W = initial weight of resin
W = weight of can lid and paper clip
W = final weight of resin plus can lid
A = 60.56 for 14.5cm lid *
VII. REPORT
Resin
Batch Number
Gel Time, Catalyst, and Catalyst Level
Relative Humidity
Ambient Temperature
Styrene Loss in g/m
Author:
Approved by:
B-14f
-------
Appendix C
Verification and Intercomparison of Compressed Gas Calibration Standards
C-i
-------
APPENDIX C. VERIFICATION AND INTERCOM?ARISON OF COMPRESSED
GAS CALIBRATION STANDARDS
Styrene and propane compressed gas calibration standards used to calibrate the Ratfisch
total hydrocarbon analyzer were purchased in 1994 and 1995 from Scott Specialty Gases in
Durham, North Carolina. The styrene calibration standards could not be used directly for
routine calibrations during styrene emission testing because of cylinder pressure limitations
associated with styrene's dew point. Instead, propane calibration standards without such pressure
limitations were used for the routine calibrations. Prior to the styrene emission testing at
Reichhold Chemical, the certified concentrations of the styrene and propane calibration standards
were verified at RT1. Additionally, the styrene and propane calibration standards were
intercompared using the total hydrocarbon analyzer to obtain a propane-to-styrene correction
factor for the styrene emission testing. This section discusses the verification and
intercomparison of the calibration standards.
The styrene in nitrogen calibration standards were verified using a Hewlett Packard
Model 5890, Series II gas chromatograph with a flame ionization detector, a gas sampling valve,
and a 10-foot long by 1/8-inch OD stainless steel column packed with 10-percent OV-101 on
Chromasorb WHP, 80/100 mesh at an oven temperature of 150 degrees Celsius (oC) isothermal.
The analytical reference standards were prepared by serial dilution of a primary standard that was
prepared by injection of liquid styrene and gaseous nitrogen into a canister. The analytical
reference standard concentrations were 249, 25.1, and 2.49 parts per million by volume (ppmv). .
Each analytical reference standard and calibration standard was analyzed three times in
succession before the next standard was analyzed. Styrene concentrations were calculated by
linear interpolation between the nearest two analytical reference standard measurements.
The propane in air calibration standards were verified using a Hewlett Packard Model
5890, Series II gas chromatograph with a flame ionization detector, a gas sampling valve, and a
6-foot long by 1/8-inch OD stainless steel column packed with n-octane/Porasil C; 80/100 mesh
at an oven temperature of 40 oC isothermal. The analytical reference standards were National
Institute of Standards and Technology (NIST) Standard Reference Materials (SRMs) containing
propane in air at concentrations of 476,94.8,48.6, and 0.99 parts ppmv. Each analytical
reference standard and calibration standard was analyzed five times in succession before the next
standard was analyzed. Propane concentrations were calculated from a least squares regression
line determined from the analytical reference standard measurements.
The results of the styrene and propane calibration standard verifications are given in the
following table:
C-l
-------
Stamped Cylinder
Number
AAL 17461
BAL3703
BALI 361
ALM036826
BAL4319
1A009797
ALM044162
ALMO 19336
AAL4968(1)
A022617
A12158
A5406
Scott- Certified
Concentration
237.1 ppm
styrene(about 175
ppm upon reanalysis)
81.5 ppmstyrene
49.5 ppm styrene
10.7 ppmstyrene
5. 12 ppm styrene
454 ppm propane
275 ppm propane
15 0.9 ppm propane
95.4 ppm propane
44.8 ppm propane
26.9 ppm propane
15.95 ppm propane
RTI-Verified
Concentration
(yearofRTI
verification)
184 ppm (95)
85.0 ppm (94)
87.9 ppm (95)
53. 4 ppm (94)
53. 9 ppm (95)
9.83 ppm (95)
5.42 ppm (94)
5.33 ppm (95).
453.6 ppm (95)
278.0 ppm (95)
153. 4 ppm (95)
95.77 ppm (93)
96.9 ppm (95)
45.03 ppm (95)
27.02 ppm (95)
15.99ppm.(95)
Percent Difference
-22.4
+4.3
+7.9
+7.9'
+8.9
-8.1
+5.9
+4.1
-0.1
+1.1
+1.7
+0.4
+1.6
+0.5
+0.4
+0.3
(1) This calibration standard was not used for routine calibrations during the styrene emission
testing, but was used during the evaluation of the total hydrocarbon analyzer.
In general, the Scott-certified concentrations and the RTI-verified concentrations for four
of five styrene calibration standards agreed to within +/- 10 percent. However, the agreement
was -22.4 percent for AAL 17461. This calibration standard was returned to Scott Specialty
Gases for reanalysis and its concentration was found to have shifted since its first analysis. It
was judged to be unstable and was not used for routine calibrations. Three other styrene
calibration standards (i.e., BAL1361, BAL3703, and BAL4319) had been verified by RTI in
1994. The reasonably good agreement between RTFs 1994 and 1995 values for these three
calibration standards supports the beliefs that all styrene calibration standards were measured
accurately and that the styrene concentrations are stable over the time period between the
verifications.
0-2
-------
The Scott-certified propane concentrations and the RTI-verified concentrations agreed to
within +/- 2 percent for all seven propane calibration standards. One propane calibration .
standard (i.e. AAL4968) had been verified by RTI in 1993. The good agreement between RTI's
1993 and 1995 values for this calibration standard supports the beliefs that all propane calibration
standards were measured accurately and that the propane concentrations are stable over the time
period between the verifications. The 1995 RTI-verified concentrations for the propane
calibration standards were used in the routine calibrations during the styrene emissions testing.
The styrene and propane calibration standards were intercompared using the Ratfisch
total hydrocarbon analyzer to obtain a propane-to-styrene correction factor for the styrene
emission testing. The calibration standards were delivered to the analyzer in a manner very
similar to that employed during routine calibrations. The analyzer operating parameters were
identical to those used during styrene emission testing. Two styrene calibration standards, four
propane calibration standards, and zero air were measured twice on Range 2 (0 to 200 ppmv
styrene). Two other styrene calibration standards, three other propane standards, and zero air
were measured twice on Range 1 (0 to 20 ppmv styrene). Least squares regression analysis of 2-
minute mean voltages yielded the following slopes for the styrene and propane regression lines:
Styrene Slope
(Volts/ppmv)
Propane Slope
(Volts/ppmv)
Slope Ratio
Range 2
0.04928
0.01834
2.686
Range 1
0.50550
0.18305
2.762
A styrene molecule has 8 carbon atoms and a propane molecule has 3 carbon atoms. As a
first approximation, one would expect the propane-to-styrene correction factor to equal the ratio
of the carbon atoms (i.e., 8/3 = 2.667). The measured slope ratios are in good agreement with
this theoretical value, particularly for Range 2.
C-3
-------
Appendix D
Summary of Calibration Data, Calibration Error Tests, and Drift Checks
D-i
-------
APPENDIX D.
SUMMARY OF CALIBRATION DATA, CALIBRATION ERROR
TESTS, AND DRIFT CHECKS
The total hydrocarbon analyzer was calibrated prior to each test run and a calibration drift
check was done at the end of each test run. This section presents a summary of the calibration
data, including calibration error tests and drift checks. The calibration requirements of EPA
Method 25A were met or exceeded for the styrene emission tests. The method specifies that the
calibration error must be less than +/- 5 percent of the calibration standard's concentration. The
mid-level calibration error was less than +/-1 percent for all test runs and the low-level
calibration error was less than +/- 2 percent for all test runs. The method specifies that the zero
and calibration drifts must be less than +/- 3 percent of scale. The zero and calibration drifts
were less than or equal to +/- 1 percent full scale (% FS) for all test runs.
The propane calibration standards that RTI had verified and intercompared with styrene
calibration standards were used for routine calibrations. The RTI-verified concentrations were
used as the values for these calibration standards. For most test runs, the styrene measurements
were made using Range 2 of the analyzer. These measurements were recorded by an Omega
Engineering model OM-170 data logger and a Hewlett-Packard model 7132A strip chart
recorder. The discussion of calibration error tests and drift checks in this section is based on
those data recorded by the strip chart recorder. These data have an approximate resolution of 1/4
%FS.
The total hydrocarbon analyzer's response to the high-level calibration standard
throughout the styrene emission testing is summarized in the following table. The analyzer was
operated on Range 2 for all test runs, except for those Range 1 test runs marked by asterisks in
the table. The analyzer's zero and span pots were not adjusted during the entire 5-week testing
period. This table demonstrates that the analyzer's calibration remained very stable during the
testing period. All 53 measurements of the high-level calibration standard on Range 2 fell
between 87 and 90.5 % FS.
Test Run
PI
P2
P3
P4
P5
High-Level
Cal. Std.
Response (%
FS)
89.75
90.50
90.00
89.50
89.75
Test Run
Gl
G2
G3
G4
G5
High-Level
Cal. Std.
Response (%
FS)
89.25
88.75
89.25
89.25
88.50
Test Run
Rl
R2
R3
R4* .
R5
High-Level
Cal. Std.
Response (%
FS)
88.75
89.00
88.75
89.50
88.50
D-l
-------
Test Run
P6
P7
P8
P9
P10
Pll
P12
High-Level
Cal. Std.
Response (%
FS)
88.50
89.00
. 89.50
88.50
89.25
89.00
89.00
Test Run
G6
G7
G8
G9
G10
Gil
G12
G13
G14
G15
G16
G17
G18
High-Level
Cal. Std.
Response (%
FS)
88.75
90.00
90.50
90.00
90.00
90.25
89.75
89.50
89.50
88.50
88.75
88.50
88.75
Test Run
R6
R7
R8
R9
RIO
Rll
R12
R13
R14
R15
R16*
R17
R18*
R19
R20
R21
R22
R23
R24
R25
High-Level
Cal. Std.
Response (%
FS)
87.75
88.00
88.25
88.50
88.00
88.25
88.00
88.00
88.50
88.50
89.00
87.00
88.00
88.75
89.00
87.75
87.50
88.00
88.25
87.75
* Analyzer operated on Range 1 (0 to 20 ppm styrene).
EPA Method 25A specifies that calibration error tests be conducted immediately prior to
each test run. After the analyzer's calibration equation was determined by measurements of the
high-level calibration standard and zero air, the linearity of the calibration curve was tested by
measurements of the mid-level and low-level calibration standards. The calibration error is
D-2
-------
calculated as the difference between the analyzer's actual response to these calibration standards
and the response that is predicted from the calibration equation. The method specifies that the
calibration error must be less than +/- 5 percent of the calibration standard's concentration. The
following table demonstrates that the mid-level calibration error was less than +/- 1 percent for
all test runs and that the low-level calibration error was less than +/- 2 percent for all test runs.
Other data not reported here supports the belief that the analyzer's calibration curve is a straight
line.
Test
Run
PI
P2
P3
P4
P5
P6
P7
P8
P9
P10
Pll
P12
Gl
G2
G3
G4
G5
G6
G7
G8
Mid-Level Cal.
Error (percent)
-1.30
-0.35
0.19
-0.60
0.02
0.05
-0.50
-0.15
0.05
0.12
-0.50
-0.50 .
-0.33
-0.23
-0.33
-0.50
0.05
-0.23
0.02
0.18
Low-Level Cal.
Error (percent)
-1.71
-0.97
0.28
-0.71
.-0.96
0.31
-0.20
-0.71
-0.45
-0.46
-1.72
-0.96
-0.46
-0.71
-1.21
-1.70
-0.45
0.05
-0.96
0.03
Test
Run
Rl
R2
R3
R4*
R5
R6
R7
R8
R9
RIO
Rll
R12
R13
R14
R15
R16*
R17
R18*
R19
R20
Mid-Level
Cal. Error
(percent)
-0.40
-0.23
0.22
0.08
0.05
-0.03
0.15
-0.13
-0.40
0.60
-0.13
-0.03
-0.03
-0.57
-0.13
0.71
-0.27
-0.74
-0.23
0.39
Low-Level
Cal. Error
(percent)
-1.20
-0.70
0.05
. -0.73
-0.45
-0.44
-0.70
-0.96
-1.21
0.06
-0.96
-1.20
-0.44
-0.95
-0.95
-0.43
-0.94
-0.39
0.05
-0.20
D-3
-------
Test
Run
G9
G10
Gil
G12
G13
G14
G15
G16
G17
G18
Mid-Level Cal.
Error (percent)
-0.42
-0.42
-0.52
-0.59
-0.60
-0.60
-0.40
-0.23
-0.40
-0.23
Low-Level Cal.
Error (percent)
-0.96
-1.70
-1.46
-1.45
-0.71
-1.46
-0.45
-0.71
-1.21
-0.71
Test
Run
R21
R22
R23
R24
R25
Mid-Level
Cal. Error
(percent)
-0.03
0.53
0.15
-0.13
-0.48
Low-Level
Cal. Error
(percent)
-1.21
-0.69
-0.70
-0.96
-1.21
* Analyzer operated on Range 1 (0 to 20 ppm styrene).
EPA Method 25A specifies that zero and calibration drift checks be conducted
immediately following the completion of each test run. The same mid-level calibration standard
and zero air that were measured during the calibration are to be remeasured at the end of the test
run. The method specifies that the zero and calibration drifts must be less than +/- 3 percent of
scale. The following table demonstrates that the zero and calibration drifts were less than or
equal to +/- 1 % FS for all test runs.
Test Run
PI
P2
P3
P4
P5
P6
Zero Drift (%
FS)
0.00
0.00
0.00
0.00
0.00
0.00
Mid-Level
Cal. Drift (%
FS)
0.50
0.25
0.50
0.25
0.00
-0.50
Test Run
Rl
R2
R3
R4*
R5
R6
Zero Drift (%
FS)
0.00
-0.25
0.00
0.00
0.00
0.00
Mid-Level
Cal. Drift (%
FS)
0.25
0.00
-0.25
0.25
-0.50
0.50
D-4
-------
Test Run
P7
P8
P9
P10
Pll
P12
Gl
G2
G3
G4
G5
G6
G7
G8
G9
G10
Gil
G12
G13
G14
G15
G16
G17
G18
Zero Drift (%
FS)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
-0.25
0.00
0.00
0.00
0.00
-0.25
-0.25
0.25
-0.25
0.00
0.00
0.00
0.00
0.00
0.00
Mid-Level
Cal. Drift (%
FS)
0.50
-0.50
0.25
-0.50
0.00
0.50
-0.25
0.25
0.25
-0.25
0.00
0.25
0.25
-0.50
0.50
0.00
-0.25
0.00
0.00
-0.50
-0.50
-0.25
0.25
0.00
Test Run
R7
R8
R9
RIO
Rll
R12
R13
R14
R15
R16*
. R17
R18*
R19
R20
R21
R22
R23
R24
R25
Zero Drift (%
FS)
0.00
0.00
0.00
0.00
0.25
0.00
0.00
0.00
0.00
-0.25
0.00
-0.75
0.00
0.00
0.00
-0.25
0.00
0.00
0.00
Mid-Level
Cal. Drift (%
FS)
0.00
0.00
0.25
-0.25
0.00
0.25
0.00
0.25
-0.75
-0.25
0.25
0.00
0.50
-1.00
0.00
0.00
0.00
0.25
0.75
* Analyzer operated on Range 1 (0 to 20 ppm sryrene).
D-5
-------
Appendix E
A Summary of Emission Measurements, Gravimetric Measurements, and Calculated Emission
Quantities and Emission Factors for the Test
E-i
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
EXPERIMENTAL RUNS
6/6/95
6/6/95
7/7/95
7/7/95
10:22
14:55
12:19
14:12
RF1-EXP
GF1-EXP
EXP1
EXP2
PILOT EXPERIMENT
6/7/95
6/8/95
6/8/95
6/8/95
6/9/95
6/9/95
6/12/95
6/12/95
6/12/95
6/13/95
6/13/95
6/13/95
14:50
10:01
12:11
14:40
10:20
14:45
10:36
13:47
15:44
10:37
12:57
15:21
Average (12 runs)
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
P11
P12
M1 -Normal spraying (6 runs)
Material/
container #
RF1
GF1
GF1
GF1
GF1 #10
GF1#1
GF1 #1
GF1#1
GF1#1
GF1*1
GF1 #1/#2
GF1#2
GF1#2
GF1*2
GF1*2
GF1*2
M2-Controlled spraying (6 runs)
A2-High air velocity (6 runs)
A1 -Low air velocity (6 runs)
M1/A1 (3 runs)
M2/A1 (3 runs)
M1/A2 (3 runs)
M2/A2 (3 runs)
Equip.
