United States                  EP A- 600 / 2 - 88-061
                                  October 1988
&EPA     Research and
           Development
           EVALUATION OF

           PERCHLCRCETHYLENE EMISSIONS

           FROM DRY CLEANED FABRICS
           Prepared for
           Office of Toxic Substances
           Prepared by
           Air and Energy Engineering Research
           Laboratory
           Research Triangle Park NC 27711

-------
                   RESEARCH REPORTING SERIES


 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into nine series. These nine broad cate-
 gories were established to facilitate further development and application of en-
 vironmental technology. Elimination of traditional  grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The nine series are:

     1. Environmental Health Effects Research

     2. Environmental Protection Technology

     3. Ecological Research

     4. Environmental Monitoring

     5. Socioeconomic Environmental Studies

     6. Scientific and Technical Assessment Reports (STAR)

     7. Interagency Energy-Environment Research and Development

     8. "Special" Reports

     9. Miscellaneous Reports

 This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
 NOLOGY series.  This series describes research performed to develop and dem-
 onstrate instrumentation, equipment, and methodology to  repair or prevent en-
 vironmental degradation from point and non-point sources of pollution. This work
 provides the new or improved technology required for the control and treatment
 of pollution sources to meet environmental quality standards.
                        EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.

This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                         EPA-600/2-88-061
                                         October 1988
EVALUATION OF PERCHLOROETHYLENE EMISSIONS FROM
              DRY CLEANED FABRICS
                      by

               Bruce A. Tichenor
               Leslie E. Sparks
              Merrill D. Jackson

     U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
               Indoor Air Branch
       Research Triangle Park, NC 27711
                 Prepared for:

     U.S. Environmental Protection Agency
      Office of Research and Development
             Washington, DC 20460

-------
                            Abstract

     A study was conducted to evaluate the emissions of
perchloroethylene* (perc) from dry cleaned fabrics to determine:
a) how the introduction of fresh dry cleaning into a home affects
the indoor concentration of p.erc, and b) the effectiveness of
"airing out" for reducing perc emissions.  Small chamber tests
were conducted to determine perc emission characteristics for
three fabrics at several temperatures and air exchange rates.
Test house studies were conducted to determine the indoor
concentration of perc due to the placement of dry cleaned
clothing in the house.  Based on the study results, and assuming
that test conditions were representative of normal dry cleaning and
consumer practices, the following conclusions were reached:

     1) Emissions from freshly dry cleaned clothing cause
        elevated levels of perchloroethylene in residences.

     2) For the three fabrics tested, "airing out" of dry cleaned
        clothing by consumers will not be effective in reducing
        perchloroethylene emissions.

     It is emphasized that these conclusions are based on the
results of the study reported.  Significant variations in dry
cleaning practices and/or in the mix of fabrics and clothing
being cleaned could provide different results and conclusions.
" Perchloroethylene is the common name for tetrachloroethylene;
  C2C14.  It has a molecular weight of 166.  This document
  reports concentrations in ug/ra3 at standard conditions (0°C,
  1 atmosphere).  To convert to ppb, multiply by 0.135.
                              11

-------
                            CONTENTS

Abstract 	  ii

Figures 	  1 v

Tables 	   v

Acknowledgment 	  vi

     Executive Summary 	   1

     1.  Introduction 	   3
        A.  Study objectives 	   3
        B.  Factors affecting emissions and indoor
           concentrations 	   3
        C.  Previous studies 	   5

     2.  Study Design 	   6
        A.  Fabric/clothing selection 	   6
        B.  Emission factor determination (small chamber
           testing) 	   7
        C.  Evaluation of perc residuals
           ( solvent extraction)	   7
        D.  IAQ model analysis 	   7
        E.  Evaluation of indoor concentrations
           (test house )  	   7
        F.  Quality assurance/quality control	   8

     3.  Experimental Procedures 	  10
        A.  Fabric treatment 	  10
        B.  Small chamber testing 	  10
        C.  Residual analyses 	  13
        D.  IAQ test house 	  14

     4.  Results 	  16
        A.  Chamber tests 	  16
        B.  Perc residuals 	  21
        C.  Indoor concentrations (test house) 	  22
        D.  IAQ model analysis 	  32

     5.  Discussion and Conclusions 	  35
        A.  Emission factors 	  35
        B.  Residuals 	  37
        C.  Indoor concentrations 	  37
        D.  Model results 	  37
        E.  Conclusions 	.-	  37

     6.  References 	  38
                              111

-------
Appendix A. Quality Assurance/Quality Control Results 	 39

     1.  Calibration Audit for Perchloroethylene  	 39

     2.  Standard Rotation 	;	 40

     3.  QC for Chamber Laboratory  	 40

     4.  QC for Test House 	 42
                              FIGURES

 Number                                                       Page

   1    Small chamber (166-L) emissions testing facility 	 11

   2    Small chamber (53-L) emissions testing facility 	 12

   3    IAQ test house 	 15

   4    Perc emissions from dry cleaned polyester/rayon
        (modeling of small chamber data) 	 18

   5    Perchloroethylene in closet 	 25

   6    Perchloroethylene in bedroom 	 26

   7    Perchloroethylene in den 	 27

   8    Perc in bedroom and den — bag off (test 1) 	 28

   9    Perc in bedroom and den -- bag on 	 29

   10    Perc in bedroom and den — aired out 	 30

   11    Perc in bedroom and den -- bag off (test 2) 	 31

   12    Effect of airing out 	 36
                              IV

-------
                             TABLES

Number                                                       Page

  1    Perc emissions from various fabrics 	 19

  2    Effect of air exchange rate on perc emissions 	 20

  3    Effect of temperature on perc emissions 	 21

  4    IAQ test house sampling frequency -- perc tests 	 23

  5    IAQ test house results 	 24

A-l    Audit gas analysis -- chamber laboratory 	 39

A-2    Audit gas analysis — test house 	 40

A-3    Perc standards rotated between residuals laboratory
       and chamber laboratory 	 40

A-4    Variability in Ro, k, and t(l/2) for
       polyester/wool 	 41

A-5    Variability in Ro, k, and t(l/2) for 100% wool 	 41

A-6    Variability in Ro, k, and t(l/2) for
       polyester/rayon	 41

-------
                         ACKNOWLEDGMENT

     The authors acknowledge the substantial contributions of the
following individuals:  Zhishi Quo for developing and executing
the techniques used to  analyze the small chamber data;  Mark Mason
and Michelle Plunket for conducting the small chamber tests;
Susan Rasor for performing the IAQ test house experiments; and
Glenn Smith for conducting the perc residuals evaluations.  All
of these individuals are employees of the Acurex Corp.  and
performed their work under EPA Contract No.  68-02-4701.
                              VI

-------
                        EXECUTIVE SUMMARY


     The Indoor Air Branch (EPA's Air and Energy Engineering
Research Laboratory -- EPA/AEERL) conducted a short term study to
evaluate the emissions of perchloroethylene (perc) from dry
cleaned fabrics.  Specifically, the study was designed to answer
two questions:

     A) To what extent does the residual perc in dry cleaned
        fabric increase the concentration of perc in residential
        environments?

     B) How effective is "airing out" in reducing indoor perc
        concentrations?

     A study consisting of five components was conducted:
1) Fabric/Clothing Selection; 2)  Emission Factor Determination
(Small Chamber Testing); 3) Evaluation of Perc Residuals (Solvent
Extraction); 4) Indoor Air Quality (IAQ) Model Analysis; and
5) Evaluation of Indoor Concentrations (Test House).

     The following results were achieved:

Emission Factors

     Emission factors for perchloroethylene from dry cleaned
fabrics were determined by testing in small environmental test
chambers under controlled conditions.  Evaluation of the data
from these tests provided the following:

     - A preliminary screening evaluation showed that wide
variations in initial emission factor, Ro,  and emission factor
half-lives, t(l/2), occurred between different fabrics.  Thus,
the type of fabric is important in determining indoor emissions
of perc from dry cleaned clothes.

