EPA/600/A-92/145
                                                    92-139.15
Using a Flame lonization Detector  (FID) to Continuously
  Measure Toxic Organic Vapors in a Paint Spray Booth

                  Jamie K. Whitfield,
          U.S.  Environmental  Protection Agency
 Air and Energy Engineering Research Laboratory. MD-61,
            Research Triangle Park, NC 27711

             Gary  B.  Howe, Bruce A.  Pate,
              Research Triangle  Institute
                     P.O.  Box 12194
            Research  Triangle Park,  NC  27709

                 and Joseph  D. Wander
  USAF  Civil Engineering  Laboratory (-HQ AFCESA/RAVS)
               Tyndall  AFB, FL 32403-6001

-------
                                                        92-139.15

INTRODUCTION

     U.S. Air Force  (AF) industrial operations procuring and
maintaining equipment and facilities are the greatest sources of
volatile organic compound (VOC) and hazardous air pollutant  (HAP)
emissions at AF installations. These emissions are subject to
stringent environmental regulations that have been evolving for
two decades, and are expected to continue their evolution for
decades to come. The USAF has stated its commitment to complying
with the new environmental standards, but two considerations
prevail: (1) maintaining the current standard of workplace
safety, and (2) keeping the cost of the emission control as low
as possible.

Program History

     The AF Civil Engineering Laboratory  (CEL) and Environmental
Protection Agency's  (EPA's)  Air and Energy Engineering Research
Laboratory  (AEERL) began a study of AF air compliance issues in
1986. The initial study showed that paint spray booths at several
sites tested would not comply with potentially more stringent
emission regulations unless emissions levels were lowered.1  To
combat this problem, the report suggested that exhaust
recirculation be considered as a means of decreasing the volume
of exhaust air requiring treatment and, therefore, lowering the
pollutant emissions. Unfortunately, this innovative
recommendation met opposition in the AF industrial hygiene  (IH)
community.

     The primary objection raised by AF IH personnel references
29 Code of Federal Regulations (CFR) 1910.107 (d) (9),2  which
explicitly states that recirculation of emissions from paint
spray booths is prohibited.  Upon further consideration,
Occupational Safety and Health Administration (OSHA) issued a
letter3 to  EPA stating  that  the regulation in question pertains
to fire safety and not to personnel toxic exposures. Instead,
OSHA referenced 29 CFR 1910.10004  for toxic  exposures, which does
not prohibit recirculation.  29 CFR 1910.107  (d)   (9)  was not
intended to prohibit recirculation or institution of innovative
technologies.  Rather, it ensures that booth operations remain
within safe levels. Thus,  under the OSHA de minimus rule,  a booth
may use recirculation if the booth irrefutably complies with 29
CFR 1910.1000. Based on this compromise,  a demonstration study
was developed at Travis Air Force Base (AFB), California.

     The CEL/AEERL team conducted two site studies of spray
booths, and will conduct more studies at Travis  AFB in June 1992.
The first study was conducted at McClellan AFB,  CA.  It determined
that fluidized-bed incineration and an adsorption/desorption
preconcentration process will effectively remove VOCs

-------
                                                        92-139.15

from exhaust gas. The disadvantage of both systems is that the
estimated cost of treating the full circulating stream may be
exorbitant5  depending on the  volumetric  flow  to  be treated.

     A second study was conducted at Hill AFB,  Utah. The spatial
distribution of volatile, metal, and isocyanate species was
measured in a horizontal-flow, waterfall booth.  The results show
that the toxics concentrate near the bottom of the booth and in
the painter's breathing zone.6 The  concentration in the  painter's
breathing zone appears to result from localized eddy currents in
the air stream, which are caused by the presence of the painter's
body as an obstacle. Armstrong Laboratory's IH newsletter has
cited excessive exposures to toxic metals, under just such
conditions,  in a context that implicitly recommends wearing a
respirator while painting in a horizontal-flow booth.7

     Travis AFB serves as the test site for two studies to be
conducted. Both studies are designed such that workplace safety
standards are not sacrificed and comply with OSHA regulations and
directives.  For example, during the study, an informed volunteer
painter will be equipped with a supplied air respirator and skin
protection.  This gear is commonly required by AF shops conducting
spray painting operations. Also, air toxic measurements will be
taken inside and outside the respirator and,  as an added safety
measure,  the painter will be medically evaluated before and after
painting tests.