RE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1 .
GE1
Application
Method
Normal
Normal
Enclosed
Part. Encl.
Controlled
Normal
Controlled
Normal
Controlled
Normal
Controlled
Normal
Controlled
Controlled
Normal
Normal
,
Air Vel.
High
High
Low
Low
Low
High
High
High
Low
Low
High
Low
High
Low
High
Low
Avg. sqrt
. (delta P)
0.3693
0.3657
0.3376
0.3464
0.3582
0.3619
0.3605
0.3464
0.3535
0.3476
0.3559
0.3524
0.3464
0.3647
0.3623
0.3582
0.3557
0.3548
0.3565
0.3556
0.3558
0.3527
0.3588
0.3569
0.3543
Exhaust flow
rate, cfm
9124
9054
8510
8681
8909
8980
8953
8681
8818
8704
8864
8796
8681
9034
8987
8909
8860
8843
8876
..
8858
8862
8803
8920
8883
8833
Avg. cone.
ppm
6.94
7.54
1.06
4.90
4.18
7.60
5.39
7.75
5.14
6.62
5.28
7.85
6.57
5.64
6.49
6.61
6.26
7.15
5.37
6.51
6.01
7.03
4.99
7.28
5.75
Background
cone., ppm
0.59
0.23
0.48
0.42
0.32
0.47
0.33
0.36
0.42
0.41
0.29
0.32
0.65
0.53
0.45
0.59
0.43
0.43
0.42
0.43
0.43
0.44
0.43
0.43
0.42
Avg. net
cone., ppm
6.35
7.30
0.58
4.48
3.85
7.14
5.05
7.39
4.72
6.21
4.99
7.53
5.92
5.11
6.04
6.02
5.83
6.72
4.94
6.09
5.57
6.59
4.56
6.86
5.32
Test run
duration, min
88.8
56.2
102.2
31.2
76.3
68.2
68.8
66.8
77.8
76.3
78.7
68.2
66.9
70.0
71.9
72.3
71.8
70.6
73.1
70.2
73.5
72.3
74.7
69.0
71.5
Total emissions
by THC, g
631
456
62
149
322
536
382
526
397
506
427
554
422
396
479
476
452
513
391
462
442
512
372
514
410
m
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
EXPERIMENTAL RUNS
6/6/95
6/6/95
7/7/95
7/7/95
10:22
14:55
12:19
14:12
RF1-EXP
GF1-EXP
EXP1
EXP2
PILOT EXPERIMENT
6/7/95
6/8/95
6/8/95
6/8/95
6/9/95
6/9/95
6/12/95
6/12/95
6/1 2/95
6/13/95
6/13/95
6/13/95
14:50
10:01
12:11
14:40
10:20
14:45
10:36
13:47
15:44
10:37
12:57
15:21
Average (12 runs)
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
P11
P12
Ml -Normal spraying (6 runs)
Material/
container ft
RF1
GF1
GF1
GF1
GF1 #10
GF1#1
GF1#1
GF1 #1
GF1#1
GF1 #1
GF1 #1/#2
GF1#2
GF1#2
GF1#2
GF1#2
GF1#2
M2-Controlled spraying (6 runs)
A2-High air velocity (6 runs)
A1 -Low air velocity (6 runs)
M1/A1 (3 runs)
M2/A1 (3 runs)
M1/A2 (3 runs)
M2/A2 (3 runs)
Equip.
RE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
Avg. spraying
cone., ppm
65.74
41.31
1.05
45.25
29.73
43.71
42.14
65.24
39.44
59.03
48.30
74.04
52.82
41.80
64.60
58.44
51.61
60.84
42.37
52.80
50.41
63.84
36.99
57.85
47.75
Avg. net
spraying, ppm
65.14
41.08
0.57
44.83
29.41
43.24
41.80
64.87
39.01
58.62
48.01
73.71
52.17
41.27
64.15
57.85
51.18
60.41
41.95
52.37
49.98
63.39
36.56
57.42
47.33
Spraying
time, min
4.5
5.0
2.0
2.2
4.2
6.5
3.0
3.7
3.2
3.5
3.0
3.4
3.2
3.2
3.0
3.2
3.6
3.9
3.3
3.7
3.4
3.4
3.5
4.4
3.1
Emissions
during
application
(by THC)
328
228
1
103
134
310
138
253
134
219
157
271
176
145
214
204
196
245
147
208
184
231
138
259
157
Std. Dev. of
application
emissions
(by THC)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
57
37
15
59
51
28
5
39
16
Wt. loss from
part during
i curing stage
(by MB)
264
193
44
50
167
190
221
203
224
242
L 235
217
213
221
235
218
216
218
214
216
215
226
204
209
223
Std. Oev. of
curing loss
from part
(by MB)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
20
18
22
16
23
12
26
19
9
Wt. loss during
application and from
overspray, g
(by MB)
622
285
21
102
213
417
221
212
188
264
231
282
176
236
300
224
247
283
211
260
235
257
212
310
209
m
rb
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
EXPERIMENTAL RUNS
6/6/95
6/6/95
7/7/95
7/7/95
10:22
14:55
12:19
14:12
RF1-EXP
GF1-EXP
EXP1
EXP2
PILOT EXPERIMENT
6/7/95
6/8/95
6/8/95
6/8/95
6/9/95
6/9/95
6/12/95
6/12/95
6/12/95
6/13/95
6/13/95
6/13/95
14:50
10:01
12:11
14:40
10:20
14:45
10:36
13:47
15:44
10:37
12:57
15:21
Average (12 runs)
P1
P2
P3
P4
PS
P6
P7
P8
P9
P10
P11
P12
Ml -Normal spraying (6 runs)
Material/
container It
RF1
GF1
GF1
GF1
GF1 #10
GF1#1
GF1#1
GF1#1
GF1#1
GF1#1
GF1 #1/#2
GF1#2
GF1#2
GF1#2
GF1#2
GF1#2
M2-Controlled spraying (6 runs)
A2-High air velocity (6 runs)
A1-Low air velocity (6 runs)
M1/A1 (3 runs)
M2/A1 (3 runs)
M1/A2 (3 runs)
M2/A2 (3 runs)
Equip.
RE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
Total emissions
byMB.g
886
478
65
152
380
607
442
415
412
506
466
499
389
457
535
442
463
501
424
476
449
482
416
519
432
Ratio of two
measurement
methods
(MB7THC)
1.40
1.05
1.05
1.02
1.18
1.13
1.16
0.79
1.04
1.00
1.09
0.90
0.92
1.15
1.12
0.93
1.03
0.98
1.09
1.03
1.03
0.94
1.12
1.01
1.06
Material
used, g
9287
1564
1732
1737
1564
2261
1695
2067
1964
2204
1900
2056
2064
2021
2077
2049
1994
2119
1868
2011
1976
2103
1850
2135
1886
Transfer
eff. %
90.1
75.5
99.9
94.2
79.8
66.1
80.0
74.0
85.1
75.7
81.2
73.8
85.0
82.6
74.5
77.7
78.0
73.6
82.3
76.8
79.1
75.7
82.5
71.5
82.1
Styrene
content, %
38.33
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
38.75
EF, % available styrene
byTHC
17.7
75.2
9.2
22.2
53.1
61.2
58.2
65.7
52.2
59.2
58.0
69.6
52.7
50.6
59.5
60.0
58.3
62.5
54.1
59.2
57.4
62.9
52.0
62.1
56.3
by MB
24.9
78.9
9.7
22.6
62.7
69.3
67.3
51.8
54.1
59.2
63.3
62.6
48.6
58.4
66.5
55.7
60.0
60.9
59.1
61.1
58.8
59.2
59.4
62.5
59.7
Std. Dev.. % AS
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
5.4
3.8
2.9
3.9
6.5
4.7
1.0
2.6
2.5
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
6.2
6.0
6.2
8.0
3.2
2.8
3.5
7.7
8.0
Total mold
surface, m2
2.28
2.28
0.32
0.45
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
m
CO
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
m
Date
Time
Test run #
EXPERIMENTAL RUNS
6/6/95
6/6/95
7/7/95
7/7/95
10:22J
14:55
12.19
14:12
RF1-EXP
GF1-EXP
EXP1
EXP2
PILOT EXPERIMENT
6/7/95
6/8/95
6/8/95
6/8/95
6/9/95
6/9/95
6/12/95
6/12/95
6/12/95
6/13/95
6/13/95
6/13/95
14.50
10:01
12:11
14:40
10:20
14:45
10:36
13:47
15:44
10:37
12:57
15:21
Average (12 runs)
PI
P2
P3
P4
P5
P6
P7
P8
P9
P10
P11
P12
Ml -Normal spraying (6 runs)
Material/
container #
RF1
GF1
GF1
GF1
GF1 #10
GF1#1
GF1#1
GF1#1
GF1#1
GF1#1
GF1 #1/#2
GF1#2
GF1#2
GF1#2
GF1#2
GF1#2
M2-C6ntrolled spraying (6 runs)
A2-High air velocity (6 runs)
AI-Low air velocity (6 runs)
M1/A1 (3 runs)
M2/A1 (3 runs)
M1/A2(3runs)
M2/A2(3runs)
Equip.
RE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
GE1
EF. g/m2
byTHC
277
200
194
332
141
236
168
231
175
222
188
244
185
174
210
209
199
225
172
203
194
225
163
226
180
by MB
389
210
205
339
167
267
194
182
181
222
205
219
171
201
235
194
203
220
186
209
197
212
183
228
190
Std. Oev., g/m2
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
30.2
12.6
15.2
24.8
34.2
14.2
15.6
10.9
8.8
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
27.6
27.4
14.4
32.7
19.6
12.6
13.9
34.8
14.1
EF, g/g material
byTHC
0.068
0.291
0.036
0.086
0.206
0.237
0.225
0.254
0.202
0.230
0.225
0.270
0.204
0.196
0.231
0.232
0.226
0.242
0.210
0.229
0.223
0.244
0.201
0.241
0.218
by MB
0.095
0.306
0.038
0.088
0.243
0.268
0.261
0.201
0.210
0.230
0.245
0.243
|_ 0.188
0.226
0.258
0.216
0.232
0.236
0.229
0:237
0.228
0.229
0.226
0.242
0.231
Std. Dev., g/g
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
0.021
0.015
0.011
0.015
0.025
0.018
0.004
0.010
0.010
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
0.024
0.023
0.024
0.031
0.012
0.011
0.014
0.030
0.031
Catalyst
Ratio, %
1.33
1.40
NC
NC
1.66
1.68
1.65
1.79
1.63
1.68
1.74
1.75
1.74
1.73
1.69
1.76
1.71
1.73
1.69
.72
.70
.73
.67
.72
.71
Glass/'rtesin
Ratio, %
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
GELCOAT EXPERIMENT
6/14/95
6/14/95
6/14/95
6/15/95
6/15/95
6/15/95
6/16/95
6/16/95
6/16/95
6/19/95
6/19/95
6/19/95
6/21/95
6/21/95
6/21/95
6/22/95
6/22/95
6/22/95
10:11
13:46
16:06
11:38
14:03
15:59
10:23
12:49
15:33
10:36
12:47
15:23
10:46
13:18
16:27
11:14
13:41
15:55
Average (18 runs)
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10
G11
G12
G13
G14
G15
G16
G17
G18
GF1 -Regular gel coat (9 runs)
Material/
container #
GF1#2
GF1 #2/#3
GF1 #2/#3
GF2#1
GF1#3
GF1 #3
GF2#1
GF1 #3/#4
GF1 #3/#4
GF1#4
GF2#1/#2
GF2#1/#2
GF2#2
GF2#2
GF1 #4
GF2#2/#3
GF2 #2/#3
GF2#3
GF2-Low VOC gel coat (9 runs)
1
GE1-AAA ext mix gun (6 runs)
GE2-HVLP int mix gun (6 runs)
GE3-HVLP ext mix gun (6 runs)
GF1/GE1 (3 runs)
GF1/GE2 (3 runs)
GF1/GE3(3runs)
GF2/GE1 (3 runs)
GF2/GE2 (3 runs)
GF2/GE3 (3 runs)
Equip.
GE3
GE2
GE1
GE1
GE1
GE1
GE3
GE3
GE2
GE3
GE2
GE2
GE2
GE1
GE2
GE1
GE3
GE3
Application
Method
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Air Vel.
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Avg^sqrt
(delta P)
0.3507
0.3464
0.3403
0.3428
0.3427
0.3440
0.3586
0.3571
0.3536
0.3500
0.3559
0.3473
0.3527
0.3493
0.3452
0.351 1
0.3535
0.3484
0.3494
0.3478
0.351 1
0:3450
0.3502
0.3530
0.3423
0.3484
0.3526
0.3477
0.3519
0.3535
Exhaust flow
rate, cfm
8764
8680
8563
8610
8609
8633
8916
8887
8819
8752
8866
8700
8804
8739
8660
8775
8820
8722
8740
8707
8772
8655
8755
8810
8602
8720
8801
8708
8790
8819
Avg. cone.
ppm
5.80
5.89
6.94
3.80
5.92
5.80
3.38
5.60
5.65
5.60
3.77
3.47
3.77
3.36
5.81
3.47
3.80
3.80
4.76
5.89
3.62
4.88
4.73
4.66
6.22
5.79
5.67
3.54
3.67
3.66
Background
cone., ppm
0.55
0.59
0.47
0.39
0.51
0.68
0.31
0.63
0.58
0.46
0.59
0.72
0.55
0.51
0.56
0.47
0.58
0.60
0.54
0.56
0.53
0.50
0.60
0.52
0.55
0.58
0.55
046
0.62
0.50
Avg. net
cone., ppm
5.25
5.30
6.46
3.41
5.41
5.13
3.07
4.97
5.07
5.14
3.18
2.74
3.22
2.85
5.25
3.00
3.21
3.19
4.21
5.33
3.10
4.38
4.13
4.14
5.67
5.21
5.12
3.09
3.05
3.16
Test run
duration, min
70.0
70.0
60.3
75.7
70.4
71.2
80.7
70.5
70.2
69.6
85.1
85.0
83.7
87.5
60.7
85.0
85.7
82.4
75.8
68.1
83.4
75.0
75.8
76.5
67.3
67.0
70.0
82.7
84.6
82.9
Total emissions
byTHC.g
395
395
409
273
403
387
271
382
386
384
294
249
291
267
339
274
298
282
332
387
278
336
326
335
400
373
387
272
278
284
m
cii
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
GELCOAT EXPERIMENT
6/14/95
6/14/95
6/14/95
6/15/95
6/15/95
6/15/95
6/16/95
6/16/95
6/16/95
6/19/95
6/19/95
6/19/95
6/21/95
6/21/95
6/21/95
6/22/95
6/22/95
6/22/95
10:11
13:46
16:06
11:38
14:03
15:59
10:23
12:49
15:33
10:36
12:47
15:23
10:46
13:18
16:27
11:14
13.41
15:55
Average (18 runs)
.
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10
G11
G12
G13
G14
G15
G16
G17
G16
GF1 -Regular gel coat (9 runs)
Material/
container 9
GF1#2
GF1 #2/#3
GF1 #2/#3
GF2#1
GF1#3
GF1#3
GF2#1
GF1 #3/#4
GF1 #3/#4
GF1#4
GF2#1/#2
GF2 #1/#2
GF2#2
GF2#2
GF1#4
GF2 #2/#3
GF2 #2/#3
GF2#3
GF2-Low VOC gel coat (9 runs)
GE1-AAA ext mix gun (6 runs)
GE2-HVLP int mix gun (6 runs)
GE3-HVLP ext mix gun (6 runs)
GF1/GE1 (3 runs)
GF1/GE2 (3 runs)
GF1/GE3(3runs)
GF2/GE1 (3 runs)
GF2/GE2 (3 runs)
GF2/GE3(3runs)
Equip.