     - Based on the screening study and on the prevalence of
fabrics used in dry cleaned clothing, three fabrics were selected
for investigation: 55% polyester/45% wool;  100% wool;  and
50% polyester/50% rayon.

     - The air exchange rate showed no effect on the emission
factor or decay rate for the three fabrics investigated.  This
suggests that the emissions are limited by the diffusion of perc
within the fabric and are not controlled by evaporative
processes.  This also suggests that increasing the ventilation by
airing out the clothes will not speed up the emission of perc.

     - Since the three fabrics tested had emission factor half-
lives of about a day, airing the clothes out for a few hours
before hanging them in the home will do little to reduce the
indoor perc concentrations.  For fabrics with faster perc decay
rates, airing out may be more practical.

-------
     - Temperature had a major impact on the emission factors and
decay rates.  Increases in temperature caused higher initial
emission factors and lower half-lives.  Thus, exposing the
clothing to higher temperatures prior to bringing them home shows
promise as a means of reducing in-home exposure to perc.
Residuals

     No acceptable data were developed on the perc residuals
within the fabric.  The solvent extraction procedure, using
methylene chloride, failed to produce reliable results.  A fully
tested "standard method" is needed.
Indoor Concentrations

     All the test house experiments showed that the introduction
of dry cleaned clothing caused elevated levels of perc in the
house.  Differences in concentration between the tests were
probably due to differences in the amount of perc retained at the
dry cleaner.


Model Results

     The IAQ model, using emissions data developed in the small
chambers, predicted indoor perc concentrations which compared
favorably with those measured in the test house.  The effect of
perc "sinks" in the test house was also demonstrated.


Conclusions

     Based on the study results, and assuming that test conditions
are representative of normal dry cleaning and consumer practices,
the following conclusions are reached:

     1 ) Emissions from freshly dry cleaned clothing cause
        elevated levels of perchloroethylene in residences.

     2) For the three fabrics tested, "airing out" of dry cleaned
        clothing by consumers will not be effective in reducing
        perchloroethylene emissions.

     It is emphasized that these conclusions are based on the
results of the study reported herein.  Significant variations in
dry cleaning practices and/or in the mix of fabrics and clothing
being cleaned could provide different results and conclusions.

-------
1.  INTRODUCTION


A.  Study Objectives

     The Indoor Air Branch (EPA's Air and Energy Engineering
Research Laboratory — EPA/AEERL) conducted a short term study to
evaluate the emissions of perchloroethylene" (perc) from dry
cleaned fabrics.  Specifically, the study was designed to answer
two questions:

     A) To what extent does the residual perc in dry cleaned
        fabric increase the concentration of perc in residential
        environments?

     B) How effective is "airing out" in reducing indoor perc
        concentrations?
B.  Factors Affecting Emissions and Indoor Concentrations

     A number of factors may affect the amount of residual perc
in dry cleaned fabrics, including:

     Type of fabric r Brodmann (1975) reports that the residual
perc after cleaning varies widely between fabrics.  In general,
synthetic fabrics retained more perc than natural fabrics.
Brodmann's evaluation of perc residuals immediately and 24 hours
after treatment in a coin operated machine indicates:

                         % Perc Retained"*     X Perc Retained
    Fabric Type            (Immediate)          (after 24 hrs)
Arnel (triacetate)
Acetate (diacetate)
Polypropylene
Spun Dacron 54
Spun Dacron 64
Polyester Double Knit
Nylon 66
0.80
0.46
0.82
0-18
0.07
0.09
0.09
0.41
0.21
0.20
0.12
0.07
0.05
0.04
"X of fabric weight

All other fabrics tested (Orion, Acrilan, wool, Antron, cotton,
and fiberglass) showed immediate residuals of < 0.02X and no
detectable residuals after 24 hours.
  Perchloroethylene is the common name for tetrachloroethylene;
  C2C14.  It has a molecular weight of 166.  This document
  reports concentrations in ug/m3 at standard conditions (0°C,
  1 atmosphere).  To convert to ppb,  multiply by 0.135.

-------
     Type of treatment - Two types of dry cleaning machines are
in use:  "transfer" and "dry-to-dry."  Data on the perc residuals
from these two machines are very limited.  Pressed fabric should
have lower perc residuals than unpressed fabric.

     Variability of treatment - Perc residuals may be affected by
the drying cycle and the aeration step, as well as by such
factors as age and condition of cleaning equipment, amount of
material cleaned per load, operator technique, and solvent
purity/age.  While data on the effect of such factors are
unavailable, such variables could cause differences in perc
residuals between cleaning establishments and between loads at
the same establishment.

     Several additional factors are important with respect to the
emission rate of perc from dry cleaned fabrics:

     Environmental variables - Temperature, humidity, ventilation
(air exchange), and the concentration of perc in the air may
affect the rate at which residual perc is emitted from fabrics.

     Type of fabric - The rate of perc emissions varies between
fabric types.  For example, Brodmann's data (shown above)
indicate that Spun Dacron 64 retained 100% of the residual perc
over a 24 hour period, Nylon 66 retained less than 50%, and
polypropylene retained less than 25%.

     Storage/handling parameters - A number of variables
associated with in-home storage or handling of dry cleaned
fabrics may affect the perc emission rate:

     - Plastic storage bag retained or removed.  Keeping the
plastic bag on may reduce the rate of emission, but not the total
emitted.

     - The amount of dry cleaned fabric stored in a closet.  The
larger the amount of material stored in a given closet, the
greater the total emissions.  The emission factor
(e.g., mg/mz-hr) may be lower, however, due to the effect of
vapor pressure suppression of evaporation.  This would not occur
if the emissions are limited by in-fabric diffusion.

    - Pre-storage "airing out."  Hanging the dry cleaned fabrics
outdoors or in a well ventilated area prior to in-home storage
may reduce indoor perc emission rates.

     - Time since treatment.  The amount of residual  perc and
subsequent emissions to the indoor environment will vary
depending on the time between cleaning and placement  in the home.
Storage at the dry cleaners and transportation time will impact
this variable.

-------
     Finally, several additional factors may affect the indoor
perc concentrations:

     Air exchange rate - The air exchange rate (amount of outside
air infiltration) determines the dilution and flushing indoors.
For a given amount of dry cleaning, the higher the air exchange
rate, the lower the indoor perc concentration.

     HVAC system - The operation of the HVAC (heating,
ventilating, air conditioning) system in the home affects the
mixing and movement of air-  Air in residences is generally well-
mixed when the HVAC fan is operating.  This would cause the perc
concentrations to be fairly consistent from room to room, except
in the closet where the dry cleaning is stored.

     Air movement - The amount of air movement between the closet
and the adjoining room and between that room and the rest of the
home will affect the perc concentrations throughout the
residence.  Factors such as HVAC operation and open or closed
doors are important in affecting air movement.

     Sink effects - Materials in the home may adsorb perc at
higher concentrations and gradually release it over time.  Such
an effect would lower initial concentrations but extend the
exposure time.  Factors such as the amount of clothing in the
closet (in addition to the dry cleaned items) could impact the
sink effect.
C.  Previous Studies

     In addition to the study by Brodmann (1975) on perc
residuals in dry cleaned fabrics, several other references deal
with the issue of perc exposure from dry cleaning.  Several
industry sponsored articles are available dealing with safe
handling of perc in dry cleaning establishments and methods for
reducing occupational exposure (Fisher, 1976; HSIA, 1986; IFI,
1987).   Fisher (1976) also provides data on perc concentrations
measured in dry cleaning plants.  Data on non-occupational perc
exposure from dry cleaning are limited to studies conducted on
alveolar (breath) air from people exposed to dry cleaners
(Verberk and Scheffers, 1980; Wallace, 1988).  The authors could
find no published data on perc concentrations inside residences
where freshly dry cleaned clothes were introduced.

-------
2.  STUDY DESIGN

     In the design of any experimental program, constraints of
time and resources limit the scope of the investigation.  For the
study described herein, these constraints were:

     - Study completion by May 1988.  (The study was initiated in
       March.)

     - No additional resources.

