Risk Screen on Substitutes for Ozone-Depleting Substances for

Adhesive, Aerosol Solvent, and Solvent Cleaning Applications

Proposed Substitute:

n-Propyl Bromide

U.S. Environmental Protection Agency

Office of Air and Radiation

Stratospheric Protection Division

April 2006

Table of Contents

  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc127943257"  Risk Screen:
Adhesive Applications	  PAGEREF _Toc127943257 \h  5  

  HYPERLINK \l "_Toc127943258"  1.	Introduction	  PAGEREF _Toc127943258
\h  5  

  HYPERLINK \l "_Toc127943259"  2.	Summary of Results	  PAGEREF
_Toc127943259 \h  6  

  HYPERLINK \l "_Toc127943260"  3.	Toxicity Reference Values for
Substitutes	  PAGEREF _Toc127943260 \h  6  

  HYPERLINK \l "_Toc127943261"  4.	Atmospheric Modeling	  PAGEREF
_Toc127943261 \h  7  

  HYPERLINK \l "_Toc127943262"  5.	Occupational Exposure Analysis	 
PAGEREF _Toc127943262 \h  7  

  HYPERLINK \l "_Toc127943265"  6.	General Population Exposure Analysis	
 PAGEREF _Toc127943265 \h  8  

  HYPERLINK \l "_Toc127943266"  7.	Volatile Organic Compound Analysis	 
PAGEREF _Toc127943266 \h  9  

  HYPERLINK \l "_Toc127943267"  8.	References	  PAGEREF _Toc127943267 \h
 9  

  HYPERLINK \l "_Toc127943269"  Risk Screen: Aerosol Solvent
Applications	  PAGEREF _Toc127943269 \h  11  

  HYPERLINK \l "_Toc127943270"  1.	Introduction	  PAGEREF _Toc127943270
\h  11  

  HYPERLINK \l "_Toc127943271"  2.	Summary of Results	  PAGEREF
_Toc127943271 \h  12  

  HYPERLINK \l "_Toc127943272"  3.	Toxicity Reference Values for
Substitutes	  PAGEREF _Toc127943272 \h  12  

  HYPERLINK \l "_Toc127943273"  4.	Atmospheric Modeling	  PAGEREF
_Toc127943273 \h  13  

  HYPERLINK \l "_Toc127943274"  5.	Occupational Exposure Analysis	 
PAGEREF _Toc127943274 \h  13  

  HYPERLINK \l "_Toc127943277"  6.	General Population Exposure Analysis	
 PAGEREF _Toc127943277 \h  14  

  HYPERLINK \l "_Toc127943278"  7.	Volatile Organic Compound Analysis	 
PAGEREF _Toc127943278 \h  14  

  HYPERLINK \l "_Toc127943279"  8.	References	  PAGEREF _Toc127943279 \h
 14  

  HYPERLINK \l "_Toc127943281"  Risk Screen: Solvent Cleaning
Applications	  PAGEREF _Toc127943281 \h  17  

  HYPERLINK \l "_Toc127943282"  1.	Introduction	  PAGEREF _Toc127943282
\h  17  

  HYPERLINK \l "_Toc127943283"  2.	Summary of Results	  PAGEREF
_Toc127943283 \h  18  

  HYPERLINK \l "_Toc127943284"  3.	Toxicity Reference Values for
Substitutes	  PAGEREF _Toc127943284 \h  18  

  HYPERLINK \l "_Toc127943285"  4.	Atmospheric Modeling	  PAGEREF
_Toc127943285 \h  19  

  HYPERLINK \l "_Toc127943286"  5.	Occupational Exposure Analysis	 
PAGEREF _Toc127943286 \h  19  

  HYPERLINK \l "_Toc127943293"  6.	General Population Exposure Analysis	
 PAGEREF _Toc127943293 \h  25  

  HYPERLINK \l "_Toc127943294"  7.	Volatile Organic Compound Analysis	 
PAGEREF _Toc127943294 \h  25  

  HYPERLINK \l "_Toc127943295"  8.	References	  PAGEREF _Toc127943295 \h
 25  

  HYPERLINK \l "_Toc127943297"  Attachment A: Determination of an AEL	 
PAGEREF _Toc127943297 \h  27  

  HYPERLINK \l "_Toc127943298"  1.	AEL Derivation	  PAGEREF
_Toc127943298 \h  27  

  HYPERLINK \l "_Toc127943299"  2.	Discussion of Relevant Literature	 
PAGEREF _Toc127943299 \h  28  

  HYPERLINK \l "_Toc127943304"  3.	AEL Determination	  PAGEREF
_Toc127943304 \h  31  

  HYPERLINK \l "_Toc127943305"  4.	Benchmark Dose Methods and Analysis	 
PAGEREF _Toc127943305 \h  32  

  HYPERLINK \l "_Toc127943311"  5.	Background on nPB AEL Benchmark Dose
Analysis	  PAGEREF _Toc127943311 \h  40  

  HYPERLINK \l "_Toc127943314"  6.	Selected Output from BMDS	  PAGEREF
_Toc127943314 \h  43  

  HYPERLINK \l "_Toc127943315"  7.	References	  PAGEREF _Toc127943315 \h
 45  

  HYPERLINK \l "_Toc127943317"  Attachment B: Derivation of a Reference
Concentration	  PAGEREF _Toc127943317 \h  47  

  HYPERLINK \l "_Toc127943318"  1.	Reference Concentration Derivation	 
PAGEREF _Toc127943318 \h  47  

  HYPERLINK \l "_Toc127943319"  2.	References	  PAGEREF _Toc127943319 \h
 48  

  HYPERLINK \l "_Toc127943321"  Attachment C: Evaluation of the Global
Warming Potential	  PAGEREF _Toc127943321 \h  49  

  HYPERLINK \l "_Toc127943322"  1.	Summary	  PAGEREF _Toc127943322 \h 
49  

  HYPERLINK \l "_Toc127943323"  2.	Discussion of Results	  PAGEREF
_Toc127943323 \h  49  

  HYPERLINK \l "_Toc127943324"  3.	References	  PAGEREF _Toc127943324 \h
 50  

  HYPERLINK \l "_Toc127943326"  Attachment D: Occupational Exposure
Analysis for Adhesive Applications	  PAGEREF _Toc127943326 \h  51  

  HYPERLINK \l "_Toc127943327"  1.	Introduction	  PAGEREF _Toc127943327
\h  51  

  HYPERLINK \l "_Toc127943328"  2.	Occupational Exposure Modeling	 
PAGEREF _Toc127943328 \h  51  

  HYPERLINK \l "_Toc127943333"  3.	Occupational Exposure Study
Evaluations	  PAGEREF _Toc127943333 \h  54  

  HYPERLINK \l "_Toc127943339"  4.	References	  PAGEREF _Toc127943339 \h
 58  

  HYPERLINK \l "_Toc127943341"  Attachment E: Occupational Exposure
Analysis for Aerosol Solvent Applications	  PAGEREF _Toc127943341 \h  59
 

  HYPERLINK \l "_Toc127943342"  1.	Occupational Exposure Modeling	 
PAGEREF _Toc127943342 \h  59  

  HYPERLINK \l "_Toc127943349"  2.	Occupational Exposure Study
Evaluations	  PAGEREF _Toc127943349 \h  63  

  HYPERLINK \l "_Toc127943355"  3.	References	  PAGEREF _Toc127943355 \h
 68  

  HYPERLINK \l "_Toc127943357"  Attachment F: General Population
Exposure Assessment for n-Propyl Bromide	  PAGEREF _Toc127943357 \h  69 


  HYPERLINK \l "_Toc127943358"  1.	Introduction	  PAGEREF _Toc127943358
\h  69  

  HYPERLINK \l "_Toc127943359"  2.	Approach	  PAGEREF _Toc127943359 \h 
69  

  HYPERLINK \l "_Toc127943360"  3.	Modeling System	  PAGEREF
_Toc127943360 \h  70  

  HYPERLINK \l "_Toc127943361"  4.	Results	  PAGEREF _Toc127943361 \h 
71  

  HYPERLINK \l "_Toc127943362"  5.	References	  PAGEREF _Toc127943362 \h
 73  

 Risk Screen on Substitutes for Ozone-Depleting Substances

Adhesive Applications

Introduction

Ozone-depleting substances (ODS) are being phased out of production in
response to a series of diplomatic and legislative efforts that have
taken place over the past decade, including the Montreal Protocol and
the Clean Air Act Amendments of 1990 (CAAA).  The U.S. Environmental
Protection Agency (EPA), as authorized by Section 612 of the CAAA, has
developed a program to evaluate the human health and environmental risks
posed by alternatives to ODS.  The main purpose of EPA’s program,
called the Significant New Alternatives Policy (SNAP) program, is to
identify acceptable and unacceptable substitutes for ODS in specific end
uses.

EPA’s decision on the acceptability of a substitute is based largely
on the findings of a screening assessment of potential human health and
environmental risks posed by the substitute in specific applications. 
EPA has already screened a large number of substitutes in many end uses
within all of the major ODS-using sectors, including refrigeration and
air-conditioning, solvent cleaning, foam blowing, aerosols, fire
suppression, adhesives, coatings and inks, and sterilization.  The
results of these risk screens are presented in a series of background
documents that are available in EPA’s docket.  The reader is referred
to this reference for an additional discussion of different
methodologies used to conduct risk screens.

The purpose of this risk screen is to supplement EPA’s background
document (EPA 1994) on the adhesive applications of n-propyl bromide
(nPB) or 1-bromopropane, which is used as an alternative to ODS such as
HCFC-141b and has been shown to exhibit toxicity upon inhalation.  The
potential human health and environmental risks posed by nPB are examined
in this risk screen.

  REF _Ref127933100 \h  Table 1.1  provides the chemical details on nPB.

Table   STYLEREF 1 \s  1 .  SEQ Table \* ARABIC \s 1  1 : Chemical
Details for n-Propyl Bromide

Name	Chemical Formula	CAS #

n-Propyl Bromide

(1-Bromopropane, nPB)	C3H7Br	000106-94-5



Occupational exposure and general population analyses were performed to
ensure that use of nPB in the application listed above does not pose
unacceptable risks to workers or the general public.  Consumer exposure
modeling was not performed because no consumer applications are proposed
for this substitute.  

The remainder of this risk screen is organized as follows: 

Section 2 of this report summarizes the results of the risk screen for
nPB;

Section 3 presents the toxicity values used for the risk screen.  

Section 4 presents atmospheric modeling and potential environmental
risks; 

Section 5 presents occupational exposure analysis; 

Section 6 presents the general population exposure analysis; and 

Section 7 assesses the emissions of volatile organic compounds (VOCs).

Summary of Results 

Use of nPB is not recommended for SNAP approval in adhesive
applications, unless adequate exposure controls, including extensive
local ventilation, are in place at the point of use in order to meet the
recommended AEL of 17 ppm.  This requirement stems from the fact that
exposure modeling and actual monitoring data obtained from three
facilities using nPB-containing adhesives show that the recommended
8-hour acceptable exposure limit (AEL) of 17 ppm is exceeded in almost
all cases for workers spraying the adhesive, even when spray booths were
installed and local ventilation was significantly improved.  Therefore,
extensive ventilation and other exposure controls would need to be
present in order to meet the acceptable exposure level throughout the
day.  Additionally, atmospheric analyses indicate that the use of nPB is
preferable to the continued use of methyl chloroform and a modeling
analysis shows that general-population exposure would be below the
recommended reference concentration (RfC).