GE3
GE2
GE1
GE1
GE1
GE1
GE3
GE3
GE2
GE3
GE2
GE2
GE2
GE1
GE2
GE1
GE3
GE3
Avg. spraying
cone., ppm
48.22
44.48
50.51
21.88
49.01
45.34
23.46
44.31
45.04
43.40
28.43
29.10
34.91
30.12
48.31
26.75
31.41
29.39
37.45
46.51
28.38
37.27
38.38
36.70
48.29
45.94
45.31
26.25
30.81
28.09
Avg. net
spraying, ppm
47.66
43.89
50.04
21.50
48.50
44.66
23.14
43.68
44.46
42.94
27.84
28.37
34.35
29.61
47.75
26.28
30.83
28.79
36.90
45.95
27.86
36.76
37.78
36.17
47.73
45.37
44.76
25.80
30.19
27.59
Spraying
time, min
3.0
3.5
3.0
4.0
3.0
3.0
3.2
3.0
3.2
2.9
3.1
3.1
3.0
2.9
3.2
3.5
3.0
3.0
3.1
3.1
3.2
3.2
3.2
3.0
3.0
3.3
3.0
3.5
3.1
3.1
Emissions
during
application
(byTHC)
154
164
158
91
154
142
80
143
155
134
94
94
112
92
-162
99
100
93
. ,
123
152
95
123
130
117
l_ 151
160
143
94
100
91
Std. Dev. of
application
emissions
(byTHC)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
30
9
8
29
31
27
7
4
8
4
9
8
Wt. loss from
part during
curing stage
(by MB)
220
206
204
153
212
220
161
211
205
214
159
157
163
165
201
155
157
160
185
210
159
185
182
187
212
204
215
158
160
159
Std. Dev. of
curing loss
frornpart
(by MB)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
26
6
4
28
22
28
7
2
4
5
2
2
Wt. loss during
application and from
overspray, g
(by MBl
128
200
200
74
143
190
112
261
193
146
168
189
211
167
186
147
153
126
166
183
150
154
191
154
178
193
178
129
189
130
m
CD
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
GELCOAT EXPERIMENT
6/14/95
6/14/95
6/14/95
6/15/95
6/15/95
6/15/95
6/16/95
6/16/95
6/16/95
6/19/95
6/19/95
6/19/95
6/21/95
6/21/95
6/21/95
6/22/95
6/22/95
6/22/95
10:11
13.46
16:06
11:38
14:03
15:59
10.23
12:49
15:33
10:36
12:47
15:23
10:46
13:18
16:27
11:14
13:41
15:55
Average (18 runs)
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10
G11
G12
G13
G14
G15
G16
G17
G18
GF1 -Regular gel coat (9 runs)
Material/
container 9
GF1#2
GF1 #2/#3
GF1 #2/#3
GF2#1
GF1#3
GF1#3
GF2#1
GF1 #3/#4
GF1 #3/#4
GF1#4
GF2#1/#2
GF2#1/#2
GF2#2
GF2#2
GF1#4
GF2#2/#3
GF2#2/ป3
GF2#3
GF2-Low VOC gel coat (9 runs)
GE1-AAA ext mix gun (6 runs)
GE2-HVLP int mix gun (6 runs)
GE3-HVLP ext mix gun (6 runs)
I
GF1/GE1 (3 runs)
GF1/GE2(3runs)
GF1/GE3 (3 runs)
GF2/GE1 (3 runs)
GF2/GE2 (3 runs)
GF2/GE3 (3 runs)
Equip.
GE3
GE2
GE1
GE1
GE1
GE1
GE3
GE3
GE2
GE3
GE2
GE2
GE2
GE1
GE2
GE1
GE3
GE3
Total emissions
by MB, g
348
406
404
227
355
410
273
472
398
360
327
346
374
332
387
302
310
286
351
393
309
338
373
342
390
397
393
287
349
290
Ratio of two
measurement
methods
(MB/THC)
0.88
1.03
0.99
0.83
0.88
.06
.01
.24
.03
0.94
.11
.39
.28
.24
.14
.10
.04
.01
1.07
1.02
. 1.11
1.02
1.16
1.02
0.98
.07
.02
.06
.26
.02
Material
used, g
1723
1808
1756
1929
1765
1817
1876
1891
1787
1723
1940
1940
2245
1933
1776
2128
2275
1963
1904
1783
2025
1888
1916
1909
1779
1790
1779
1997
2042
2038
Transfer
eff. %
84.6
' 81.4
79.0
88.8
84.3
82.9
87.3
80.0
82.4
84.7
83.3
81.9
81.1
81.9
82.0
87.4
85.7
85.4
83.6
82.4
84.8
84.1
82.0
. 84.6
82.1
81.9
83.1
86.0
82.1
86.1
Styrene
content, %
38.75
38.75
38.75
25.35
38.75
38.75
25.35
38.75
38.75
38.75
25.35
25.35
25.35
25.35
38.75
25.35
25.35
25:35
32.05
38.75
25.35
32.05
32.05
32.05
38.75
38.75
38.75
25.35
25.35
25.35
EF, % available styrene
byTHC
59.2
56.4
60.2
55.8
58.9
54.9
56.9
52.1
55.7
57.6
59.8
50.6
51.2
54.6
49.2
50.9
51.7
56.6
55.1
56.0
54.2
55.9
53.8
55.7
58.0
53.8
56.3
53.7
53.9
55.1
by MB
52.1
58.0
59.4
46.4
51.9
58.2
57.4
64.4
57.5
53.9
66.5
70.4
65.7
67.8
56.2
56.0
53.8
57.5
58.5
56.8
60.2
56.6
62.4
56.5
56.5
57.2
56.8
56.7
67.5
56.2
Std. Dev., % AS
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
3.4
3.3
3.1
3.0
3.8
2.8
2.2
3.2
3.0
2.1
4.2
2.4
by MB
NA
NA
NA
. NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
6.1
3.7
7.4
6.6
5.4
4.0
3.3
0.7
5.4
8.7
2.0
1.7
Total mold
surface, m2
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
L 2.28
2.28
2.28
2.28
2.28
2.28
228
2.28
2.28
2.28
m
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
m
00
Date
Time
Test run #
GELCOAT EXPERIMENT
6/14/95
6/14/95
6/14/95
6/15/95
6/15/95
6/15/95
6/16/95
6/16/95
6/16/95
6/19/95
6/19/95
6/19/95
6/21/95
6/21/95
6/21/95
6/22/95
6/22/95
6/22/95
10:11
13:46
16:06
11:38
14:03
15:59
10:23
12:49
15:33
10:36
12:47
15.23
10:46
13:18
16:27
11:14
13:41
15:55
Average (18 runs)
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10
G11
G12
G13
61 4
G15 .
G16
G17
G18
GF1 -Regular get coat (9 runs)
Material/
container #
,
GF1#2
GF1 #2/#3
GF1 #2/#3
GF2#1
GF1 #3
GF1#3
GF2#1
GF1 #3/#4
GF1 #3/#4
GF1#4
GF2#1/#2
GF2#1/#2
GF2#2
GF2#2
GF1JW
GF2 #2/#3
GF2#2/#3
GF2#3
GF2-Low VOC gel coat (9 runs)
. .
GE1-AAA ext mix gun (6 runs)
GE2-HVLP int mix gun (6 runs)
GE3-HVLP ext mix gun (6 runs)
GF1/GE1 (3 runs)
GF1/GE2 (3 runs)
GF1/GE3 (3 runs)
GF2/GE1 (3 runs)
GF2/GE2 (3 runs)
GF2/GE3 (3 runs)
Equip.
GE3
GE2
GE1
GE1
GE1
GE1
GE3
GE3
GE2
GE3
GE2
GE2
GE2
GE1
GE2
GE1
GE3
GE3
EF,ฃ
byTHC
174
174
180
120
177
170
119
168
169
169
129
109
128
117
149
120
131
124
146
170
122
147
143
147
176
164
170
119
122
124
/m2
by MB
153
178
177
100
156
180
120
207
175
158
144
152
164
146
170
132
136
125
154
173
135
149
164
150
171
174
173
126
153
127
Std. Dev., g/m2
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
25.1
8.4
6.4
28.4
23.2
23.1
4.2
10.9
2.5
1.3
9.1
4.8
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
25.1
15.7
17.9
27.5
12.4
29.1
10.8
3.5
24.5
19.3
8.4
6.6
EF, g/g material
byTHC
0.229
0.218
0.233
0.141
0.228
0.213
0.144
0.202
0.216
0.223
0.152
0.128
0.130
0.138
0.191
0.129
0.131
0.144
O.T77
0.217
0.137
0.180
0.172
0.179
0.225
0.208
0.218
0.136
0.137
0.140
by MB
0.202
0.225
0.230
0.118
0.201
0.226
0.146
0.250
0.223
0.209
0.169
0.178
0.167
0.172
0.218
L 0.142
0.136
0.146
0.186
0.220
0.152
0.181
0.196
0.181
0.219
0.222
0.220
0.144
0.171
0.142
Std. Dev., g/g
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
L_ NA
NA
0.041
0.013
0.008
0.045
0.038
0.040
0.009
0.012
0.012
0.005
0.011
0.006
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
0.038
0.014
0.019
0.042
0.026
0.042
0.013
0.003
0.021
0.022
0.005
0.004
Catalyst
Ratio, %
1.86
1.88
1.82
1.68
1.92
1.87
1.65
1.80
1.82
1.80
1.86
1.75
1.87
1.86
1.89
1.80
1.71
1.78
1.81
1.85
1.77
1.83
1.85
1.77
1.87
1.86
1.82
1.78
1.83
1.71
Glass/Resin
Ratio. %
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
RESIN EXPERIMENT
6/23/95
6/23/95
6/23/95
6/26/95
6/26/95
6/26/95
6/27/95
6/27/95
6/27/95
6/28/95
6/28/95
6/28/95
6/29/95
6/29/95
6/29/95
6/30/95
6/30/95
6/30/95
7/5/95
7/5/95
7/5/95
7/6/95
7/6/95
7/6/95
7/7/95
11:06
14:30
16:35
10:42
13:27
16:05
10:46
13:14
15:51
10:21
12:45
15:20
10:22
12:20
14:36
10:14
13:13
15:51
11:53
14:02
16:19
10:12
14:34
16:41
10:10
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
RIB
R19
R20
R21
R22
R23
R24
R25
Material/
container #
RF6#1
RF1#3
RF1#3
RF1#4
RF1#5
RF3#1
RF1 #5*6
RF2HM
RF4#1
RF1#6
RF6#1/2
RF6 #1/2/3
RF2#2
RF2#2
RF1#6/7
RF1#4/8
RF402
RF1 #4/8
RF3#2
RF3#2/3
RF1#8/7
RF1#8/#9
RF4#3
RF1#8/ป9
RF4#4
Average (25 runs, w "Normal" run H 10)
Average (24 runs, w/o "Normal" run RIO)
I- , .
Equip.
RE1
RE2
RE3
RES
RE1
RE1
RE1
RE1
RE4
RE1
RE1
RE1
RE1
RE1
RE1
RE3
RE4
RE2
RE1
RE1
RE1
RE1
RE4
RE2
RE4
RE1/RF1-AAA ext mix gun (6 runs, w/ run R10)
RE1 /RF1 -AAA ext mix gun (5 runs, w/o run R1 0)
RE2-Flow coater (3 runs) |
RES-Pressure-fed roller (3 runs)
RE4-AAA ext mix gun for BPO system (see RF4)
RF2-L.OW styrene Resin (3 runs)
RF3-Styrene suppressed resin (3 runs)
RF4-BPO-Catalyzed Resin (4 funs)
RF4-BPO-Catalyzed. Resin (2 runs, slow gel)
RF4-BPO-Catalyzed Resin (2 runs, fast gel)
RFS-Water emulsified resin (Not tested)
RF6-Styrene suppressed resin plus wax (3 runs)
Application
Method
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Normal
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Controlled
Air Vel.
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Avg. sqrt
(delta P)
0.3478
0.3464
0.3316
0.3401
0.3366
0.3366
0.3427
0.3391
0.3295
0.3369
0.3354
0.3369
0.3403
0.3366
0.3438
0.3366
0.3380
0.3346
0.3440
0.3354
0.3327
0.3415
0:3452
0.3384
0.3451
0.3388
0.3389
0.3391
0.3395
0.3392
0.3361
0.3387
0.3387
0.3395
0.3338
0.3452
NA
0.3400
Exhaust flow
. rate, cfm
8710
8683
8399
8563
8495
8495
8613
8543
8358
8501
8471
8501
8566
8495
8633
8495
8521
8457
8637
8471
8419
8589
8661
8492
8658
8537
8539
8542
8550
8544
8486
8535
8534
8550
8440
8659
NA
8561
Avg. cone.
ppm
3.21
3.97
4.15
3.63
6.66
4.08
5.83
5.59
5.21
7.95
3.22
2.55
5.28
5.43
5.97
4.51
7.61
4.52
3.92
3.91
6.32
6.15
5.85
4.38
6.15
5.04
4.92
6.48
6.18
4.29
L 4.10
5.44
3.97
6.20
6.41
6.00
NA
2.99
Background
cone., ppm
0.35
0.45
0.46
0.45
0.39
0.43
0.77
0.43
0.46
0.46
0.42
0.36
0.27
0.40
0.37
0.49
0.51
0.44
0.29
0.41
0.45
0.41
0.30
0.43
0.62
0.43
0.43
0.48
0.48
0.44
0.46
0.37
0.38
0.47
0.48
0.46
NA
0.38
Avg. net
cone., ppm
2.86
3.52
3.70
3.17
6.27
3.65
5.05
5.16
4.76
7.49
2.81
2.18
5.01
5.04
5.59
4.03
7.10
4.07
3.63
3.51
5.87
5.74
5.55
3.96
5.53
4.61
4.49
6.00
5.71
3.85
3.63
5.07
3.59
5.73
5.93
5.54
NA
2.62
Test run
duration, min
103.1
81.3
75.0
83.8
67.3
78.0
75.7
71.8
152.4
81.5
91.3
95.2
76.5
75.1
76.5
79.1
102.6
69.3
75.6
74.7
77.2
75.9
88.9
78.0
85.1
83.6
83.7
75.7
74.5
76.2
79.3
74.5
76.1
107.2
127.5
87.0
NA
96.5
Total emissions
byTHC.g
316
305
286
279
440
296
404
389
743
636
267
217
403
394
453
332
762
293
291
272
468
459
524
322
500
402
392
477
445
306
299
395
286
632
752
512
NA
266
m
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
RESIN EXPERIMENT
6/23/95
6/23/95
6/23/95
6/26/95
6/26/95
6/26/95
6/27/95
6/27/95
6/27/95
6/28/95
6/28/95
6/28/95
6/29/95
6/29/95
6/29/95
6/30/95
6/30/95
6/30/95
7/5/95
7/5/95
7/5/95
7/6/95
7/6/95
7/6/95
7/7/95
11:06
14:30
16:35
10:42
13:27
16:05
10:46
13:14
15:51
10:21
12:45
15:20
10:22
12:20
14:36
10:14
13:13
15:51
11:53
14:02
16:19
10.12
14:34
16:41
10.10
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
Material/
container tf
RF6#1
RF1#3
RF1#3
RF1#4
RF1#5
RF3#1
RF1 #5/#6
RF2#1
RF4#1
RF1#6
RF6#1/2
RF6 #1/2/3
RF2#2
RF2#2
RF1 #6/7
RF1 #4/8
RF4#2
RF1 #4/8
RF3#2
RF3#2/3
RF1 #6/7
RF1 #8/#9
RF4#3
RF,1 #8/#9
RF4#4
Average (25 runs; w "Normal" run RIO)
Average (24 runs, w/o "Normal" run R10)
Equip.
