HAP

5.0

BASELINE EMISSION ESTIMATES AND METHODOLOGY

This chapter presents estimates of baseline hazardous air pollutant (HAP) emissions for the production and ancillary operations used by facilities included within the reinforced plastic composites source category and the methodology used to estimate baseline HAP emissions. The baseline HAP emission estimates are estimates of the amount of organic HAP emitted by facilities in the reinforced plastic composites source category as they currently exist prior to the application of any controls required by the proposed rule. The primary organic HAP emitted by these facilities (excluding the cleaning process) is styrene. Methyl methacrylate is also emitted, but to a much lesser extent. The primary organic HAP emitted in the cleaning process is methylene chloride. We determined baseline HAP emission estimates presented in this chapter primarily from resin or gel coat usage and emission factors that are a function of the HAP content of the resin or gel coat. Emissions from the various operations within the reinforced plastic composites source category are at least partially dependent on process variables other than the resin or gel coat HAP content. Therefore the actual emissions at a specific facility may be higher or lower than estimated by the average HAP emission estimation equations presented in this chapter. For open molding and centrifugal casting operations, the emission estimation equations are the same as the point value equations described in Chapter 4. In the Chapter 4 discussion, we state that point value equations should not be used as emission factors. However, for purposes of developing source category emission estimates, we determined that the best approach was to use the same data sets and equation forms to estimate baseline emissions and emission reductions resulting from the application of the regulatory alternatives described in Chapter 4.

5.1

METHOD USED TO ESTIMATE BASELINE EMISSIONS

5.1.1

Model Plant vs. Facility-Specific Approach Nationwide baseline emissions of a rule may be estimated using model plants, or by

calculating emissions for each identified facility in the source category. A model plant approach relies on the development of “typical” plants that reflect such parameters as plant size, process 5-0

configuration, emission rates and characterization (e.g. flow, concentration) and emission reduction techniques. The number of model plants depends on the diversity of the facilities within the source category. Baseline emissions are then estimated for each individual model plant. Then nationwide emissions are estimated by determining the numbers of each model plant that produces a distribution that represents the total source category, and multiplying the number of model plants times individual model plant emissions. Facilities in this industry may have a variety of different operations present and we determined it would be difficult to develop model plants that could be considered to be typical of the industry as a whole. However, detailed resin and gel coat use data were available for a large number of facilities, and these data were sufficient to calculate emissions. We also used these data to calculate emissions from individual facilities where we did not have raw material use data based on industry averages. We estimated nationwide baseline emissions by then summing the emissions from each individual facility.

5.1.2

Basic Approach We calculated baseline HAP emissions for the reinforced plastic composites source

category by using the following sequence of steps. First, we collected facility-specific information from the 1993 industry screening surveys, telephone contact reports, site visit reports, facility permits, and other available information sources. From these data sources, we determined the production and ancillary operations used by each facility, the resin and gel coat usage, resin and gel coat HAP content(s), and method(s) of control used for each specific operation. Most of the facilities reported data for years other than 1997, the base year for our analysis. In order to adjust the resin/gel coat use reported by a particular facility to 1997, we used industry growth estimates developed for each specific process. More details on methods used to estimate industry growth may be found in the memorandum "Growth in the Reinforced Plastics Industries", Dated July 7, 2000. This memorandum is in the project docket. After adjusting a facility's resin use to 1997, we estimated the corresponding gel coat use based on the typical resin to gel coat ratios.

5-1

We developed HAP emission estimation equations for open molding operations, centrifugal casting operations, continuous lamination/continuous casting operations, pultrusion, SMC manufacturing, BMC manufacturing/mixing operations and storage. Emissions from closed molding were estimated based on published emission factors. We estimated the appropriate operation specific emissions for each facility based upon each facility's resin and gel coat usage, resin and gel coat HAP content(s), method(s) of control, and the applicable HAP emission equation or factor for the production and ancillary operations located at the facility. Next, we obtained total HAP emissions for a specific operation by aggregating each individual facility's emissions for that operation. Finally, we determined the total baseline HAP emissions for the reinforced plastic composites source category by aggregating the total HAP emissions from each operation included in the reinforced plastic composites source category. In developing baseline emission estimates, we assumed no emission reductions due to use of vapor suppressants. Though vapor suppressants have been shown to reduce emissions for certain operations, the amount of emission reduction is facility specific. None of the facilities reporting the use of vapor suppressants provided test data to substantiate the performance of the suppressants.

5.2

BASELINE EMISSION ESTIMATES Table 5-1 presents the baseline HAP emissions for each operation type, each aggregated

operation, and the entire source category. Appendix A presents the baseline HAP emissions by facility within each aggregated or individual operation. The information presented in Table 5-1 indicates that the baseline HAP emissions for the reinforced plastic composites source category are approximately 22,200 tons per year. The baseline HAP emissions from open molding operations constitute approximately 80% of the total baseline HAP emissions from the source category. The remaining 20% of the total baseline HAP emissions are primarily from centrifugal casting, equipment cleaning, continuous lamination/casting, and pultrusion. We did not develop separate emission estimates for

5-2

operations that produce class 1 smoke and flame products, or for white/off-white gel coats verses other colors. It is assumed that emissions from these operations are included in the appropriate

5-3

TABLE 5-1. BASELINE HAP EMISSIONS FOR THE REINFORCED PLASTIC COMPOSITES SOURCE CATEGORY a

Baseline HAP

Baseline HAP

Emissions Operation Type

Aggregated Operation (tpy)

Open Molding

Manual Resin Application

620

Mechanical Resin Application

12511

Emissions

Individual Operation c

(% of SC)

(tpy)

(% of SC) b

Manual Resin Application - Corrosion-Resistant

189

0.9%

Manual Resin Application - Non-Corrosion Resistant

377

1.7%

Manual Resin Application - Tooling / Mold Manufacturing

55

0.2%

Mechanical Resin Application - Corrosion-Resistant

739

3.3%

Mechanical Resin Application - Non-Corrosion Resistant

1925

8.7%

9497

42.7%

350

1.6%

Filament Winding - Corrosion-Resistant

319

1.4%

Filament Winding - Non-Corrosion Resistant

159

0.7%

Gel Coat Application - Pigmented Production

3469

15.6%

Gel Coat Application - Clear Production

597

2.7%

Gel Coat Application - Tooling / Mold Manufacturing

111

0.5%

b

2.8%

56.3%

(Unfilled) Mechanical Resin Application - Non-Corrosion Resistant (Filled) Mechanical Resin Application - Tooling / Mold Manufacturing Filament Winding

479

Gel Coat Application

4178

Totals for the Open Molding Subcategory

17888

5-4

2.2%

18.8%

80.0%

Baseline HAP

Baseline HAP

Emissions Operation Type

Aggregated Operation (tpy)

Closed Molding

Compression / Injection

(% of SC) b

423

1.9%

11

0.1%

Totals for the Closed Molding Subcategory

434

2.0%

Centrifugal Casting

1054

4.7%

Molding Resin Transfer Molding (RTM)

Centrifugal Casting

Centrifugal Casting - Corrosion-Resistant Centrifugal Casting - Non-Corrosion Resistant

Totals for the Centrifugal Casting Subcategory

1054

4.7%

731

3.3%

Polymer Casting

227

1.0%

Pultrusion

277

1.2%

SMC Manufacturing

356

1.6%

BMC Manufacturing/Mixing

452

2.0%

Equipment Cleaning

807

3.6%

95

0.4%

22221

100.0%

Continuous Lamination / Casting

Storage of HAP Containing Materials Totals for the Reinforced Plastic Composite Source Category

a

Facility and operation specific information can be found in Appendix A of this document. The base year is 1997.