Toxicity Reference Values for Substitutes

To assess potential health risks from exposure to this substitute for
ODS in the adhesive end use, EPA identified the relevant toxicity
threshold values, including an AEL, excursion limit, and RfC for
comparison to modeled and actual exposure concentrations.  The AEL
represents the maximum 8-hour time weighted average (TWA) at which a
worker can be exposed regularly without adverse effects.  The excursion
limit was derived from the AEL based on the following recommendation of
the American Conference of Governmental Industrial Hygienists (ACGIH):
“Excursions in worker exposure levels may exceed 3 times the
[threshold limit value] TLV-TWA for no more than a total of 30 minutes
during a workday” (ACGIH 2004).  The excursion limit is a recommended
level established as a guideline for good practice in industrial hygiene
and is used by industrial hygienists as a marker to reduce the
possibility that the 8-hour TWA limit will be exceeded.  The RfC was
used to assess risks to the general population from exposure to ambient
air releases.  The AEL, excursion limit, and RfC are shown in   REF
_Ref127933132 \h  \* MERGEFORMAT  Table 3.1 .  Additional information on
the derivation of the AEL and RfC is outlined in   REF _Ref127942248 \h 
\* MERGEFORMAT  Attachment A  and   REF _Ref127942192 \h  \* MERGEFORMAT
 Attachment B .

Table   STYLEREF 1 \s  3 .  SEQ Table \* ARABIC \s 1  1 : Toxicity
Threshold Values   SEQ Table \* ARABIC  

Chemical	AEL	Excursion Limit	RfC

nPB	17 ppm	51 ppma	1 ppm

a According to ACGIH (2004), “[U]nder no circumstances should [worker
exposure levels] exceed five times the [threshold limit value] TLV-TWA,
provided that the TLV-TWA is not exceeded.”

Atmospheric Modeling 

This section presents an assessment of the potential risks to
atmospheric integrity posed by the use of nPB in the adhesive sector. 
The ozone depletion potential (ODP), global warming potential (GWP), and
atmospheric lifetime (ALT) of the proposed substitute are presented in  
REF _Ref127933166 \h  Table 4.1 .

Table   STYLEREF 1 \s  4 .  SEQ Table \* ARABIC \s 1  1 : ODP, GWP, and
ALT for nPB

Substitute	ODP	100-Year GWP	ALT (days)

nPB	0.016 to 0.019a	1.57b	19

a In the U.S., depending on the latitude at which this short-lived
halocarbon is emitted.

b For further information on the derivation of the GWP for nPB, please
see   REF _Ref127942283 \h  \* MERGEFORMAT  Attachment C .

EPA believes that the substitute is substantially less harmful to the
ozone layer than the continued use of methyl chloroform (CH3CCl3), which
has an estimated ODP of 0.1, 100-year GWP of 144, and ALT of 5 years
(WMO 2002).

Occupational Exposure Analysis

To determine whether the use of the substitute in adhesive applications
would pose an unacceptable risk to workers’ health, occupational
exposure modeling and evaluations of occupational exposure monitoring
data from three facilities using nPB were performed.

Occupational Exposure Modeling

In general, nPB usage in adhesive applications is highly emissive. 
Significant occupational exposure can occur when a spray gun is used to
aerosolize the adhesive.  Exposure levels depend on multiple factors
including the ventilation in the room, the size of the room, the amount
of nPB being used, and the worker’s proximity to the spray gun.  The
aerosol mist is present in the work area throughout the production
process, and workers typically spend 8 hours per day applying adhesive.

A box model was used to estimate concentrations of nPB that might be
present in the air of several hypothetical adhesive application
facilities.  In general, the setting in which nPB is applied varies
considerably depending on the size of the operation and the type of
application.  Adhesives are typically applied in either 1) an open-top
workbench spray area with side panels and some minor local ventilation,
or 2) an open room with no mist containment (i.e., supplemental
ventilation systems are assumed not to be present) (Swanson et al.
2002). 

Four emissions scenarios were considered in this analysis: 

Emissions from a facility with average ventilation and average adhesive
use (S1);

Emissions from a facility with average ventilation and high adhesive use
(S2);

Emissions from a facility with poor ventilation and average adhesive use
(S3);

Emissions from a facility with poor ventilation and high adhesive use
(S4).

The model inputs were based on information provided in the report
prepared for EPA’s Design for the Environment Program, “Alternative
Adhesives Technologies: Foam Furniture and Bedding Industries”
(Swanson et al. 2002).  This report was prepared using data gathered
through facility site visits, publicly available chemical data, and
input from industry experts.  Additional information on the assumptions
used to construct these scenarios, as well as the methodology used to
generate exposure estimates, is presented in   REF _Ref127942304 \h 
Attachment D , Section   REF _Ref127942331 \r \h  2 .  The results of
the analysis are outlined in   REF _Ref127933217 \h  Table 5.1 .  The
8-hour exposure concentrations from each scenario exceed the recommended
8-hour AEL of 17 ppm.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  1 : Eight-Hour
Exposure Concentration (ppm) to nPB tc \l1 "TABLE 3. EIGHT-HOUR EXPOSURE
CONCENTRATION (ppm) TO NPB 

Ventilation	

Average Adhesive Use	

High Adhesive Use



Average ventilation	

60.3 (S1)	

603 (S2)



Poor ventilation	

253 (S3)	

2,533 (S4)



Occupational Exposure Study Evaluations  

Monitoring data for nPB exposure were available from three adhesive
application facilities.  These facilities are Custom Products, Inc., in
Morrisville, NC; STN Cushion Company in Thomasville, NC; and Marx
Industries in Sawmills, NC (NIOSH 2002a, 2002b, 2003).  The exposure
data were obtained in the immediate vicinity of the actual spray
application and in surrounding areas for long-term exposures on an
8-hour TWA.  Although short-term exposure data were available for some
facilities, they were not included in this analysis.

The monitoring values for those workers actively engaged in spraying the
nPB-containing adhesive ranged from 5.4 ppm to 381.2 ppm; the wide range
is a result of variations in both the amount of adhesive applied and the
ventilation conditions.  Monitoring was done at two different times for
all three facilities.  Before the second sampling sessions at STN
Cushion Company and Custom Products, the spray stations were enclosed or
replaced; no action was taken to improve the spray stations at Marx
Industries between sampling sessions.  

When comparing the monitoring data to the modeled concentrations for the
four hypothetical situations discussed previously, Custom Products
(prior to spray station replacement) and Marx Industries had exposure
concentrations that are most comparable with the estimated values of
Scenario 3, with average adhesive use and poor ventilation.  STN
Cushion Company had exposure data that were comparable to the estimated
values for Scenario 1, with average adhesive use and average
ventilation.  After spray station improvements, both STN Cushion
Company and Custom Products had exposure concentrations that were, in
general, lower than those of Scenario 1 but still exceeded the AEL in
most cases.  In each scenario, the majority of nPB exposures in the
adhesives industry did not meet the 17 ppm exposure limit. 

The spray station improvements did result in significant reductions in
worker exposure.  For example, at the STN Cushion Company, there was a
72 percent reduction in worker exposure, and at Custom Products, there
was an 89 percent reduction in worker exposure.  Nevertheless, the
8-hour TWA exposures still exceeded the AEL for the majority of the
workers.  While actions to further reduce exposure are possible—such
as a decrease in unused space in spray booths to promote
ventilation—it is not clear whether these improvements will result in
the significant reduction needed to meet the 17 ppm AEL.  Furthermore,
it is not clear whether facilities would be able to reduce nPB exposure
for all employees, especially the most active sprayers. 

General Population Exposure Analysis

This section presents the results of the general population exposure
modeling to assess the potential risks to the general population from
exposure to ambient air releases of the substitute.  Releases occurring
during the use of nPB in two typical adhesive facilities—a stand-alone
average-size adhesive application facility (Scenario 1) and an urban
row-house-type warehouse (Scenario 2)—were modeled.

Information regarding the assumptions used to construct the scenarios
and the methodology used to generate exposure estimates are presented in
  REF _Ref127942377 \h  Attachment F .  EPA’s SCREEN3 air dispersion
model was used to assess dispersion of emissions to estimate the highest
1-hour concentration from the two scenarios.  

The vented emission concentrations from the modeled Scenario 1
facilities were lower than the non-vented emissions and were well below
the RfC of 1 ppm.  None of the Scenario 2 facilities modeled had an
exposure concentration that exceeded the RfC.  

Volatile Organic Compound Analysis

nPB is considered a VOC under CAAA regulations (40 CFR 51.100(s)), and
therefore emissions should be controlled.  

References

ACGIH.  2004.  TLVs and BEIs: Threshold Limit Values for Chemical
Substances and Physical Agents, Biological Exposure Indices.  American
Conference of Governmental Industrial Hygienists.  Cincinnati, OH.

EPA.  1994.  SNAP Technical Background Document: Risk Screen on the Use
of Substitutes for Class I Ozone-Depleting Substances: Adhesives,
Coatings, and Inks.  U.S. Environmental Protection Agency. 

NIOSH.  2003.  NIOSH Health Hazard Evaluation Report: HETA
#99-0260-2906; Marx Industries, Inc.; Sawmills, NC.  National Institute
for Occupational Safety and Health.  June 2003.  Available online at  
HYPERLINK "http://www.cdc.gov/niosh/hhe/reports/pdfs/1999-0260-2906.pdf"
 http://www.cdc.gov/niosh/hhe/reports/pdfs/1999-0260-2906.pdf . 

NIOSH.  2002a.  NIOSH Health Hazard Evaluation Report: HETA #
98-0153-2883; Custom Products, Inc.; Mooresville, NC.  National
Institute for Occupational Safety and Health.  November 2002.  Available
online at   HYPERLINK
"http://www.cdc.gov/niosh/hhe/reports/pdfs/1998-0153-2883.pdf" 
http://www.cdc.gov/niosh/hhe/reports/pdfs/1998-0153-2883.pdf . 

NIOSH.  2002b.  NIOSH Health Hazard Evaluation Report: HETA
#2000-0410-2891; STN Cushion Company; Thomasville, NC.  National
Institute for Occupational Safety and Health.  August 2002.  Available
online at   HYPERLINK
"http://www.cdc.gov/niosh/hhe/reports/pdfs/2000-0410-2891.pdf" 
http://www.cdc.gov/niosh/hhe/reports/pdfs/2000-0410-2891.pdf . 

Swanson, M.B., J.R. Geibig, and K.E. Kelly.  2002.  Alternative
Adhesives Technologies: Foam Furniture and Bedding Industries, Final
Draft.  Volume 2: Risk Screening and Comparison.  Chapter 4: Exposure
Assessment.  Produced by the University of Tennessee Center for Clean
Products and Clean Technologies under a grant from EPA’s Design for
the Environment Branch, Office of Pollution and Prevention and Toxics. 
June 2002.  Available online at   HYPERLINK
"http://eerc.ra.utk.edu/ccpct/aap1.html" 
http://eerc.ra.utk.edu/ccpct/aap1.html .

WMO. 2002.  Scientific Assessment of Ozone Depletion: 2002.  World
Meteorological Organization, Global Ozone Research and Monitoring
Project, Report No. 47.

Risk Screen on Substitutes for Ozone-Depleting Substances

Aerosol Solvent Applications

Introduction

Ozone-depleting substances (ODS) are being phased out of production in
response to a series of diplomatic and legislative efforts that have
taken place over the past decade, including the Montreal Protocol and
the Clean Air Act Amendments of 1990 (CAAA).  The U.S. Environmental
Protection Agency (EPA), as authorized by Section 612 of the CAAA, has
developed a program to evaluate the human health and environmental risks
posed by alternatives to ODS.  The main purpose of EPA’s program,
called the Significant New Alternatives Policy (SNAP) program, is to
identify acceptable and unacceptable substitutes for ODS in specific end
uses.