RE1
RE2
RE3
RE3
RE1
RE1
RE1
RE1
RE4
RE1
RE1
RE1
RE1
RE1
RE1
RE3
RE4
RE2
RE1
RE1
RE1
RE1
RE4
RE2
RE4
RE1/RF1-AAA ext mix gun (6 runs, w/ run RIO)
RE1/RF1-AAA ext mix gun (5 runs, w/o run R10)
RE2-Flow coater (3 runs)
RE3-Pressure-fed roller (3 runs)
RE4-AAA ext mix gun for BPO system (see RF4)
RF2-Low styrene Resin (3 runs)
RF3-Styrene suppressed resin (3 runs)
RF4-BPO-Catalyzed Resin (4 runs)
RF4-BPO-Catalyzed Resin (2 runs, Slow gel)
RF4^BPO-Catalyzed Resin (2 runs, fast gel)
RF5-Water emulsified resin (Not tested)
RF6-Styren* suppressed resin plus wax (3 runs)
Avg. spraying
cone., ppm
40.88
7.36
7.31
6.77
58.52
56.47
65.93
56.12
73.79
88.66
44.67
32.34
52.43
51.23
60.23
8.08
56.67
9.21
53.18
57.64
57.03
58.25
50.51
8.82
58.01
44.81
42.98
64.77
59.99
8.46
7.39
53.26
55.77
59.74
65.23
54.26
NA
39.30
Avg. net
spraying, ppm
40.53
6.91
6.86
6.32
58.13
56.04
65.16
55.69
73.34
88.20
44.26
31.98
52.16
50.83
59.86
7.60
56.16
8.77
52.89
57.24
56.56
57.83
50;21
8.39
57.39
44.37
42.55
64.29
59.51
8.02
6.93
52.90
55.39
59.27
64.75
53:80
NA
38.92
Spraying
time, min
4.5
23.0
19.2
21.4
3.5
3.3
3.0
3.3
3.5
3.4
3.3
2.9
3.5
3.6
3.6
27.9
6!1
13.9
3.0
3.0
3.7
3.4
3.6
15.3
3.2
7.5
7.7
3.4
3.4
17.4
22.8
3.5
3.1
4.1
4.8
3.4
NA
3.6
Emissions
during
application
(by THC)
196
169
136
142
212
193
207
195
266
310
153
97
192
191
228
221
356
126
166
180
216
207
192
134
193
195
190
230
214
143
166
192
180
252
311
193
NA
149
Std. Dev. of
application
emissions
(by THC)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
55
55
36
8
19
39
2
11
67
45
0
NA
41
Wt. loss from
part during
curing stage
(by MB)
71
92
135
107
193
97
187
162
451
250
85
76
175
174
204
121
314
166
78
85
210
216
262
189
262
174
171
2.10
202
149
121
170
87
322
383
262
NA
77
Std. Dev. of
curing loss
from part
(by MB)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
87
87
20
11
41
11
6
8
77
69
0
NA
6
Wt. loss during
application and from
overspray, g
(by Md)
213
243
132
173
241
205
237
199
243
357
162
127
199
192
238
215
418
99
184
188
218
215
226
100
226
210
204
2ง1
230
147
173
197
192
278
331
226
NA
167
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
Date
Time
Test run #
RESIN EXPERIMENT
6/23/95
6/23/95
mam
6/28/95
6/26/95
6/26/95
6/27/95
6/27/95
6/27/95
6/28/95
6/28/95
6/28/i5
6729/95
6/29/95
6/28/95
6/30/95
6/30/95
6/30/95
7/5/95
7/S/95
7/5/95
7/6/95
7/6/95
7/8/9S
7/7/95
11:06
14:30
16:35
10:42
13:27
16:05
10:46
13:14
15:51
10:21
12:45
15:20
10:22
12:20
14:36
10:14
13:13
15:51
11:53
14:02
16:19
10:12
14:34
16:41
10:10
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
Material/
container #
RF6*1
RF1i3
RF1#3
RF1#4
RF1#5
RF3S1
RF1#S/#6
RF2HH
RF4#1
RF1#6
RF6#1/2
RF6 #1/2/3
RF2#2
iRF2#2
lRF1#6/7
RF1#4/8
RF4#2
RF1 #4/8
RF3#2
RF3#2/3
RF1W7
RF1#8/#9
RF4#3
RF1ซWปi
RF4#4
Average (25 runs, w "Normar run R10)
Average (24 runs, w/o "Normar run R10)
Equip.
RE1
RE2
RE3
RE3
RE1
RE1
RE1
RE1
RE4
RE1
RE1
RE1
RE1
RE1
RE1
RE3
RE4
RE2
RE1
RE1
RE1
RE1
RE4
RE2
RE4
RE1/RF1 -AAA ext mix gun (6 runs, w/ run R10)
RE1/RF1-AAA ext mix gun (5 runs, w/o run R10)
RE2-Flow coaler (3 runs)
RE3-Pressure-fed roller (3 runs)
RE4-AAA ext mix gun for BPO system (see RF4)
RF2-Low styrene Resin (3 runs)
RF3-Styrene suppressed resin (3 runs)
RF4-BPO-Catalyzed Resin (4 runs)
RF4*PO-estaIyzed Resin (2 runs, stow gel)
RF4-BPO-Catatyzed Resin (2 runs, fast gel)
RF5-Watef emulsified resin (Not tested)
RF6-Styrene suppressed resin plus wax (3 runs)
Total emissions
by MB, g
284
335
267
280
434
302
424
361
694
607
247
203
374
366
442
336
732
265
262
273
428
431
48S
289
488
384
375
461
432
296
294
367
279
601
713
488
NA
245
Ratio of two
measurement
methods
(MB/THC)
0,90
1,10
0.93
1,00
0.99
1.02
1.05
0.93
0.93
0.95
0.93
.0.94
0.93
0.93
0.98
1.01
0.96
090
0.90
1.00
0.91
0.94
0.93
0.90
0.98
0.96
0.96
0.97
0.97
0.97
0.98
0.93
0.97
0.95
. 0,95
0.95
NA
0.92
Material
used, g
7712
5445
4919
5041
5978
5663
6160
5979
6116
6133
5152
4872
6566
6870
7000
5328
7467
. 5371
6689
6423
6958
7256
5606
6040
5698
6098
6096
6581
6670
5619
5096
6472
6258
6222
6792
5652
NA
5912
Transfer
eff. %
92.7
96.2
97.2
97.3
90.8
91.7
91.0
91.4
88.7
80.5
89.9
92.7
91.4
93.2
92.6
96.7
79.5
97.7
92.9
92.9
92.2
93.5
91.7
97.9
91.9
92.2
92.7
90.1
92.0
97.3
97.1
92.0
92.5
88.0
84.1
91.8
NA
91.8
Styrene
content, %
43.29
38.33
38:33
38.33
38.33
43.45
38.33
35.34
42.61
38.33
43.29
43.29
35.34
35.34
38.33
38J3
42.61
38.33
43.45
4345
38:33
38.33
42.6f
38.33
42.61
31.87
39.93
38.33
38.33
38.33
38,33
35.34
43.45
42.61
42.61
. 42.61
NA
43.29
I
EF, % available styrene
byTHC
9.5
14.6
15.2
14.5
19.2
12.0
17.1
18.4
28.5
27,1
12.0
10.3
.17.4
16.2
16.9
16.3
23.9
14.2
10.0
9.8
17.5
16.5
21.9
13.9
20.6
16.5
16.1
19.1
17.5
14.2
L 15.3
17.3
10.6
23.7
26.2
21.3
NA
10.6
by MB
8.5
16.1
14.2
14.5
18.9
12.3
18.0
17.1
26.6
25.8
11.1
9.6
16.1
15.1
16.5
16.5
23.0
12.9
9.0
. 9.8
16.0
15.5
20.4
12.5
20.1
15.8
15.4
18.5
17.0
13.8
15.0
16.1
10.4
22.5
24.8
20.3
NA
9.7
Std. Dev., % AS
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
4.9
4.9
3.7
0.9
0.3
0.7
0.9
1.0
3.0
2.3
0.7
NA
1.0
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
4.7
4.7
3.5
1.3
1.6
1.0
0.8
1.4
2.6
1.8
0.2
NA
1.1
Total mold
surface, m2
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
2.28
NA
2.28
m
SUMNEW.XLS
-------
Emission Measurements, Gravimetric Measurements, and Calculated Emissions and Factors
m
Date
Time
Test run #
RESIN EXPERIMENT
6/23/95
6/23/95
6123/95
6/26/95
6/26/95
6/26/95
6/27/95
6/27/95
6/27/95
6/28/95
6/28/95
6/28/95
6/29/95
6/29/95
6/29/95
6/30/95
6/30/95
6/30/95
7/5/95
7/5/95
7/5/95
7/6/95
7/6/95
7/6/95
7/7/95
11:06
14:30
16:35
10:42
13:27
16:05
10:46
13:14
15:51
10:21
12:45
15:20
10:22
12:20
14:36
10:14
13:13
15:51
11:53
14:02
16:19
10:12
14:34
16:41
10:10
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
R21
R22
R23
R24
R25
Material/
container #
RF6#1
RF1#3
RF1#3
RF1#4
RF1#5
RF3HM
RF1 #5/06
RF2#1
RF4#1
RF1#6
RF6HM/2
RF6 #1/2/3
RF2#2
RF2#2
RF1 #6/7
RF1#4/8
RF4#2
RFKM/8
RF3#2
RF302/3
RF1#677
RF1,#8/#9
RF4#3
RF1 #8/#9
RF4*4
Average (25 run's,, w "Normal" run R1 0)
Average (24 runs, w/o "Normal" run R 10)
Equip.
RE1
RE2
RE3
RE3
RE1
RE1
RE1
RE1
RE4
RE1
RE1
RE1
RE1
RE1
RE1
RES
RE4
RE2
RE1
RE1
RE1
RE1
RE4
RE2
RE4
RE1/RF1-AAA ซxt mix gun (6 runs, w/ run R10)
RE1/RF1-AAA ext mix gun (5 runs, w/o run R10)
RE2-Flow coaler (3 runs)
RE3-Pressure-fed roller (3 runs)
RE4-AAA ext mix gun for BPO system (see RF4)
RF2-Low styrene Resin (3 runs)
RF3-Styrene suppressed resin (3 runs)
RF4-BPO-Catalyzed Resin (4 runs)
RF4-BPO-Catalyzed Resin (2 runs, slow gel)
RF4-BPO-Catalyzed Resin (2 runs, last gel)
RF5-Water emulsified resin (Not tested)
RF6-Styrene suppressed resin plus wax (3 runs)
I
EF.c
byTHC
138
134
125
123
193
130
177
170
326
279
117
95
177
173
199
146
334
129
127
119
205
201
230
141
219
175
172
209
195
134
131
173
126
277
330
225
NA
117
I/m2
by MB
125
147
117
123
190
132
186
158
304
266
108
89
164
161
194
147
321
116
115
120
188
189
214
127
214
167
165
202
189
130
129
161
122
263
313
214
NA
107
1
Std. Dev., g/m2
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
,_ NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
62.0
62.0
32.5
9.8
5.1
10.3
17
4.5
52.9
4.0
5.4
NA
17.7
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
58.8
58.8
28.7
2.7
12.7
13.1
2.3
7.4
49.7
8.3
0.0
NA
14.5
I
EF, g/g material
byTHC
0.041
0.056
0.058
0.055
0.074
0.052
0.066
0.065
0.122
0.104
0.052
0.044
0.061
. 0.057
0.065
. 0.062
0.102
0.055
0.043
0.042
0.067
0.063
0.094
0.053
0.088
0.065
0.064
0,073
0.067
0.055
0.059
0.061
0.046
0.101
0.112
0.091
NA
0.046
by MB
0.037
0.062
0.054
0.056
0.073
0.053
0.069
0.060
0.113
0.099
0.048
0.042
0.057
0.053
0.063
0.063
0.098
0.049
0.039
0.043
0.062
0.059
0.087
0.048
0.086
0.062
0.061
0.071
0.065
0.053
0.058
0.057
0.045
0.096
0.106
0.086
NA
6.042
I
Std. Dev., g/g
byTHC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
. NA
NA
NA
NA
0.020
0.020
0.014
0.004
0.001
0.003
0.003
0.004
0.013
0.010
0.003
NA
0.004
by MB
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
0.019
0.019
0.013
0.005
0.006
0.004
0.003
0.006
0.011
0.008
0.001
NA
0.005
Catalyst
Ratio, %
1.50
1.45
1.65
1.61
1.46
1.45
1.46
1.47
1.67
1.45
1.44
1.48
1.42
1.41
1.47
1.69
2.50
1.51
1.44
1.46
1.45
1.47
3.12
1.52
3.09
1.67
1.67
1,46
1.46
1.49
1.65
1.43
1.45
2.60
2.09
3.11
NA
1.47
Glass/Resin
Ratio, %
29.2
32.0
33.6
33.8
33.0
34.6
32.4
33.2
34.3
32.2
33.4
33.3
34.8
32.9
32.6
32.4
35.4
32.7
30.0
31.9
32:7
31.1
37.5
30.6
38.1
33.1
33.1
32.3
32.4
31.8
33.3
33.6
32.2
36.3
34.9
37.8
NA
32.0
SUMNEW.XLS
-------
Appendix F
Statistical Analyses of Test Results
Table of Contents
Section Page No.
F.I Pilot Experiment F-l
F.2 Gelcoat Experiment F-8
F.3 Resin Experiment F-l5
F.4 Summary F-23
F-i
-------
List of Tables
Table No. Page No.
1 Analysis of variance results for pilot study F-3
2 Estimated mean differences and 95% confidence interval estimates
for normal vs. controlled application methods in the pilot experiment F-4
3 Confidence limits for pilot study: high vs. low air velocity F-5
4 Within-cell standard deviationspilot study data F-5
5 Analysis of covariance results for pilot study F-7
6 Analysis of variance results for gelcoat experiment F-9
7 Confidence limits for gelcoat study: regular vs. low VOC gelcoats F-10
8 Confidence limits for gelcoat study: different spray guns F-l 1
9 Within-cell standard deviationsgelcoat experiment F-l2
10 Analysis of covariance results for gelcoat experiment F-l 4
11 Analysis of variance results for resin experiment F-l 7
12 Estimated mean differences and confidence interval estimates for percent
available styrene: resin experiment F-18
13 Estimated mean differences and confidence interval estimates for emission
factor (g/m2): resin experiment F-l 9
14 Estimated mean differences and confidence interval estimates for emission
factor (g/g): resin experiment F-20
15 Within-treatment standard deviationsresin experiment F-21
16 Analysis of covariance results for resin experiment F-22
17 Summary of results of specific hypothesis tests F-23
F-ii
-------
F.I Pilot Experiment.
Before executing the main experiment, a pilot study was conducted to help establish
"standard" conditions under which the subsequent experiments should be conducted. A
secondary purpose was to gain insight into the magnitude of measurement error variability that
might be anticipated in the main experiment. The pilot study was run as a factorial experiment
involving two factors: M = spraying method, and A = air velocity. The experiment consisted of
12 trials namely, three replicates for each of the following conditions:
(a)
(b)
(c)
(d)
Normal spraying method (Ml), low air velocity (Al)
Controlled spraying method (M2), low air velocity (Al)
Normal spraying method (Ml), high air velocity (A2)
Controlled spraying method (M2), high air velocity (A2)
A single gelcoat formulation (GF1) and a single application equipment (GE1) were used
throughout the pilot experiment. The 12 trials were run in a random order.
Although it was recognized that the pilot study would provide only a limited amount of
statistical information, the resultant data were subjected to statistical analysis in order to get some
idea of the impact of the two factors on percent available styrene (%AS) and other related
measures. The specific outcome measures that were analyzed are the following:
Variable
X
Zl
Z2
Yl
Y2
Y3
Y4
Y5
Y6
Y7
Y8
Description
Transfer Efficiency (%) (standardized to have mean 0)
Total Emissions (g), by THC
Total Emissions (g), by MB
Emissions during the application stage (g), by THC
Curing Loss from Part (g), by MB
EF (% Available Styrene), by THC
EF (% Available Styrene), by MB
EF (g/m2), by THC
EF (g/m2), by MB
EF(g/g),byTHC
EF(g/g),byMB
-------
The x variable was considered a possible coovariate for the Zl and Z2 variables; as a result, the
X variable was standardized to have a zero mean.
For each variable, an initial analysis of variance (ANOVA) of the following form was performed:
SOURCE OF DEGREES OF
VARIATION FREEDOM
M (methods) 1
A (air flow) 1
M x A (interaction) 1
Error 8
Total 11
In each case, the interaction term appeared statistically nonsignificant; hence the interaction term
was dropped to produce an ANOVA as follows:
SOURCE OF DEGREES OF
VARIATION FREEDOM
M (methods) 1
A (air flow) 1
Error 9
Total 11
The results of the above analyses are summarized in Table 1 for the X variable and each
of the Ys (the two Z variables are considered subsequently). The upper portion of the table gives
means of the pertinent variables for each combination of the M and A factors. These rows are
followed by rows giving the means separately for each level of each factor. The lower portion of
the table provides information relating to the statistical tests. The first row gives the foot mean
squared error (RMSE); this is equivalent to the pooled within-cell variance. This is used to test
whether, there is a METHOD by AIR VELOCITY interaction. The test results, summarized in
the next row, show these interactions to be statistically nonsignificant (n.s.). Consequently, a
new RMSE for testing the main effects of the factors is constructed (by pooling of the former
RMSE and the interaction mean square); these RMSEs, based on 9 degrees of freedom (df), are
displayed in the next row of Table 1. Results of the tests for main effects are given in the last
two rows of the table. The effect of AIR VELOCITY is statistically nonsignificant, although the
low velocity exhibited lower estimated values than the high velocity for each of the Y variables.