     Also incorporated into the Travis AFB study is the design of
a fail-safe diversion system for converting the recirculating
flow pattern back to "straight-through" ventilation. This relies
upon the use of a flame ionization detector (FID)  to monitor the
concentration levels within the booth. The information obtained
from using this instrument for measurement gives a total
hydrocarbon analysis, as opposed to a speciated compound
analysis. However,  the compound of interest can be identified by
calculating its percent composition of the total,  assuming
complete volatilization of the paint system's components.

     The Travis AFB study will provide an experimental basis for
evaluating the impacts of exhaust modifications on air toxic
emission rates and on attainment of IH standards.

Project Objective

     This project,  which is a component of the justification
process for the experimental evaluation of flow-splitting and
exhaust recirculation, is the final step in preparation for
evaluation of these techniques. Although the immediate purpose is
to establish that the Travis AFB test does not present an
extraordinary health risk to the volunteer painter and other

-------
                                                        92-139.15

personnel participating on the testing team in the booth, the study
will  also  apply  to  any  other  volatile  mixture  of  defined
composition in a paint spray booth.

     Specifically, the objective is to demonstrate experimentally
that the response of a FID is adequate for determining, reliably,
that the concentration of toxic compounds will remain, at all
times, less than 25% of the short term exposure limit  (STEL).
This FID system must activate when:  (1) the instantaneous
concentration exceeds 25% of the STEL8" for the coating and
catalyst, NSN 8010-01-336-3036,  used exclusively during the
demonstration, (2) a concentration of 23% of the STEL is
maintained for 15 consecutive minutes, or (3) a concentration of
20% of the STEL has been measured during a total of 180
nonconsecutive minutes during a single painter's shift. This
mechanism ensures that the exposure standards specified in 29 CFR
1910.1000 are not exceeded at 15-minute or 8-hour time-weighted
levels during the test.

     The purpose of generating these data is to confirm that,
when accurately set and properly maintained and calibrated, the
FID can reliably detect any excursions above an arbitrarily
selected threshold, sound an alarm to alert site personnel, and
trigger devices that reconfigure the ventilation system and/or
disable the spray gun and intercede before harmful exposure
develops during the spray booth operation.

EXPERIMENTAL DESIGN

Continuous Analyzer

     A total hydrocarbon FID from Ratfisch Instruments, model
RS-55CA,  was used for the project. The analyzer is configured
with four measuring ranges: 0-100, 0-1000, 0-10,000,  and
0-100,000 parts per million as carbon  (ppmC). Sample gas was
introduced into the hydrogen flame through a sample capillary at
a low flow rate.  The FID response was based on the concentration
of the VOCs in the sample gas and the type of compounds present.
The analyzer output consisted of both a 0 to 1 V analog recorder
signal and a front panel display.

Test Mixture Composition

     This project specifically addresses the health risks
attending the use of the polyurethane coating and aliphatic
isocyanate catalyst,  NSN 8010-01-336-3036. Lists of the hazardous
components for the specified coating and catalyst were identified
in the manufacturer's Material Safety Data Sheets (MSDSs).  The
weight percentage of each component was used to calculate a
combined weight in the mixture (3:1 ratio by weight of coating to

-------
                                                                92-139.15

    catalyst), from  which the weight percentage  was calculated based
    on  the VOC content only. The  calculated  composition of  the
    mixture is shown in Table I.
        TABLE 1.  COMBINED COATING/CATALYST SOLVENT COMPOSITION8

Compound

ri-Butyl acetate
Ethyl-3-ethoxypropionate
Toluene
Xyleneb
Methyl isobutyl ketone
Methyl ethyl ketone
Ethylbenzene
PG.ME acetate*
2, 4-Pentanedione