EPA’s decision on the acceptability of a substitute is based largely
on the findings of a screening assessment of potential human health and
environmental risks posed by the substitute in specific applications. 
EPA has already screened a large number of substitutes in many end uses
within all of the major ODS-using sectors, including refrigeration and
air-conditioning, solvent cleaning, foam blowing, aerosols, fire
suppression, adhesives, coatings and inks, and sterilization.  The
results of these risk screens are presented in a series of background
documents that are available in EPA’s docket.  The reader is referred
to this reference for an additional discussion of different
methodologies used to conduct risk screens.

The purpose of this risk screen is to supplement EPA’s background
document (EPA 1994) on the aerosol solvent applications of n-propyl
bromide (nPB) or 1-bromopropane, which is used as an alternative to ODS
such as HCFC-141b and has been shown to exhibit toxicity upon
inhalation.  The potential human health and environmental risks posed by
nPB are examined in this risk screen.

  REF _Ref127935234 \h  Table 1.1  provides the chemical details on nPB.

Table   STYLEREF 1 \s  1 .  SEQ Table \* ARABIC \s 1  1 : Chemical
Details for n-Propyl Bromide

Name	Chemical Formula	CAS #

n-Propyl Bromide

(1-Bromopropane, nPB)	C3H7Br	000106-94-5



Occupational exposure and general population analyses were performed to
ensure that use of nPB in the application listed above does not pose
unacceptable risks to workers or the general public.  Consumer exposure
modeling was not performed because no consumer applications are proposed
for this substitute.

The remainder of this risk screen is organized as follows: 

Section 2 of this report summarizes the results of the risk screen for
nPB;

Section 3 presents the toxicity values used for the risk screen.  

Section 4 presents atmospheric modeling and potential environmental
risks; 

Section 5 presents occupational exposure analysis; 

Section 6 presents the general population exposure analysis; and 

Section 7 assesses the emissions of volatile organic compounds (VOCs).

Summary of Results 

Use of nPB is recommended for SNAP approval in aerosol solvent
applications, provided that a spray booth or local ventilation is
present at the point of use.  This requirement stems from the fact that
exposure modeling indicates that use of nPB in aerosol solvent
applications may pose a significant risk to human health if adequate
ventilation is not present.  Specifically, in various in-situ exposure
modeling scenarios and in three occupational exposure studies simulating
inadequate ventilation, the recommended 8-hour acceptable exposure limit
(AEL) of 17 ppm was often exceeded.  However, site-specific data show
that the installation of spray booths and local ventilation can decrease
worker exposure levels to below the AEL.  In addition, atmospheric
analyses indicate that use of nPB is preferable to the continued use of
HCFC-141b.  The modeled exposure levels in the general-population
exposure screening are all below the recommended reference concentration
(RfC). 

Toxicity Reference Values for Substitutes

To assess potential health risks from exposure to this substitute for
ODS in the aerosol solvent sector, EPA identified the relevant toxicity
threshold values, including an AEL, excursion limit, and RfC for
comparison to modeled and actual exposure concentrations.  The AEL
represents the maximum 8-hour time weighted average (TWA) at which a
worker can be exposed regularly without adverse effects.  The excursion
limit was derived from the AEL based on the following recommendation of
the American Conference of Governmental Industrial Hygienists (ACGIH):
“Excursions in worker exposure levels may exceed 3 times the
[threshold limit value] TLV-TWA for no more than a total of 30 minutes
during a workday” (ACGIH 2004).  The excursion limit is a recommended
level established as a guideline for good practice in industrial hygiene
and is used by industrial hygienists as a marker to reduce the
possibility that the 8-hour TWA limit will be exceeded.  The RfC was
used to assess risks to the general population from exposure to ambient
air releases.  The AEL, excursion limit, and RfC are shown in   REF
_Ref127935725 \h  Table 3.1 .  Additional information on the derivation
of the AEL and RfC is outlined in   REF _Ref127942401 \h  Attachment A 
and   REF _Ref127942192 \h  Attachment B .

Table   STYLEREF 1 \s  3 .  SEQ Table \* ARABIC \s 1  1 : Toxicity
Threshold ValuesTable   SEQ Table \* ARABIC  Table   SEQ Table \* ARABIC
 

Chemical	AEL	Excursion Limit	RfC

nPB	17 ppm	51 ppma	1 ppm

a According to ACGIH (2004), “[U]nder no circumstances should [worker
exposure levels] exceed five times the [threshold limit value] TLV-TWA,
provided that the TLV-TWA is not exceeded.”

Atmospheric Modeling 

This section presents an assessment of the potential risks to
atmospheric integrity posed by the use of nPB in the aerosol solvent
sector.  The ozone depletion potential (ODP), global warming potential
(GWP), and atmospheric lifetime (ALT) of the proposed substitute are
presented in   REF _Ref127935773 \h  Table 4.1 .

Table   STYLEREF 1 \s  4 .  SEQ Table \* ARABIC \s 1  1 : ODP, GWP, and
ALT for nPB

Substitute	ODP	100-Year GWP	ALT (days)

nPB	0.016 to 0.019a	1.57b	19

a In the U.S., depending on the latitude at which this short-lived
halocarbon is emitted.

b For further information on the derivation of the GWP for nPB, please
see   REF _Ref127942421 \h  \* MERGEFORMAT  Attachment C .

EPA believes that the substitute is substantially less harmful to the
ozone layer than the continued use of HCFC-141b, which has an estimated
ODP of 0.1, 100-year GWP of 713, and ALT of 9.3 years (WMO 2002).

Occupational Exposure Analysis

To ensure that the use of the substitute in aerosol solvent applications
does not pose an unacceptable risk to workers, occupational exposure
modeling and evaluations of three occupational exposure studies were
performed for the proposed substitute.

Occupational Exposure Modeling

Aerosol nPB usage in solvent applications is highly emissive. 
Significant occupational exposure can occur when aerosol solvents are
used to clean or treat a piece of equipment.  Exposure levels depend on
multiple factors that include room ventilation, room size, the amount of
nPB being used, and worker proximity to the spray apparatus.  Therefore,
workers applying nPB solvent in this type of application can be exposed
to varying concentrations.  The occupational exposure modeling attempts
to quantify nPB exposure resulting from several likely in-situ use
scenarios.

The analysis is based on a simple box-model approach that draws
assumptions from spray tests performed by aerosol solvent manufacturing
companies.  Information on the assumptions and application scenarios
used to create the exposure model can be found in   REF _Ref127942434 \h
 Attachment E , Section   REF _Ref127942467 \r \h  1 .  These
application scenarios provide exposure estimates of very high
concentrations because intense use and low ventilation were assumed. 
The modeled exposures reached a maximum of 32.6 ppm and a minimum of 1
ppm, depending on the concentration of nPB in the solvent, the level of
solvent use, and the ventilation of the area.

The results of the exposure modeling suggest that when using solvent
mixtures with nPB concentrations greater than or equal to 45 percent and
thereby using more than 450 g of nPB per day, ventilation is required to
keep the occupational exposure levels below the recommended AEL exposure
limit.  But, if  adequate ventilation exists and care is taken to
monitor and reduce the frequency of highly concentrated use,
occupational exposure to nPB would not pose a threat to workers’
health. 

Occupational Exposure Study Evaluations

In addition to the occupational exposure modeling, three occupational
exposure studies for aerosol solvent use were evaluated to assess
possible exposure issues for the use of nPB as an aerosol solvent.   
REF _Ref127942492 \h  Attachment E , Section   REF _Ref127942484 \r \h 
2  provides further details on the results of the occupational study
evaluations.

The exposure studies were conducted for a reasonable application end
use, and the poor or worst-case ventilation characteristics during
monitoring produced nPB (or surrogate) exposure results that were well
above the AEL and excursion limit for worker safety.  The 8-hour
exposures ranged from a maximum of 194 ppm to a minimum of 5.5 ppm.  The
15-minute exposures ranged from a maximum of 1,100 ppm to a minimum of
11 ppm.  The measured exposure concentrations depended on the location
where the samples were taken, the amount of solvent used, and the degree
of ventilation.  Despite the high exposures measured in these studies,
it was shown that when sufficient ventilation was present, it was
possible to meet occupational safety limits for the use of nPB. 
Therefore, users of more traditional solvents will most likely have to
increase ventilation rates in order to meet the low AEL and excursion
limit for nPB.

It should also be noted that in addition to adequate ventilation,
preferably through the use of locally-ventilated booths, appropriate
worker personal protective equipment (PPE) and procedural best
management practices to reduce worker exposure are also critical.

General Population Exposure Analysis

This section presents the results of the general population exposure
modeling to assess the potential risks to the general population from
exposure to ambient air releases of the substitute.  Releases occurring
during the use of nPB in two typical adhesive facilities—a stand-alone
average-size adhesive application facility (S1) and an urban
row-house-type warehouse (S2)—were modeled.

The adhesives end use was chosen as the worst-case scenario to model
because adhesive use is more emissive than the aerosol solvent use. 
Thus, EPA expects that the greatest nPB exposure to the general
population will result from adhesive operations.

Information regarding the assumptions used to construct the scenarios
and the methodology used to generate exposure estimates are presented in
  REF _Ref127942505 \h  Attachment F .  EPA’s SCREEN3 air dispersion
model was used to assess dispersion of emissions to estimate the highest
1-hour concentration from the two scenarios.  

The vented emission concentrations from S1 facilities were lower than
the non-vented emissions and are well below the RfC of 1 ppm.  None of
the S2 facilities modeled had an exposure concentration that exceeded
the RfC.

Volatile Organic Compound Analysis

nPB is considered a VOC under CAAA regulations (40 CFR 51.100(s)), and
therefore emissions should be controlled.

References

ACGIH.  2004.  TLVs and BEIs: Threshold Limit Values for Chemical
Substances and Physical Agents, Biological Exposure Indices.  American
Conference of Governmental Industrial Hygienists.  Cincinnati, OH.

EPA.  1994.  SNAP Technical Background Document: Risk Screen on the Use
of Substitutes for Class I Ozone-Depleting Substances: Aerosol Solvents.
 U.S. Environmental Protection Agency.  

WMO. 2002.  Scientific Assessment of Ozone Depletion: 2002.  World
Meteorological Organization, Global Ozone Research and Monitoring
Project, Report No. 47.

Risk Screen on Substitutes for Ozone-Depleting Substances

Solvent Cleaning Applications

Introduction

Ozone-depleting substances (ODS) are being phased out of production in
response to a series of diplomatic and legislative efforts that have
taken place over the past decade, including the Montreal Protocol and
the Clean Air Act Amendments of 1990 (CAAA).  The U.S. Environmental
Protection Agency (EPA), as authorized by Section 612 of the CAAA, has
developed a program to evaluate the human health and environmental risks
posed by alternatives to ODS.  The main purpose of EPA’s program,
called the Significant New Alternatives Policy (SNAP) program, is to
identify acceptable and unacceptable substitutes for ODS in specific end
uses.

EPA’s decision on the acceptability of a substitute is based largely
on the findings of a screening assessment of potential human health and
environmental risks posed by the substitute in specific applications. 
EPA has already screened a large number of substitutes in many end uses
within all of the major ODS-using sectors, including refrigeration and
air-conditioning, solvent cleaning, foam blowing, aerosols, fire
suppression, adhesives, coatings and inks, and sterilization.  The
results of these risk screens are presented in a series of background
documents that are available in EPA’s docket.  The reader is referred
to this reference for an additional discussion of different
methodologies used to conduct risk screens.