Thus, the proposed study was limited to available resources
and facilities of the Indoor Air Branch and was to be completed in
a short time span.  Given these constraints, it was obvious that a
comprehensive analysis of perc emissions from dry cleaned fabrics
which included consideration of all relevant variables  (see
above) was not possible.  The following study plan was developed
to best meet the study objectives within the imposed resource and
time limits.

     A study consisting of fiv« components was conducted:

     1) Fabric/Clothing Selection;

     2) Emission Factor Determination (Small Chamber Testing);

     3) Evaluation of Perc Residuals (Solvent Extraction);

     4) Indoor Air Quality (IAQ) Model Analysis; and

     5) Evaluation of Indoor Concentrations (Test House).


A.  Fabric/Clothing Selection

     Since the available data on perc residuals in fabrics are
over 10 years old, changes in types of fabrics and dry cleaning
technologies may have occurred which would affect the perc
emissions.  Thus, prior to the selection of the test fabrics, a
"screening study" was conducted.  One yard (0.9 m) pieces of 12
fabrics were purchased and dry cleaned.  Using fabric pieces
instead of clothing allowed appropriate sizes to be cut for use
in the small chambers. These samples were placed in small test
chambers and preliminary data on perc emission rates were
obtained.  These data were then used to select three fabrics for
further evaluation.

-------
     For the test house studies, clothing made of the three
selected fabrics was used.  A mix of clothing consisting of a
man's suit, a woman's skirt, and two blouses was evaluated in the
test house.

     Treatment was by commercial cleaning at a single facility.
The fabric samples and the clothing were handled in a manner
consistent with normal dry cleaning operation, including
pressing.  Protective plastic bags from the cleaners enclosed all
cleaned material prior to testing.


B.  Emission Factor Determination (Small Chamber Testing)

     Small environmental research chambers (six 53-L chambers;
and two 166-L chambers) were used to develop data on emission
factors for perc from dry cleaned fabrics.  Three fabrics (see
above) were evaluated.  The effect of air exchange (air changes
per hour [ACH]) on emission factors was investigated for each
fabric to determine the effect of "airing out."  The effect of
temperature on emission factors was also evaluated.  All testing
was conducted with a water vapor content equivalent to a 50%
relative humidity at 20°C.  This test program was designed to
provide emission factors for perc (ug/mz-hr), information on the
rate of decay of the emissions for the three fabrics, and
information on the effects of air exchange and temperature.


C.  Evaluation of Perc Residuals (Solvent Extraction)

     Testing was conducted to determine the amount of perc
within the fabrics.  The method selected involved solvent
extraction using methylene chloride followed by gas
chromatography.  Residual testing was conducted before and
after emissions testing in the small chambers and the test house.


D.  IAQ Model Analysis

     The chamber emission factor data were used in an IAQ model
to  predict expected perc concentrations in the test house based
on  available data on the air exchange and air movement in the
test house.  These results were used to design the test house
experiments.  In addition the model enabled an evaluation of the
"sink" effect.
E.  Evaluation of Indoor Concentrations (Test House)

     Based on the results of the chamber tests and subsequent
model analyses,  four 2-week test house experiments were
conducted.  Each test consisted of 1 week of testing followed by
1 week of data evaluation and house "airing" in preparation for
the next test.  Indoor air samples were collected at three

-------
locations in the house: a) the closet where the clothing was
hung, b) the adjacent bedroom, and c) the den.  Sampling
frequency was determined based on the IAQ model analysis of the
small chamber data.  A set of background samples were collected
prior to each test.

     The following four tests were conducted:

     Test A - Cleaned clothes hung in empty closet in corner
              bedroom.  Plastic wrap removed. HVAC system on.
              Closet door closed.  All other interior doors open.

     Test B - Same as Test A, except plastic wrap not removed.

     Test C - Same as Test A, except clothing  "aired out" for 4
              hours prior to being hung in the closet.

     Test D - Replicate of Test A.


F. Quality Assurance/Quality Control

     The following QA/QC procedures were designed for the study
to ensure the production of data with known, acceptable quality.


1. Calibration Audit for Perchloroethylene

     A calibration audit was performed for the analysis of
perchloroethylene by gas chromatography (GC).  Two audit gas
cylinders (one high and one low concentration) were obtained for
analysis by the GC in the chamber laboratory and the GC at the
test house.  The following goals applied: accuracy, ± 20%;
precision, ± 15%.

2. Standard Rotation

     A perchloroethylene standard rotation procedure was
implemented between the chamber laboratory and the laboratory
where perchloroethylene residual was analyzed.  One perc standard
from each laboratory was analyzed by the other.  The objective
for analytical accuracy was +. 20%; for precision, ± 15%.

3. Chamber Laboratory

     QC measures for the chamber laboratory included
documentation and control of the variability of sample size, test
chamber temperature, relative humidity, air flow, and precision
and accuracy of measurements of the perchloroethylene
concentration.  Variabilities of sample size, temperature,
relative humidity, and air flow were determined for each test.
The QA/QC goals for the above parameters were ± 5%, ± 1°C, ± 15%,
and ± 5%, respectively.

-------
    The allowable range for the recovery of the internal
standard, present in every sample, was ± 20%.  The goal for  the
relative standard deviation (RSD) between duplicate samples  was
1 10%.

4. Residual Analysis

     For GC calibration, five standard perchloroethylene
solutions were prepared, spanning the range of perc
concentrations in the fabric sample extracts; for each standard,
the goal for the RSD for quadruplicate injections was <. 15%.

     Duplicate fabric samples were used; for each sample
duplicate extract injections were made; the goal for the RSD
between duplicate injections was <. 15%.

     In each automatic sampler run (6-10 samples), two
standards were included, one before and one after the samples.
For the run to be accepted, the standards must have a relative
error less than 10% from known concentration.

5. Test House

     The gas chromatograph for perchloroethylene analysis was
calibrated every morning of each sampling day prior to sampling.
The GC performance was checked every 3 hours during the sampling
day by injecting at least two liquid perc standards with
different concentrations.

     Duplicate samples were taken from each sampling location.
The QA/QC objective for the precision of the duplicate samples
was + 15%.

-------
3.  EXPERIMENTAL PROCEDURES

A.  Fabric Treatment

     All fabrics (bolt material and clothing) were cleaned at a
local dry cleaning establishment.  The cleaner used a dry-to-dry
machine with a 35 Ib (16 kg) capacity operating at 140°F (60°C).
Both utility and legger presses were available, with steam traps
at 300°F (150°C).  The cleaned fabrics were picked up from the
cleaners within 1 hour of being cleaned.


B. Small Chamber Testing

1. Facilities and Equipment

     AEERL's Indoor Air Source Characterization Laboratory
contains two chamber test systems: one with a pair of 166 L test
chambers, the other with six 53 L chambers.  Each system consists
of the following components: a clean air conditioning and
delivery system, an incubator containing the environmental test
chambers, sampling manifolds, and sample collection adsorbers
using Tenax and charcoal.  A permeation system for quality
control is included.  The environmental variables are monitored
and controlled by a microcomputer-  Organic analyses are
conducted by thermal desorption, concentration via purge and
trap, and gas chromatography (GC) using flame ionization
detection (FID).  A. separate microcomputer provides GC data
analysis.  All data are input to spreadsheets for further
analysis (Tichenor and Mason, 1988).  Figure 1 is a schematic
drawing of the 166-L chamber system; Figure 2 shows the 53-L
system in less detail.

2. Testing/Measurement Methods

     Within 1 hour of being picked up at the dry cleaners,
fabrics were cut to size, hung on wire racks, and placed in the
test chambers.  The first sample was collected within 30 minutes
of the start of testing.  Several samples were collected on the
first day, a sampling frequency of one per day was continued
until the end of the test period.  Most tests were concluded
within 5 days.