For the METHOD effect, statistically signficant differences were found for several of the
variables: the controlled spraying, as contrasted with the normal spraying, yielded a higher
transfer efficiency (0.01 level), and lower average spraying emissions, percent available styrene,
F-2
-------
1
TABLE 1. Analysis of Variance Results for Pilot Study: Variables X, Yl, Y2,..., and Y8
Means, by variable:
Appli-
cation Air
method velocity
Hypothesis Testing Results
RMSE for testing
interaction (8 df):
Test for interaction:
RMSE for testing main
effects (9 df):
Test for METHOD:
Test for AIR VELOCITY:
Xa
Controlled
Controlled
Normal
Normal
Controlled
Normal
_
-
High
Low
High
Low
_
-
High
Low
3
3
3
3
6
6
6
6
4.1
4.5
-6.4
-2.2
4.3
-4.3
-1.2
1.2
3.16
n.s.
3.17
**
n.s.
* = statistically significant at the 0.05 level
** = statistically significant at the 0.01 level
*** = statistically significant at the 0.001 level
'Overall mean for X prior to standardization was 79.96.
Yl
Y2 Y3
Y4
Y5
Y6
157.0 223.0 56.3 59.7
137.7 204.0 52.0 58.4
259.0 209.3 62.1 -62.5
231.3 225.7 62.9 59.2
147.3 213.5 54.1 59.1
245.2 217.5 62.5 60.9
208.0 216.2 59.2 61.1
184.5 214.8 57.5 58.8
180.3 190.0
163.3 183.0
225.7 228.0
225.0 211.7
171.8 186.5
225.3 219.8
203.0 209.0
194.2 197.3
.n.s.
n.s. n.s.
n.s.
n.s. n.s.
***
n.s.
n.s.
n.s.
**
n.s.
n.s. n.s.
*** *
n.s. n.s.
Y7
Y8
0.218 0.231
0.201 0.226
0.241 0.242
0.244 0.229
0.210 0.229
0.242 0.236
0.229 0.237
0.223 0.228
31.50 21.74 3.71 7.34 15.76 25.85 0.014 0.028
n.s n.s
29.79 22.89 3.80 6.95 15.59 24.52 0.015 0.027
n.s
n.s. n.s.
-------
and emission factor (EF), as measured by THC. When measured via the mass balance (MB)
approach, only the EF variable appeared statistically significant (at the 0.05 level).
Point estimates and confidence interval estimates for the differences due to application
methods and air velocities are shown in Tables 2 and 3, respectively.
In Tables 2 and 3, the construction of the confidence intervals is based on an assumption
that the data for a given combination of METHOD and AIR VELOCITY are approximately
normally distributed with a common measurement error variability (the same assumption applies
to the hypothesis tests previously described, and similar assumptions apply to all other
confidence intervals described herein). Cell-specific estimates of the measurement error
variablity (or standard deviations) are shown in Table 4. The 95 percent confidence interval on
the difference is related to a pairwise hypothesis test (conducted at a significance level of 0.05) in
that if the estimated confidence interval does not include zero, then the corresponding test of no
difference in mean %AS will be rejected.
TABLE 2. Estimated Mean Differences and 95% Confidence Interval Estmates for
Normal Vs. Controlled Application Methods in the Pilot Experiment
Variable and comparison
X
Zl
Z2
Yl
(g)
Y2
Y3
Y4
Y5
Y6
Y7
Y8
- Transfer efficiency (%)
- Total emissions by THC (g)
- Total emissions by MB (g)
- Application emissions by THC
- Curing loss from part by MB (g)
- EF as % AS by THC
- EFas%ASbyMB
- EF as g/m2 by THC
- EF as g/m2 by MB
-EF(g/g)byTHC
-EF(g/g)byMB
Lower 95%
confidence
limit
-12.79
76.40
3.88
58.92
-25.90
3.44
-7.29
33.14
1.31
0.0134
-0.0282
Esitmated
mean
difference
-8.65
121.83
76.33
97.83
4.00
8.40
1.78
53.50
33.33
0.0325
0.0069
Upper 95%
confidence
limit
-4.51
167.27
148.79
136.75
33.90
13.36
10.86
73.86
65.35
0.0516
0.0420
Significant
**
***
' *
***
n.s
**
n.s
***
*
*
n.s.
* = Statistically significant at the 0.05 level.
** = Statistically significant at the 0.01 level.
= Statistically significant at the 0.001 level.
n.s. = Not statistically significant at the 0.05 level.
F-4
-------
TABLE 3. Confidence Limits for Pilot Study: High Vs. Low Air Velocity
Variable and comparison
X - Transfer efficiency %
Zl - Total emissions by THC (g)
Z2 - Total emissions by MB (g)
Yl - Application emissions by
THC (g)
Y2 - Curing loss from part by MB
(g)
Y3 - EF as % AS by THC
Y4 - EF as % AS by MB
Y5 - EF as g/m2 by THC
Y6 - EF as g/m2 by MB
Y7 - EF (g/g) by THC
Y8 -EF(g/g)byMB
Lower 95%
confidence
limit
-6.45
-25.27
-46.12
-15.41
-28.56
-3.20
-6.73
-11.53
-20.35
-0.0122
-0.0260
Estimated mean
difference
-2.32
20.17
26.33
23.50
1.33
1.77
2.35
8.83
11.67
0.0069
0.0091
Upper 95%
confidence
limit
1.82
65.60
98.79
62.41
31.23
6.73
11.43
29.20.
43.69
0.0260
0.0442
Significant
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s
n.s.
n.s.
n.s.
n.s
n.s = Not statistically significant at the 0.05 level.
TABLE 4. Wi thin-cell Standard Deviations - Pilot Study Data
METHOD AIRVEL
Zl
Z2 Yl
Y2
Y3 Y4 Y5
Y6 Y7
Y8
Controlled
Controlled
Normal
Normal
High
Low
High
Low
2.6
2.7
4.7
2.0
24.7
43.0
30.4
39.3
39.4
38.7
97.0
35.1
19.0
6.4
48.3
35.2
11.1
32.1
23.2
14.2
3.1
1.3
3.2
5.8
9.8
4.3
9.4
3.5
10.8
19.3
13.8
17.7
17.3
17.1
42.9
15.4
0.012 0.038
0.005 0.017
0.012 0.036
0.022 0.013
F-5
-------
For the two total emission variables (Zl and Z2), an initial analysis of covariance
(ANACOVA) of the following form was performed:
SOURCE OF VARIATION DEGREES OF FREEDOM
X (transfer efficiency) 1
M (methods) 1
A (air flow) 1
M x A (interaction) 1
Error 7
Total 11
Since in each case the interaction term appeared statistically nonsignificant, that term was
dropped to produce the following:
SOURCE OF VARIATION DEGREES OF FREEDOM
X (transfer efficiency) 1
M (methods) 1
A (air flow) 1
Error 8
Total 11
The results are summarized in Table 5. The format of the .table is similar to that of Table 1 in
that the upper portion shows means by cells and factor levels and the lower portion gives RMSEs
and results of hypothesis tests. In this case, the hypothesis tests are performed on means of the Z
variables after adjustment for the covariate X (assuming a linear relaionship between a Z and X).
(The ANOVA results for the Z variables (i.e., ignoring adjustments for X) are shown in the lower
left portion of the table while the ANACOVA results are given in the lower right portion.) The
adjusted means are given in the upper righthand part of the table. Total emissions by THC
(variable Zl) appear to be affected (0.05 significance level) by spraying method, while AIR
VELOCITY and X appear to have nonsignificant effects on Zl. Neither factor A or M nor the
covariate'X appears to affect variable Z2. As noted previously (see Table 1), the METHOD
factor appears to affect the covariate X and this situation can lead to problems with
interpretation. For instance, it is not clear whether METHOD affects X only, which in turn
induces differences in the Zl variate, or whether (at least part of) the effect of METHOD on Zl a
direct effect (i.e., not related to the difference in X levels).
F-6
-------
TABLE 5. Analysis of Covariance Results for Pilot Study: Variables Zl and Z2
MEANS
MEANS ADJUSTED FOR X
METHOD AIRVEL
N
Zl
Z2
Zl
Z2
Controlled High 3
Controlled Low 3
Normal High 3
Normal Low 3
Controlled - 6
Normal - 6
High 6
Low 6
Hypothesis Testing Results:
RMSE for testing
interaction (8 df):
(7 dp:
Test for interaction:
RMSE for testing main
effects (9 df):
(and covariate) (8 df):
Test for METHOD:
Test for AIR VELOCITY:
Test for covariate X:
4.1 410.3
4.5 371.7
-6.4 513.7
-2.2 512.0
4.3 391.0
-4.3 512.8
-1.2 462.0
1.2 441.8
35.1
n.s.
4.8
***'
n.s.
432.3
416.3
519.0
482.3
424.3
500.7
475.7
449.3
58.5
n.s.
55.5
*
n.s.
413.7
375.4
508.4
510.2
389.1
514.7
462.5
441.3
37.4
n.s.
36.9
*
n.s.
n.s.
477.9
466.8
447.7
457.6
469.5
455.5
463.6
461.4
50.1
n.s.
47.2
n.s.
n.s.
n.s.
* = statistically significant at the 0.05 level
** = statistically significant at the 0.01 level
*** = statistically significant at the 0.001 level
F-7
-------
F.2 Gelcoat Experiment.
This experiment is aimed at evaluating how styrene emissions are affected by gelcoat
formulation (factor GF) and type of equipment (factor GE). The factor combinations (two
formulations and three equipment types) were each run in triplicate, with trials run in a random
order (i.e., a 2x3 factorial experiment with three replications embedded in a completely random
design). The analysis, which was conducted for the X variate and each of the previously defined
Y variates, entailed conducting an analysis of variance (ANOVA) having the following structure:
SOURCE OF VARIATION DEGREES OF FREEDOM
GF(gelcoats) 1 .
GE (equipments) 2
GF x GE (interaction) 2
Error 12
Total . 17
In each case, the interaction term appeared statistically nonsignificant; hence the interaction term
was dropped to produce an ANOVA as follows:
SOURCE OF VARIATION DEGREES OF FREEDOM
GF (gelcoats) 1
GE. (equipments) 2
Error 14
Total 17
The results are summarized in Table 6. The format of this table is similar to that described for
Table 1: the upper portion gives means of the pertinent variables for each combination of the GF
and GE factors, the middle portion shows the means separately for each level of each factor, and
the lower portion gives the pertinent RMSEs and hypothesis testing results. With one exception
(a difference in variable Yl for GE2 and GE3, at the 0.05 level of significance), the effect of
EQUIPMENT is statistically nonsignificant. For the gelcoat FORMULATIONS effect,
statistically signficant differences were found for the X variable (0.05 level) and for several of
the Y variables (Yl, Y2, Y5, and Y6, all highly significant). Statistically significant differences
for the GF and GE factors were not found for the percent available styrene variables (Y3 and
Y4).
Point estimates and confidence interval estimates for the differences due to two gel coat
formulations and three spray guns are shown in Tables 7 and 8, respectively. The with-in cell
standard deviations for gel coat experiment are shown in Table 9.
F-8
-------
TABLE 6. Analysis of Variance Results for Gelcoat Experiment: Variables X, Yl, Y2,..., Y8
MEANS
GF
GF1
GF1
GF1
GF2
GF2
GF2
GF1
GF2
_
.
-
EQUIP
GE1
GE2
GE3
GE1
GE2
GE3
.
-
GE1
GE2
GE3
N
3
3
3
3
3
3
9
9
6
6
6
Xa
-1.5
-1.6
-0.5
2.5
-1.5
2.6
-1.2
1.2
0.5
-1.5
1.1
Yl
151.3
160.3
143.7
94.0
100.0
91.0
151.8
95.0
122.7
130.2x
117.3x
Y2
212.0
204.0
215.0
157.7
159.7
159.3
210.3
158.9
184.8
181.8
187.2
Y3
58.0
53.8
56.3 .
53.8
53.9
55.1
56.0.
54.2
55.9
53.8
55.7
Y4
56.5
57.2
56.8
56.7
67.5
56.2
60.2
56.8
56.6
62.4
56.5
Y5
175.7
164.0
170.3
119.0
122.0
124.7
170.0
121.9
147.3
143.0
147.5
Y6
171.0
174.3
172.7
126.0
153.3
127.0
172.7
135.4
148.5
163.8
149.8
Y7
0.225
0.208
0.218
0.136
0.137
0.140
0.217
0.137
0.180
0.179
0.172
Y8
0.219
0.222
0.220
O.J44
0.171
0.142
0.220
0.152
0.181
0.196
0.181
Hypothesis Testing Results:
RMSE for testing
interaction (12 df):
Test for interaction:
RMSE for testing main
effects (14 df):
Test for FORMULATIONS:
Test for EQUIPMENT:
2.26
n.s.
2.28
n.s.
8.39
n.s.
7.97
***
4.95
n.s.
5.41
***
n.s.
3.62
n.s.
3.51
n.s.
n.s.
5.58
n.s.
5.88
n.s.
n.s.
7.94 17.33
n.s.
n.s.
8.15 17.31
n.s.
***
n.s.
* = stalistically significant at the 0.05 level
** = statistically significant at Ihe 0.01 level
*** = statistically significant at the 0.001 level
"Overall mean for X prior to standardization was 83.56
0.012
n.s.
0.011
***
n.s.
0.017
n.s.
0.017
***
n.s.
-------
TABLE 7. Confidence Limits for Gelcoat Study: Regular Vs. Low Voc Gelcoats
Variable and comparison
X -
Zl
Z2
Yl
(g)
Y2
Y3
Y4
Y5
Y6
Y7
Y8
Transfer efficiency %
- Total emissions by THC (g)
- Total emissions by MB (g)
- Application emissions by THC
- Curing loss from part by MB (g)
- EF as % AS by THC
- EF as % AS by MB
- EF as g/m2 by THC
- EF as g/m2 by MB
-EF(g/g)byTHC
-EF(g/g)byMB
Lower 95%
confidence limit
-4.70
90.44
46.63
48.72
45.98
-1.76
-9.27
39.87
19.72
0.0680
0.0505
Estimated
mean
difference
-2.39
109.00
86.78
56.78
51.44
1.79
-3.32
48.11
37.22
0.0796
0.0678
Upper 95%
confidence
limit
-0.08
127.56
126.93
64.83
56.91
5.34
2.62
56.36
54.73
0.0912
0.0851
Significant
*
***
***
*** '
***
n.s.
n.s.
***
***
***
***
* = Statistically significant at the 0.05 level.
** = Statistically significant at the 0.01 level.
*** = Statistically significant at the 0.001 level.
n.s. = Not statistically significant at the 0.05 level.
F-10
-------
TABLE 8. Confidence Limits for Gelcoat Study: Different Spray Guns
Variable and comparison
Lower 95%
confidence
limit
Esitmated
mean
difference
Upper 95%
confidence
limit
Significant
Comparison between AAA vs. HVLP (internal mix)
X - Transfer efficiency %
Zl - Total emissions by THC (g)
Z2 - Total emissions by MB (g)
Yl - Application emissions by THC (g)
Y2 - Curing loss from part by MB (g)
Y3 - EF as % AS by THC
Y4 -EFas%ASbyMB
Y5 - EF as g/m2 by THC
Y6 - EF as g/m2 by MB
Y7 -EF(g/g)byTHC
Y8 -EF(g/g)byMB
-0.79
-12.90
-83.84
-17.37
-3.69
-2.28 .
-13.05
-5.76
-36.77
:0.0063
-0.0363
Comparison between AAA vs. HVLP (external mix)
X - Transfer efficiency %
Zl - Total emissions by THC (g)
Z2 - Total emissions by MB (g)
Yl - Application emissions by THC (g)
Y2 - Curing loss from part by MB (g)
Y3 - EF as % AS by THC
Y4 - EF as % AS by MB
Y5 - EF as g/m2 by THC
-3.39
-22.56
-52.34
-4.53
-9.03
-4.15
-7.18
-10.26
2.03
9.83
-34.67
-7.50
3.00
2.07
-5.77
4.33
-15.33
0.0080
-0.0151
-0.57
0.17
-3.17
5.33
-2.33
0.20
0.10
-0.17
4.86
32.56
14.50
2.37
9.69
6.41
1.51
14.43
6.11
0.0222.
0.0062
2.26
22.90
46.00
15.20
4.36
4.55
7.38
9.93
n.s.
n.s.
n.s. .
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s. .
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
F-ll
(cont.)
-------
TABLE 8. Continued
Variable and comparison
Y6
Y7
Y8
- EF as g/m2 by MB
- EF (g/g) by THC
- EF (g/g) by MB
Lower 95%
confidence
limit
-22.77
-0.0127
-0.0212
Esitmated
mean
difference
-1.33
0.0016
0.0000
Upper 95%
confidence
limit
20.11
0.0158
0.0213
Significant
n.s.
n.s
n.s.