The purpose of this risk screen is to supplement EPA’s background
document (EPA 1994) on the solvent cleaning applications of n-propyl
bromide (nPB) or 1-bromopropane, which is used as an alternative to ODS
such as CFC-113 and methyl chloroform and has been shown to exhibit
toxicity upon inhalation.  The potential human health and environmental
risks posed by nPB are examined in this risk screen.

  REF _Ref127935926 \h  \* MERGEFORMAT  Table 1.1  provides the chemical
details on nPB.

Table   STYLEREF 1 \s  1 .  SEQ Table \* ARABIC \s 1  1 : Chemical
Details for n-Propyl Bromide

Name	Chemical Formula	CAS #

n-Propyl Bromide

(1-Bromopropane, nPB)	C3H7Br	000106-94-5



Occupational exposure and general population analyses were performed to
ensure that use of nPB in the application listed above does not pose
unacceptable risks to workers or the general public.  Consumer exposure
modeling was not performed because no consumer applications are proposed
for this substitute.

The remainder of this risk screen is organized as follows: 

Section 2 of this report summarizes the results of the risk screen for
nPB;

Section 3 presents the toxicity values used for the risk screen.  

Section 4 presents atmospheric modeling and potential environmental
risks; 

Section 5 presents occupational exposure analysis; 

Section 6 presents the general population exposure analysis; and 

Section 7 assesses the emissions of volatile organic compounds (VOCs).

Summary of Results 

Use of nPB is recommended for SNAP approval in solvent cleaning
applications.  This recommendation assumes that adequate exposure
controls, such as secondary cooling coils for vapor degreasers, are in
place at the point of use and that an effort is made to limit the amount
of time employees spend at the solvent cleaning stations.  It is also
recommended that users receive training in the proper use of solvent
cleaning equipment.  These recommendations stem from the fact that
exposure modeling indicates that use of nPB in solvent cleaning
applications may pose a significant risk to human health if adequate
exposure controls are not present.  Specifically, in some in-situ
exposure modeling scenarios based on historic data, the recommended
8-hour acceptable exposure limit (AEL) of 17 ppm was often exceeded. 
However, other modeled exposure data and site-specific worker exposure
data show that exposure levels can be held to levels significantly below
the AEL.  In addition, analyses of atmospheric effects indicate that use
of nPB is preferable to the continued use of CFC-113 and methyl
chloroform.  The modeled exposure levels in the general-population
exposure screening are all below the recommended reference concentration
(RfC).

Toxicity Reference Values for Substitutes

To assess potential health risks from exposure to this substitute for
ODS in the solvent cleaning sector, EPA identified the relevant toxicity
threshold values, including an AEL, excursion limit, and RfC for
comparison to modeled and actual exposure concentrations.  The AEL
represents the maximum 8-hour time weighted average (TWA) at which a
worker can be exposed regularly without adverse effects.  The excursion
limit was derived from the AEL based on the following recommendation of
the American Conference of Governmental Industrial Hygienists (ACGIH):
“Excursions in worker exposure levels may exceed 3 times the
[threshold limit value] TLV-TWA for no more than a total of 30 minutes
during a workday” (ACGIH 2004).  The excursion limit is a recommended
level established as a guideline for good practice in industrial hygiene
and is used by industrial hygienists as a marker to reduce the
possibility that the 8-hour TWA limit will be exceeded.  The RfC was
used to assess risks to the general population from exposure to ambient
air releases.  The AEL, excursion limit, and RfC are shown in   REF
_Ref127935952 \h  Table 3.1 .  Additional information on the derivation
of the AEL and RfC is outlined in   REF _Ref127942527 \h  Attachment A 
and   REF _Ref127942192 \h  Attachment B .

Table   STYLEREF 1 \s  3 .  SEQ Table \* ARABIC \s 1  1 : Toxicity
Threshold ValuesTable   SEQ Table \* ARABIC  Table   SEQ Table \* ARABIC
 

Chemical	AEL	Excursion Limit	RfC

nPB	17 ppm	51 ppma	1 ppm

a According to ACGIH (2004), “[U]nder no circumstances should [worker
exposure limits] exceed five times the [threshold limit value] TLV-TWA,
provided that the TLV-TWA is not exceeded.”

Atmospheric Modeling 

This section presents an assessment of the potential risks to
atmospheric integrity posed by the use of nPB in the solvent cleaning
sector.  The ozone depletion potential (ODP), global warming potential
(GWP), and atmospheric lifetime (ALT) of the proposed substitute are
presented in   REF _Ref127935978 \h  Table 4.1 .

Table   STYLEREF 1 \s  4 .  SEQ Table \* ARABIC \s 1  1 : ODP, GWP, and
ALT for nPB

Substitute	ODP	100-Year GWP	ALT (days)

nPB	0.016 to 0.019a	1.57b	19

a In the U.S., depending on the latitude at which this short-lived
halocarbon is emitted.

b For further information on the derivation of the GWP for nPB, please
see   REF _Ref127942546 \h  \* MERGEFORMAT  Attachment C .

EPA believes that the substitute is substantially less harmful to the
ozone layer than the continued use of methyl chloroform, which has an
estimated ODP of 0.1, 100-year GWP of 144, and ALT of 5 years, and
CFC-113, which has an estimated ODP of 1.0, 100-year GWP of 6030, and
ALT of 85 years (WMO 2002).

Occupational Exposure Analysis

To ensure that the use of the substitute in solvent cleaning
applications does not pose an unacceptable risk to workers, occupational
exposure modeling and an evaluation of three occupational exposure
studies were performed for the proposed substitute.

Occupational Exposure Modeling

nPB can be used as a cleaning solvent for metal, precision, and
electronics cleaning.  There are three types of equipment typically used
in solvent cleaning applications: cold cleaners, open top vapor
degreasers, and conveyorized vapor degreasers.  Exposure levels that
occur during the use of the three equipment types depend on multiple
factors including room ventilation, room size, the amount of nPB being
used, and worker proximity to the cleaning apparatus.  Therefore,
workers performing nPB solvent cleaning activities can be exposed to
varying concentrations.  The occupational exposure modeling evaluated
here attempts to quantify nPB exposure resulting from the most likely
use scenarios.

The analysis is based on the surrogate method used in EPA’s SNAP
Technical Background Document (1994), which uses exposure data for a
solvent in current use (called the analog solvent) as the basis for
estimating inhalation exposures to the substitute solvent.  The exposure
data used as the basis for the extrapolation reflect the exposures of
workers engaged in the same processes (e.g., electronics and metal
cleaning) as the end uses of interest.  The surrogate method
extrapolates on the basis of the vapor pressures (in mmHg) of the analog
and substitute solvent.  The formula used to estimate exposures for pure
substances is: 

Cs * MWs	=	Ca * MWa	*	Ps

24.45

24.45

Pa



Where:	MWs	= molecular weight of the substitute in g/mol

	MWa	= molecular weight of the analog, in g/mol

	Cs	=  estimated concentration for the substitute, in ppm (by volume)

	Ca	=  actual exposure concentration for the analog, in ppm (by volume)

	Ps	=  vapor pressure of substitute, in mm Hg

	Pa	=  vapor pressure of analog, in mm Hg

Exposure data for several solvents, including CFC-113, methyl
chloroform, and trichloroethylene, were used to extrapolate inhalation
exposure concentrations in the SNAP Background Document (1994).  This
exposure data were obtained from several NIOSH Health Hazard Evaluations
from the 1970’s and 1980’s, which contain actual measured exposures
of workers engaged in electronics or metal cleaning, and are presented
in   REF _Ref127936110 \h  Table 5.1  below.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  1 : Exposure Data
for Electronics and Metal Cleaning for Analog Solvents

Analog Solvent	Exposures Reported for Analog Solvent (ppm)	NIOSH Study
Number

	Minimum	Maximum

	Electronics Cleaning



	Methyl Chloroform	14	57	81-370-1050

Methyl Chloroform	14a	83-164-1377

CFC-113	7	26	81-370-1050

CFC-113	0	54	83-164-1377

CFC-113	5	90	83-089-1329

Metal Cleaning



	Methyl Chloroform	12	17	76-24-350

Methyl Chloroform	1	5	83-296-1491

Methyl Chloroform	9	132	83-170-1346

Trichloroethylene	11	12	80-87-708

Trichloroethylene	15	28	76-24-350

Trichloroethylene	4	10	83-296-1491

a Only one value was given.  

In addition to the data available from the NIOSH studies, occupational
exposure data received for another solvent used in similar applications
(HFE-7100) were used to model exposure to nPB during solvent cleaning
uses.  The average 8-hour TWA exposures for the related compound are
reported in   REF _Ref127936145 \h  Table 5.2  for the following three
scenarios that could take place during vapor degreaser use:

The first scenario tested exposures at the lip of the degreaser.  This
is a worst-case scenario of worker exposure and would occur in the event
that an operator was to reach down into the machine during its
operation;

The second case tested worker exposure in the operator zone during
typical degreasing operations; and

The third test scenario tested worker exposure levels during a spill. 
In this test, a measured quantity of the chemical was spilled—first
when the chemical was at room temperature and later when it had reached
its boiling point—and operator exposure was measured.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  2 : Exposure Data
for HFE-7100 During Electronics Cleaning Operations

Test Scenario	Time Period of Exposure	Occupational Exposure (ppm)

Exposure at Lip of Machine	8-hour TWA	331

Exposure in Operator Zone	8-hour TWA	23

Exposure During a Spill	8-hour TWA	80a, 131b

a Exposure at room temperature.

b Exposure at boiling point.

Note: The value reported for exposure during a spill is an 8-hour TWA. 
According to the manufacturer, the highest short-term (acute) exposure
to HFE-7100 is less than 400 ppm during routine operations or in the
case of a moderate spill.

No data describing the exposures of workers to solvents used in
precision cleaning were specifically identified.  However, because many
of the processes and methods used for precision cleaning are similar to
those used in metal and electronics cleaning, worker exposures to
solvents during precision cleaning would not differ substantially from
those observed during metal and electronics cleaning.  Therefore one can
assume a similar outcome during precision cleaning.

Using the molecular weights and vapor pressures of the four analog
solvents and the one surrogate solvent (nPB), which are presented in  
REF _Ref127936178 \h  Table 5.3 , extrapolated exposures for nPB were
calculated using the formula presented above.  These extrapolated
exposures are displayed in   REF _Ref127936198 \h  Table 5.4  below.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  3 : Molecular
Weights and Vapor Pressures of Analog and Surrogate Solvents

Solvent	Molecular Weight (g/mol)	Vapor Pressure 

(mm Hg)

nPB	123	146

Methyl Chloroform	133	100

CFC-113	187	285

Trichloroethylene	131	58

HFE-7100	250	200



Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  4 : Extrapolated
Inhalation Exposures for nPBa

Analog Solvent	nPB Exposure Concentration (ppm)

	Minimum	Maximum

Electronics Cleaning	 	 

Methyl Chloroform	23	91

Methyl Chloroform	22b

CFC-113	6	20

CFC-113	0.2	42

CFC-113	4	70

HFE-7100: Exposure at Lip of Machine	491b

HFE-7100: Exposure in Operator Zone	34b

HFE-7100: Exposure During a Spill	119c, 194d

Metal Cleaning	 	 

Methyl Chloroform	19	28

Methyl Chloroform	2	8

Methyl Chloroform	15	209

Trichloroethylene	29	33

Trichloroethylene	39	75

Trichloroethylene	11	26

a Bold values are above the recommended AEL of 17 ppm.

b Only one value was given.

c Exposure at room temperature.

d Exposure at boiling point.

The results of the exposure modeling show that of the 24 exposure
concentrations, seven (29%) are below the recommended AEL of 17 ppm.