5-5

Emissions

Individual Operation c

(tpy)

(% of SC) b

17

0.1%

1036

4.7%

b

Values shown in the table may not add up to 100 percent due to rounding.

c

Emissions from operations producing class 1 smoke and flame products and pigmented gel coat - non white colors were not calculated separately. These emission are included in the totals for other operations. The specific operation depends on where the facility listed that resin or gel coat use.

5-6

other operation categories depending on which how the facilities reported use of these resins and gel coats.

We did not have specific data that we believed would allow for a meaningful separate

estimation of emissions from these operations.

5.3

DERIVATION OF AVERAGE EMISSION ESTIMATION EQUATIONS We developed the average HAP emission estimation equations for production operations,

ancillary operations, and control technologies by reviewing several types of data sources, including: (1) emission studies conducted on behalf of the EPA, (2) emission studies conducted by organizations representing the reinforced plastic composites industry, (3) emission tests or studies conducted at facilities in the reinforced plastic composites industry, (4) industry responses to Section 114 Screening Surveys and other correspondence from industry, and (5) EPA's Compilation of Air Pollutant Emission Factors (AP-42). We developed emission estimation equations from emission studies or testing in those instances where the data from such studies or testing were superior in quantity and/or quality to the data contained in the AP-42 or other sources. In general, we used the emission factors published in the AP-42 for those processes for which no data from emission studies, testing or other facility-specific data were available. Table 5-2 summarizes the HAP emission estimation equations for the various production operations, ancillary operations, and control technologies within the plastic composites source category. Table 5-2 also identifies the data sources used to derive these equations. The HAP emission estimation equations shown for each operation are applicable for both filled and unfilled resin systems. For open molding operations (manual resin application, mechanical resin application, filament winding, and gel coat application), the point value equations previously discussed were used to estimate emissions. We originally developed point value equations as a method to rank facilities and not to determine emission factors. This is because the only independent variable in the point value equations is HAP content, and there are factors other then HAP content that affect HAP emissions. However, we did not have sufficient data to accurately quantify other variables.

5-7

TABLE 5-2. AVERAGE HAP EMISSION ESTIMATION EQUATIONS BY PROCESS (Concluded) TABLE 5-2. AVERAGE HAP EMISSION ESTIMATION EQUATIONS BY PROCESS Process

Specific Condition / Control Technology

HAP Emission Estimation Equationa,b,c

Reference

EF (lbs/ton) = 0.028 x (% HAP) 2.275

Manual Resin Application (Hand Lay-Up w/ Bucket & Tool)

All Manual Resin Application

Mechanical Resin Application: Atomized (Spray-Up w/ HVLP, Airless, Airless-Air-Assisted)

All Atomized Mechanical Resin Application

Mechanical Resin Application: Non-Atomized (Pressure-Fed Rollers/Flow Coaters)

All Non-Atomized Mechanical Resin Application

Filament Winding

All Filament Winding

EF (lbs/ton) = 1.675 x (% HAP) 1.225

6,7

Gel Coat Application

All Gel Coating

EF (lbs/ton) = 0.890 x (% HAP) 1.675

1

Compression Molding

2% of Available HAP

8

Injection Molding

2% of Available HAP

8

Resin Transfer Molding (RTM)

2% of Available HAP

Centrifugal Casting

55.8% of Available HAP

Continuous Casting

1 EF (lbs/ton) = 0.028 x (% HAP) 2.425

1 EF (lbs/ton) = 0.028 x (% HAP) 2.275

1

8

Particle Line, Active Ventilation of the Line, Laminate Thickness (T) = 250 Mils

5-8

9

4.6% of Available HAP 11- 13

TABLE 5-2. AVERAGE HAP EMISSION ESTIMATION EQUATIONS BY PROCESS (Concluded) Process

Continuous Lamination

HAP Emission Estimation Equationa,b,c

Reference

Sheet Line, Passive Ventilation of the Line, Laminate Thickness (T) = 500 Mils VP = Vapor Pressure of HAP @ Line Temperature

0.6 EF(tons/yr) = [(Total Styrene x VPS) x (120 / T) x 0.00156 + (Total MMA x VPMMA) x (120 / T)0.6 x 0.0014] x 1.25

14 - 17

Active Ventilation of the Line, Laminate Thickness (T) = 60 Mils, 75 Mils VP = Vapor Pressure of HAP @ Line Temperature

EF(tons/yr) = [(Total Styrene x VPS) x (120 / T)0.6 x 0.0023 + (Total MMA x VPMMA) x (120 / T)0.6 x 0.0014] x 1.1

Active Ventilation of the Line, Laminate Thickness (T) = 90 Mils, VP = Vapor Pressure of HAP @ Line Temperature

EF(tons/yr) = [(Total Styrene x VPS) x (120 / T)0.6 x 0.0023 + (Total MMA x VPMMA) x (120 / T)0.6 x 0.0014] x 1.15

Passive Ventilation of the Line, Laminate Thickness (T) = 90 Mils, VP = Vapor Pressure of HAP @ Line Temperature

EF(tons/yr) = [(Total Styrene x VPS) x (120 / T)0.6 x 0.00156 + (Total MMA x VPMMA) x (120 / T)0.6 x 0.0014] x 1.25

Specific Condition / Control Technology

Polymer Casting Pultrusion

14 - 17

14 - 17

2% of Available HAP

8

Open Resin Bath

3.26% of Available HAP

21

Resin Bath w/ Wet Area Enclosure (WAE) and Resin Drip Collection System (RDCS)

1.33 % of Available HAP

SMC Manufacturing BMC Manufacturing/Mixing

14 - 17

21 2.67% of Available HAP

Open Mixing/BMC Manufacturing with Passive Ventilation of the Mixing Vessel

0.50% of Available HAP

Open Mixing/BMC Manufacturing with Active Ventilation of the Mixing Vessel

3.10% of Available HAP

5-9

22 23 23

TABLE 5-2. AVERAGE HAP EMISSION ESTIMATION EQUATIONS BY PROCESS (Concluded) Process

Specific Condition / Control Technology

HAP Emission Estimation Equationa,b,c

Closed/Covered Mixing/BMC Manufacturing Using Mixer Covers with No Visible Gaps with No Active Ventilation of the Mixing Vessel

0.25% of Available HAP

Equipment Cleaning

HAP Containing Cleaners

100.0% of Available HAP

Storage of HAP Containing Materials

All storage containers

72.49 grams HAP per hour

Reference

23

23

a

The equations shown for the open molding operations of manual resin application, mechanical resin application, and filament winding have units of pounds of HAP emitted per ton of "neat resin plus" consumed. "Neat resin plus" means the neat resin plus any additional HAP added to the neat resin. The formula for the percent HAP in the neat resin plus (and used in the open molding equations for manual resin application, mechanical resin application, and filament winding shown above) is:

%HAP of Neat Resin Plus =

b

The equations shown for the open molding operations of gel coat application have units of pounds of HAP emitted per ton of “neat gel coat plus” consumed. "Neat gel coat plus" means the neat gel coat plus any additional HAP added to the neat gel coat. The formula for the percent HAP in the neat gel coat plus (and used in the open molding equation for gel coat application shown above) is:

%HAP of Neat Gel Coat Plus =

c

Total Amount of HAP x 100% (Total Amount of Neat Resin) + (Total Amount of HAP Added)

Total Amount of HAP x 100% (Total Amount of Neat Gel Coat) + (Total Amount of HAP Added)

The equations shown for continuous lamination were used for the majority of the facilities that use the continuous lamination operation. However, facility specific equations and/or source test data were

used to estimate HAP emissions for those facilities that use the continuous lamination operation with specific conditions that differ substantially from those listed in the table for continuous lamination.