Wt. % (NSNC 8010-01-336-3036)
Coating Catalyst Combined

5
5
1
1
5
5
0.1
1
5

5
20
0
5
35
0
0
0
0

5
8.75
0.75
2
12.5
3.75
0.075
5
0.75
3.75

wt. % of voc
content
13.4
23.4
2
5.4
33.5
10
0.2
2
10

a3:1 mixture  (by weight)
bSubstituted  xylene for "C8 & CIO aromatic hydrocarbon" from MSDS for
 03GN240CAT.
cNational  Stock Number
dPropylene glycol monomethyl ether  (PGME) acetate is the common name for  1-
 Methoxy isopropyl  acetate.
   Calculation of Exposure Limit

         The STEL for each component is shown in Table  2. Where available,
   the STEL given corresponds to the OSHA permissible exposure  limit  (PEL).
   For compounds not having an OSHA PEL-STEL, the STEL as assumed  to  be 1.5
   times  the manufacturer-recommended  Threshold Limit  Value (TLV)-time
   weighted average  (TWA).

         Application of the ACGIH rule*b for additive effects to the  data
   in Table 2 resulted in a calculated STEL for the  coating  and-catalyst,
   NSN 8010-01-336-3036,  of 350 mg/m3.  This value is based on the
   assumption that volatilization is complete and uniform; although  this
   is not strictly correct, toxicity of the less-volatile constituents is
   greater, so the calculated value underestimates the  "true"  exposure
   limit.  (Comparable STELs were calculated for several other
   polyurethane coatings presently in AF inventory,  but a significantly
   lower value was calculated for MIL-P-23377 primer.)

-------
                                                             92-139.15
          TABLE 2.   COMBINED COATING/CATALYST STEL  CALCULATION
Compound
n-Butyl acetate
Ethyl-3 ethoxypropionate
Toluene
Xylene
Methyl isobutyl ketone
Methyl ethyl ketone
Ethylbenzene3
PGME. acetateb
2, 4-Pentanedione
Wt. % (of
VOC)
13.4
23.4
2.0
5.4
33.5
10.0
0.2
2.0
10.0
STEL,
aig/m3
950
447
560
655
300
885
545
810
123
Source of STEL
OSHA PEL
Manufacturer
OSHA PEL
OSHA PEL
OSHA PEL
OSHA PEL
OSHA PEL
Manufacturer
Manufacturer
 Mixture STEL=1/(0.134/950+0.234/447+0.02/560+0.054/655+0.335/300+0.1
    /885+0.002/545+0.02/810+0.1/123) = 350 mg/m3

 "Omitted from mixture  since weight %  is  insignificant.
 bPropylene glycol monomethyl ether (PGME)  acetate  is  the  common  name
 for 1-Methoxy isopropyl acetate.
Test Mixture Preparation

          Nine mixtures were prepared for characterizing the analyzer
response in the 0-1000 ppmC range and three mixtures were prepared for
the 0-100 ppmC range. The highest concentration mixture was prepared
by vaporizing a known mass of a liquid mixture and flushing it into an
evacuated aluminum cylinder with a known volume of dry air. The liquid
mixture was prepared by adding a given mass of each compound to a
glass vial, based on the weight percentages shown in Table 1.

          Subsequent mixtures were prepared by withdrawing a portion of
the gaseous mixture and refilling the cylinder with dry air. A bourdon
tube vacuum/pressure gauge was used to measure accurately cylinder
pressures for determining test mixture concentrations.

Test Mixture Analysis

     During normal  field operation,  the  Ratfisch  RS-55CA analyzer  uses
an internal pump to draw a high volumetric flow rate of sample through
a sample port, past a tee at the FID, and then through a back-pressure
regulator before exhausting through the bypass port.  The back-pressure
regulator is used to adjust the sample gas pressure,  which controls
the flow rate of sample gas through a capillary and into the FID.  The
normal pressure is 2 psig (13.8 kPa).

     Since  the sample gas  mixture  in this study was contained in an
aluminum cylinder at approximately 40 psig (275.8 kPa),  operation of
the analyzer was modified. The internal sample pump was not operated,
but was left in the sample flow path. To conserve sample test mixture,

-------
                                                             92-139.15

the bypass port was fitted with a shutoff valve. When this valve was
closed, the sample back-pressure regulator was not functional.  The
sample gas pressure was regulated with an in-line single-stage
pressure regulator installed at the outlet of the test mixture
cylinder. Similar regulators were used to set the gas pressures for
the propane calibration gases and the zero air.