To ensure that occupational exposure does not exceed the recommended
AEL, the following exposure controls that are appropriate to the
particular solvent cleaning application can be employed:

Use covers on cold-cleaning and vapor degreasing equipment when not in
use;

Increase freeboard height and install a freeboard refrigeration device
on the vapor degreaser for maximum condensation;

Use automatic hoists or conveyorized systems;

Reduce drag-out losses of solvent by keeping the item in the vapor zone
long enough to drain and dry any entrapped or remaining solvent;

Limit air movement over the degreaser, which “can increase both
diffusion and convection losses by creating turbulence in the
degreaser” (Center for Emissions Control 1992);

Use a lip-vent exhaust system, which is designed to reduce solvent
exposure to workers by capturing vapors and venting them out of the
room.  Lip-vent exhaust systems do significantly increase solvent
consumption and atmospheric emissions, unless vapor recovery equipment
is employed;

Use moderate general ventilation for the solvent cleaning room or the
entire facility to decrease solvent exposure to workers outside the
immediate solvent cleaning area; care should be taken to ensure that
this ventilation does not create turbulence over the degreaser;

Follow industrial hygiene practices and train workers to follow
practices;

Place signs in the work place warning workers of the symptoms that may
occur from over-exposure to nPB;

Provide training in the proper use of cold cleaning and vapor degreasing
equipment.

It should be noted that more recent occupational exposure data from
facilities using nPB show that nPB exposures in solvent cleaning
applications were significantly below the AEL.  Therefore, assuming
adequate emissions controls exist and care is taken to limit cleaning
activities throughout the day, occupational exposure to nPB would not
pose a threat to workers’ health. 

Occupational Exposure Study Evaluation

In addition to the occupational exposure modeling, three occupational
exposure studies for solvent cleaning were evaluated to assess possible
exposure issues for the use of nPB in solvent cleaning applications.

Trilithic, Inc.

Monitoring data for nPB exposure were available from a NIOSH Health
Hazard Evaluation for Trilithic, Inc., which manufactures
instrumentation and components for the radio frequency and microwave
communications industry.  Employees perform degreasing activities using
an enclosed cold vapor degreaser, on average, two to three times per
week.

  REF _Ref127936242 \h  Table 5.5  presents the 8-hour TWA exposure data
measured at the Trilithic facility.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  5 : nPB 8-Hour TWA
Exposure Concentration Results at Trilithic, Inc.

Job Title	Location	Concentration (ppm)	Number of Times Using Degreaser
During Shift

Assembler	Components	0.02	0

Assembler	Components	0.18	1

Assembler	Components	0.01a	0

Assembler	Components	0.02a	0

Assembler	Tech Station I	0.02	0

Assembler	Tech Station I	0.02	0

Assembler	Tech Station I	0.02	0

Assembler	Custom Filters	0.08	1

Assembler	Custom Filters	0.02	0

Assembler	Custom Filters	0.02	0

Assembler	Filters	0.17	1

Assembler	Filters	0.04	1

Assembler	Filters	0.02a	0

Assembler	Filters	0.05	1

Assembler	Filters	0.01a	0

Assembler	Tunable	0.02a	0

Assembler	Tunable	0.02a	0

Assembler	Tunable	0.02	0

Assembler	Tunable	0.08	1

Assembler	Engineering Support	0.63	2

a Concentrations were below the minimum quantifiable concentration (MQC)
of 0.02 ppm and above the minimum detectable concentration (MDC) of
0.004 ppm and are considered semi-quantitative.

Prior to the test results above, the degreaser was moved to an enclosed
room equipped with a local exhaust ventilation system to vent vapors
from the room.  As shown above, all of the 8-hour exposure
concentrations were significantly below the recommended AEL of 17 ppm,
even for the employee who used the degreaser twice in one work day.

Albemarle Corporation

Albemarle Corporation, a producer of nPB-based solvents, runs a product
stewardship program that includes monitoring of worker exposure for its
customers.  Over 400 exposure samples from workers using vapor
degreasers for metals, electronics, and precision cleaning taken as part
of the program were summarized and submitted to EPA (EPA 2003).  The
data are summarized in   REF _Ref127936291 \h  Table 5.6  below.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  6 : Summary of nPB
8-Hour TWA Exposure Concentration Data from Albemarle Corporation

Concentration Range (ppm)	Number of Samples in Range	Percentage of
Samples in Range

<1	109	26.5%

1 to 5	139	33.8%

6 to 10	55	13.4%

11 to 15	26	6.3%

16 to 20	24	5.8%

21 to 25	14	3.4%

26 to 30	0	0.0%

31 to 40	11	2.7%

41 to 50	8	1.9%

51 to 60	1	0.2%

61 to 70	2	0.5%

71 to 80	3	0.7%

81 to 90	1	0.2%

91 to 100	1	0.2%

100 to 125	5	1.2%

126 to 150	3	0.7%

151 to 175	1	0.2%

176 to 200	1	0.2%

201 to 250	5	1.2%

251 to 300	0	0.0%

>300	2	0.5%



As indicated in   REF _Ref127936291 \h  Table 5.6 , more than 80 percent
of the measured exposure concentrations were below the recommended AEL
of 17 ppm.

Company A

Company A, a distributor of nPB-based solvents, provides worker exposure
badges to its customers and ensures that they are analyzed for worker
exposure concentrations (EPA 2003).  Company A provided EPA with the raw
data for 94 worker exposure samples, which are summarized in   REF
_Ref127936322 \h  Table 5.7  below.

Table   STYLEREF 1 \s  5 .  SEQ Table \* ARABIC \s 1  7 : Summary of nPB
8-Hour TWA Exposure Concentration Data from Company A

Concentration Range	Number of Samples in Range	Percentage of Samples in
Range

<1	6	6.4%

1 to 5	25	26.6%

6 to 10	18	19.1%

11 to 15	10	10.6%

16 to 20	10	10.6%

21 to 25	6	6.4%

26 to 30	6	6.4%

31 to 40	4	4.3%

41 to 50	8	8.5%

51 to 60	1	1.1%



As presented in   REF _Ref127936322 \h  Table 5.7 , approximately 63
percent of the exposures were below 15 ppm.  Sixty-eight percent of the
exposures were at or below the recommended AEL of 17 ppm (EPA 2003).

Conclusion

Exposure results for employees using an enclosed cold vapor degreaser
were significantly below the recommended AEL of 17 ppm.  Additionally,
80 percent of exposure results from over 400 employees using vapor
degreasers were below the AEL, and 68 percent of exposure results from
92 employees using nPB as a solvent were below the AEL.  Although the
modeling results indicate that nPB exposure during vapor degreasing
could possibly be above the AEL, occupational exposure data from
facilities using nPB indicate that nPB exposures during solvent cleaning
applications can be kept well below the recommended AEL of 17 ppm.

General Population Exposure Analysis

This section presents the results of the general population exposure
modeling to asses the potential risks to the general population from
exposure to ambient air releases of the substitute.  Releases occurring
during the use of nPB in two typical adhesive facilities—a stand-alone
average-size adhesive application facility (S1) and an urban
row-house-type warehouse (S2)—were modeled.

The adhesives end use was chosen as the worst-case scenario to model
because adhesive use is more emissive than the solvent cleaning use. 
Thus, EPA expects that the greatest nPB exposure to the general
population will result from adhesive operations.

Information regarding the assumptions used to construct the scenarios
and the methodology used to generate exposure estimates are presented in
  REF _Ref127942576 \h  Attachment F .  EPA’s SCREEN3 air dispersion
model was used to assess dispersion of emissions to estimate the highest
1-hour concentration from the two scenarios.  

The vented emission concentrations from S1 facilities were lower than
the non-vented emissions and are well below the RfC of 1ppm.  None of
the S2 facilities modeled had an exposure concentration that exceeded
the RfC.  

Volatile Organic Compound Analysis

nPB is considered a VOC under CAAA regulations (40 CFR 51.100(s)), and
therefore emissions should be controlled.  

References

ACGIH.  2004.  TLVs and BEIs: Threshold Limit Values for Chemical
Substances and Physical Agents, Biological Exposure Indices.  American
Conference of Governmental Industrial Hygienists.  Cincinnati, OH.

Center for Emission Control.  1992.  Solvent Cleaning (Degreasing): An
Assessment of Emission Control Options.  Washington, D.C.  November
1992. 

EPA.  2003.  Summary of Data on Workplace Exposure to n-Propyl Bromide. 
EPA’s summary of exposure data from nPB suppliers and NIOSH.  May 21,
2003.  Available online as document number EPA-HQ-OAR-2002-0064-0015 at
http://www.regulations.gov/.

EPA.  1994.  SNAP Technical Background Document: Risk Screen on the Use
of Substitutes for Class I Ozone-Depleting Substances: Solvent Cleaning.
 U.S. Environmental Protection Agency.

MnTAP.  2005.  “Reducing Solvent Emissions from Vapor Degreasers.” 
Minnesota Technical Assistance Program Fact Sheet.  University of
Minnesota.  Available online at   HYPERLINK
"http://www.mntap.umn.edu/mach/70-VaporDegreasers.htm" 
http://www.mntap.umn.edu/mach/70-VaporDegreasers.htm .  

NIOSH.  2002.  NIOSH Health Hazard Evaluation Report: HETA
#2000-0233-2845; Trilithic, Inc.; Indianapolis, IN.  National Institute
for Occupational Safety and Health.  January 2001.  Available online at 
 HYPERLINK
"http://www.cdc.gov/niosh/hhe/reports/pdfs/2000-0233-2845.pdf" 
http://www.cdc.gov/niosh/hhe/reports/pdfs/2000-0233-2845.pdf . 

WMO. 2002.  Scientific Assessment of Ozone Depletion: 2002.  World
Meteorological Organization, Global Ozone Research and Monitoring
Project, Report No. 47.

Attachment   SEQ Attachment \* ALPHABETIC  A 

Determination of an AEL for n-Propyl Bromide and Its Documentation

AEL Derivation

Recommended AEL: 		17 ppm (8-hour Time Weighted Average)

Basis and Endpoints: 	A decrease in the number of estrous cycles in a
3-week period prior to mating following 7 weeks of exposure

Study: 	An Inhalation Two-Generation Reproductive Toxicity Study of
1-Bromopropane in Rats (WIL 2001)

Protocol:	Whole-body inhalation, 6 hours/day, 7 days/week for 70 days
prior to mating, during mating, gestation, lactation for two generations

Concentrations:			0, 100, 250, 500, or 750 ppm

BMDL:				162 ppm (mean number of estrous cycles in 3 weeks)

NOAEL:			100 ppm

LOAEL:			250 ppm (estrous cycle, sperm motility and hepatic effects)

BMDL [adj]:			(162 ppm × 6 hours / 8 hours × 7 days / 5 days = 170
ppm)

BMDL [HEC]:			170 ppm

Uncertainty Factors:	10 (composite factor of 3 for animal-to-human
extrapolation and 3 for within-human variability to account for
differences in individual sensitivity)

Results from BMD Analysis: 	Section   REF _Ref127940580 \r \h  4 
presents the results from the benchmark dose analyses conducted on
estrous cycle data.

Discussion of Relevant Literature

  SEQ CHAPTER \h \r 1 ICF has performed a re-evaluation of the
literature on n-propyl bromide (1-bromopropane, nPB) for the purpose of
assessing potential reproductive toxicity in females.  This
re-evaluation was prompted by the publication of several peer-reviewed
studies (e.g., Sekiguchi 2002; Yamada et al. 2003;   SEQ CHAPTER \h \r 1
Ichihara et al. 2002, 2004 a, b) examining changes in estrous/menstrual
cycle parameters in nPB-exposed groups.  ICF previously evaluated a
broad range of studies for numerous endpoints in order to assess
quantitatively the most sensitive level at which adverse health effects
occur and to develop an acceptable exposure limit as an 8-hour, time
weighted average for the workplace (ICF 2002).  Of the health effects
examined previously, ICF found liver effects (centrilobular
vacuolation), nervous system effects, and reproductive effects to be of
concern.  In 2002, ICF determined male reproductive effects (sperm
motility) to be the most sensitive effect requiring protection at the
lowest concentration.