     Samples were collected by pulling a portion of the chamber
air stream through tandem glass cartridges filled with Tenax and
Tenax/charcoal sorbents.  Sampling was conducted at 100 cm3/min.
Sampling time varied from 5 to 100 minutes providing sample
volumes of 0.5 to 10 L.  The sorbent cartridges were
thermally desorbed at 220°C to the Tenax/charcoal concentrator
column of a purge-and-trap unit.  The concentrator column was
rapidly heated, and the collected compounds were desorbed to the
analytical column of a gas chromatograph (GC).  Perchloroethylene
was identified by retention time and quantified by FID response.
                               10

-------
COMPRESSOR
                                                      CONTROL
         COOUNG COILS
HEATED LINES
                  UMNO FAN
           166 LITER
         TEST CHAMBER
                                                SAMPLING
                                                MANIFOLDS
           166 LITER
         TEST CHAMBER
          INCUBATOR
                   I PUMP | |  PUMP |
             Figure 1.  Small chamber (166-L) emissions  testing facility

-------
 CLEAN

  AIR

SYSTEM



TEST CHAMBERS



1

2

3





4

5

6

^•MM
••M^B
I
I
I
L.
                                            SORBENT
                                         I"
                                         I
                                         I
                                         I
                                            SORBENT
                   GAS

                 CHROM.
          Figure 2.  Small chamber (53-L) emissions testing facility.

-------
     The GC was calibrated by loading Tenax sample cartridges
with known amounts of perc.  Liquid standards were prepared by
serial dilution of a gravimetrically determined primary standard.
Two microliters of stock solutions, ranging in concentration from
16 to 1600 ng/ul,  were injected through the hot port of the
purge-and-trap concentrator-  Volatilized compounds were swept
by a helium purge to the Tenax sample cartridges.  The cartridges
were then analyzed in the same manner as the chamber samples.  A
linear least squares fit of the FID response (in area counts)
provided a response factor which was used to convert FID response
to mass units.
C. Residual Analyses

1. Fabric Preparation

     Upon receipt of the fabric (bolt material or clothing) to be
sampled, sections of cloth were cut out and placed in brown-glass
screw-cap jars with Teflon seals.   The lids were sealed with
Teflon tape, and the jars were stored at 4°C.

     Before cutting samples to be extracted, the jars were warmed
to room temperature.  Two rectangular samples were cut from each
fabric, about 3 by 10 cm (approx.  0.5 g).  The actual size of the
samples was determined by the thickness of the cloth.  Thicker
cloth required smaller samples to reduce crowding in the vial.
The dimensions of the fabric were measured to the nearest
millimeter, the samples were folded in half lengthwise, rolled,
and placed in pre-weighed 15 ml vials.  The vials were capped and
weighed, and the mass of the fabric calculated.

     For spiked samples, a drop of tetrachloroethylene was added
and the vials weighed again.

2. Extraction Procedure

     Five railliliters of methylene chloride (dichloromethane) was
added by volumetric pipette to the vials containing the fabric
samples.  This quantity of solvent was sufficient to cover the
sample completely.  The vials were tightly capped, placed in an
ultrasonic bath, and sonicated for 1 hour, during which time the
temperature of the bath rose from 25 to 40°C.  The vials were
remove and allowed to cool to room temperature, and the solvent
was decanted into 7.5 ml vials.  The vials were capped, sealed
with Teflon tape, and stored at 4°C.

3. Analysis

     Aliquots of the sample extract were placed in 0.75 ml vials
and sealed.  The vials were loaded into an automatic sampler
(preceded and followed by known concentration QC samples).  The
samples were analyzed for perchloroethylene by GC with an FID.
                               13

-------
D. IAQ Test House

1. Facilities and Equipment

     A single-story frame house with three bedrooms, two baths,
den, kitchen, and a living/dining combination  is leased for use
as an IAQ test house (Figure 3).  The house is 8 years old,
has a forced-air gas heating/electric air conditioning system,
and is insulated to be energy efficient  (Jackson, et al..  1987).
A garage along one end of the house provides a convenient
instrument room.  Available instruments  included a GC with ECD
(electron capture detector).  This instrument was used for
determination of perc concentrations.

2. Testing/Measurement Methods

     A standard set of clothing consisting of a woman's wool
skirt, two polyester/rayon blouses, and  a man's two-piece wool
blend suit was dry cleaned in a commercial facility.  The clothes
were transported to the house in the standard plastic bag
provided by the dry cleaners.  The clothes were placed in the
closet of the corner bedroom and the closet doors were closed.
All other interior doors were open.  The house was closed, and
the sampling initiated within 15 minutes.  The house was
maintained at a temperature of 20°C (68°F).  The HVAC fan was
operated in the normal mode, and the fan operating times were
recorded.

     Samples were collected at three locations for each test:
a) the closet in which the clothes were  placed, b) the corner
bedroom  (adjacent to the closet), and c) the den.  Samples were
collected by gastight syringes for immediate injection into the
GC.  Samples were taken in the center of each room at a height of
160 cm from the floor.  Samples were also collected at heights of
15 and 198 cm in the closet to check for stratification.  A
system was design to allow sampling while the closet doors
remained closed.  This system also returned the air to the closet
to prevent the loss of perchloroethylene by dilution.

     The air exchange rate for the house was determined on the
first day of each test by use of SFs tracer gas.  The gas was
released at the start of the each run and collected in Tedlar
bags at hourly intervals until eight samples were taken.  The SFs
was analyzed by GC with ECD.

     After each test, all the windows in the house were opened and
the house was allowed to air out for a minimum of 4 days and
then was reclosed for 4 hours before the start of the next run.
A background check was performed to ensure that the the perc from
the previous run was below detection limits.  The detection limit
for the sampling and analysis system used in the study was  1
ug/m3 .
                               14

-------
               Master
               bedroom
 Dry

Cleaning
          Clos   Clos
DOS
                                Master
                                bath
                                Bath
                                    Clos
            Return
             dr
           Comer
           bedroom
                A Clos
Utility
               Middle
               bedroom
                                            Den
              Kitchen
            I	I
                                        Living Room
                                              x  : cios
                                                                           Instruments
                                        Garage
                                                                               •• Registers
                                     Figure 3.  IAQ test  house.

-------
4. RESULTS


A. Chamber Tests

1. Data Analysis

     Models have been developed to analyze the results of the
chamber tests in order to provide emission rates (Dunn and
Tichenor, 1988).  The simplest model (i.e., neglecting sink and
vapor pressure effects) assumes: a) the chambers are ideal
continuous stirred tank reactors (CSTRs),  and b) the change in
emission factor can be approximated by a first order decay, as
shown in Equation (1):

                      R = Roe-kt                            (1)

Where, R  = Emission factor, mg/m2-hr
       Ro = Initial emission factor, mg/m2-hr
       e  = Natural log base
       k  = First order rate constant,  hr*1
       t  = Time, hr

The mass balance for the chamber over a small time increment, dt,
is:

      Change in mass = Mass emitted - Mass leaving chamber

This can be express as:

                VdC = ARoe-ktdt - QCdt                      (2)

Where, V = Chamber volume, m3
       C = Chamber concentration, mg/m3
       A = Area of source, m2
       Q = Flow through chamber, m3/hr

Equation (2) can be rearranged:

              dC/dt + (Q/V)C = (A/V)Roe-kt                   (3)

Equation (3) is a linear, non-homogeneous differential equation.
Given that C = 0 when t = 0, the solution to Equation (3) is:

               C = ARo (e-kt - e-Nt )/V(N - k)                 (4)

Where, N = Air exchange rate,  hr~1, and is equal to Q/V

      Using a non-linear regression curve fit routine,
implemented on a microcomputer, values of Ro and k can be
obtained by fitting the concentration vs.  time data from the
chambers to Equation (4).  To conduct such analyses,  initial
estimates of Ro and k are required.  A good initial estimate of ^
is :
                               16

-------
                      k = Ne( k - N > t • a *                         (5)

Where, tmax is the time of maximum concentration, Cmax.

Equation (5) is obtained by substituting C [Equation (4)] into
Equation (3) and setting dC/dt = 0 at t = tmax.  Once an estimate
of k is achieved from Equation (5), Ro can be estimated from
Equation (4).  Figure 4 illustrates the curve fitting process for
a polyester/rayon fabric chamber test at an air exchange rate of
1 hi"1; the solid line is the "best fit" of Equation (4), and the
data points are shown as solid squares.