5-10

Also, facility specific data other than HAP content was not available. Therefore we determined the point value equations were the best resource available to calculate emissions from open molding. We had different data bases for each open molding operation, but we used the same basic statistical analysis design. We designed each statistical analysis to derive an emission estimation equation with the lowest possible standard error of y-estimate subject to the following criteria (constraints): (1) zero HAP emissions at zero percent resin HAP content, (2) increasing HAP emissions with increasing resin HAP content, and (3) data points for a specific HAP content are balanced so that the HAP emission estimation equation gives equal weight to each combination of parameters tested. In addition, for manual and mechanical resin applications, we added an additional criteria: the HAP emission estimation equation for manual resin application yields lower emissions than the HAP emission estimation equation for mechanical resin application when HAP emissions are rounded to the nearest tenth. The statistical analyses conducted to derive the HAP emission estimation equations for these four operations are found in electronic format in the docket.

5.3.1

Derivation of Emission Estimation Equations for Uncontrolled Production and Ancillary

Operations 5.3.1.1 Manual Resin Application.1 Table 5-3 presents HAP emission data for uncontrolled manual resin application available from an emission study conducted by the CFA. The CFA Phase I study examined the effect of certain process parameters on HAP emissions from manual resin application. Specifically, the CFA Phase I study examined the effects of resin HAP content, laminate thickness, gel time, and air flow across the mold. During the CFA Phase I study, 16 separate combinations of the above parameters were tested in order to evaluate the effect of each parameter on HAP emissions from manual resin application. The resins used in the study had HAP contents of 35 and 42%. The resins were applied to a laminate thickness of 0.041 or 0.088 inches with gel times of 15 or 30 minutes. The air flow across the mold was either 50 or 100 feet per minute. A total of twenty test runs consisting of sixteen runs with differing parameters and four duplicate runs were conducted during the study.

5-11

TABLE 5-3. CFA PHASE I STUDY: MANUAL RESIN APPLICATION EMISSIONS a Styrene Styrene Resin Monome Applied Surface Amount Vapor Thicknes Gel Time Air Flow of Filler Suppress Emissions Emissions r (lb/ton (% of ant s Run No. Content (minutes) (feet/min (% Filler) resin) Resin) (% VS) (inches) (% ) Styrene)

a

1

35%

0.041

30

100

0.0%

0.0%

6.2%

124

2

42%

0.088

15

50

0.0%

0.0%

4.9%

97

3A

42%

0.041

15

50

0.0%

0.0%

7.2%

143

3B

42%

0.041

15

50

0.0%

0.0%

7.6%

152

4A

35%

0.088

30

100

0.0%

0.0%

4.2%

83

4B

35%

0.088

30

100

0.0%

0.0%

4.6%

92

5

42%

0.041

30

50

0.0%

0.0%

8.1%

162

6

35%

0.088

15

100

0.0%

0.0%

3.6%

72

7

35%

0.041

15

100

0.0%

0.0%

5.3%

105

8

42%

0.088

30

50

0.0%

0.0%

5.7%

114

9

35%

0.041

15

50

0.0%

0.0%

5.0%

100

10

42%

0.088

30

100

0.0%

0.0%

6.0%

121

11A

42%

0.041

30

100

0.0%

0.0%

8.9%

179

11B

42%

0.041

30

100

0.0%

0.0%

8.9%

177

12A

35%

0.088

15

50

0.0%

0.0%

3.4%

68

12B

35%

0.088

15

50

0.0%

0.0%

3.6%

72

13

35%

0.088

30

50

0.0%

0.0%

4.2%

83

14

42%

0.041

15

100

0.0%

0.0%

7.1%

141

15

35%

0.041

30

50

0.0%

0.0%

6.3%

126

16

42%

0.088

15

100

0.0%

0.0%

4.9%

99

The complete data set for the CFA Phase I manual resin emission study is found in electronic format in the docket.

5-12

The CFA Phase I study indicated that HAP content and the laminate thickness of the resin were the primary factors affecting HAP emissions, with gel time having a minor effect and air flow across the mold having no significant effect. We derived a HAP emission estimation equation by conducting a statistical analysis (as previously described) on the twenty data points obtained by the CFA Phase I study discussed above. The HAP emission estimation equation for manual resin application derived from this statistical analysis is shown below. EF (lbs HAP/ton resin consumed) = 0.028 x (%HAP)2.275

We used this HAP emission estimation equation in conjunction with facility-specific resin HAP content values to calculate the baseline HAP emissions for all vapor suppressed and nonvapor suppressed manual resin application operations. 5.3.1.2 Mechanical Resin Application with Atomized Spay. 1-5 HAP emission data for mechanical resin application using atomized spray are available from five separate emission studies conducted by Research Triangle Institute (RTI), the National Marine Manufacturers Association (NMMA), and the Composites Fabricators Association (CFA). Each of these studies examined the effects of parameters on HAP emissions resulting from uncontrolled mechanical resin application.

Of the five studies, the CFA Phase I study was the most extensive and balanced in examining the effects that certain process parameters have on HAP emissions from mechanical resin application. Therefore, we used the CFA Phase I study to derive the atomized spray mechanical resin application emission estimation equation. The CFA Phase I study examined the effects of resin HAP content, laminate thickness, gel time, resin application rate, and air flow across the mold. During the CFA Phase I study, 16 separate combinations of the above parameters were tested in order to evaluate the effect of each parameter on HAP emissions. The resins used in the study had HAP contents of 35 and 42%. The resins were applied to a laminate thickness of 0.040 or 0.080 inches with gel times of 15 or 30 minutes. The resins were applied at a rate of 2 or 4 pounds per minute with air flow across the 5-13

mold at either 50 or 100 feet per minute. A total of twenty test runs consisting of sixteen runs with differing parameters and four duplicate runs were conducted during the study. Table 5-4 presents the results of these test runs. The CFA Phase I study indicated that HAP content of the resin is the primary factor affecting HAP emissions, with resin application rate having a minor effect and laminate thickness, gel time, and air flow across the mold having insignificant effects. We derived a HAP emission estimation equation by conducting a statistical analysis on the twenty data points obtained by the CFA Phase I study discussed above. The HAP emission estimation equation for mechanical resin application with atomized spray derived from this statistical analysis is shown below. EF (lbs HAP/ton resin consumed) = 0.028 x (%HAP)2.425

We used this HAP emission estimation equation in conjunction with facility-specific resin HAP content values to calculate the baseline HAP emissions for all vapor suppressed and nonvapor suppressed mechanical resin application operations with atomized spray. The use of pressure-fed rollers/flow coaters in mechanical resin application operations is considered an emission control technique. The derivation of emissions from mechanical resin application using pressure-fed rollers/flow coaters is discussed in section 5.3.2.2.2 of this chapter. 5.3.1.3 Filament Winding.6,7 HAP emissions data for uncontrolled filament winding are available from an emission study conducted at the Dow Chemical Company. This study examined the impact of the process parameters of resin HAP content, use of vapor suppressants, mandrel size (diameter), and ambient temperature on HAP emissions. The Dow study was conducted inside a total temporary enclosure (TTE) and used a total hydrocarbon analyzer to measure the HAP emissions. The resins used in the study were two Bis A epoxy vinyl esters with a viscosity of 300-400 centipoise and HAP contents of 33 and 48%. The resins were applied by filament winding a woven reinforcement 5 feet in length to create a laminate thickness of 0.25 inches. Each resin was tested at different mandrel sizes (6 and 33 inches), different levels of vapor suppressants (0% and 1.5%), and different ambient temperatures (73 and 85°F). A total of nine uncontrolled (i.e., non-vapor suppressed) test runs