Ichihara et al.

ICF performed a comprehensive, critical evaluation of all published
occupational studies on nPB by Ichihara et al. (1999, 2002, 2004a,
2004b).  Workers were examined using neurological, electrophysiological,
hematological, biochemical, neurobehavioral and postural sway tests. 
ICF concluded that these studies were limited because of:  small sample
sizes; inadequate exposure characterization (e.g., a single
time-weighted-average sample from each individual); very short study
duration (2-3 days); incomplete information on how interview data were
collected and validated; co-occurring exposures to other toxicants,
including 2-bromopropane; and failure to adjust for numerous confounding
variables.  Therefore, ICF found that the Ichihara studies were
insufficient for assessing the potential neurological, hematological and
reproductive health effects (e.g., amenorrhea, adverse effects on the
peripheral nerves and the central nervous system, and anemia) of nPB.

Sekiguchi 

Sekiguchi (2002) observed an increase in estrous cycle length in female
rats exposed via inhalation to 1000 ppm nPB for 3 weeks.  However, this
finding was not statistically significant, possibly because duration of
exposure was too short to induce a significant effect.

Yamada et al.  

In another animal study, Yamada et al. (2003), Wistar female rats (N =
9/dose group) were exposed to 0, 200, 400 or 800 ppm nPB for 8
hours/day, 7 days/week for 12 weeks.  The Yamada et al. (2003) study
found that, relative to controls, treated females exhibited (1) a
significant decrease in the number of estrous cycles, due to prolonged
diestrous, at 400 and 800 ppm; (2) a change in the subtype distribution
of ovarian follicles with dose-dependent statistically significant
decreases in antral follicles at 200 and 400 ppm, and in growing
follicles at 400 ppm; and (3) an increase in the number of primordial
follicles at 400 ppm that was not statistically significant.  It should
be noted that in this study, rats exposed to 800 ppm became “seriously
ill” and were euthanized at week 8.  Therefore, ovarian follicle
subtype numbers in this group were reported, but excluded from
statistical analysis.  The decreases in antral follicles at 200 ppm
could be a possible indicator of reduced ovulated oocytes and
potentially the critical effect in the Yamada et al. (2003) study. 
However, changes in distribution of follicular subtype as a critical
effect are insufficient in the absence of additional data. 
Statistically significant decreases in both antral follicles and corpora
lutea would provide strong evidence of a significant toxicological
effect.  However, data on corpora lutea were not provided by Yamada et
al. (2003).  In the absence of data on corpora lutea, it is not possible
to interpret the antral follicle findings; therefore, the use of estrous
cycle changes instead of decreases in follicles as the critical effect
in the Yamada study is scientifically justifiable and defensible. 
Estrous cycle irregularities in the Yamada et al. (2003) study were
first observed at approximately 2-3 weeks following initiation of
exposure in the 800 ppm group and at around 7-9 weeks in the 400 ppm
group. 

WIL 

Subsequently, ICF reviewed the WIL (2001) multi-generation
reproductive/developmental toxicity study and noted findings on female
reproductive toxicity similar to those of Yamada et al. (2003).  In this
study, male and female Sprague-Dawley F0 and F1 rats (N = 25/dose group)
were exposed via inhalation to 0, 100, 250, 500 and 750 ppm for 70 days
prior to mating and during mating, gestation, and lactation.  Primordial
follicles and corpora lutea were only counted in the controls and
high-dose groups of the F0 and F1 females, and therefore, no
dose-response could be established for these endpoints.  However,
significant increases in primordial follicles and decreases in corpora
lutea were noted in both generations.  An increase in estrous cycle
length was also observed in each exposure group, relative to controls,
prior to mating.

WIL (2001), however, did not statistically evaluate the estrous cycle
data.  In addition, the estrous cycle data reported by WIL (2001)
excluded females that showed no evidence of cycling (1 and 2 animals for
the 500- and 750-ppm F0 groups, respectively).  Therefore, ICF
re-evaluated the individual data by (1) counting the number of estrous
cycles within the three-week period prior to mating; (2) including
females who did not complete an estrous cycle; and (3) conducting a
statistical analysis of dose-response using an under-dispersed Poisson
regression model.  Using the new data, ICF performed statistical
analyses on two estrous cycle measures: (1) mean cycle length; and (2)
mean number of estrous cycles occurring during the 3-week period prior
to mating, both following 7 weeks of nPB exposure.

For evaluation of differences in the mean number of estrous cycles
occurring within the specified time period, data were assessed using an
under-dispersed Poisson regression model.  Since the number of cycles is
a whole number, a Poisson model, appropriate for count data, was chosen
to model the observed counts.  The simple Poisson model with a variance
equal to the mean fit the data poorly, and so a better fitting
under-dispersed Possion model was chosen instead; this model has a
variance equal to the mean multiplied by a scale factor.  By
comparison, the regression model assumes that the logarithm of the mean
is a linear function of the dose.

Females who did not cycle (acyclic) during this period were considered
to be in prolonged diestrous, and therefore were included in the
analysis.  Statistically significant decreases in the mean number of
estrous cycles were observed at ( 250 ppm in F0 females and at 500 ppm
in F1.females.  (It should be noted that F0 females exposed to 750 ppm
produced no live litters and therefore, there was no F1 generation at
this exposure level.)  Female reproductive toxicity findings are
presented in   REF _Ref127930840 \h  Table A.1  and   REF _Ref127936870
\h  Table A.2 .

Table A.  SEQ Table_A. \* ARABIC  1 : Female Reproductive Endpointsa

Endpoint	0 ppm	100 ppm	250 ppm	500 ppm	750 ppm

F0 Final Body Wt (g)	331(20.7b	330(22.3	327(24.8	332(38.3	319(25.5

F1 Final Body Wt (g)	321(27.3	325(28.1	318(26.7	309(29.5	-

F0 Ovaries (g)	0.1227(0.0259	0.1265(0.0240	0.1152(0.02360	0.1119(0.01514
0.09575(0.02798**

F0 Relative Ovaries

 (g/100 g)	0.037(0.0078	0.038(0.0068	0.035(0.0072	0.034(0.0056
0.031(0.0079**

F1 Ovaries (g)	0.1131(0.0155	0.1077(0.0317	0.1056(0.02791	0.1062(0.02302
-

F1 Relative Ovaries (g/100 g)	0.035(0.0027	0.022(0.0032	0.022(0.0042
0.021(0.0045	-

F0 Fertility index (%)

N (Number of animals at exposure level)	92.0

N =25	100

N =25	88.0

N =25	52.0**

N =25	0.0**

N =25

F1 Fertility index (%)

N	100.0

N =25	84.0

N =25	80.0

N =25	100.0

N =25	-

F0 Mating index (%)

N	96.0

N =25	100

N =25	100

N =25	84.0

N =25	68.0*

N =25

F1 Mating index (%)

N	88.0

N =25	68.0

N =25	64.0

N =25	72.0

N =25	-

F0 Evidence of mating w/out delivery (no.)	1	0	3	10	17

F1 Evidence of mating w/out delivery (no.)	3	4	4	8	-

Number of F0 Implantation sites

N	15.3(2.53

N = 23	14.3(3.09

N = 25	13.8(4.23

N =22	9.0(4.54**

N =11	NA

Number of F1 Implantation sites

N	15.5(2.11

N =22	15.8(3.29

N =17	13.5(4.12

N =16	9.8(4.93**

N =17

	F0 Number of born

N	15.0(2.42

N =23	13.6(3.09

N =25	12.5(4.27

N =22	8.5(4.41**

N =11	NA

F1 Number of born

N	14.9(1.97

N =22	15.1(3.35

N =17	13.1(4.12

N =16	8.6(4.51**

N =17	-

F0 Number of Unaccounted Implantation Sites

N	0.3(0.57

N =23	0.7(0.95

N =25	1.3(1.36**

N =22	0.5(0.69

N =11	NA

F1 Number of Unaccounted Implantation Sites

N	0.5(0.86

N =22	0.6(1.22

N =17	0.4(0.63

N =16	1.2(1.09

N =17

	a Data were provided on pp. 123-124, 207, 272-275, and 356 of the study
report

b Mean ± standard deviation

Table A.  SEQ Table_A. \* ARABIC  2 : Estrous Cycle Data

Endpoint	0 ppm	100 ppm	250 ppm	500 ppm	750 ppm

F0 Estrous cycle length (days)b

N	4.2(0.49a

N =25	4.5(1.05

N =25	4.7(0.90

N =25	5.5(2.17*

N =23	5.6(1.79*

N =22

F1 Estrous cycle length (days)b

N	4.5(1.25

N =24	4.5(0.91

N =24	4.9(1.43

N =22	5.1(1.68

N =21	-

F0 Mean no. of estrous cycles within 3 weeksc

N	3.96(0.54

N =25	3.84(0.62

N =25	3.52(0.65*

N =25	2.88(1.17**

N =25	2.56(1.26**

N =25

F1 Mean no. of estrous cycles within 3 weeksc

N	3.64(1.15

N =25	3.68(1.11

N =25	2.88(1.36

N =25	2.68(1.35*

N =25	-

F0 Mean no. of estrous cycles within 3 weeks excluding  acyclic femalesd

N	3.96(0.54

N =25	3.84(0.62

N =25	3.52(0.65**

N =25	3.00(1.02***

N =24	2.78(1.04***

N =23

F1 Mean no. of estrous cycles within 3 weeks excluding  acyclic femalesd

N	3.64(1.15

N =25	3.68(1.11

N =25	3.13(1.10

N =23	2.91(1.12*

N =23	-

a Mean ± standard deviation 

b Statistical analysis calculated by ICF

c Calculated by ICF using an under-dispersed Poisson log-linear
regression model 

d Calculated by ICF excluding females that did not have a full estrous
cycle assuming an under-dispersed Poisson log-linear regression model. 

* Significantly different from control, p<0.05.

** Significantly different from control, p<0.01.

*** Significantly different from control, p<0.001.

To gain support for the selection of the estrous cycle endpoint, an
expert panel was formed to consider the occupational exposure limit for
nPB and the toxicological significance of a statistically significant
decrease in the number of estrous cycles within a given time period. 
Dr. George Daston, Dr. Ulrike Luderer, and Dr. Jodi Flaws were asked to
serve on the panel because of their expertise on this subject and the
experience of the first two scientists as panel members for the Center
for the Evaluation of Risks to Human Reproduction’s 2001 review of the
reproductive toxicity of nPB (CERHR 2001).  The peer reviewers offered
constructive comments that were, in general, supportive of the critical
endpoint selected for derivation of the industrial AEL and of the
modeling approach used by ICF (2004).

AEL Determination

The WIL (2001) study was selected as the principal study for the AEL
determination because it (1) was a well-conducted experiment performed
in accordance with standard test guidelines for multi-generation
reproductive toxicity; (2) was documented using Good Laboratory Practice
procedures; (3) underwent an independent audit; (4) had sufficiently
large sample sizes; and (5) provided raw data for review and analysis. 
As identified earlier, the Yamada et al. (2003) study was not used in
the AEL determination because (1) limited data were provided, precluding
audit and independent analysis; (2) the study sample sizes were small;
(3) the high dose greatly exceeded the maximum tolerated dose, and
animals in this group had to be euthanized at week 8 because of severe
illness; and (4) fewer reproductive parameters were measured than in the
WIL (2001) study.