     The  FID was  operated with normal settings for hydrogen fuel (6 psig
or 41.4 kPa)  and combustion air (11 psig or 75.8 kPa).

     Each of the  test  mixtures  and calibration  gas mixtures  was  sampled
in the analyzer,  and the FID response was recorded on a strip chart
recorder. The in-line pressure regulator was adjusted so that the same
sample pressure was used in all analyses.

RESULTS

Propane Calibration and VOC Mixture Analysis Results

     The  analyzer was  first  calibrated  for  the  0-1000 ppmC range.  With
zero air passing into the sample inlet port, the zero potentiometer
was adjusted to produce a zero output from the FID.  Then,  with a
standard containing 525 ppmC propane in air passing into the sample
inlet port,  the span potentiometer was adjusted to produce an analyzer
output reading corresponding to 525 ppmC. Three additional
concentrations of the standard containing propane in air were also
analyzed.

     After changing  the  analyzer range  setting  to 0-100  ppmC, the  zero
and span were  adjusted while sampling zero air and a standard
containing 46  ppmC propane in air  standard,  respectively.  Two
additional concentrations of the standard containing propane in air
were also analyzed in this range.

     To illustrate the response of characteristics of the  analyzer to both
propane  and the  VOC  test  mixture,  plots  of  analyzer  response  versus
concentration  for both  the  1000 and the  100  ppmC ranges a^re  shown  in
Figures  1 and  2.   The  concentrations  for  the propane  and  the  solvent
mixture are expressed in milligrams per  cubic  meter. Linear regression
analysis was performed  for each set of data, and the results  are  shown
in each figure,  respectively.

     Both of these figures  display straight-line fits to data  generated
by passing the solvent mixture,  representative of the volatile  fraction
of the coating and catalyst, and  the calibration gas,  propane,  through
the RS-55CA FID.  The results of the linear regression  analyses of the two
figures further support the  statement that  FID  response to  the  organic
analytes was linear over  the entire range measured  and well below the
STEL calculated for the coating and catalyst,  NSN  8010-01-336-3036,  or
any other paints  for which the STEL was estimated.

-------
   800
O 700
E

S 600
oJ

§500
O.

| 400
k_
Q)
£ 300
TO

< 200
   100


     0
                                                                92-139.15
       -Propane
         Linear Regression
                Slope     = 1.68
                y-intercept = -6.59
                R squared = 0.9998
                         I
                                I
                                            Solvent Mixture
                                            Linear Regression
                                                  Slope     =1.078
                                                  y-intercept = 14.87
                                                  R squared = 0.994
                100     200      300      400       500

                              Concentration, mg/m3
                                                            600
                                                                  700
            Figure 1. Ratflsch RS-55CA Calibration 0-1000 ppmC Range.
o

a.
o.

-------
                                                             92-139.15

DISCUSSION

     The test at Travis AFB will  use a  servo-actuated set of'flap valves
to   convert   a  horizontal-flow  spray   booth   from  straight-through
 (single-pass) circulation  into a partially recirculating system. During
recirculation,  a portion  of the  exhaust  stream will be filtered,  mixed
with an equal•volume of fresh air, and returned via an intake filter to
the  circulating air stream. For  the  test,  the exhaust  stream  will be
released  without treatment. However,  the concentrations of  air  toxic
materials will be measured and the acquired data will be used to estimate
the  cost  and  effect   of  application  of  available  technologies  for
purification of the reduced exhaust stream prior to releases.

     Similar measurements  will also be made on the recirculated stream to
evaluate  the impact  on  IH safety standards.   Calculations  indicate
that, as  long as the particulate control  system (dry filters) is
maintained and  operated properly, the  increase in toxic exposure will
be negligible in comparison to ambient spray booth concentrations.
However,  no calculation can exclude all potential risk of overexposure
to personnel in the spray  booth. A prudently conservative attitude
dictates  that the experiment be  conducted under circumstances that
minimize  both the risk  of  exposure and the extent to which
overexposure could occur.  Risk minimization is the purpose  of
incorporating the FID  into the study.