5-14

TABLE 5-4. CFA PHASE I STUDY: MECHANICAL RESIN APPLICATION EMISSIONS a Resin Surface Amount Vapor Styrene Monome Applied Resin Styrene Flow Air Flow of Filler Suppress Emissions Emissions Gel Thickne r Run Rate (feet/mi (% filler) Time ss (lb/ton (% of No. Content . n) (inches) (minutes (lb/min) resin) (% resin) (% VS) ) styrene) 1 2 3A 3B 4A 4B 5 6 7 8 9 10 11A 11B 12A 12B 13 14 15 16

35% 42% 42% 42% 35% 35% 42% 35% 35% 42% 35% 42% 42% 42% 35% 35% 35% 42% 35% 42%

0.040 0.080 0.040 0.040 0.080 0.080 0.040 0.080 0.040 0.080 0.040 0.080 0.040 0.040 0.080 0.080 0.080 0.040 0.040 0.080

30 15 15 15 30 30 30 15 15 30 15 30 30 30 15 15 30 15 30 15

4 2 2 2 4 4 2 4 4 2 2 4 4 4 2 2 2 4 2 4

100 50 50 50 100 100 100 50 50 100 100 50 50 50 100 100 50 100 50 100

a

0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

6.6% 14.2% 13.3% 14.0% 6.6% 6.8% 15.1% 6.5% 8.3% 16.0% 8.2% 9.5% 12.1% 12.8% 6.4% 5.7% 6.5% 10.6% 7.2% 10.9%

131 284 266 281 133 137 303 130 166 321 164 189 242 257 129 115 130 213 144 218

The complete data set for the CFA Phase I mechanical resin emission study is found in electronic format in the docket.

5-15

consisting of eight runs with differing parameters and one duplicate run were conducted during the study. Table 5-5 presents the results of these test runs.

TABLE 5-5. DOW CHEMICAL COMPANY STUDY: FILAMENT WINDING EMISSIONSa

a

Run No.

Monomer Content (% Styrene)

Mandrel Diameter (Inches)

Ambient Temperature (°F)

Styrene Emissions (% of Resin)

Styrene Emissions (lb/ton resin)

1 2 6 7 9 13 14 16 18 (Duplic.)

33 48 48 48 33 48 33 33 48

6 6 33 33 6 6 33 33 33

73 85 85 73 85 73 73 85 85

4.93% 8.58% 10.28% 11.77% 5.90% 8.34% 7.21% 6.29% 11.96%

98.6 171.6 205.6 235.4 118.0 166.8 144.2 125.8 239.2

The complete data set for the Dow Chemical Company filament winding emission study can be found in Appendix B.

The results of the Dow Chemical Company study for the non-vapor suppressed resin runs indicated that the HAP emissions from filament winding are dependent primarily on HAP content of the resin. The study also indicated that mandrel size and ambient temperature have substantially smaller, secondary effects. We derived a HAP emission estimation equation by conducting a statistical analysis on the nine data points non-vapor suppressed resins obtained from the Dow Chemical Company study discussed above. The HAP emission estimation equation for filament winding derived from this statistical analysis is shown below. EF (lbs HAP/ton resin consumed) = 1.675 x (%HAP)1.225

5-16

We used this HAP emission estimation equation in conjunction with facility-specific resin HAP content values to calculate the baseline HAP emissions for all vapor suppressed and nonvapor suppressed filament winding operations. 5.3.1.4 Gel Coating.1,2,4,5 HAP emission data for the application of gel coats using a spray gun are available from four separate emission studies conducted by RTI, NMMA, and CFA (Phase I and Phase II studies). Each of these studies examined the effects that certain gel coating process parameters have on HAP emissions. The CFA Phase II study focused mainly on the effect of mold size and type on HAP emissions. Of the four studies conducted, the CFA Phase I study was the most extensive and balanced in examining the effects that certain process parameters have on HAP emissions from gel coating. The Phase II study looked mainly at the impact of mold size and type on emissions. We could not incorporate these variables into our emission estimation equation because we had no data on mold size and type from the facilities in our data base. Therefore, the studies conducted by RTI and NMMA and the CFA Phase II study were not used to on our development of an emission estimation equation for gel coating using spray guns. The CFA Phase I study examined the effects of gel coat HAP content, applied thickness of gel coat, gel time, gel coat application rate, and air flow. During the CFA Phase I study, 16 separate combinations of the above parameters were tested in order to evaluate the effect of each parameter on HAP emissions from gel coating. The gel coats used in the study were polyester gel coats with HAP contents of 35 and 40%. The gel coats were applied to a thickness of 0.018 or 0.024 inches with gel times of 10 or 20 minutes. The gel coats were applied at a rate of 2 or 4 pounds per minute with air flow across the mold at either 50 or 100 feet per minute. A total of twenty test runs consisting of sixteen runs with differing parameters and four duplicate runs were conducted during the study. Table 5-6 present the results of these test runs. The CFA Phase I study indicated that HAP content and the applied thickness of the gel coat are the primary factors affecting HAP emissions, with gel time, gel coat application rate, and air flow having insignificant effects. We derived a HAP emission estimation equation for gel coating using spray guns by conducting a statistical analysis on the twenty data points obtained from the CFA Phase I study 5-17

discussed above. The HAP emission estimation equation for gel coating derived from this statistical analysis is as follows. EF (lbs HAP/ton gel coat consumed) = 0.890 x (%HAP)1.675

5-18

TABLE 5-6. CFA PHASE I STUDY: GEL COATING EMISSIONS a

Run No.

1 2 3A 3B 4A 4B 5 6 7 8 9 10 11A 11B 12A 12B 13 14 15 16 a

Styrene Styrene Surface Monomer Applied Gel Coat Gel Coat Content Thickness Gel Time Flow Rate Air Flow Emissions Emissions (Inches) (Minutes) (lb/min) (feet/min) (% of Gel (lb/ton Gel (% Coat) Styrene) Coat) 35% 40% 40% 40% 35% 35% 40% 35% 35% 40% 35% 40% 40% 40% 35% 35% 35% 40% 35% 40%

0.018 0.024 0.018 0.018 0.024 0.024 0.024 0.024 0.018 0.018 0.018 0.024 0.018 0.018 0.024 0.024 0.024 0.018 0.018 0.024

20 10 10 10 20 20 20 10 10 20 10 20 20 20 10 10 20 10 20 10

4 2 2 2 4 4 2 4 4 2 2 4 4 4 2 2 2 4 2 4

100 50 50 50 100 100 50 100 100 50 50 100 100 100 50 50 50 100 50 100

18.5% 18.5% 21.5% 23.4% 16.8% 14.4% 19.6% 15.1% 17.5% 24.3% 20.1% 21.0% 24.8% 24.8% 16.1% 15.6% 15.6% 22.7% 17.2% 20.8%

371 371 431 469 336 288 393 302 351 486 403 420 495 497 322 311 313 453 344 415

The complete data set for the CFA Phase I gel coating emission study is found in electronic format in the docket.