Comparison between HVLP (int) vs. HVLP (external mix)
X -
Zl
Z2
Yl
Y2
Y3
Y4
Y5
Y6
Y7
Y8
Transfer efficiency %
- Total emissions by THC (g)
- Total emissions by MB (g)
- Application emissions by THC (g)
- Curing loss from part by MB (g)
- EF as % AS by THC
- EF as % AS by MB
- EF as g/m2 by THC
- EF as g/m2 by MB
-EF (g/g) by THC
- EF (g/g) by MB
-5.43
-32.40
-17.67
2.97
-12.03
-6.21
-1.41
-14.60
-7.44
-0.0206
-0.0061
-2.60
-9.67
31.50
12.83
-5.33
-1.87
5.87
-4.50
14.00
-0.0064
0.0151
0.23
13.06
80.67
22.70
1.36
2.48
13.15
5.60
35.44 _
0.0078
0.0363
n.s.
n.s.
n.s.
*
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
* = Statistically significant at the 0.05 level.
n.s. = Not statistically significant at the 0.05 level.
TABLE 9. Within-cell Standard Deviations - Gelcoat Experiment
GF EQUIP X Zl Z2 Yl Y2 Y3 Y4 Y5 Y6
Y7 Y8
GF1
GF1
GF1
GF2
GF2
GF2
GE1
GE2
GE3
GE1
GE2
GE3
2.7
0.5
2.7
3.6
1.1
1.0
11.4
30.1
7.0
3.8
25.2
13.6
30.2
9.5
68.4
54.1
23.6
18.8
8.3
4.7
10.0
4.4
10.4
10.1
8.0
2.6
4.6
6.4
3.1
2.1
2.8
4.0
3.7
2.6
5.1
2.9
4.0
0.9
6.6 .
10.7
2.5
2.1
5.1
13.2
3.2
1.7
11.3
6.0
13.1
4.0
29.8
23.6
10.1
8.2
0.010
0.015
0.014
0.007
0.013
0.008
0.016
0.003
0.026
0.027
0.006
0.005
F-12
-------
For the two Z variables, an initial analysis of covariance (ANACOVA) of the following
form was performed:
SOURCE OF VARIATION DEGREES OF FREEDOM
X (transfer efficiency) 1
GF (gelcoats) 1
GE (equipments) 2
GF x GE (interaction) 2
Error 11
Total 17
Since in each case the interaction term appeared statistically nonsignificant, that term was
dropped to produce the following:
SOURCE OF VARIATION DEGREES OF FREEDOM
X (transfer efficiency) 1
GF (gelcoats) 1
GE (equipments) . 2
Error 13
Total 17
Table 10 provides the ANOVA and ANACOVA results for the two Z variables. Type of
gelcoat FORMULATION appears to have a significant effect on both variables. Neither Zl nor
Z2 exhibits differences attributable to types of EQUIPMENT.
F-13
-------
TABLE 10. Analysis of Covariance Results for Gelcoat Experiment: Variables Zl
and Z2
GF
GF1
GF1
GF1
GF2
GF2
GF2
GF1
GF2
MEANS MEANS ADJUSTED FOR X
EQUIP
GE1
GE2
GE3
GE1
GE2
GE3
_
-
GE1
GE2
GE3
N
3
3
3
3
3
3
9
9
6
6
6
X
-1.5
-1.6
-0.5
2.5
-1.5
2.6
-1.2
1.2
0.5
-1.5
1.1
Zl
399.7
373.3
387.0
271.3
278.0
283.7
386.7
277.7
335.5
325.7
335.3
Z2
389.7
397.0
393.3
287.0
349.0
289.7
393.3
308.6
338.3
373.0
341.5
Zl
399.7
373.4
387.0
271.3
278.0
283.6
385.3
279.0
336.1
323.9
336.6
Z2
368.1
373,5
386.7
322.7
327.9
326.8
375.9
326.0
345.5
350.4
356.9
Hypothesis testing results:
RMSE for testing
interaction (12 df): . 17.9 39.8
(lldf): 18.7 23.9
Test for interaction: . n.s. ' n.s. n.s. n.s.
RMSE for testing main
effects (14 df): 18.4 39.7
(andcovariate)(13 df): ' ' . 18.9 . 22.4
Test for FORMULATIONS: *** *** *** **
Test for EQUIPMENT: n.s, n.s. n.s. . n.s.
Test for covariate X: n.s. ***
* = statistically significant at the 0.05 level.
** = statistically significant at the 0.01 level.
*** = statistically significant at the 0.001 level.
F-14
-------
F.3 Resin Experiment.
This experiment was aimed at evaluating how styrene emissions are affected by type of
resin formulation (factor RF) and type of equipment (factor RE). The factor combinations used
in the experiment and the associated sample sizes were as follows:
Treatment Level of RE Level of RF n
1
2
3
4
5
6
7
8
RE1
RE1
RE1
RE1
RE2
RE3
RE4
RE4
RF1
RF2
RF3
RF6
RP1
RF1
RF4FG
RF4SG
5
3
3
3
3
3
2
2
This design deviates slightly from the orignally planned design. Resin RF4 was modified to
shorten gel time after two test runs had shown that resin drains to the flange before it cures.
These two RF4 test runs are noted with SG for slow gel. The modified RF4 resin has a faster gel
time that allows resin to cure before it drains. RF 5 resin was not tested because the resin
manufacturer decided to drop out from the testing. The actual runs yield 16 degrees of freedom
for error. Separate analyses are used to test for equipment differences and for different resin
formulations.
The ANOVA for equipment comparisons among RE1, RE2, and RE3 (for resin RF1) is
as follows:
SOURCE OF VARIATION DEGREES OF FREEDOM
RE (equipments) 2
Error 16 (assumes pooling of variances from within all treatments)
For resin formulation comparisons (i.e., excluding treatments 5 and 6) the following
ANOVA applies:
SOURCE OF VARIATION DEGREES OF FREEDOM
RF (resins) 5
Error 16 (assumes pooling of variances from within all treatments)
F-15
-------
It should be noted that any comparison involving RF4SG or RF4FG versus one of the other
treatments is not only a comparison of resin formulations but also of application method.
Table 11 presents the ANOVA results for the X and "all" of the Y variables. The upper
portion of the table gives means for each treatment (combination of equipment and formualtion)
while the lower portion shows the results of testing hypotheses. Nonsignificant results were
obtained for X and Yl with regard to differences due to resin formulations (while using a fixed
type of equipment, RE1); the remaining Y variables all exhibited highly significant effects due to
these formulations. Type of equipment appeared to be a significant factor for all of seven of
these variables. Point estimates of mean differences and 95 percent confidence interval estimates
are given in Table 12 for EF in percent available styrene (% AS) as measured by THC and MB.
Similar information is presented in Table 13 for EF in g/m2 as measured by THC arid MB.
Similar information is presented in Table 14 for EF in g/g as measured by THC and MB. The
with-in cell standard deviations for resin experiment are shown in Table 15.
For the two Z variables, an initial analyses of covariance were run by incorporating the
transfer efficiency variable X into the ANOVA model. The results are given in Table 16. Both
type of resin formulation and type of equipment appears to have a significant effect on both
variables.
F-16
-------
TABLE 11. Analysis of Variance Results for Resin Experiment: Variables X, Yl, Y2,..., Y8
MEANS
TREATMENT N
Yl
Y2
Y3
Y4
YS
Y6
Hypothesis Testing Results:
1.90
***
30.82 31.50
***
RMSE for testing
effects (16 df):
Test for Treatments
Test for Resin
Formulations
(Trt 1 vs. 2
vs. 3 vs; 4):
Test for Equipments
(Trt 1 vs. 5 vs. 6): ** * *:
* = statistically significant at the 0,05 level.
** = statistically significant at the 0.01 level.
*** = statistically significant at the 0.001 level.
"Overall mean for X prior to standardization was 92.65.
***
1.28
***
1.52
***
11.14
11.17
***
***
n.s.
n.s.
***
***
***
***
***
***
Y7
0.0053
***
***
Y8
1 RE1
2RE1
3RE1
4RE1
5RE2
6RE3
7RE4
8RE4
RF1
RF2
RF3
RF6
RF1
RF1
RF4FG
RF4SG
5
3
3
3
3
3
2
2
-0.6
-0.7
-0.2
-0,9
4.6
4.4
-0.9
-8.6
214.0
.192.7
179.7
148.7
143.0
166.3
192.5
311.0
202.0
170.3
86.7
77.3
149.0
121.0
262.0
382.5 :
17.4
17.3
10.6
10.6
14.2
15.3
21.3
26.2
17.0
16.1
10.4
9.7
13.8
15.1
20.3
24.8
195.0
173.3
125.3
116.7
134.7
131.3
224,5
. 330.0
189.4
161.0
122.3
107.3
130.0
129,0
214.0
312.5
0.067
0.061
0.046
0.046
0.055
0.059
0.091
0.112
0.065
0.057
0.045
0.042
0.053
0.058
0.086
0.106
0.0061
***
-------
TABLE 12. Estimated Mean Differences and Confidence Interval Estimates for Percent
Available Styrene (% AS): Resin Experiment
Variable and comparison
Lower 95% Estimated
confidence
limit
mean
difference
Upper 95%
confidence
limit
Significant
Y3 - EF, % AS (by THC)
Equipment Comparisons
RE1RF1 -RE2RF1
RE1RF1 -RE3RF1
RE2RF1 -RE3RF1
Resin Formulation Comparisons
RE1RF1 -RE1RF2
RE1RF1 -RE1RF3
RE1RF1 -RE1RF6
RE1RF2 -RE1RF3
RE1RF2 -RE1RF6
RE1RF3 -RE1RF6
Comparison of MEKP and BPO systems
RE1RFI -RE4RF4FG
RE1RF1 -RE4RF4SG
1.220
0.120
-3.322
-1.880
4.853
4.853
4.512
4.512
-2.222
-6.086
-11.036
3.207
2.107
-1.100
0.107
6.840
6.840
6.733
6.733
0.000
-3.810
-8.760
5.194
4.094
1.122
2.094
8.827
8.827
8.955
8.955
2.222
-1.534
-6.484
**
*
***
***
***
***
**
***
Y4 - EF, % AS (by MB)
Equipment Comparisons
RE1RF1 .-RE2RF1
RE1RF1 -RE3RF1
RE2RF1 - RE3RF 1
Resin Formulation Comparisons
RE1RFI -RE1RF2
RE1RF1 -RE1RF3
RE1RF1 -RE1RF6
RE1RF2 -RE1RF3
RE1RF2 -RE1RF6
RE1RF3 -RE1RF6
Comparison of MEKP and BPO systems
RE1RF1 -RE4RF4FG
RE1RF1 -RE4RF4SG
0.796
-0.437
-3.322
-1.470
4.263
4.896
3.105
3.739
-1.995
-5.963
-10.513
3.147
1.913
-1.100
0.880
6.613
7.247
5.733
6.367
0.633
-3.270
-7.820
5.497
4.264
1.122
3.230
8.964
9.597
8.361
8.995
3.261
-0.577
-5.127
*
***
***
+**
***
*
***
Comparisons significant at the 0.05 level are indicated by '*'
* = statistically significant at the 0.05 level.
** = statistically significant at the 0.01 level.
*** = statistically significant at the 0.001 level.
F-18
-------
TABLE 13. Estimated Mean Differences and Confidence Interval Estimates for Emission
Factor (g/m2): Resin Experiment
Variable and comparison
Lower 95%
confidence
limit
Estimated
mean
difference
Upper 95%
confidence
limit
Significant
Y5 - EF, g/m2 (by THC)
Equipment Comparisons
RE1RF1 -RE2RF1
RE1RF1 -RE3RF1
RE2RF1 -RE3RF1
Resin Formulation Comparisons
RE1RF1 -RE1RF2
RE1RF1 -RE1RF3
RE1RF1 -RE1RF6
RE1RF2 -RE1RF3
RE1RF2 -RE1RF6
RE1RF3 -RE1RF6
Comparison of MEKP and BPO systems
RE1RF1 -RE4RF4FG
RE1RF1 -RE4RF4SG
43.094
46.428
-15.940
4.428
52.428
61.094
28.726
37.393
-10.607
-49.250
-154.750
60.333
63.667
3.333
21.667
69.667
78.333
48.000
56.667
8.667
-29.500
-135.000
77.572
80.906
22.607
38.906
86.906
95.572
67.274
75.940
27.940
-9.750
115.250
***
***
*
*
***
***
***
**
*ป*
Y6 - EF, g/m2 (by MB)
Equipment Comparisons
RE1RF1 -RE2RF1
RE1RF1 -RE3RF1
RE2RF1 -RE3RF1
Resin Formulation Comparisons
RE1RF1 -RE1RF2
RE1RF1 -RE1RF3
RE1RF1 -RE1RF6
RE1RF2 -RE1RF3
RE1RF2 -RE1RF6
RE1RF3 -RE1RF6
Comparison of MEKP and BPO systems
RE1RF1 -RE4RF4FG
RE1RF1 -RE4RF4SG
42.104
43.104
-18.338
11.104
49.771
64.771
19.329
34.329
-4.338
-44.415
-142.915
59.400
60.400
1.000
28.400
67.067
82.067
38.667
53.667
15.000
-24.600
-123.100
76.696
77.696
20.338
45.696
84.363
99.363
58.004
73.004
34.338
-4.785
-103.285
***
***
**
***
***
***
***
*
***
Comparisons significant at the 0.05 level are indicated by "*'.
* = Statistically significant at the 0.05 level.
** = Statistically significant at the 0.01 level.
*** = Statistically significant at the 0.001 level.
n.s. = Not statistically significant at the 0.05 level.
F-19
-------
TABLE 14. Estimated Mean Differences and Confidence Interval Estimates for Emission
Factor (g/g): Resin Experiment
Variable and comparison
Lower 95%
confidence
limit
Esimtated
mean
difference
Upper 95%
confidence
limit
Significant
Y7 - EF, g/g (by THC)
Equipment Comparisons
RE1RF1 -RE2RF1
RE1RF1 -RE3RF1
RE2RF1 -RE3RF1
Resin Formulation Comparisons
RE1RF1 -RE1RF2
RE1RF1 -RE1RF3
REIRF1 -RE.1RF6
RE1RF2 -RE1RF3
RE1RF2 -RE1RF6
RE1RF3 -RE1RF6
Comparison of MEKP and BPO
systems
RE1RF1 -RE4RF4FG
RE1RF1 -RE4RF4SG
0.004
0.000
-0.013
-0.003
0.013
0.013
0.006
0.006
-0.009
-0.033
-0.054
0.012
0.008
-0.004
0.006
0.021
0.021
0.015
0.015
0.000
-0.024
-0.045
0.020
0.016
0.005
0.014
0.029
0.029
0.024
0.025
0.009
-0.014
-0.035
**
*
***
***
**
**
***
***
Y8 - EF, g/g (by MB)
Equipment Comparisons
RE1RF1 -RE2RF1
RE1RF1 -RE3RF1
Resin Formulation Comparisons
RE1RF1 -RE1RF2
RE1RF1 -RE1RF3
RE1RF1 -RE1RF6
RE1RF2 -RE1RF3
RE1RF2 -RE1RF6
RE1RF3 -RE1RF6
Comparison of MEKP and BPO
systems
RE1RF1 -RE4RF4FG
RE1RF1 -RE4RF4SG
0.003
-0.002
-0.015
-0.001
0.011
0.014
0.001
0.004
-0.008
-0.032
-0.051
0.012
0.007
-0.005
0.008
0.020
0.023
0.012
0.015
0.003
-0.021
-0.041
0.022
0.017
0.006
0.018
0.030
0.032
0.022
0.025
0.013
-0.010
-0.030
*
***
***
*
**
***
***
Comparisons significant at the 0.05 level are indicated by '*'.
* = Statistically significant at the 0.05 level.
** = Statistically significant at the 0.01 level.
*** = Statistically significant at the 0.001 level.
n.s. = Not statistically significant at the 0.05 level.