     All of the test runs were analyzed using this procedure, and
the results (Ro and k) for perc are presented in Tables 1 and 2.
Values for Ro/k are also presented; they represent the total
available emissions (or source strength) per unit area for the
material being tested.  Total emissions are estimated by
integrating Equation (1) from time zero to infinity.  Finally,
the half-life of the emission factor, t(l/2), is also provided;
t(l/2) is the amount of time required for the emission factor to
be reduced by 50%.

2. Fabric Selection

     A preliminary screening study was conducted on 12
different fabrics (cleaned/unpressed).  The fabrics were
investigated in the 53-L test chambers under the following
conditions: air exchange rate = 1.0 hr*l,  temperature = 20°C,
relative humidity = 50%, and sample area (one side) = 0.168 ra2 .
The sample area was selected to provide a chamber loading (area
of sample/volume of chamber) similar to what was expected in the
test house closet.

     It is emphasized that the results from this preliminary
screening study are useful in a qualitative sense for comparing
the emission characteristics of the fabrics tested.  Only one
short term test was conducted on each fabric, and only two or
three data points were collected.   Thus, the results of the curve
fit procedure described above should be used with caution.  These
results are presented in Table 1.
                               17

-------
          4000 -
oo
     Kl

      E
     3000 -
      01
      D
      C
      o
c:

-------
          Table 1. Perc Emissions from Various Fabrics
Ro
Fabric (ug/m2-hr)
50% Polyester/50% Rayon-
Rayon
Polyester Knit
Acetate
Acrylic Knit
Wool Blend (I)"
Wool Blend (II)
Cotton
Linen
65% Polyester/35% Cotton
85% Rayon/15% Flax
Silk
220
55
430
6700
56
1200
990
440
570
350
180
(No
k
(hr-
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
data
1 )
030
034
031
009
039
033
080
140
076
210
150
- S
Ro/k
(mg/m2 )
7
1
14
740
1
36
12
3
7
1
1
ampl
.3
.6


.4


.1
.5
.7
.2
ing
Ul/2
(hr)
23
20
23
77
18
21
8
5
9
3
4
errors
)






.7
.0
.1
.3
.6
)
  •Selected for further testing.

     The results (Table 1) show that the amount of perc held on
fabric surfaces varies greatly depending on fabric type.  The
total amount of perc initially available for emission (Ro/k)
ranges from 1,2 to 740 mg/m2.  Also, some fabrics hold perc
longer; t(l/2) values range from 3.3 to 77 hours.  Again, the
reader is cautioned to use these results only in a qualitative
way to compare one fabric to another.  Use of these data to
calculate emissions should be avoided.

     Based on these preliminary results and an evaluation of the
prevalence of materials used in clothing on the market that is
normally dry cleaned, the following three fabrics were selected
as the test materials:

     - A 5_5% Polyester - 45% Wool Blend.  This fabric had a
relatively high emission rate and is widely used in men's suits.

     - A 50% Polyester - 50% Rayon Blend.  This fabric had a
relatively low emission rate and is widely used in women's
clothing.

     - A 100% Wool.  Wool was not evaluated, but it is widely
used and commonly sent to dry cleaners.

     The highest emitter, acetate, was not selected due to its
low volume use in dry cleaned clothing.
                               19

-------
3. Effect of Air Exchange Rate

     The three selected fabrics (cleaned/pressed) were
investigated in the 53-L test chambers under the following
conditions: air exchange rates = 0.25, 1.0, and 2.0 hr" x ;
temperature = 20°C; relative humidity = 50%; and sample area
(one side) = 0.168 m2.   The results are shown in Table 2.
     Table 2. Effect of Air Exchange Rate on Perc Emissions
Fabric Air
(
Polyester/
Wool

100% Wool


Polyester/
Rayon

Exchange
I hr- M
0.25
1.0
2.0
0.25
1.0
2.0
0.25
1.0
2.0
Ro
(ug/m2 -hr)
1500
2400
800
930
1200
800
560
1100
470
k
( hr- * )
0.028
0.045
0.028
0.041
0.028
0.052
0.022
0.038
0.027
Ro/k
(mg/m2 )
54
54
29
23
43
15
26
28
17
(hr)
25
15
25
17
25
13
32
18
26
     The results provided in Table 2 show variations of Ro, k,
Ro/k, and t(l/2) for each fabric at the three air exchange rates
tested.  Some of this variation is due to experimental error;
estimates of the sampling and analysis errors are presented in
Appendix A.  Another source of the variations is the amount of
perc retained at the time the fabrics were picked up from the dry
cleaners.  Four separate dry cleaning "loads" were required over
a 5 week period to complete the tests reported in Table 2.
Whatever the cause of the variations, the data do not suggest
that higher air exchange rates have a significant effect on the
rate of decay of the emission rate, k, or on the half-life of the
emission factor, t(l/2).


4. Effect of Temperature

     Two fabrics,  55% polyester/45% wool and 100% wool, were
selected to evaluate the effect of temperature on the emissions
of perc from dry cleaned fabrics.   Tests were conducted in the
166-L chambers under the following conditions: temperatures = 30
35,  40, and 45°C;  air exchange rate = 1.0 hr'1 ;  and sample area =
0.5 m2 .  The humidity control system was set to  supply water
vapor at a rate equivalent to 50% relative humidity at 20°C.  The
results for Ro and t(l/2), along with the previous data for 20°c
are shown in Table 3.
                               20

-------
        Table 3. Effect of Temperature on Perc Emissions
Fabric Temperature
(°C)
Polyester/
Wool



100X Wool




20
30
35
40
45
20
30
35
40
45
Ro
(ug/m2 -hr)
2400
8600
5400
9200
16000
1200
3100
3200
8100
14000
t(
(
15
4
2
2
1
25
7
1
0
0
1/2)
hr)

.5
.9
.5
.8

.1
.3
.6
.5
     Table 3 shows that temperature has a major effect on the
emission characteristics of perc.  Increasing temperatures caused
higher initial emission factors, Ro,  and lower emission factor
half-lives, t(l/2).
B. Perc Residuals

     The data on the residual perc determined by solvent
extraction using methylene chloride are not reported herein.
Several problems were encountered: 1) mechanical breakdowns of
the automatic sampler caused extensive delays in analyzing the
extracts, and no data are available on how such delays might
affect the final results; 2) the recoveries of spiked perc from
samples of different materials were quite variable and often too
low (some variability in recoveries for samples of the same
material was also observed); 3) the amount of residual perc
extracted was inconsistent with values determined from the
chamber data; and 4) the amount of residual perc from the
clothing used in the test house experiments (immediately after
cleaning vs. after removal from the test house) showed little
relationship to the actual amount of perc emitted in the test
house as determined by a mass balance using measured
concentrations.

     The reasons for the failure of the solvent extraction
procedure to provide useful data are unclear-  A search of the
literature did not uncover a "standard method" for extracting
perc from dry cleaned fabrics, although solvent extraction by
methylene chloride is mentioned by Brodmann (1975).  Subsequent
to initiating the study, discussions with representatives of the
dry cleaning industry suggested that extraction using carbon
disulfide is more common.
                               21

-------
C.  Indoor Concentrations (Test House)

     Evaluations of the impact of dry cleaned clothes on
indoor concentrations of perc were conducted in the IAQ test
house.  For each test, the following clothing was dry cleaned and
brought to the test house:

     - A two piece man's suit (55% polyester/45% wool blend),

     - A woman's skirt (100% wool), fully lined (acetate), and

     - Two women's blouses (50% polyester/50% rayon).

The total area of this mix of clothing, including linings,
padding, pockets, and accounting for seam overlap, is 8.6 m2.

     Four tests were conducted as described in Section 3.D:

     a) "Bag Off" - the plastic bag was removed prior to placing
         the clothes in the closet (March 22 - 31);

     b) "Bag On" - the plastic bag was not removed (April 6  -
         13);

     c) "Aired Out" - the plastic bag was removed, and the
         clothes were hung in an open carport for 4 hours
         prior to being placed in the closet (April 18 - 26);

     d) A repeat of the first "Bag Off" test (May 2 - 10).