5-19

We used this HAP emission estimation equation in conjunction with facility-specific gel coat HAP content values to calculate the baseline HAP emissions for all gel coating using a spray gun. 5.3.1.5 Closed Molding and Polymer Casting. 8 Closed molding processes include compression molding, injection molding, and resin transfer molding. The HAP emission factors for closed molding in AP-42 are 1 to 3% of available HAP. This emission factor range is also considered to be applicable to polymer casting, which is a semi-closed process. In addition, data are available from one emission test on an injection molding process at Glastic Corporation's Jefferson plant. The testing was done in 1993 following the protocol outlined in "The Measurement Solution: Using a Temporary Total Enclosure for Capture Efficiency Testing," (EPA-450/4-91-020). The test results show an emission factor of 1.1% of available HAP for injection molding, which falls within the range reported by the AP-42. The factors in AP-42 appear reasonable. These are closed or semi-closed processes, which should reduce HAP evaporation from the resin. Therefore we would expect the emission factors for these processes to be significantly lower than the factors for the open molding processes previously discussed. The average HAP emission factor selected for closed molding and polymer casting is 2% of available HAP, which represents the AP-42 midpoint for these operations. We used this average HAP emission factor to calculate the baseline HAP emissions for all closed molding and polymer casting operations. 5.3.1.6 Centrifugal Casting. 10 HAP emission data for centrifugal casting are available from a mass balance emission study conducted by the Fibercast Company. Fibercast collected mass balance emission data for five pipe sizes ranging from 2 to 12 inches manufactured on eight of their centrifugal casting machines. There were six samples taken from a 2-inch pipe size machine, five samples taken from a 4-inch pipe size machine, and four samples each taken from 6-inch, 8inch, and 12-inch pipe size machines. One sample each from the 2-inch and 4-inch pipe sizes was discarded as invalid because a leak in the pipes' end cap was detected. The average emission factors obtained from these samples were 87.2, 58.1, 50.9, 42.1, and 40.6% of available HAP for the 2-inch, 4-inch, 6-inch, 8-inch, and 12-inch pipes, respectively. 5-20

We selected an average HAP emission factor of 55.8% of available HAP for centrifugal casting, which was obtained by averaging the five average emission factors discussed above. We used this average HAP emission factor to calculate the baseline HAP emissions for all centrifugal casting operations. The individual average emission factors for each pipe size could not be used to calculate baseline HAP emissions because the facilities using centrifugal casting did not report the different pipe diameters they produce when responding to the 1993 industry screening survey. 5.3.1.7 Continuous Lamination/Continuous Casting. 11 - 20 For continuous lamination/casting three types of HAP emission data are available: (1) facility-specific source test data; (2) facility-specific emission factors; and (3) emission estimation equations. Facility-specific source test data are available from the Lasco Panel Products’ continuous lamination line and from one of the two continuous casting lines at the International Paper facility. The source test data from the Lasco Panel Products facility were obtained on the wide line at the facility in 1997. For International Paper, source test data were available for the particle line. Facility-specific emission factors to estimate wet out area and oven emissions are available from the Krane Kemlite-Joilet, the Krane Kemlite-Jonesboro, and the W. H. Brady facilities. These factors, which were provided by the companies, are used by the companies in their State permits. Emission estimation equations derived from emissions data are available from the source test conducted at Lasco Panel Products in 1997. Appendix B presents the emission estimation equations and a summary of the data used to develop these equations. The variables used in these emission estimation equations are: (1) the amount of available styrene and methyl methacrylate (MMA), (2) the vapor pressures of the styrene and MMA, (3) the typical thickness of the product (e.g., 90 mils), and (4) the mode of air management (either none or active ventilation over the wet out area). The vapor pressures of the styrene and MMA were calculated using Antoine’s equation. The temperature used in Antoine’s equation was an estimate of the average temperature of the resin over the length of the wet out area. This average temperature was estimated as a function of the temperature at which the resin is applied, the length of the wet out area, and an estimated increase in temperature per linear foot of wet out area. Table 5-7 summarizes the estimated uncontrolled emissions and the emission estimation methodology used for each continuous lamination and casting facility and the lines at that facility. 5-21

Three facilities control their emissions with add-on control devices. The controlled emissions are discussed in section 5.3.2.1. The emissions presented in Table 5-7 are estimated 1997 emissions, using estimated 1997 resin usage. As seen in Table 5-7, three methodologies were used -- facility-specific source test data, facility-specific emission factors, and emission estimation equations. Facility-specific source test data were used to estimate emissions from the Lasco Panel Products’ continuous lamination lines and from one of the two continuous casting lines at the International Paper facility. The source test data from the Lasco Panel Products facility were used to estimate both wet out area and oven emissions for the wide line and then were extrapolated to the two narrow lines at that facility. For the International Paper facility, source test data were used to derive the emissions from the particle line. The data show emissions from four areas (mixing, wet out area, oven, and storage). Using data in the source test report, we estimated the efficiency of the carbon adsorber used to control mixing, wet out area, and oven emissions and applied this efficiency to the tested controlled emission levels to derive an estimated uncontrolled emission rate (pounds per hour) from the wet out area and the ovens. We then multiplied this estimated uncontrolled emission rate by an estimate of the number of hours of operation in 1997 to estimate uncontrolled emissions from the wet out area and the ovens. Using data from the Lasco test, we assigned 80 percent of these emissions to the wet out area and 20 percent to the ovens. For the Krane Kemlite-Joilet, the Krane Kemlite-Jonesboro, and the W. H. Brady facilities, facility-specific emission factors were used to estimate their wet out area and oven emissions. These factors, which were provided by the companies, are used by the companies in their State permits. Wet out area emissions from the Krane Kemlite-Joilet and the W.H. Brady facilities were estimated by dividing the wet out area and oven emissions by a factor of 0.15, which was derived from the Lasco Panel Products test results for non-gel coat runs using a typical thickness of 90 mils. For the Krane Kemlite-Jonesboro facility, wet out area emissions were estimated by dividing the wet out area and oven emissions by a factor 1.0894, which was derived from the Lasco Panel Products facility test results considering both non-gel coat and gel coat runs. Different factors were used to reflect individual operations at these three facilities (e.g., the Jonesboro facility conducts gel coating while the Joilet facility does not). 5-22

TABLE 5-7. UNCONTROLLED EMISSION ESTIMATIONS (1997) FOR CONTINUOUS LAMINATION AND CONTINUOUS CASTING Uncontrolled Emissions (tons per year) Facility Enduro, Ft. Worth

Line

Wet Area Only

Wet Area and Oven

19.11

21.97

--

Methodology Equations 2 and 4

Glasteel, Collierville

--

54.59

60.05

Equations 1 and 4

International Paper, Odenton

Sheet

19.24

24.05

Equations 7 and 8

Particle

5.33

6.66

Derived from source test data

Facility Total

24.57

30.71

--

48.92

56.26

Equations 2 and 4

LP-3

8.35

9.60

LP-6

6.59

7.58

Wet out area: (wet out area plus oven emissions) / 1.15 Wet out area and oven: 0.0032 x available HAP

LP-7

5.65

6.50

Facility Total

20.59

23.67

5w

70.87

77.20

4n

17.72

19.30

Facility Total

88.59

96.50

1w

113.00

141.25

Equations 7 and 8

2n

56.50

70.62

Equations 7 and 8

3n

28.25

35.31

Equations 7 and 8

4n

28.25

35.31

Equations 7 and 8

Facility Total

226.00

282.49

87.76

95.61

Extrapolated from 1997 source test data and applied to estimate of 1997 throughput

2n

43.89

47.81

Extrapolated from source test data on Line 1w

3n

43.89

47.81

Facility Total

175.54

191.23

Resolite, Zelienople

--

38.79

42.67

Equations 1 and 4

W.H. Brady, Milwaukee

--

31.36

36.06

Wet out area: (wet out area plus oven emissions) / 1.15 Wet out area and ovens: 0.47 x available HAP

Kalwall, Bow Krane Kemlite, Joilet

Krane Kemlite. Jonesboro

Kemlite-Sequentia, Grand Junction

Lasco Panel Products, Florence

1w

Wet out area: (wet out area plus oven emissions) / 1.0895 Wet out area and oven: 0.04 x available HAP

NOTE 1: Equation numbers refer to equations presented in Appendix B.