Regulatory guidance for interpreting the biological significance of
estrous cycle irregularities in the absence of additional reproductive
effects is not conclusive.  Although EPA’s Guidelines for Reproductive
Toxicity Risk Assessment (USEPA 1996) notes that these effects are to be
considered indicative of potential reproductive toxicity, NRC (2001)
concludes that changes in the distribution of estrous cycle length alone
are not a reliable predictor of reproductive toxicity.  In the WIL
(2001) study, statistically significant changes in the number of estrous
cycles in a three-week period prior to mating, due to an increase in the
cycle length and especially the diestrous phase, were an early precursor
to a functional reproductive effect that occurred with increasing dose,
as noted in   REF _Ref127930840 \h  Table A.1  and   REF _Ref127936870
\h  Table A.2 .

Further, the relevance of this endpoint was discussed with Drs. Sally
Darney and Ralph Cooper, reproductive toxicity experts at EPA’s NHERL
(National Health and Environmental Effects Research Lab at Research
Triangle Park) (Birgfeld 2004).  Both scientists agreed that estrous
cycle length is a relevant endpoint to use to determine an AEL and that
performing a BMD analysis on estrous cycle length might be possible. 
Dr. Darney further suggested that the data be transformed into number of
estrous cycles within a three-week period to avoid problems in the data
generated by acyclic rats.  Therefore, estrous cycle length is
considered to be the critical effect because it is the effect occurring
at the lowest concentration along a continuum of adverse reproductive
outcomes that increase in frequency and severity at higher doses.  The
number of estrous cycles in a 3-week period prior to mating was used
instead of the estrous cycle length in order to allow inclusion of the
data from acyclic females, for which an estrous cycle length is not
defined.

Although nPB inhalation exposure produces reproductive toxicity in both
male and female rats, a weight-of-evidence comparison of the
reproductive hazard findings suggested that females in the F0 generation
were slightly more sensitive than the F1 females (with regard to estrous
cycle length), the F0, or the F1 males (see   REF _Ref127936870 \h 
Table A.2 ), ensuring that this endpoint will be sufficiently
protective.  After considering the appropriate uncertainty factors for
each endpoint, the reproductive endpoints for both males and females
result in lower AELs than those for neurotoxicity or liver effects (ICF
2002).  Further, as seen in   REF _Ref127930840 \h  Table A.1  and   REF
_Ref127936870 \h  Table A.2 , the F0 females show a much clearer
dose-response for many of the reproductive endpoints than do the F1
females.  The pattern of exposure in the F0 generation resembles that
occurring in occupational exposure scenarios and is relevant to
occupational exposure in humans.  Therefore, the BMDL for F0 female
reproductive toxicity was used as the point of departure for development
of the AEL. 

  SEQ CHAPTER \h \r 1 Benchmark Dose Methods and Analysis 

Background

The Benchmark Dose (BMD) analysis utilized in the determination of the
AEL for nPB was based on the mean number of estrous cycles for the F0
generation within the 3-week period during which estrous cycle
measurements were made, prior to mating (  REF _Ref127936964 \h  Table
A.3 ).  Based on a weight-of-evidence hazard characterization and
biological relevance, this endpoint in the F0 generation was considered
to be the most sensitive reproductive endpoint for the study.  Acyclic
females were included in the analysis. 

Table A.  SEQ Table_A. \* ARABIC  3 : F0 Mean Number of Estrous Cycles
in Female Sprague-Dawley Rats Administered n-Propyl Bromide via
Inhalation for 70 Daysa

Dose (ppm)	F0 Animals

(N)	Estrous Cycles

(mean)

0	25	3.96

100	25	3.84

250	25	3.52

500	25	2.88

750	25	2.56

a Estrous cycles measured for three-week period prior to mating

N, number of animals per group

The data sets considered for dose-response modeling are discrete and
categorical, since the number of estrous cycles is a whole number.  The
EPA’s Benchmark Dose Software (BMDS) (Version 1.3.2, EPA 2000a) was
used to accomplish all of the model fitting and BMD estimation.  This
most recent version of BMDS is designed for either dichotomous data,
with only two possible responses, or for continuous data with infinitely
many responses.  The BMDS model required that the data be treated as
both approximately continuous and approximately normally distributed. 
The continuous endpoints of interest with respect to toxicity were
quantitatively summarized by group means and measures of variability
(standard errors or standard deviations).  The models used to represent
the dose-response behavior of those continuous endpoints are those
implemented in EPA’s BMDS.  These models were the power models, the
Hill models, and the polynomial models, including the linear model.  The
BMDS methods and models applied to continuous endpoints are presented in
Section   REF _Ref127940647 \r \h  5 .

Definition of the Benchmark Response (BMR) and Corresponding BMD and
BMDL

BMDs were implicitly defined as follows:

μ(BMD) - μ(0)	=	0.1	(Eq. 1)

μ(0)



	

In other words, the BMR was defined as a 10 percent change in mean. 
BMDLs were defined as the 95 percent lower bound on the corresponding
BMD.  Confidence intervals were calculated using a profile likelihood
method.  

Dr. George Daston, a member of the expert panel, recommended that the 10
percent difference in mean for this endpoint not be used as a BMR
because of its difference from a typical 10 percent change in the
probability of a response for a quantal variable (ICF 2004).  It was
stated that the 10 percent change in the mean number of estrous cycles
might still be within the range of normal values for this endpoint in
female rats.  Dr. Daston further suggested that, ideally, ICF should
find a scientific consensus on the normal range for this value, if such
agreement exists.  However, no scientific consensus exists with regard
to this value at present.  Solicitation was not pursued given that such
a discussion among reproductive experts would likely not yield a
definitive answer.  It would, of necessity, rely on archives of
historical data and would therefore be a major science policy
undertaking.  Further, research areas for a few laboratories indicate
that measuring estrous cycles in various animal species is ongoing (IZW
2005).  These points illustrate that data are not readily available and
would require a significant research effort to further investigate. 
Finally, the cycle ranges of the control rats in the nPB studies cited
are more relevant for comparison than what is considered to be the
normal range for all rats.  As Dr. Luderer discussed, estrous cycle
lengths at the 250, 500, and 750 ppm doses were outside the normal range
found in the WIL (2001) study (see discussion below).  In the absence of
an agreed normal range, Dr. Daston listed several other options for
determining a biologically significant change, such as using the mean
response equal to 0.5 control standard deviations from the control mean
(ICF 2004).  However, these options were not consistent with EPA
guidance, which recommends the use of one standard deviation, and were
thus not used.

Alternate approaches, based on data variability, were considered and
discarded.  There are several reasons why these alternate approaches
were not used and why ICF is not basing the AEL on a BMDL value based on
one standard deviation of variability.  First, the number of estrous
cycles is not truly a continuous variable, but instead is a categorical
variable with whole number values.  It does not neatly fit into the two
types of variables addressed by EPA’s Benchmark Dose Technical
Guidelines: quantal and continuous.  Secondly, the control data failed
tests for normality of distribution (Shapiro-Wilk, Anderson-Darling,
Cramer–von Mises, and Kolmogorov-Smirnov; p<0.05).  This suggested
that the standard formula used to calculate the standard deviation,
which is based upon assumptions of normality, would not be a good (i.e.,
statistically efficient) estimate of the true standard deviation of the
data.  Rather than attempt to calculate a more precise standard
deviation estimate based upon some alternate distribution function for
the data, which also would have been inconsistent with the normal
approximation used for the dose-response and BMD modeling, the choice
was made to base the BMR upon a 10 percent relative change. 

The 10 percent response is scientifically justified for several reasons.
 The dose-response curve in the WIL study indicates that the number of
estrous cycles decreases in F0 females with increasing dose (  REF
_Ref127936870 \h  Table A.2 ).  The differences from controls are
statistically significant at 250, 500 and 750 ppm.  The 500-ppm dose
included one acyclic female.  At 750 ppm the rats were completely
incapable of producing an F1 generation.  Therefore, the dose-response
curve is very steep from animals with normal estrous cycle lengths at
the low dose to ones that cannot produce any offspring at all at the
high dose.  

These data indicate that the number of estrous cycles is a sensitive
early indicator of reproductive success.  The decrease in the mean
number of estrous cycles at 250 ppm was 11 percent, which is slightly
higher than the 10 percent change modeled using BMD according to Eq. 1
above.  The next higher dose, 500 ppm, began to produce acyclic females.
 Thus, a dose somewhere between 100 and 250 ppm is the maximum one that
will cause a decrease in the number of estrous cycles without disrupting
the ovarian cycle of any of the animals in the study.  Therefore, a 10
percent difference in the mean number of estrous cycles in a 3-week
period is a protective endpoint to calculate, because all of the female
animals would still exhibit a normal range of cyclicity (e.g.,
comparable to controls) and reproduction should not be impaired.  

Therefore, without additional information on the exact dose at which
this decrease in estrous cycles prevents estrous cycling altogether and
thus prevents reproduction for at least one animal, the 10 percent level
was chosen as appropriate.  It is appropriate because of its consistency
with BMD technical guidance and because the 10 percent level is a value
consistent with the BMDL levels chosen for other adverse outcomes
measured in the animals from the WIL study (e.g., sperm motility in
males and liver effects in males and females).  Consistent with the BMD
guidance (EPA 2000), the data have also been modeled as continuous,
using one standard deviation from the mean as an appropriate level of
change, and presented in the graph and BMDL using the linear model (the
model with the best fit) at the end of Section   REF _Ref127938264 \r \h
 4.3.1 .  The BMDL for this model was 208 ppm, which is greater than
that obtained with the models using a 10 percent change in the mean (see
analysis below).

Additional support for selection of this endpoint and the measure of
response came from Dr. Ulrike Luderer, of the peer review panel.  Dr.
Luderer indicated in her comments that there were four reasons why the
measured number of estrous cycles over a period of time was a valid
endpoint (ICF 2004):

There is a clear dose-response relationship of decreasing estrous cycles
with increasing dose;

The alterations in estrous cycle length fall outside the historic range
of estrous cycle duration for the laboratory that conducted one of the
studies (Stump 2001 [sic] refers to WIL 2001 report);

Alterations in estrous cycle length are not isolated findings, but occur
in the context of other effects on the female reproductive system; and,

Effects on estrous cycling were observed in two independent studies of
nPB exposure (Stump 2001; Yamada et al. 2003).

These arguments from Dr. Luderer provide significant support for ICF’s
approach.

Benchmark Dose-Response Modeling Results and Choice of BMDL

The results of the benchmark modeling are presented in   REF
_Ref127932261 \h  Table A.4 .  Background material regarding the models
and software used, as well as an extensive explanation of the decision
tree (below) and the criteria for deciding which models provided the
best fit, are provided in Section   REF _Ref127938309 \r \h  5 .  The
following guidelines were used in the selection of BMDLs for each data
set:  

Models with an unacceptable fit (including consideration of local fit in
the low-dose region) were excluded.  Visual fit, particularly in the
low-dose region, was assessed for models that had acceptable global
goodness of fit.

If the BMDL values for the remaining models for a given endpoint were
within a factor of three, no model dependence was assumed, and the
models were considered indistinguishable in the context of the precision
of the methods.  The models were then ranked according to the AIC, and
the model with the lowest AIC was chosen as the basis for the BMDL. 

If the BMDL values were not within a factor of three, some model
dependence was assumed, and the lowest BMDL was selected as a reasonable
conservative estimate, unless it was an outlier compared to the results
from all of the other models.  Note that when outliers are removed, the
remaining BMDLs may then be within a factor of three, and so the
criteria given in item two would be applied.