     As designed, the recirculating booth will include the FID
immediately behind the  intake filters. The FID will continuously
•measure the concentration  of combustible  airborne volatile  substances
and report to two triggering devices.  As in the case of the solvent
mixture,  no other sources  of toxic substances exist, so any additional
contributions by non-halogenated combustibles will overestimate the
exposure.

     The  first   triggering  device  will  respond  immediately  to  any
instantaneous   excursion   above   350   mg/m3,   the   calculated  STEL.
Overexposure would  then be limited to the  air  volume  passing into and
through  the  intake plenum and  booth  during  the  fraction  of  a second
required  for the servos to switch the  flap valves to a  fresh air intake.

     The second  triggering device, a computer, will  store  a record of the
FID outputs  at  a fixed sampling rate, and  convert  to straight-through
ventilation if either of the following two criteria is met: 1) a period
of 30 consecutive seconds is maintained at 87.5  mg/m3 (25% of the STEL),
or 2)  a total of 180  nonconsecutive minutes  at  87.5 mg/m3  (25% of the
STEL) is  accumulated during a single painter's shift.

     The results in Figure 1 include,  as  the lowest point in the range,
105 mg/m3, and the  results in Figure 2 include,  as  the highest point in
the range, 77 mg/m3.   Thus, an extrapolation is made to cover the range
from 77 to  105  mg/m3 that  includes  the point 87.5  mg/m3,  or 25% of the
STEL. The results in Figures 1 and  2  demonstrate that  the FID exhibits
sensitivity and linearity of response adequate to discriminate organics
accurately below 87.5 mg/m3.

-------
                                                             92-139.15
CONCLUSIONS
     This  study  has demonstrated  linear  and similar  responses  of  a
Ratfisch RS-55CA  FID to  a  solvent mixture  identical to the  volatile
organics in the coating and  catalyst,  NSN 8010-01-336-3036,  and to the
calibration  gas,   propane,  used  in  field  calibrations  of  the  FID.
Sensitivity and linearity  have been shown  to extend from 715 to  45 mg/m3,
which  brackets  the  calculated  STEL  and  lower  action  thresholds.
Monitoring  is  maintained constantly,  and,  under  field  conditions,
equilibration  occurs rapidly and  analysis  and output  transpire  in
milliseconds. As a  trigger for fail-safe  conversion from recirculation
mode into a straight-through booth configuration,  the FID may confidently
be  expected  to initiate  a  corrective  response  before  a  transient
elevation of VOC concentrations overexposes area personnel.


REFERENCES

1.   J. Ayer  and D.  Wolbach,  Volatile Organic Compound and Particulate
Emission Studies of Air Force Paint Booth Facilities;
Phase  I,  ESL-TR-87-55;  EPA-600/2-88-071  (NTIS  ADA19803Z),   Air  Force
Engineering  and Services  Center, Tyndall  AFB, 1988.

2.   29 CFR Ch. 17,  1910.107,  Occupational Safety and Health Standards;
Spray  Finishing  Using Flammable  and  Combustible Materials:  Federal
Register, June, 1974, pp. 183-185.

3.   T.J.  Shepich,  letter to  S.R. Wyatt, U.S. EPA, 16 January 1990.
Reproduced as Appendix G  of ESL-TR-89-35;  EPA-600/2-90-051 (NTIS
ADA198092),  (Reference 5).

4.   29 CFR Ch. 17,  1910.1000, Occupational Safety and Health
Standards; Toxic and Hazardous Substances; Air Contaminants; Federal
Register, June  27,1974, pp.  598-599.

5.   D.H.  Ritts, C. Garretson, C. Hyde, et al..  Evaluation of
Innovative Volatile Organic Compound  and  Hazardous  Air Pollutant
Control Technologies for  U.S.  Air Force Paint Sprav Booths. ESL-
TR-89-51; EPA-600/2-90-059  (NTIS ADA242508),  Air  Force Engineering and
Services Center, Tyndall  AFB,  FL, 1990.

6.   J. Ayer and C.  Hyde, VOC Emission Reduction Study at the Hill Air
Force Base Building 515 Painting Facility.  ESL-TR-89-35; EPA-600/2-90-051
(NTIS ADA198092),  Air Force Engineering  and Services Center, Tyndall AFB,
FL,  1990.