5-23

The third methodology used is the application of emission estimation equations derived from emissions data obtained from the source test conducted at Lasco Panel Products in 1997. Appendix B presents the emission estimation equations and a summary of the data used to develop these equations. These equations were applied based on matching the types of ventilation used at facilities with similar ventilation scenarios tested at Lasco Panel Products. We then used facility specific information, where available, in the selected equations. Facility specific information used included HAP content by type of HAP, length of wet out area, temperature of resin as applied, and typical thickness of product. 5.3.1.8 Pultrusion. 21 HAP emission data for uncontrolled pultrusion operations are available from a 1994 emissions test conducted at the Glasforms, Inc., facility located in San Jose, California. The facility constructed an enclosure around a pultrusion line to capture HAP emissions and then measured those emissions using Bay Area Air Quality Management District (BAAQMD) Method ST-7. The line speeds during the testing did not exceed 18 inches per minute and the line tested produces products such as round rods, tubing, and flat bar. Two different air flows, one representing natural draft conditions (approximately 100 feet per minute) and the other representing forced air conditions (approximately 200 feet per minute), were tested for uncontrolled pultrusion operations. The testing indicated emission factors of 3.26% and 12.48% of available HAP for the natural draft and forced air conditions, respectively. We selected an average HAP emission factor of 3.26% of available HAP for uncontrolled pultrusion operations. We based this selection on the assumption that the emission factor for the natural draft conditions is more representative of the industry than the emission factor for forced air conditions. We used this average HAP emission factor to calculate the baseline HAP emissions for all uncontrolled pultrusion operations. The use of wet area enclosures with pultrusion processes is considered a control technique and is discussed in Section 5.3.2.3 of this chapter. 5.3.1.9 Sheet Molding Compound (SMC) Manufacturing. 22 HAP emission data for SMC manufacturing are available from a 1992 emissions test conducted at Applied Composites Corporation. At Applied Composites Corporation, SMC manufacturing occurs inside a totally

5-24

enclosed (negative pressure) building with several stacks emitting HAP pollutants from the various steps that constitute SMC manufacturing. Emissions from each of these stacks were measured under maximum production conditions and an empirically-derived emission factor of 2.67% of available HAP was obtained. We selected an average HAP emission factor of 2.67% of available HAP for SMC manufacturing. We used this average HAP emission factor to calculate the baseline HAP emissions for all SMC manufacturing operations. 5.3.1.10 Cleaning. Based on information provided in response to the 1993 ICR screening survey, methylene chloride is the most predominant HAP cleaner used in the reinforced plastic composites industry. Methylene chloride is a HAP with a boiling point of 40.1 degrees Celsius, which will evaporate completely when it is exposed to the atmosphere. We selected an emission factor of 100% of available HAP from cleaning operations. We assumed that all of the HAP cleaner that is purchased is eventually being emitted to the atmosphere. 5.3.1.11 BMC Manufacturing/Mixing. 9,23,24 HAP emission data are available for BMC manufacturing and for mixing. For BMC manufacturing, HAP emission data are available from one mass balance emission study (Fibercast Company) and one carbon tube test (Cytec Industries). The Fibercast Company conducted a mass balance emission study on a BMC mix composed of vinyl ester resin, polyethylene filler, clay filler, glass fibers, peroxide, pigment, and a mold release agent. Fibercast actively (forced) vented the BMC mixer used for this emission study. The preand post-mixing mass of the BMC mixture was determined using an electronic scale. The difference between the pre- and post-mixing mass was assumed by Fibercast to be completely due to the evaporation of styrene and indicated an emission factor of 1.5% of available HAP. Cytec Industries, Inc., conducted carbon tube testing for HAP on two types of actively vented BMC mixers while manufacturing BMC for four separate products. Average HAP emission factors for each product were calculated and ranged from 1.31 to 3.40% of available HAP. HAP emission data for open mixing with no active ventilation are available from a CFA evaluation of HAP emissions from the production of cultured marble sinks manufactured using 5-25

mixing and polymer casting operations. The test evaluation was conducted using a small open top mixer and a sink mold. The test results indicated that HAP emissions from the mixing operation alone were approximately 0.5% of available HAP. HAP emission data for the production of “densified” materials using vacuum mixing are also available from a CFA evaluation of the HAP emitted during a standard vacuum mixing cycle. The evaluation test encompassed an entire mixing cycle, including the opening of the vessel for the addition of materials at three separate stages of the mixing cycle. The test results indicated that HAP emissions from the vacuum mixing operation were approximately 3.1% of available HAP. We selected a HAP emission factor of 0.5% of available HAP to estimate the baseline HAP emissions from mixing operations conducted with open mixing vessels and no active ventilation of the mixing vessel. We selected this HAP emission factor because the CFA mixing test data from mixing resin and filler in an open mixer is representative of the practice of open mixing with no active ventilation used by facilities in the reinforced plastic composites production industry. For mixing operations with closed or covered mixing vessels and no active ventilation of the mixing vessel, we assumed a HAP emission factor of 0.25% of available HAP. This HAP emission factor assumes that HAP emissions from mixing operations conducted using closed or covered mixing vessels with no active ventilation of the mixing vessel are approximately ½ of HAP emissions from mixing operations conducted using open mixing vessels with no active ventilation of the mixing vessel. We assumed that the use of closed or covered mixing vessels with no active ventilation reduces HAP emissions due to the formation of a layer of HAP vapor between the resin/gel coat matrix and the mixer cover that reduces the vapor pressure differential between the HAP in the resin/gel coat matrix and the blanket layer of HAP vapor above the resin/gel coat matrix. Since the vapor pressure differential is the physical mechanism behind the movement of HAP from the resin/gel coat matrix to the open area above the resin/gel coat matrix (and subsequently to the atmosphere), we assumed that this lower vapor pressure differential reduces the emission of HAP. We used a HAP emission factor of 3.1% of available HAP to estimate baseline HAP emissions from vacuum mixing and mixing vessels (open or closed) actively venting the mixing

5-26

vessel to the atmosphere or an add-on control device. Active ventilation of an open mixing vessel entails the capture of HAP emissions from the open mixing vessel through the use of hoods and ducting located above (although not directly attached to) the open mixing vessel. Active ventilation of a closed mixing vessel entails the capture of HAP emissions from the closed or covered mixing vessel through the use of ducting attached directly to the mixing vessel. We selected a HAP emission factor of 3.1% of available HAP for vacuum mixing and open or closed actively vented mixing vessel because the CFA test data for vacuum mixing is considered representative of these types of mixing practices. This factor also falls within the range of test results at Cytec. 5.3.1.12 Storage of HAP Containing Materials.23 HAP emission data are available from an emission study conducted by the CFA. The CFA measured HAP emissions from open 5 gallon pails and 55 gallon drums of HAP containing material. Table 5-8 presents the results of this emission study.