     The daily sampling frequency for the four tests is shown  in
Table  4.  Early  in each test, 24 hour coverage of the test house
was provided.  After the first couple of days, sampling occurred
during normal working hours.  Limited sampling occurred on
weekends.

     The results of the test house evaluations are shown in  Table
 5  and  in Figures 5 - 11.  Figures 5-7 show the average daily
concentrations for all four tests as measured in the closet,
bedroom, and den, respectively.  The closet values are the
average of the three sampling elevations, since no consistent
stratification was observed in the closet.  Figures 8-11
provide the average daily values for the bedroom and den for the
 four  tests.  Note that, in all cases, the level of perc dropped
to near or below the detection limit within a day after the
clothing was removed from the test house.

     The results show two consistent patterns: a) on any given
day,  the concentrations are highest  in the closet, followed  by
the bedroom, with the den having the lowest concentrations;  and
b) the concentrations in all rooms generally decrease over time.
                                22

-------
Table 4. IAQ Test House Sampling Frequency -- Perc Tests
Test
Bag Off
(1)








Bag On







Aired Out








Bag Off
(2)







Day
1
2
3
4
5
6
7
8
9'
10
1
2
3
4
5
6
7"
8
1
2
3
4
5
6
7
8"
9
1
2
3
4
5
6
7
Bm
9
No. of Samples
16
13
4
4
1
1
4
5
3
1
15
15
4
3
3
6
6
1
8
25
20
5
1
1
1
3
1
12
26
17
5
5
1
2
1
1
Sampling Interval
(hr)
0.5
1.5
2.25
2.25
24
24
1.25
1.5
1.25
24
1.25
1.0
2.0
4.0
4.0
0.75
0.5
24
0.75
1.0
1.0
1.8
24
24
24
3.8
24
1.0
1.0
1.0
1.0
1.0
24
6.0
24
24
"Clothes removed from house.
                          23

-------
            Table 5. IAQ Test House Results
                     Average Perc Concentration (ug/m3
Test
Bag Off
(1)








Bag On







Aired Out








Bag Off
(2)







Day
1
2
3
4
5
6
7
8
9'
10
1
2
3
4
5
6
7»
8
1
2
3
4
5
6
7
8*
9
1
2
3
4
5
6
7
8"
9
Closet
521
325
310
273
258
158
125
52
13
ND
1240
498
438
331
307
393
290
ND
2900
2740
425
371
393
307
720
146
3
1420
745
606
353
218
557
170
168
ND
Bedroom
60
58
57
46
45
25
15
14
9
ND
139
61
44
32
31
27
18
ND
195
124
50
55
43
37
85
24
3
119
104
52
39
22
16
19
11
ND
Den
36
29
32
25
26
13
8
6
6
ND
37
32
25
16
18
14
9
ND
79
83
24
33
26
27
40
15

51
46
28
26
15
16
12
5
ND
"Clothes removed from house;  ND = Not Detected
                          24

-------
         3000
to
o»
    ro
     o>
     3
     O
2000 -i
         1000 -
                                     Bog Off (1)
                                      Bog On
                               V77X  Aired Out
                                      Bag Off (2)
                                      Model Prediction
                 Day 1
                    Day 2
Day 3       Day 4
      Time  (day)
Day 5
Day 6
Day 7
                                Figure 5.  Perchloroethylene in closet

-------
          200
          150 -
to
en
     o»
     3
     O
O


O
O
     100 -
                Day 1
                                                                        Bag Off (1)


                                                                        Bag On


                                                                  V77X  Aired Out


                                                                        Bag Off (2)


                                                                        Model Prediction
                       Day 2
Day 3      Day 4


      Time (day)
Day 5
Day 6
Day 7
                               Figure 6.   Perchloroethylene in bedroom.

-------
NJ
     en
     3
     o
    S3
     o

     o
    u
                                                                               Bag Off (1)



                                                                               Bag On


                                                                               Aired Out


                                                                               Bag Off (2)
                                                                               Model Prediction
                 Day 1
Day 2
Day 3       Day 4


      Time (day)
Day 5
Day 6
Day 7
                                  Figure 7.   Perchloroethylene in den,

-------
ro
CO
     ro
     o
C

u
C
           10
               3/22    3/23
                        Figure 8.  Perc in bedroom and den -- bag off (test 1).

-------
to
10
     O
     33
c
o
c
140
130
120
110 H
100
 90
 80 -
 70
 60
 50 -
 40
 30 -
 20 -
 10
  0

                  4/6
                             1
                      4/7
                                                    I
                                                                   POM  Bedroom
                                                                   V77X  D«n
                                   4/9
                                     Data
4/10     4/11      4/12      4/13
                              Figure 9'.  Perc in bedroom  and den -- bag on.

-------
CJ
o
     n


      o>
c
o
53

|

•
O
c
            200
            180
                                                                           Bedroom


                                                                     EZ3  Den
                 4/18    4/19     4/20
                                     4/21    4/22


                                          Date
4/23    4/24    4/25    4/26
                           Figure  10.•  Perc in bedroom and  den -- aired out.

-------
IO


0>
§
8
     120
     110 -
     100
                                                                ESSS  Bedroom

                                                                Y77A  Dan
            5/2
5/9     5/10
                   Figure  11.  Perc in bedroom and den  --  bag off (test 2).

-------
     A wide variation in perc concentrations was observed between
the four tests for the first couple of days of each experiment.
Since one would not expect that keeping the bag on or airing out
the clothes would cause increases in emissions, these differences
are not believed to be due to the experimental variables.
Rather, it is assumed that the differences are due to the amount
of perc retained in the clothes at the dry cleaners.  At any
rate, under all test conditions, elevated levels of perc were
measured in the test house when freshly dry cleaned clothing was
placed in the closet.  Also, the perc dropped to near or below
the detection limit after the clothes were removed from the
house.
D. IAQ Model Analysis

     The EPA IAQ model (Sparks,  et al..  1988) was used to
determine the consistency of the test house and chamber data.
The model was also used to estimate sink effects.

1. Model description

     The model estimates the effects of heating, ventilating, and
air conditioning (HVAC), air cleaning,  room to room air
movement, and natural ventilation on pollutant concentrations.
The model has been used in a similar fashion with experiments
conducted in the EPA test house using moth crystal cakes composed
of para-dichlorobenzene (Sparks, et al. . 1988).  The agreement
between small chamber emission factors,  model predictions, and
test house data was very good.  Predicted weight loss of the moth
crystal cakes was within 5% of the measured weight loss.
Predicted room concentrations of para-dichlorobenzene was within
10% of the measured values.

     The EPA IAQ model is based on conducting mass balances of
pollutant and air flow between multiple well-mixed model rooms.
This model was selected because data from the EPA test house
showed that pollutant concentrations within a room do not vary
significantly with position in the room (Jackson, et al.. 1987).

     The model uses a menu-driven "fill in the form" data-input
user interface.  This interface is easy to use and is self
prompting.  The user is able to change the input parameters
quickly and easily, and several conditions can be rapidly
analyzed.  The results of the model calculations are displayed as
plots of concentration versus time for the various rooms.  The
plots require that a graphics adapter and color monitor be
installed on the computer.
                               32

-------
2. Source Terms

     The source term used to model  the  perc  emission was based on
the small chamber data and  is  in  the  form:

                         E( t)  = Roe-kt A                     (6)

Where, E(t) = Emission rate, ug/hr, at  time  t
       Ro   = Initial emission factor,  ug/m2-hr
       k    = First order rate constant,  hr~l
       A    = Area of clothes, m2

For the test house studies, Ro =  1600 ug/mz-hr,  k  =  0.03 hr~1,
and A = 8.6 m2.   These values  were  obtained  by using a  weighted
average of the chamber test results based on the measured areas
of the fabric types for the clothing used in the test house
experiments.