F-20
-------
TABLE 15. Within-treatment Standard Deviations - Resin Experiment
TRT N X Zl Z2 Yl Y2 Y3 Y4 Y5 Y6 Y7 Y8
RE1RF1
RE1RF2
RE1RF3
REIRF6
RE2RF1
RE3RF1
RE4RF4FG 2 0.1 17.0 . 0,0 0.7 0.0 0.9 0.2 7.8 0.0 0.004 0.001
RE4RF4SG 2 6.5 13.4 26.9 63.6 96.9 3.3 2.5 5.7 12.0 0.014 0.011
5
3
3
3
3
3
1.1
1.0
0.7
1.6
0.9
0.3
25.0
7.1
12.7
49.5
14.6
28.8
6,8
6.6
20.7
40.6
35.6
36.7
8.7
2.1
13.5
49.6
22.9
47.4
11.9
7.2
9.6
7.1
50.7
14.0
1.0
1.1
1.2
1.3
0.4
0.9
1.4
1.0
1.7
1.3
2.0
1.3
11.0
3.5
5.7
21.5
6.0
12.7
3.0
3.0
8.7
18.0
15.7
15.9
0.004
0.004
0.005
0.006
0.001
0.004
0.005
0.004
0.007
0.006
0.008
0.005
F-21
-------
TABLE 16. Analysis of Covariance Results for Resin Experiment: Variables Zl and Z2
MEANS MEANS ADJUSTED FOR X
TREATMENT N
1RE1
2RE1
3RE1
4RE1
5RE2
6RE3
7RE4
8RE4
RFT
RF2
RF3
RF6
RF1
RF1
RF4FG
RF4SG
5
3
3
3
3
3
2
2
Hypothesis Testing Results:
RMSE for testing
effects (16 df):
(and covariate) (15 df):
Test for Treatments:
Test for Resin
Formulations
(Trt 1 vs. 2
vs. 3 vs. 4):
Test for Equipments
(Trt 1 vs. 5 vs. 6):
Zl
25.46
***
***
***
Z2
25.44
***
.***
***
Zl
***
***
***
Z2
-0.6
-0.7
-0.2
-0.9
4.6
4.4
-0.9
-8.6
444.8
395.3
286.3
266.7
306.7
299.0
512.0
752.5
431.8
367.0
279.0
244.7
296.3
294.3
488.0
713.0
444.4
394.9
286.2
266.0
309.9
302.1
511.4
746.5
428.7
363.8
278.2
240.3
319.0
316.0
483.8
671.0
26.26 24.45
***
***
***
* = statistically significant at the 0.05 level
** = statistically significant at the 0.01 level
*** = statistically significant at the 0.001 level
'Overall mean for X prior to standardization was 79.96.
F-22
-------
F.4 Summary
Table 17 provides a summary of the hypothesis testing results.
TABLE 17. Summary of Results of Specific Hypothesis Tests
Comparison
Ml vs. M2
A1 vs. A2
GF1 vs. GF2
GE1 VS.GE2
GE1 VS.GE3
GE2 vs. GE3
RF1 vs. RF2
RF1 vs. RF3
RF1 vs.RF6
RF2 vs. RF3
RF2 vs. RF6
RF3 vs. RF6
REl vs.RE2
REl vs.RE3
RE2 vs. RE3
REl
REl
REl
REl
REl
REl
RF1
RF1
RF1
RE1RF1 vs. RE4RF4FG
RE1RF1 vs. RE4RF4SG
RE4RF4FG vs. RE4RF4SG
X ZI Z2 Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8
** *** * *** ** *** * *
* *** *** *** *** *** *** *** ***
*
* ** . * **
*** *** *** *** *** *** *.** *** ***
*** *** * *** *** *** *** *** *** ***
*** *** ** *** *** *** *** ** *
*** *** ** *** *** *** *** ** **
** *** *** ** * ** * *** *** ** *
** *** *** ** * *** *** *
** * ******* *** ***
*** *** *** *** *** *** *** *** *** *** ***
*** *** *** *** *** *** *** *** *** ** **
* = statistically significant at the 0.05 level.
** = statistically significant at the 0.01 level.
*** = statistically significant at the 0.001 level.
F-23
-------
Appendix G
THC Analyzer Evaluation: Sampling Line Loss and Pressure Effect
Table of Contents
Section . PageNO,
G.I Sampling Line Loss G-l
G,2 Pressure Effect G-2
G-i
-------
G.I Sampling Line Loss
At the beginning of the project, questions were raised about whether styrene would be
lost in the sampling line due to its reactive nature. RTI investigated potential sampling line
losses prior to the styrene emission testing at Reichhold Chemical. Styrene and propane
calibration standards were delivered through various sampling lines connected to the analyzer's
calibration and sample ports in a manner very similar to that employed during routine
calibrations. The analyzer operating parameters were identical to those used during styrene
emission testing. The results of this study indicate that there are no significant pressure effects or
styrene losses associated with the sampling line.
In an initial test, two different lengths (4 feet versus 25 feet)of 1/8-inch ID PFA Teflon
tubing were compared. In a second test, 4 feet of 1/8-inch ID PFA Teflon tubing was compared
to 12 feet of 1/4-inch ED tetrafluoroethylene (TFE) Teflon tubing.
In the initial test, two styrene calibration standards, four propane calibration standards,
and zero air were measured on Range 2 (0 to 200 ppmv styrene) after being delivered through the
4-foot and 25-foot lengths of 1/8-inch ID sampling lines. Two other styrene calibration
standards, three other propane calibration standards, and zero air were measured on Range 1 (0 to
20 ppmv styrene) after being delivered through the two sampling lines. Least squares regression
analysis of 2-minute mean voltages yielded slopes for the 25-foot data regressed against the 4-
foot data. The results of this test are given in the following table:
Styrene Slope
(Volts/Volt)
'Propane Slope
(Volts/Volt)
Styrene- to-Propane Slope
Ratio
Range 2
1.0697
(+/- 0.0070)
1.0485
(+/-0.0150)
1.0203
Range 1
1.1115
(+/- 0.0266)
1.0430
(+/- 0.0032)
1.0657
A styrene or propane slope equal to 1.0000 would indicate equal or no losses for propane
or styrene in the two different sampling lines. The measured slopes are greater than 1.0000,
which could mean that the 25-foot sampling line had lower losses than the 4-foot sampling line.
This possible conclusion is counterintuitive. Instead, it was hypothesized that the higher total
hydrocarbon analyzer responses in the 25-foot sampling line were associated with some pressure
effect in the analyzer.
G-l
-------
A styrene-to-propane slope ratio of 1.0000 would indicate that any styrene losses in the
sampling lines were equal to any propane losses. No conclusions were drawn from these data
concerning styrene losses due to questions about possible pressure effects.
The sampling line loss measurements were repeated in a second test using a different
sampling line. A 12-foot length of 3/16-inch ID TFE tubing, which more realistically simulates
the actual sampling line, was compared to the 4-foot length of 1/8-inch ID PFA tubing. This test
used the same calibration standards as was used in the previous test. The results of the least
squares regression analysis of these test data are given in the following table:
Styrene Slope
(Volts/Volt)
Propane Slope
(Volts/Volt)
Styrene-to-Propane Slope
Ratio
Range 2
0.9972
(+/-0.0186)
1.0134
(+/- 0.0020)
0.9840
Range 1
1.0326
(+/- 0.0072)
1.0228
(+/- 0.0064)
1.0095
These results indicate that there are no significant pressure effects associated with the two
sampling lines that were tested and that there is no significant styrene losses in the longer
sampling line.
G.2 Pressure Effect
During routine calibrations of the total hydrocarbon analyzer, RTI observed that the
analyzer's response is very sensitive to changes in the sample pressure at the inlet of the FID's
capillary. The analyzer's operating manual states that it is very critical to maintain exactly the
same sample pressure during calibrations as during analysis of actual samples, it may be
necessary to adjust the back-pressure regulator to maintain the sample pressure.
RTI carefully maintained the sample pressure at 3.0 psig during calibrations and routine
sampling and noted that the sample pressure shifted by approximately 0.1 psig when the analyzer
was switched between these two operating modes. The back-pressure regulator was adjusted as
needed. RTI found that a +/- 0.1 psig sample pressure shift during a calibration produced an
approximately +/- 6.5 percent shift in the analyzer's response.
Some questions remained concerning possible residual pressure effects. The sample
pressure gauge is connected to the inlet of the back-pressure regulator, which is about 9 inches
downstream of the inlet of the FID capillary. Differences in sampling line pressure drops
upstream of the sampling pump may produce pressure differences at the FID inlet even though
the sample pressure was maintained at 3.0 psig.
G-2
-------
RTI investigated these questions by delivering three propane calibration standards and
zero air through three different sampling lines to the analyzer in a manner similar to that used
during routine calibrations. The analyzer was operated on Range 2. The three following sample
lines were connected to the analyzer's three inlet ports:
(A) 8-feet x 1/8-inch ED PFA Teflon tubing with gas pressure maintained at 5 psig by
an in-line pressure regulator (normal calibration configuration);
(B) 4-feet x 1/8-inch ED PFA Teflon tubing with gas pressure maintained at 0 psig by
an atmospheric- pressure sampling manifold; and
(C) 12-feet x 3/16-inch ID TFE Teflon tubing with gas pressure maintained at 0 psig
by an atmospheric-pressure sampling manifold.
Least squares regression analysis of 2-minute
for the regression lines associated with the three sam
Sampling
Line
A
B
C
Regression Slope
(Volts/ppmv)
0.01845
0.01792
0.01969
mean voltages yielded the following slopes
pling lines:
Slope Ratio relative to Normal
Calibration Configuration
0.971
1.067
These results indicate that residual pressure effects exist even when the sample pressure
is maintained at 3.0 psig. Although Sampling Lines B and C were operating at atmospheric
pressure, the slopes differed by approximately 11 percent. It is hypothesized that the pressure
drops across the sampling lines are the cause of the difference. The theoretical pressure drop
across Sampling Line B is 0.47 psig at a flowrate of 7 liters per minute. The corresponding
pressure drop across Sampling Line C is 0.20 psig.
It is unclear how these results can be extrapolated to the 12-foot long by 1/4-inch ID PFA
Teflon tubing used as the sampling line in the styrene emission tests. The theoretical pressure
drop for this sampling line is 0.053 psig.
Immediately following the EPA performance evaluation, RTI conducted an additional
pressure effects test using the 153 ppm propane calibration standard and various sampling lines.
The results of this test is given in the following table.
-------
Sampling Line and
Configuration
Analyzer
Response to
153 ppm
Propane (Volts)
Analyzer Response
Ratio Relative to
Normal Calibration
Configuration
8-feet x 1/8-inch ID PFA Teflon tubing with
pressure maintained at 5 psig by an in-line
pressure regulator (normal calibration
configuration)
2.782
12-feet x 3/16-inch ID TFE Teflon tubing with
pressure maintained at 0 psig by an atmospheric
pressure sampling manifold
2.809
1.010
12-feet x 1/4-inch ID PFA Teflon tubing with
pressure maintained at 0 psig by an atmospheric
pressure sampling manifold (normal sampling
line)
2.968
1.067
These results are somewhat different from the previous set of results. The 3/16-inch
tubing produced an analyzer response 1 percent greater than the normal calibration configuration
whereas it was 6.7 percent greater in the previous set of results. This difference may be due to
either a variation in the magnitude of the pressure effect during the two sets of measurements or
to a variation in the concentration in the sampling lines. Nevertheless, the general trend of these
results is the same (i.e., larger-bore sampling lines at atmospheric pressure produce greater
responses than the normal pressurized calibration line).
Based on these limited measurements and those obtained during the performance
evaluation, the following conclusions may be drawn:
despite careful control of the analyzer's sample pressure, variations of the gas pressure in
the sampling line and the calibration line produce changes in the analyzer's response;
the results of the pressure effects measurements are inconsistent and more measurements
would be required to quantify definitively the magnitude of the pressure effect; and
' the observed magnitude of the pressure effect ranged between 1 and 7 percent.
The impact of the pressure effects on the total hydrocarbon measurements is not
excessive relative to the +/- 5 percent accuracy objective that is given in the quality assurance
project plan. Consequently, the total hydrocarbon measurements should not be corrected for
pressure effects of unquantified magnitude.
(Reference: F. Caplan, "Finding Air Pressure Drop Through Smooth Tubes", Plant Engineering.
January 6, 1977, pp. 74-75.)
G-4
-------
Appendix H
An RTI Technical System Audit Report
H-i
-------
RESEARCH TRIANGLE INSTITUTE
/RTi
MEMORANDUM
TO: Emery Kong, RTI Project Leader
FROM: Cynthia Salmons, RTI QA Officer
DATE: July 20, 1995
SUBJECT: Technical Systems Audit of Evaluation of Pollution Prevention
Techniques to Reduce Styrene Emissions From Open Contact Molding
Processes
On June 9, Dr. William Yeager and I conducted a technical systems audit
(TSA) of the styrene emissions project. Pre-audit activities included preparation of
an audit checklist based on the quality assurance project plan (QAPP, see
attachment) and an informal walkthrough on June 2. This memorandum
summarizes the audit findings.
We generally found that the activities were conducted in accordance with
the QAPP. The scales and total hydrocarbon analyzer (THA) were calibrated. The
actual THA calibration procedure was more extensive than that described in the
QAPP and Method 25A. Originally, use of two concentration ranges on the THA
was planned so adequate cylinders to calibrate both were obtained. -Through June
9, for all test runs involving spray guns, the higher concentration range (0-200
ppm) was used, but all of the cylinders (seven all together) were used to calibrate
and/or check the calibration of the one range before each run. In addition, the
calibration drift test was performed after each run. When the flow coater or
pressure-fed roller was used in the test run, the lower concentration range (0-20
ppm) on the THA was used for better resolution. On days with three or four runs,
this approach involved many checks with the same calibration gases. If a
calibration is performed before each run, using two of the cylinders from the series
of calibration gases for the calibration drift check immediately before, for nine
analyses of the seven cylinders, is probably not necessary. While it seems
reasonable to perform the seven point calibration at the start of the day, a four
point calibration, as called for in the method, should be sufficient before the other
runs of the day. While Method 25A is explicit about which concentrations should
be used (80-90%, 45-55%, and 25-35% of the span value, and a zero) for the four
point calibration, with this many calibrations taking place, it would seem
reasonable to switch around the cylinders used for the four point calibration,
appropriate for the high or low concentration range, in the course of the day.
After June 9, when we discussed this issue, seven point calibrations were
performed at the start of each day. Then, four point calibrations were performed
before each test run.
The scale calibrations consisted of adding 9 weights sequentially. This
30^0 Ccnwaii.s Rcac PCS: Office Box 12194 . Research Triangle Park, North Carolina 27709-2194 USA
Telephone 919 541-6000 Fax 919 541-5985
H-l
-------
procedure was then repeated with something heavy, such as the mold, on the
scale. Both scales (i.e., inside the enclosure and outside the enclosure) were
linear.
The measurement point for the exhaust flow rate (i.e., where the pitot tube
is placed) is approximately 5-6 diameters downstream from a bend. Thus, the
measurement point satisfies the requirement that there not be a disturbance less
than 2 diameters upstream of the measurement point, but is not consistent with
the recommendation of not having a disturbance less than 8 diameters upstream
of the measurement point. There is not a reasonable way of achieving the
recommended 8 diameters. Method 1 requires at least 16 points for a velocity
traverse; 24 points were used for this project. Mark Banner checked for off-axis
flow by rotating the pitot tube in the duct and making velocity head (A P)
measurements. This check did not indicate a problem with off-axis flow.
The demonstration that the spray booth meets the criteria for total "
.enclosure was not observed, but the notes from thisjwere reviewed. .A few of the :
hot-wire anemometer readings at the natural draft openings were slightly less y
than 200 fpm, and there was considerable fluctuation in the hot wire anemometer
readings. Smoke tubes were used to check the natural draft at the door, and the"
air was always observed to be flowing into the spray booth. Calculating the duct
flowrate by dividing the duct flow rate, which averaged 8,674 cfm, by the open !
area (42 square feet) indicates a natural draft opening velocity of 206 fpm. The *;
results of the styrene evaporation experiments indicate that the styrene emissions
were captured. ' ; ' -'-- -.-... --...-.... .-. ...^.....-..^.-..-..-h
We did not observe the sampling and analysis of the gelcoat and resin
because this was not performed on the day of our visit. We understand that a
Reichhold laboratory will perform the gelcoat property measurements. We also
did not observe the styrene evaporation experiment or the weekly traverse '
measurements, but we reviewed results from these activities. We also did hot see
the measurement of the equipment delivery rate/ We understand that this was
performed once per day, at the end of the day, during the early tests. Later, this
was done after each test run, except for the pressure-fed roller, for which thisfv_
determination was difficult to make and inaccurate. '',''. "'' ' / l " ''
' ' 'The manufacturer's specifications for the gelcoat and resin indicate that
styrene is the only volatile component. The 55 gallon drum was mixed and then
split into 5 gallon cans. One quart samples were taken for the laboratory
measurements from the first 5 gallon cans when the material was used the first
ytime.1' Small samples (approximately 30 grains) were taken when the material'was
last used for the analysis of non-volatile content.