TABLE 5-8. CFA EMISSION STUDY: HAP EMISSIONS FROM OPEN STORAGE CONTAINERS Measurement Description

5 Gallon Pail

55 Gallon Drum

Total Emissions Weight (g)

6.618

25.371

Emissions (g/min)

0.343

1.208

Emissions (g/hr)

20.57

72.49

Emissions (g/24 hr)

493.8

1739.7

Emissions (lbs/min)

0.0008

0.0027

Emissions (lbs/hr)

0.0453

0.1597

Emissions (lb/24 hr)

1.0876

3.8320

0.65

2.76

0.0005

0.0073

Surface Area (ft2) Emissions (g/ft2/min)

5-27

We selected a HAP emission factor of 0.016 lbs per hour per 55 gallon drum to estimate the baseline HAP emissions that occur from the storage of HAP containing materials. We selected the HAP emission factor for 55 gallon drums (as opposed to 5 gallon pails) based on information provided in response to the 1993 ICR screening survey sent to reinforced plastic composites production facilities by EPA and information received from resin and gel coat suppliers. This information indicates that the majority of facilities receive resins and gel coats in shipments of multiple 55 gallon drums. In order to use this emission factor, we had to estimate the number of 55 gallon drums of HAP containing materials used annually by each facility. Since information on the amount of HAP containing materials (primarily resin and gel coat) used by facilities is only available on a weight basis (pounds), we had to use the density of the HAP containing materials to convert the amount of HAP containing materials to a volume basis (gallons). The actual density of the materials varies depending on the type of material and the material’s HAP content. Since the actual densities of the specific resins and gel coats used by each facility were not available, we assumed approximate densities for different types of materials based on values indicated in guides from resin and gel coat manufacturers. Table 5-9 summarizes the densities used to convert pounds of HAP containing materials to gallons of HAP containing materials.

TABLE 5-9. DENSITY VALUES USED TO CONVERT MATERIALS FROM POUNDS TO GALLONS Density (lbs/gallon)

Material All Types of Resin (Corrosion Resistant, Non Corrosion Resistant, Tooling)

10.0

Pigmented Production Gel Coats

10.5

Clear Production Gel Coats

8.75

5-28

Tooling Gel Coats

10.5

We obtained the total amount of HAP containing material in gallons used by a facility by dividing the amount (in pounds) of each type of material by the assumed density (in pounds/gallon) of that material and then summing the results (in gallons) for all materials used by the facility. We obtained the number of 55 gallon drums of HAP containing materials used annually by a facility by dividing the total amount of HAP containing materials (in gallons) used annually by a facility by 55 gallons per drum. Once the number of 55 gallon drums of HAP containing materials used annually by a facility was calculated, we had to make an assumption concerning the average length of time the storage containers used by a facility are open before mixing occurs or all HAP containing material in the storage container is removed. The actual time a storage container is open is dependant on the methods used by the facility to empty the container and whether the facility opens several containers and stages them in a holding area prior to using the materials. We assumed that, on average, the storage containers used by reinforced plastic composites production facilities were open for one hour before mixing occurs or all HAP containing material in the storage container is removed. We multiplied the total number of 55 gallon drums of HAP containing materials used annually by each facility by the one hour each 55 gallon drum is assumed to be opened to estimate the annual emissions (in pounds) from the storage of HAP containing materials at each facility. We estimated the total annual baseline emissions from storage for the reinforced plastic composites production source category (in tons per year) by summing these individual facility totals (in pounds) and dividing by 2000 pounds per ton.

5.3.2

Derivation of Emission Estimation Equations for Facilities Using Control Techniques 5.3.2.1 Add-On Control Devices. We estimated HAP emissions from facilities using an

add-on control systems by multiplying the process specific emission factor or equation that best

5-29

estimates the emissions from a process prior to the emissions' destruction in the control system by 1 minus the overall control efficiency (CE) of the control system (expressed as a decimal). The overall CE of the control system is determined by multiplying the overall capture efficiency (CTE) of emissions from the entire process by the destruction efficiency (DE) of the control device. This HAP emission estimation equation is shown below.

EFpost-control system = EFpre-control system x [1 - CE] = EFpre-control system x [1 - (CTE x DE)] We obtained the capture and destruction efficiency of the control system from the 1993 industry surveys, facility permits, telephone contact reports, and site visit reports. For facilities that reported meeting EPA's five point criteria for a permanent total enclosure, we assumed a capture efficiency of 100 percent. Table 5-10 presents facility specific information on controlled emissions from three continuous lamination/casting facilities. Appendix E presents facility-specific information pertaining to the type and effectiveness of add-on control devices for all facilities. We used the above assumptions and HAP emission estimation equation in the calculation of baseline HAP emissions when facilities indicated the use of an add-on control device. 5.3.2.2 Pressure-Fed Roller/Flow Coater Application. 2,4,5 HAP emission data are available from three emissions studies conducted by RTI, CFA, and NMMA. Each of the studies examined the level of HAP emissions produced when pressure-fed rollers or flow coaters were used to apply resins in the fabrication of plastic composites. None of these studies tested pressure-fed rollers or flow coaters using a wide variety of process parameters. Two of the studies (CFA, RTI) used only one HAP content and other process parameters such as laminate thickness, resin application rate, gel time, and air flow across the mold were not varied extensively during the testing. A comparison of the results of each of the studies shows that HAP emissions from pressure-fed roller/flow coater application are dependent primarily on resin HAP content. A comparison of the HAP emissions for the tested HAP content ranges indicated that emissions from pressure-fed roller/flow coater application are somewhat similar to manual resin

5-30

application. This similarity may be attributed to similarities in the application technique. Both types of application are essentially methods of transferring resin to a mold in a non-atomized state prior to nearly identical final roll out and curing stages. As such, for the same process parameters, their emission profiles are also likely to be similar. In selecting an emission factor for pressure-fed roller and flow coater applications, we considered the data set for HAP emissions to be insufficiently balanced to use as the basis for deriving an average HAP emission factor or equation. Instead, because of the similarities in emission levels and profiles between pressure-fed roller/flow coater applications and manual

5-31

TABLE 5-10. CONTROLLED EMISSIONS (1997) FOR CONTINUOUS LAMINATION AND CONTINUOUS CASTING

Facility Enduro, Ft. Worth

International Paper, Odenton

Line

Controlled Emissions (tons per year)

Comments

3.82

Controlled emissions estimated using 100 percent capture of wet out area emissions and 95% destruction of captured emissions in an incinerator; oven emissions are uncontrolled.

12.15

Controlled emissions estimated using a 37.5% capture of wet out area emissions, 100% capture of oven emissions, and 99% reduction of captured emissions in a carbon adsorber.

0.60

Controlled emissions estimated using a 89% capture of wet out area emissions, 100% capture of oven emissions, and 99% reduction of captured emissions in a carbon adsorber.

--

Sheet

Particle

Krane Kemlite, Jonesboro

Facility Total

12.75

1

17.33

Controlled emissions estimated using 80 percent capture of wet out area emissions, 100 percent capture of oven emissions, and 95% destruction of the captured emissions in an incinerator.