3. Sink Terms

     A re-emitting sink was used  in the perc modeling.   The  rate
going to the sink was assumed  to  be:

                            Ra  = ka Cr Aa                       ( 7 )

Where, Ra = Rate to the sink,  ug/hr
       ka = Constant, m/hr
       Cr = Concentration in the  room,  ug/m3
       Aa = Area of the sink,  m2

The value of ka  was estimated  using data  from a  special  perc
experiment, data from moth  crystal  cake experiments, and a
theoretical analysis of mass transfer to  walls in  a  well stirred
reactor.  The value used for ka was 0.35  m/hr.

     The emission rate from the sink, Es  (ug/hr),  was assumed  to
be:

               Es = keMsAa(Cr  - Cc ) , when Cr > Co            (8)

                      Ea ~  0,  when  Cr 1 Cc                   (9)

Where, ke = Emission constant, m/ug-hr
       Ma = Mass collected  on  the sink, ug
       Cc = Critical concentration, ug/m3

When Cr  > Cc , emissions are possible, and when Cr  <.  Cc ,
emissions are not possible.  ke was estimated to be  5 m/ug-hr,
and Cc was estimated to be  40  ug/m3.
                               33

-------
4.  Results

     The results of the initial model runs with no sink effects
did not provide good agreement with the measured test house perc
concentrations.  The predicted concentrations were too high, and
the predicted curves did not show the changes in slope noted in
the experimental data.   The likely explanation for the
differences between predicted and measured concentrations is the
existence of a sink effect.

     Model runs were then made with a re-emitting sink as
described above.  The model results for days 1-6, using a
re-emitting sink, are shown in Figures 5-7.  Note that the
agreement between the model predictions and the measured data is
good both in magnitude  and in the shape of the decay curve.  Also
note that the model predictions,  as well as the measured data,
show plateaus of nearly constant  perc concentration,  including
slight increases, for days 3 through 6.  These plateaus and small
increases in perc concentration are caused by re-emissions from
the sink, because, without a re-emitting sink, the concentrations
would have continued to decay.

     The main effects of the re-emitting sink are:

     1. Initial peak concentrations are reduced as perc goes to
        the sink; and

     2. The rate of concentration decay is slowed as perc is
        re-emitted from the sink, and some increases in
        concentration can also be observed.
                               34

-------
5. DISCUSSION AND CONCLUSIONS


A. Emission Factors

     Emission factors for perchloroethylene from dry cleaned fabrics
were determined by testing in small environmental test chambers
under controlled conditions.   Evaluation of the data from these
tests provides the following  results:

     - A preliminary screening evaluation showed that wide
variations in initial emission factor, Ro,  and emission factor
half-lives, t(l/2), occurred  between different fabrics.  Thus,
the type of fabric is important in determining indoor emissions
of perc from dry cleaned clothes.

     - Based on the screening study and on the prevalence of
fabrics used in dry cleaned clothing, three fabrics were selected
for investigation: 55% polyester/45% wool;  100% wool; and
50% polyester/50% rayon.

     - The air exchange rate  showed no effect on the emission
factor or decay rate for the  three fabrics investigated.  This
suggests that the emissions are limited by the diffusion of perc
within the fabric and are not controlled by evaporative
processes.  This also suggests that increasing the ventilation by
airing out the clothes will not speed up the emission of perc.

     - Since the three fabrics tested had emission factor half-
lives of about a day, airing  the clothes out for a few hours
before hanging them in the home will do little to reduce the
indoor perc concentrations.

     - For fabrics with faster perc decay rates, airing out may
be more practical.  The percent emitted during airing out is
calculated by:

 % Emitted during airing = (Amount emitted/Total available)100%

The Amount emitted equals the integral of the emission rate
function [Equation (1)] over  the time aired out:

                Amount emitted =(Ro/k)(l - e'«k)            (10)

Where, a = airing out time (hr)

As defined above, the Total available = Ro/k, thus:

            % Emitted during  airing = (1 - e-«k)100%        (11)

The effect of airing out is illustrated in Figure 12 which shows
the percent of perc emitted during airing out as a function of
the time aired out for a wide range of decay rates (k).  Note
that, for the decay rates determined for the three fabrics
                               35

-------
to
en
      o

      en
      c
      en
      c


      D
      Q
 0)
-*-•
-4-»
• —

 E

bJ


O



Q.
                          2468

                             Time Aired Out  (hours)
                                                                    10
                              Figure 12.  Effect of airing out.

-------
investigated in this study (from Table 2, k ranges from 0.022 to
0.052 hr" ! ), the percent of perc emitted over a 4 hour airing out
period would range from 8 to 19%; for an 8 hour period, the range
would be 16 to 34%.

     - Temperature had a major impact on the emission factors and
decay rates.  Increases in temperature caused higher initial
emission factors and lower half-lives.  Thus, exposing the
clothing to higher temperatures prior to bringing them home shows
promise as a means of reducing in-home exposure to perc.


B.  Residuals

     No acceptable data were developed on the perc residuals
within the fabric.  The solvent extraction procedure, using
methylene chloride, failed to produce reliable results.  A fully
tested "standard method" is needed.
C.  Indoor Concentrations

     All the test house experiments showed that the introduction
of  dry cleaned clothing caused elevated levels of perc in the
house.  Differences in concentration between the tests were
probably due to differences in the amount of perc retained at the
dry cleaner.


D.  Model Results

     The IAQ model, using emissions data developed in the small
chambers, predicted indoor perc concentrations which compared
favorably with those measured in the test house.  The effect of
perc "sinks" in the test house was also demonstrated.


E.  Conclusions

     Based on the study results, and assuming that test
conditions are representative of normal dry cleaning and consumer
practices, the following conclusions are reached:

     1) Emissions from freshly dry cleaned clothing cause
        elevated levels of perchloroethylene in residences.

     2) For the three fabrics tested, "airing out" of dry cleaned
        clothing by consumers will not be effective in reducing
        perchloroethylene emissions.

     It  is emphasized that these conclusions are based on the
results of the study reported herein.  Significant variations  in
dry cleaning practices and/or in the mix of fabrics and clothing
being cleaned could provide different results and conclusions.
                               37

-------
6.  REFERENCES


Brodmann,  G.  L.,  "Retention of Chlorinated Solvents in Fabrics,"
     Journal  of American Association of Textile Chemists and
     Colorists. Vol.  7,  No. 5, 1975.

Dunn, J. E. and Tichenor,  B.  A.,  "Compensating for Sink Effects
     in Emissions Test Chambers By Mathematical Modeling,"
     Atmospheric Environment. Vol. 22,  No. 5,  pp.  885-894, 1988.

Fisher, W. E., "Safe  Handling of Perchloroethylene: Parts 1 & 2,"
     Focus on Drvcleaning.  Vol. 2, Nos. 1  &. 2 ,  International
     Fabricare Institute,  Silver Springs,  MD,  1976.

HSIA (Halogenated Solvents Industry Alliance),  The Safe Handling
     of Perchloroethylene Dry Cleaning  Solvent. Washington, DC,
     1986.

IFI (International Fabricare Institute),  "Perchloroethylene Vapor
     in Drycleaning Plants," Focus on Drvcleaning. Vol. 11,
     No. 1, 1987

Jackson, M. D., Clayton, R. K., Stephenson,  E.  E., Guyton, W. T.,
     and Bunch, J. E., "EPA's Indoor Air Quality Test House, 1.
     Baseline Studies,"  Proceedings of  the 1987 EPA/APCA
     Symposium on Measurement of Toxic  and Related Air
     Pollutants.  EPA-600/9-87-010, (NTIS No.  PB88-113402) , 1987.

Sparks, L. E., Jackson,  M.  D., and Tichenor,  B. A., "Comparison
     of EPA Test House Data with Predictions  of an Indoor Air
     Quality Model,"  presented at ASHRAE IAQ88, Atlanta, GA,
     1988.

Tichenor,  B.  A. and Mason,  M. A., "Organic Emissions from
     Consumer Products and Building Materials  to the Indoor
     Environment," JAPCA.  Vol. 38, No.  3,  pp.  264-268, 1988.