The baseline in the spray booth was measured with the THA for
approximately 5 minutes before each run. For the purposes of ending a test,
baseline was defined as a reading of less than 1 ppm styrene on the THA.
Typically, the pre-test background was between 0.4 and 0.6 ppm, and tests were
not ended until the readings reached this level. The data logger records every 2
seconds. At the end of each run, the data were downloaded to a computer, but
there was a computer software problem with this step during the period we
observed. This computer software problem was resolved later, and all data were
H-2
-------
smoothly transferred to a computer.
Bob Wright recorded the pitot tube measurements and temperature every
15 minutes on data recording sheets and in his notebook. There is a NIST-
traceable thermometer for checking these temperature measurements, but this
had not been performed yet.
We have reviewed the materials that Mark Banner provided on June 20 in
response to Bill Yeager's electronic mail request of June 12. This included results
from the styrene evaporation experiment, the weekly traverse measurements, the
checks for off-axis flow in the duct, and the tests that we observed during our
audit. These results did not indicate any problems with the project activities.
cc: Shri Kulkarni
Bill Yeager
Bob Wright
Mark Banner
Keith Leese
File: 5171-016
H-3
-------
TECHNICAL SYSTEMS AUDIT CHECKLIST
Audit Subject: Styrene Emissions Date: June 9, 1995
Auditee: RTI Auditors: Salmons, Yeager
Location: Reichhold Chemical Auditor Affil.: RTI
Personnel Present During Audit: Emery Kong, Mark Banner, Bob Wright
GENERAL
1. Is a written and approved QA Project Plan (QAPP) available to .field personnel
for this project? , , ..,,..,...,,.
Yes.
2. Are all deviations from the QAPP and methods properly documented?
The data logger is using a 2 second interval, which is more frequent than
discussed in the QAPP. This still needs to be formally documented.
3. How are corrective action procedures implemented?
None needed so far.
INITIAL AND PERIODIC PROCEDURES
4. Are there records of the comparison between the propane and styrene
standardizatons? When was the comparison performed? What were the results?
Yes. May 26, 1995. Experimental and theoretical results are in good
agreement.
5. Are there records for the comparison of direct injection and unheated sample
lines? When was this performed? What were the results?
Yes. May 30, 1995? The agreement is good when the pressure is steady.
When the pressure fluctuates, introducing gas through the sample line
leads to higher, not lower, values. A larger diameter sample line was also
used and seemed to improve agreement.
6. Was EPA Method 204 used to demonstrate that the spray booth meets the
criteria for a total enclosure?
Yes. Some hot-wire anemometer measurements at the natural draft
openings were less than 200 fpm (approx. 175). Considerable fluctuation in
H-4
-------
the hot-wire anemometer readings.
7. Was the capture efficiency of the booth determined by evaporating a known
amount of pure styrene? Was this test replicated? Within days? Among days?
Yes. Five tests on 6/5; repeated on 6/7 and on 6/29.
8. Is the measurement location for exhaust flowrate consistent with the EPA
recommendation (8 diameters downstream of bend or disturbance and 2 diameters
upstream)? Was the flow direction parallel to the axis of the duct at all points on
the traverse? Was a 3D pitot tube used to check for turbulence?
The measurement location is 5 or 6 diameters downstream of the bend so it
meets the 2 diameter criteria but not the 8 diameter recommendation. A
3D pitot tube was not used. At a later time, a check for off-axis flow was
performed by rotating the pitot tube in the duct. Off-axis flow does not
seem to be a problem.
9. Does the flowrate derived from centerline A P accurately predict the flowrate
determined by the traverse? How does the pitot tube A P at the center of the duct
compare with the A Ps on the traverse? How much does this A P vary during a
run? Are there any acceptance criteria for this variation? The precision goal for
the exhaust flow rate is RPD less than or equal to 10%.
Yes, within approximately 1 percent. Approx. 0.14 at center, approx. 0.10
at side of the traverse. Velocity varies by approx. 18%. A P is slightly
asymmetric across the duct.
10. When was.the hot wire anemometer calibrated? Has it recently been checked
against a pitot tube?
Calibration on. April 14, 1995. Will check again after June. Not checked
against a pitot tube.
11. Were the linear air velocities in the spray booth consistent with the exhaust
flow rate measurements?
Approx. 7700 crni versus 9000 in the duct. There were considerable
fluctuations in the hot-wire anemometer used to measure flow rate at the
filter face in the booth.
12. Was the exhaust flow rate determined at least once per week?
Yes. A traverse of both the spray booth and the stack was performed each
week. The exhaust flowrate was calculated based on the center line A P
measurements, which were taken every 15 minutes.
H-5
-------
13. Were the air flow velocities determined and recorded for the natural draft
openings and over the application area?
Yes.
14. Were the contents of the gelcoat or resin container mixed for 2 minutes before
sampling?
Not observed, but discussed this with technical staff.
,15. How was the 1 liter sample of gelcoat or resin taken from the container to
assure that it was representative? Was each container sampled? Were these .
determinations replicated? Within days? Among days? Does the manufacturer's"
invoice indicate that styrene is the only volatile component of the material?
Not observed. Manufacturer's specification indicates that styrene is the,.
only volatile component. 155 gallon drums are broken down into 5 gallon
cans.' 1 quart samples will be taken from the first 5 gallon can and a 30
gram sample will be taken from the last 5 gallon can. The gelcoat samples
will be sent to a Reichhold laboratory for analysis.
16. Were the material type, lot or batch number, and container number recorded
oh' the .gelcoat or resin container delivered to the laboratory?
Yes.
' *~ ' ff - . " S, ~ , , - , , -,,'* " ' " - "'- - ,?"'," ' i i >
17. Was styrene content for the gelcoat and resin materials determined with
Reidhhold, Standard Test Method No. 18-001?
Not observed.
& , i
18.,Were tne gel time, time to peak, and peak exotherm characteristics of , , . .:.,..
polyester resins measured following Reichhold Standard Test Method No. 18-050
and 18-051? " 4 ' . - - <
. s. -1 *' - - .-
Noฃpbserved.
PROCEDURES FOR EACH RUN
,*i . . ... .:'...,.
19., Was ttie/THC Calibrated before each run? '
):Yes,
ซ-. 5>v'^-"-vrปp *\ *"%"? f ^ *i *,'" M -" IT . r .... , - . jt i - - s* -s , fj,s,: " r f'*y*---
20.,, Was the calibration error test procedure for "the THC followed (i.e.,
-1* * "' f * "'^ *- ** *"
introduction of the zero and high-level gases, adjustment, analysis of the low, and
mid-level gases)? If so, were the results within 5% of expected? What were the
H-6
-------
predicted and observed readings?
During the project, the zero and span pots were not adjusted so that the
instrument drift could be monitored. As is described in the memorandum to
which this is attached, more standards were analyzed than are described in
the method. The calculated calibration was based on the high span gas and
the zero gas, which was assumed to be zero. The other standards
functioned as linearity and calibration checks. The predicted and observed
values were generally within 5%. Background, as determined from the
instrument reading prior to the start of the test, was subtracted.
21. Were the balances checked with standard reference weights? If so, how many
and what size weights were used?
Yes. Nine weights, ranging from 2 kg to 1 g. Also a box of fiberglass or a
mold was added to the scale and the weights added again to check for
linearity.
22. Was the baseline in the spray booth measured with the THC before and after
each run? For how long?
Yes. The baseline is indicated by a THC reading of less than 1 ppm.
Approx. 5 minutes.
23. Were the initial air temperature, velocity head, and relative humidity
measured and recorded?
Yes.
24. Was the equipment delivery rate measured in a remote location?
Not observed, but told this was done at the end of the day during the early
testing. Later, this was done after each test, except for the pressure-fed
roller.
25. Were the equipment setup conditions recorded?
Not observed.
26. Were the initial weights for the mold, gelcoat/resin, catalyst, fiberglass
reinforcement, protective skirt for cart, and groundcover recorded?
Yes.
27. Was the time initiated as soon as application started?
H-7
-------
Yes.
28. Were the velocity head and air temperature recorded every 15 minutes?
Yes. Recorded on a data recording sheet and in Bob Wright's notebook
except during lunch breaks.
29. Were the time and weight of gelcoat or resin recorded as soon as application
was completed?
Yes.
30. Was the weight of the wet mold at the end of application without the
protective skirt recorded?
No. Recorded with the skirt on. The skirt was not removed until gel had
cured and THC returned to baseline.
31. Are the equipment delivery rates consistent with the before and after mass
measurements for the mass balance calculations?
Not observed.
32. Was the run stopped when the THC indicated that the concentration returned
to baseline? Was the time at which this happened recorded?
Yes.
33. Was the baseline drift for the THC determined at the end of the run?
Yes.
34. Was the enclosure flushed with fresh makeup air until the baseline stabilized?
Yes.
35. At the end of the day, were the recording sheets with sampling and analytical
results collected, verified, and analyzed by the Testing Crew Chief?
Not observed, but Mark Banner had the results from the previous days and
was able to show comparisons, etc.
H-8
-------
Appendix I
EPA Performance Evaluation of Total Hydrocarbon Analyzer
I-i
-------
APPENDIX I. EPA PERFORMANCE EVALUATION OF TOTAL
HYDROCARBON ANALYZER
EPA personnel conducted a performance evaluation of the total hydrocarbon analyzer on
July 7, 1995. The purpose of the performance evaluation was to assess the accuracy of the
analyzer and to compare this assessment with the accuracy data quality objective that was given
in the quality assurance project plan. The assessment was accomplished by measurement of a
styrene perforfance evaluation sample whose certified concentration was unknown to RTI. The
measurements were performed on Range 2 of the analyzer. RTI's predicted concentration for this
sample differed by 2 percent from its certified concentration. This result meets the data quality
objective of +/- 5 percent for calibration gas accuracy that was given in the quality assurance
project plan. Additional measurements showed that pressure effects did not exceed 3 percent
during the performance evaluation. The styrene loss in the sampling line was found to be less
than 2 percent. A more detailed discussion of the EPA performance evaluation is presented
below.
Discussions between RTI and EPA personnel resulted in agreement that three
components of accuracy needed to be assessed during the performance evaluation:
(1) the accuracy of RTI's calibration standards (especially in light of the use of
propane to calibrate for styrene measurements);
(2) pressure effects associated with calibration line/sampling line differences; and
(3) styrene sampling line losses.
EPA and RTI personnel agreed .that these three accuracy components would be assessed
by measurement of RTI's propane calibration standards and EPA's performance evaluation
sample through RTFs calibration line from the cylinders directly and through RTI's sampling
line from an EPA heated sampling manifold.
The accuracy of RTFs propane calibration standards was determined from measurements
in RTI's normal calibration configuration. RTI measured its propane calibration standards and
the performance evaluation sample. RTI calculated a calibration equation from the propane
measurements and predicted a concentration for the styrene sample based on this equation. This
accuracy component was calculated as the percentage difference between the predicted and
certified concentrations.
Pressure effects were determined by comparing measurements of RTI's propane
calibration standard through the normal calibration line and through the sampling line from an
EPA heated sampling manifold. This accuracy component was calculated as the percentage
difference between the two sets of analyzer responses.
1-1
-------
Styrene sampling losses were determined from measurements of EPA's styrene
performance evaluation sample in the normal calibration configuration and through the EPA
heated sampling manifold/ RTI sampling line. This accuracy component was calculated as the
percentage difference between the two predicted styrene concentrations.
The EPA performance evaluation sample was a compressed gas calibration standard that
was prepared in June 1995 by Scott Specialty Gases. Its certified styrene concentration is 31.0
ppm as measured by Scott, but this value was not independently verified by EPA. RTI did not
know the certified concentration until after the evaluation had been completed.
EPA's sampling manifold was constructed from a 28-inch length of 2-inch OD stainless
steel tube. Stainless steel tubing fittings were welded to the tube. RTI's sampling line was
inserted into the manifold along the tube's long axis. The manifold was heated to approximately
130 oC. The static pressure inside the manifold was measured by RTI's Magnehelicฎ
differential pressure gauge and was maintained at 0.01 inches of water to ensure that it was at
near-atmospheric conditions.
The results of the evaluation are given in the following table:
Parameter
Equivalent Styrene
Concentration
THC Analyzer
Response via Cal Port
(Volts)
THC Analyzer
Response via EPA
Manifold (Volts)
Predicted Styrene
Concentration (Cal
Port)
Predicted Styrene
Concentration (EPA
Manifold)
RTI 45.03 ppm
Propane
Calibration
Standard
16.88
0.844
0.863
RTI 153.37 ppm
Propane
Calibration
Standard
. 57.51
2.848
2.765
.
EPA Styrene
Performance
Evaluation
Sample
Unknown during
analysis
1.511
1.469
30.39
29.82
RTI
Zero
Air
0.00
0.008
0.035
-0.08
-0.81
1-2
-------
Parameter
Response Change due
to Pressure Effects
(percent)
Styrene Loss in
Sampling Line
(percent)
RTI 45.03 ppm
Propane
Calibration
Standard
+2.23
RTI 153.37 .ppm.
Propane
Calibration
Standard
-2.92
EPA Styrene
Performance
Evaluation
Sample
-2.76
-1.87
RTI
Zero
Air
...
Using measurements in RTFs normal calibration configuration, the predicted
concentration of the styrene performance evaluation sample is 30.39 ppm. This value differs by -
2.0 percent from the certified value of 31.0 ppm.
The results of the pressure effects measurements are inconsistent. The analyzer's
response for one propane calibration standard increased by 2.2 percent, but response decreased
by 2.9 and 2.8 percent for the other propane calibration standard and for the styrene performance
evaluation sample, respectively. It is hypothesized that stabilization problems or some unknown
problem were biasing the EPA manifold results. In any case, pressure effects do not appear to
exceed +/- 3 percent.
The styrene loss in RTI's sampling line were less that 2 percent if one accepts all the
propane and styrene measurements as being correct.
1-3
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/R-97-018b
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation of Pollution Prevention Techniques to
Reduce Styrene Emissions from Open Contact
Molding Processes; Volume II, Appendixes
5. REPORT DATE
March 1997
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Emery Kong, Mark Banner, Robert Wright, and
Andrew Clayton
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Research Triangle Institute
P. O. Box 12194
Research Triangle Park, North Carolina 27709
11. CONTRACT/GRANT NO.
CR 818419-03
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 4/94-9/95
14. SPONSORING AGENCY CODE
EPA/600/13
IB.SUPPLEMENTARY NOTES ApPCD project officer is Geddes H. Ramsey, Mail Drop 61. 9197
541-7963. Volume I is the final report.
iซ. ABSTRACT
repor|; gives results of a study to evaluate several pollution prevention
techniques that could be used to reduce styrene emissions from open molding pro-
cesses in the fiberglass-reinforced plastics /composites (FRP/C) and fiberglass boat
building industries. Styrene emissions using standard industry techniques, mater-
ials, and equipment were evaluated in a controlled environment and compared to a
basline condition to determine the effects of these pollution prevention techniques on
styrene emissions. The study found that using controlled spraying (i.e. , reducing
overspray), low- styrene and styrene- suppressed materials, and nonatomizing appli-
cation equipment can reduce styrene emissions by from 11 to 52%. Facilities should
investigate the applicability and feasibility of these pollution prevention options to
reduce their styrene emissions. The calculated emission factors were from 1.6 to
2. 5 times the mid- range AP-42 emission factors for the corresponding gel coat and
resin application. These results indicate that facilities using AP-42 emission factors
to estimate emissions in open molding processes are likely to underestimate actual
emissions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Styrene Resins
Emission
Molding Techniques
Fiberglass-reinforced Plastics
Boats
Pollution Prevention
Stationary Sources
Boat Building
13 B
111, 11J
14G
13H
11D
13 J
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report I
Unclassified
21. NO. OF PAGES
148
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
1-4
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
. National Risk Management Research Laboratory
Technology Transfer and Support Division
Cincinnati, Ohio 45268
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE. S3OO
AN EQUAL OPPORTUNITY EMPLOYER
It your address is incorrect, please change on the above label
tear off; and return to the above address.
If you do not desire to continue receiving these technical.
reports, CHECK HERE LTJ. tear oil label, and return it to the
above address.
Publication No. EPA- 600 /R- 97-0i8b
------- |