2 4.33 Facility Total

21.66

5-32

resin application, we considered the data for manual resin application to be the best available data with which to estimate HAP emissions from pressure-fed roller and flow coater applications. Therefore, we used the same HAP emission estimation equation derived for manual resin application for estimating emissions for pressure fed roller/flow coater application as shown below. EF (lbs HAP/ton resin consumed) = 0.028 x (%HAP)2.275

We used this HAP emission estimation equation in conjunction with facility-specific resin HAP content values to calculate the baseline HAP emissions for facilities using pressure-fed roller and flow coaters. 5.3.2.3 Pultrusion Wet Area (Resin Bath) Enclosures.21 HAP emission data for pultrusion operations that are controlled using wet area enclosures are available from a 1994 emissions testing conducted at the Glasforms, Inc. facility located in San Jose, California. The facility constructed a TTE around a pultrusion line to capture HAP emissions and then measured those emissions using Bay Area Air Quality Management District (BAAQMD) Method ST-7. Three different air flows, one representing natural draft conditions and two representing forced air conditions were tested for pultrusion operations utilizing wet area enclosures. The testing indicated emission factors of 1.33%, 2.64%, and 2.49% of available HAP under natural draft conditions (approximately 80 feet per minute), forced air conditions using one exhaust fan (approximately 150 feet per minute), and forced air conditions using two exhaust fans (approximately 200 feet per minute), respectively. We selected an average HAP emission factor for pultrusion operations using wet area enclosures of 1.33% of available HAP with the exhaust rate of 80 feet per minute. We based this selection on two factors: (1) the assumption that the emission factor for natural draft air conditions is more representative of the industry than the emission factor for the forced air conditions and (2) the selection of the corresponding set of values (i.e., HAP content and flow rate) to develop the uncontrolled emission factor (see Section 5.3.1.11). We used this average

5-33

HAP emission factor to calculate the baseline HAP emissions for facilities using wet area enclosures to control HAP emissions from pultrusion operations.

5.4

REFERENCES

1.

Lacovara, R., Composite Fabricators Association and Craigie, L., Cowley, T., Wykowski, P., Webster, G., Dow Chemical Company. Phase I-Baseline Study, Final Report. September 1996. (CFA Phase I Study)

2.

Kong, E., Bahner, M., Wright, R., and A. Clayton, Research Triangle Institute. Evaluation of Pollution Prevention Techniques to Reduce Styrene Emissions from Open Molding Contact Processes. September 1995.

3.

Research Triangle Institute. Evaluation of Resin and Filler Variables Affecting Styrene Emissions from Filled Resin Spray up Applications. Unpublished.

4.

Haberlein, R., Engineering Environmental, on behalf of the Composite Fabricators Association, International Cast Polymer Institute, Society of Plastics - Composites Institute. Derivation and Verification of CFA Emission Models. (CFA Phase II Study)

5.

Stelling Engineering, P.A., Air-Tech Environmental LLC, Radian International LLC, on behalf of the National Marine Manufacturers Association. Baseline Characterization of Emissions from Fiberglass Boat Manufacturing for National Marine Manufacturers Association. August 1997.

6.

Letter and attachments from Craigie, L., Dow Chemical Company. Letter to R. Jemison, Pacific Environmental Services, Inc. and Madeleine Strum, EPA/ESD. November 10, 1997. Table of the results from the Dow Chemical Filament Winding Emission Study that specifically provides emission profile information. 3 pp.

7.

Letter and attachments from Craigie, L., Dow Chemical Company. Letter to R. Jemison, Pacific Environmental Services, Inc. December 17, 1998. Summary table of results from Dow Chemical Filament Winding Emission Study. 2 pp.

8

U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission Factors (AP-42). U.S. Environmental Protection Agency, Research Triangle Park, North Carolina. pp. 4.12.1-4.12.10.

9. Boyer, J., Molding Press Emissions Testing at Glastic Corporation Jefferson (Ohio) Plant. April 23, 1993. Unpublished. 5-34

10.

11.

Letter and attachments from Kinman, R., Fibercast Company to M. Strum, EPA/ESD. November 19, 1996. Letter in response of information requested during the July 10, 1996 site visit. 159 pp. Telecon. Stanier, C., International Paper (formerly Nevamar Decorative Surfaces Corp.) with Strum, M., EPA/ESD. February 16, 1995. Telecon discussing process conditions, product formulations, emission points, and control measures for the continuous casting and polymer casting processes at the International Paper facility located in Odenton, Maryland. 5 pp.

12.

Telecon. Stanier, C., International Paper (formerly Nevamar Decorative Surfaces Corp.) with Dunstun, W., EPA/ESD. November 21, 1995. Telecon and attachments discussing the process conditions for and HAP emissions from the continuous casting and polymer casting processes at the International Paper facility located in Odenton, Maryland. 8 pp.

13.

Nevamar Corporation. United States Patent 5,183,600. November 19, 1993. Method and Apparatus for Continuous Casting of Polymerizable Thermosetting Material. 14 pp. Docket No. II-I-49

14.

Pacific Environmental Services, Inc. Air Emissions From Reinforced Plastics Composite Manufacturing Process LASCO Panel Products. Vol. I. Prepared for U.S. Environmental Protection Agency. Research Triangle Park, NC. Publication No. EPA454/R-99-024a. August 1999. 15.

Pacific Environmental Services, Inc. Air Emissions From Reinforced Plastics Composite Manufacturing Process LASCO Panel Products. Vol. II. Prepared for U.S. Environmental Protection Agency. Research Triangle Park, NC. Publication No. EPA-454/R-99-024b. August 1999.

16.

Pacific Environmental Services, Inc. Air Emissions From Reinforced Plastics Composite Manufacturing Process LASCO Panel Products. Vol. III. Prepared for U.S. Environmental Protection Agency. Research Triangle Park, NC. Publication No. EPA-454/R-99-024c. August 1999.

17.

Letter and attachments from North, H., LASCO Panel Products to M. Toomer, EPA. May 27, 1998. Letter summarizing process conditions for and HAP emissions from the continuous lamination production lines during an emission test conducted by the EPA at the LASCO Panel Products facility in Florence, Kentucky. 2 pp.

18.

Facsimile from Margherio, P., Kemlite Company, Inc. to Strum, M., EPA/ESD. July 7, 1998. Facsimile summarizing resin usage, monomer content, and emissions data for the Kemlite Company facility in Joliet, Illinois. 2 pp.

5-35

19.

Memorandum and attachments from LaFlam, G. and M. Proctor, Pacific Environmental Services, Inc., to Strum, M., EPA/ESD. September 24, 1996. Summary of August 19, 1996 meeting between EPA and representatives of the Kemlite Company. 10 pp.

20.

Telecon. Meardon, K., Pacific Environmental Services, Inc., with Hennessy, P., W.H. Brady Company. May 29, 1998. Process and emissions information for the continuous lamination line located at the W.H. Brady facility in Milwaukee, Texas. 2 pp.

21.

McQuarrie, T. The Pultrusion Industry Council of the Society of Plastics - Composites Institute. Industry Proposal: Existing Source MACT for Pultrusion. October 1997.

22.

Applied Composites Corporation. Response to Section 114 Screening Survey for the St. Charles, Illinois facility. September, 1993.

23.

Letter and attachments from McNally, S., Composite Fabricators Association, to Madeleine Strum, EPA/ESD. December 15, 1998. Industry recommendations for polymer casting, mixing and blending, storage, equipment cleaning, acrylic sheet, and vacuum bagging/closed mold curing. 15 pp.

24.

Letter and attachments from Leitzman, R., Cytec Industries, Inc. to M. Strum, EPA/ESD. October 21, 1996. Letter in response of information requested during the July 25, 1996 site visit.

5-36