7.   J.F.  Seibert,  Ventilation Booths and Respirators, Armstrong
Laboratory Occupational and Environmental Health  Newsletter, S.
Spradling, ed., Brooks AFB, TX, Volume  15, Issue  4, 1991, p.l.

8.   American Conference  of Governmental Industrial Hygienists,
1990-1991 Threshold Limit Values for  Chemical Substances and Physical
Agents and Biological Exposure Indices, ACGIH, Cincinnati, OH, 1990.
(a)  p.4;  (b) pp. 43-45.


                                  10

-------
 AEERL-P-925
                                 TECHNICAL REPORT DATA
                          (Please read iHSlructions on tlic reverse before completing)
 1 REPORT NO.
  EPA/6QO/A-92/145
                                                         3. RECIPIENT'S ACCESSION NO.
 4. TITLE AND SUBTITLE
 Using a Flame lonization Detector (FID) to  Contin-
  uously Measure Toxic Organic Vapors in a Paint
  Spray Booth
                                                         5. REPORT DATE
                                                        6. PERFORMING ORGANIZATION CODE
 7.AUTHOR(sij.whitfield (EPA);  G.Howe and B. Pate (RTI);
  and J. Wander (USAF)
                                                         8. PERFORMING ORGANIZATION REPORT NO
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Research Triangle Institute
 P.O. Box  12194
 Research Triangle Park, North Carolina  27709
                                                         10. PROGRAM ELEMENT NO.
                                                        11. CONTRACT/GRANT NO.
                                                         CR815169-03
 12. SPONSORING AGENCY NAME AND ADDRESS
 EPA.  Office of Research and Development
   Air and Energ}r Engineering Research Laboratory
   Research Triangle Park,  North Carolina 27711
                                                        13. TYPE OF REPORT AND PERIOD COVERED
                                                         Published paper; 7/91-1/92
                                                        14. SPONSORING AGENCY CODE
                                                         EPA/600/13
 15. SUPPLEMENTARY NOTES AEERL project officer is Jamie K.  Whitfield,  Mail Drop 61,  919 /
 541-2509.  For presentation at AWMA  annual meeting, Kansas City, MO, 6/21-26/92,
 16. ABSTRACT
               paper repOrts the demon s tr ation of linear and similar responses of a
 Ratfisch RS-55CA flame ionization detector {FID) to a solvent mixture identical to
 the volatile organic compounds (VOCs) in the coating and catalyst  (NSN 8010-01-336-
 3036) and to the calibrating gas (propane) used in field calibrations of the FID.  Sen-
 sitivity and linearity have been shown to extend from 715 to 45 mg/cu m,  which brac-
 kets the calculated short-term exposure limit (STEL) and lower action  thresholds.
 Monitoring is maintained constantly  and,  under field conditions, equilibration oc-
 curs rapidly (analysis and output transpire in milliseconds).  As a trigger for fail-
 safe conversion from recirculation mode to a straight- through paint spray booth  con-
 figuration,  the FID may confidently be expected  to initiate a corrective response be-
 fore a transient elevation of VOC concentrations overexposes area personnel.
 7.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                            b.IDENTIFIERS/OPEN ENDED TERMS
                                                                      c.  COSATI Field/Group
 Pollution          Solvents
 Spray Painting
 Organic Compounds
Measurement
~   ,   ....
Conductivity
                     P.PROO.CEOBv
                     us DEPARTMENT OF COMMERCE
                     NATIONAL TECHN'CAL INFORMATION SERVICE
                     SPRNGCIE.D VA 22161
Pollution Control
Stationary Sources
Paint Spray Booths
Flame lonization De-
  tectors  (FIDs)
13 B
13 H
07C
06T
14G
                                                                            11K
 8. DISTRIBUTION STATEMENT
 Release to  Public
                                           19. SECURITY CLASS (This Report)
                                           Unclassified
                                                                     21. NO. OF PAGES
                                            20. SECURITY CLASS (This page)
                                            Unclassified
                                                                      22. PRICE
EPA Form 2220-1 (9-73)

-------