Verberk, M. M. and Scheffers, T.  M. L., "Tetrachloroethylene in
     Exhaled Air of Residences Near Dry-Cleaning Shops,"
     Environmental Research.  Vol. 21, pp.  432-437, 1980.

Wallace, L.,  Personal communication, 1988.
                               38

-------
APPENDIX A. QUALITY ASSURANCE/QUALITY CONTROL RESULTS


     As discussed in Section 2.F, Quality Assurance/Quality
Control, a number of QA/QC steps were implemented in the conduct
of the  study; the results are reported below.  Accuracy (or
bias) is reported as the average deviation from the true value:

                  Accuracy = [(m - Xo)/x0]100X            (A-l)

Where, m is the mean, and xo is the true value.

Precision is reported as Relative Standard Deviation (RSD).   RSD
(also called the Coefficient of Variation) is calculated by:

                         RSD = (s/m)100%                (A-2)

Where, s is an estimate of the standard deviation, and m is the
mean.


1. Calibration Audit for Perchloroethylene

     Two audit gas cylinders were obtained and analyzed for
perchloroethylene in the chamber laboratory and  at the test
house.  The results of the analyses are shown in Tables A-l  and
A-2.   In all cases, the goals of 20% accuracy and 15% precision
were met.


       Table A-l.  Audit Gas Analysis -- Chamber Laboratory


    Reported                Measured
      Cone.    Number of      Cone.     Accuracy    Precision
      (ppb)    Analyses       (ppb)       (%)          (%)


       400        2            452       12.8          1.0

       10-9       3            9.4       13.8         11.2
                               39

-------
          Table  A-2.   Audit Gas Analysis — Test House
Reported
Cone .
(ppb)
400
10.9

Number of
Analyses
7
4
Measured
Cone .
(ppb)
370
10.0

Accuracy
(%)
7.5
8.3

Precision
(%)
6.7
2.3
2. Standard Rotation

     A perchloroethylene standard rotation procedure was
implemented between the chamber laboratory and the laboratory
that conducted the residual analysis.   One liquid
perchloroethylene standard from each laboratory was analyzed by
the other laboratory.   The results are shown in Table A-3.   The
QC goals of 20% accuracy and 15% precision were met.


      Table A-3.   Perc Standards Rotated between Residuals
                Laboratory and Chamber Laboratory
Reported
Cone .
(mg/ml )
0.325*
0.158**
Number of
Analyses
4
2
Measured
Cone .
(mg/ml )
0.270
0.157
Accuracy
(%)
16.9
0.6
Precision
(%)
3.1
3.1
     "Prepared by residuals laboratory;  analyzed by chamber
      laboratory.
     "Prepared by chamber laboratory;  analyzed by residuals
       laboratory.


3. QC for Chamber Laboratory

     Variability of sample size,  determined by comparing the mass
of each fabric used in the individual experiments with the
average mass in all tests, ranged from 1.2 to 4.2% for each
fabric.  The average test chamber temperature was controlled to
within ± 1°C of the setpoint for all but one experiment, which
exceeded the expected range by 0.4°C.  Relative humidity was
controlled to within ± 15% of the setpoint for all experiments.
Uncertainty of chamber air flow,  determined from the difference
in measured flow at the beginning and end of each experiment, was


                               40

-------
± 4%.  Recovery of the internal standard, present in every
sample, averaged 103 ± 9.8% for 270 samples, meeting the goal of
20%.  The precision of duplicate samples averaged 7.4 ± 4.7% for
39 duplicate pairs; the goal was 10%.

     Estimates of the variability of the Ro, k, and t(l/2) for
all tests conducted at 20°C are shown in Tables A-4, A-5, and
A-6-, where the mean and standard deviation  for each of these
parameters is given.  The variability indicated in these tables
includes load to load variation at the dry  cleaners, experimental
error,  and statistical errors associated with the curve fitting
of the  data.
        Table A-4.
Variability in Ro,  k, and t(l/2) for
     Polyester/Wool
Air Ex.
(hr-1 )
0
1
2
.25
.0
.0
No. of Ro
Samples (ug/m2-hr)
6
5
1
1507 ± 529
2412 + 495
798
0
0
0
k
( hr- l )
.0277 + 0.0067
.0453 + 0.0123
.0278
t
(1/2)
(hr)
26.
16.
24.
5
3
9
+ 6
+ 3.9

        Table A-5.
Variability in Ro,  k, and t(l/2) for
        100% Wool
Air Ex.
(
0
1
2
hr-1 )
.25
.0
.0
No. of
Ro

Samples (ug/m2-hr)
2
4
2
933
1186
2021
+ 95
+ 539
+ 18
0
0
0

k


(hr-1 )
.0410
.0277
.0385
+ 0
+ 0
+ 0
.0170
-0095
.0035
20
25
18
tC
L/2)

(hr)
.4
.5
.7
+ 8
+ 3
+ 0
.4
.7
.9
        Table A-6.
Variability in Ro,  k, and t(l/2) for
     Polyester/Rayon
Air Ex.
(hr-1 )
0-25
1.0
2.0
No. of
Samples
2
2
1
Ro
(ug/m2 -hr)
562 ± 32
1072 + 254
474
k
(hr-1 )
0.0220 + 0 .0060
0.0380 + 0.0041
0.0274
t(l/2]
(hr)
34.0 +
18.5 +
25.3
1

9.3
2.0

                               41

-------
4. QC for Test House

     The GC used for perchloroethylene analysis was calibrated
every morning before any samples were taken,  and its performance
was checked every 3 hours during the day by injecting two liquid
perchloroethylene standards.

     Duplicate samples were taken from each sampling location.  A
total of 69 pairs of duplicate samples were taken.   The precision
of the duplicate samples averaged 5.6 ± 5.7%.   The  precision of
7 of the 69 duplicate pairs exceeded the QC goal of 15%,  which
occurred during the first sampling in the house and may depict
levels of perchloroethylene rising due to the  recent introduction
of the clothing into the house.   These samples had  a slight
difference in collection time.
                              42

-------
TECHNICAL REPORT DATA
(Please read latauctions on the reverse before completing)
EPA-600/2-88-061
4. TITLE AND SUBTITLE
Evaluation of Perchloroethy]
Cleaned Fabrics
5. REPORT C
Lene Emissions from Dry Cctobe
6. PERFORM
7. AUTHOR(S) 8. PERFORM
Bruce A. Tichenor, Leslie E. Sparks, and
Merrill D. Jackson
9. PERFORMING ORGANIZATION NAME ANCI
See Block 12
12. SPONSORING AGENCY NAME ANO AOOR
EPA, Office of Research an<
Air and Energy Engineering
Research Triangle Park, NC
19. SUPPLEMENTARY NOTES A.EERL D
541-2991. (*) Supported by A
18. ABSTRACT _. .
The report gives re
(perc) from dry cleaned fabi
cleaning into a house affects
of "airing out" for reducing
determine perc emission ch
and air exchange rates. Tes
concentration of perc due to
Based on study results, and
normal dry cleaning and coi
ched: emissions from fresh.
chloroethylene in residence
cleaned clothing by consum<
emissions. Significant vari;
rics and clothing being clea
I
1

17. KEY WORDS ANO DOCUMENT ANALYSIS
a. DESCRIPTORS b. IDENTIFIERS/OPEN ENO6D TS
Pollution Pollution Control
Tetrachloroethylene Stationary Sources-
Emission Perchloroethylene
Fabrics Indoor Air
Dry Cleaning
Atmosphere Contamination Control
Release to Public
Unclassified
Unclassified

3ATE
r 1988
ING ORGANIZATION CODE
ING ORGANIZATION REPORT NO.
id ELEMENT NO.
NO.
VO PERIOD COVERED
3-5/88
-:Y CODE
Drop 54, 919/
•02-4701.
)roechylene
resh dry
effectiveness
2 conducted to
emperatures
ae the indoor
le house.
entative of
s were rea-
ls of per-
out" of dry
Droethylene
he mix of fab-
. elusions.

RMS c. COSATI Field/Group
13B
06C.07C
14G
HE
13 H
06K
49
22. PRICE 	


-------