 These criteria were applied to the BMDLs reported in   REF
_Ref127932261 \h  Table A.4  for each endpoint. 

Table A.  SEQ Table_A. \* ARABIC  4 : Comparison of BMD Modeling Results
and Final Decision

Model-Variance Model	AIC	P-Value	BMD	BMDL	Sufficient Data?	Good Visual
Fit in the Low-Dose Portion of the DR?	BMDLs Within a Factor of 3?
Lowest AIC?

Linear-Homogeneous	98	0.88	201	168	Y	Y	Y	N

Linear-Heterogeneous1	75	0.42	200	162	Y	Y	Y	Y

Polynomial (2)-Homogeneous	100	0.75	178	109	Y	Y	Y	N

Polynomial (2)-Heterogeneous	77	0.29	176	122	Y	Y	Y	N

Polynomial (3)-Homogeneous	102	1.00	232	102	Y	Y	Y	N

Polynomial (3)-Heterogeneous	76	0.64	234	146	Y	Y	Y	N

Power-Homogeneous	102	0.43	201	168	Y	Y	Y	N

Power-Heterogeneous	77	0.25	200	162	Y	Y	Y	N

Hill-Homogeneous	104	0.84	233	103	Y	Y	Y	N

Hill-Heterogeneous	77	0.41	249	143	Y	Y	Y	N

1 Model that best fits the data; these results were used to derive the
AEL

Reference: WIL Study 2000

Decision Tree Based Upon EPA Benchmark Dose Technical Guidance Document 

 using a value of α =0.1 to determine a critical value using the Chi
Square test.

All models pass.

Further reject models that apparently do not adequately describe the
relevant low-dose portion of the dose-response, examining residuals and
graphs of model and data.

All models pass.  Graphs of the model fits with heterogeneous variance
are included in   REF _Ref127932517 \h  Figure A.1 .  As dose increases
the number of estrous cycles goes down and therefore the dose response
curve is downward sloping.

As the models remaining have met the default statistical criteria for
adequacy and visually fit the data, any of them theoretically could be
used for determining the BMDL.  The remaining criteria for selecting the
BMDL are adopted as defaults.

If the BMDL estimates from the remaining models are within a factor of
three of each other, then they are considered to show no appreciable
model dependence and will be considered indistinguishable in the context
of the precision of the methods. 

All BMDL estimates within a factor of three.

Models are ranked based on the values of their Akaike Information
Criterion (AIC), a measure of the deviance of the model fit adjusted for
the degrees of freedom, and the model with the lowest AIC is used to
calculate the BMDL.  If this is not unique, the simple average or
geometric mean of the BMDLs with the lowest AIC is used.

Linear model with heterogeneous variance has lowest AIC (75).  BMDL from
this model (162 ppm) is used.

The uncertainty analysis associated with the selection of the models, as
recommended by the EPA’s Benchmark Dose Technical Guidance, is
represented in   REF _Ref127932389 \h  Table A.5  below.  This guidance
addresses two types of variables: quantal and continuous.  The chosen
endpoint, the number of estrous cycles, does not represent either type
as it is not truly a continuous variable, but instead is a categorical
variable with whole number values.  Therefore, the choice was made to
base the BMR upon a 10 percent relative change as discussed above in
Section   REF _Ref127939852 \r \h  \* MERGEFORMAT  4.2 .

Table A.  SEQ Table_A. \* ARABIC  5 :  Uncertainty Analysis Recommended
for Each Type of Data

Type of Data	Uncertainty Analysis Recommended by Benchmark Does
Technical Guidance 

	Percent Change	One Standard Deviationc	Percent Change and Compare to
Standard Deviation

Quantal (Dichotomous) Data

Sensitive studies (e.g., reproductive, developmental, epidemiology)a	(



All other quantal datab	(



Continuous Data

Minimal level of change in significant endpoint

	(

Level of adverse response known

	(

Level of adverse response unknown

(

	Type of Data	Uncertainty Analysis Recommendations Not Covered by
Benchmark Dose Technical Guidance

Categorical variables with whole number valuesd

	(

a Reproductive and developmental studies typically use five percent
change uncertainty, and epidemiology studies typically use one percent
change.

b All other quantal data typically uses 10 percent change uncertainty.

c An uncertainty of one standard deviation is generally only applied to
normally distributed data.  Source: EPA. 2000a.  Benchmark Dose
Technical Guidance Document.  EPA/630/R-00/001.  External Review Draft. 
Risk Assessment Forum, Washington, DC.  October 2000.

d In the absence of EPA guidance on how to assign uncertainty to
categorical, whole number data, an expert decision was made to base the
BMR upon a 10 percent relative change and compare these results to one
standard deviation uncertainty, as discussed above in Section   REF
_Ref127939872 \r \h  \* MERGEFORMAT  4.2 .

Figure A.  SEQ Figure_A. \* ARABIC  1 : Graphs of Heterogeneous Model
Fits to Data

Dose is measured in parts per million and response is measured as the
mean number of estrous cycles in 3 weeks.  As dose increases the number
of estrous cycles every 3-weeks goes down and therefore the dose
response curve is downward sloping.

 

Figure A.  SEQ Figure_A. \* ARABIC  2 :  Graph of Linear Model Fit to
Data Using One Standard Deviation

Dose = Parts per million

Response = Mean number of estrous cycles in 3 weeks

 

Discussion

From the goodness-of-fit summary statistics and the visual plots, the
two best candidates are the linear and quadratic model for the mean with
a heterogeneous model for the variance.  They have similar AIC values,
although the linear model is preferred on the basis of having a lower
AIC.  In general, if a simpler model gives as good a fit or a better fit
to the data than a more complicated model, then the simpler model is
pref牥敲⹤†潂桴洠摯汥⁳獡畳敭琠敨瘠牡慩据⁥獩愠
洠汵楴汰⁥愨灬慨 景琠敨洠慥⁮慲獩摥琠⁯桴⁥潰敷
⁲桲Ɐ眠敨敲愠灬慨愠摮爠潨愠敲攠瑳浩瑡摥映潲⁭桴
⁥慤慴‮吠敨氠湩慥⁲潭敤⁬潦⁲桴⁥敭湡愠獳浵獥琠
慨⁴桴⁥敭湡渠浵敢⁲景攠瑳潲獵挠捹敬⁳獩漠⁦桴⁥
潦浲ഠ

Effect = β0 + β1dose

The quadratic model for the mean assumes that the mean number of estrous
cycles is of the form 

Effect = β0 + β1dose + β2dose2

There are several statistical tests that can be used to compare the
linear and quadratic models, all leading to the same conclusion that the
linear model is preferred:

The AIC comparison compares the AIC values for the two models, where AIC
equals -2L + 2p, L is the log-likelihood, and p is the number of unknown
parameters.  Models with too many parameters are penalized by the AIC
statistic.  The better fitting models have the lower AIC values so that
the linear model, with AIC = 75, is preferred to the quadratic model,
with AIC = 77.  Note that the size of the AIC itself does not have a
statistical interpretation; only differences between AIC values are
meaningful.  Note also that there is no general statistical procedure
for comparing two AICs to see if the models are statistically
significantly different.

β2 equals zero, a standard statistical test is used to examine the
quadratic coefficient and test whether it is statistically significantly
different from zero.  Because the Wald 95 percent confidence interval
for β2 ranges from -2.0E-6 to +3.3E-6, which includes zero, the
quadratic coefficient is not statistically significantly different from
zero.  Thus the more complicated quadratic model does not show a
significant improvement over the linear model and so the linear model is
preferred.

β2 equals zero, another standard statistical test is a chi-square test
based on twice the difference in log-likelihoods, -2 (L, as presented in
Section   REF _Ref127932574 \r \h  4.2 .  Since -2 (L = 0 (to one
decimal place), which is less than the critical value 3.84, the 95th
percentile of a chi-square distribution with one degree of freedom, the
chi-square test is not statistically significant.  Again, the more
complicated quadratic model does not show a significant improvement over
the linear model and so the linear model is preferred. 

By adding an extra parameter that is not statistically significant, the
other parameters are less precisely estimated, so that the predicted
values for the quadratic model are less precise than for the linear
model.  Although the quadratic model has a lower BMD, this predicted
dose value is more uncertain than for the linear model.  This is
reflected in a wider confidence interval for the BMD for the quadratic
model. 

Background on nPB AEL Benchmark Dose Analysis

BMDS Models

The models used to represent the dose-response behavior of those
continuous endpoints are those implemented in EPA’s BMDS.  These
models were the power models, the Hill models, and the polynomial
models, including the linear model.  These mathematical models fit to
the data are defined here.  In all cases, μ(d) indicates the mean of
the response variable following exposure to “dose” d.

The power model is represented by the equation

		μ(d) = γ + βdα

where the parameter α >0. 

The Hill model is given by the following equation:

		μ(d) =  γ + (vdn) / (dn + kn)

where the parameter k is greater than 0 and n is greater than 1 (USEPA
2000b).The polynomial model is defined as:

		μ(d) = β0 + β1d + ... + βndn

where the degree of the polynomial, n, was set to less than the number
of dose groups in the experiment being analyzed.  The linear model is a
special case of the polynomial model where n=1.

In the case of continuous endpoints, one must assume something about the
distribution of individual observations around the dose-specific mean
values defined by the above models.  The assumptions imposed by BMDS
were used in this analysis, i.e., individual observations were assumed
to vary normally around the means with heterogeneous variances given by
the following equation:

		σi2 = α[μ(di)]ρ

where both α and ρ were parameters estimated by the model.  Also
fitted were homogeneous variance models where ρ=0.  As discussed above,
the data used for these analyses were discrete, integer-valued, and the
normal distribution was used as an approximation to their distribution. 
The validity of this approximation to the joint probability distribution
uses the central limit theorem of statistics. 

Given the above assumptions about variations around the means, maximum
likelihood methods, were applied to estimate all of the parameters,
where the log-likelihood to be maximized is (except for an additive
constant) given by

		L = Σ [(Ni/2) ln(σi2) + (Ni - 1)si2/2σi2 + Ni{mi - μ(di)}2/2σi2]

where Ni is the number of individuals in group i exposed to dose di, and
mi and si are the observed mean and standard deviation for that group,
respectively.  The summation runs over i from 1 to k (the number of dose
groups).

Goodness-of-Fit Analysis

The BMDS software provides three or four different Tests of Fit that the
user may use to determine an appropriate model for fitting data.  These
Tests of Fit are based on asymptotic theories of the likelihood ratio. 
The likelihood ratio represents the ratio of two likelihood values, many
of which are given in the BMDS output.  Statistical theory proves that
-2*log(likelihood ratio) converges to a Chi-Square random variable as
the sample size gets large and the number of dose levels gets large. 
These values can in turn be used to obtain approximate probabilities to
make decisions about model fit.

Each of the ten fitted models has a likelihood value.  The BMDS program
uses these values to create ratios from two models that form a
meaningful test.  Suppose the user wishes to test two models for fit, A
and B.  One assumption that is made for these tests is that the "true"
model is in fact B, but it can be simplified in such a way that the
simplified model describes the data as well as B.  Also suppose A is a
much simpler model in that it has much fewer parameter values (the goal
is to simplify the model as much as possible without losing information
about the data).  Assume each model has a maximum likelihood value, call
them L(A) and L(B).  A ratio can be formulated easily:  L(A)/L(B). 
(Note: The model with a higher number of parameters is always in the
denominator of this ratio).  Now, using the theory, -2*log{ L(A)/L(B)}
approaches a Chi-Square random variable.  This can be simplified by
using the fact that the log of a ratio is equal to the difference of the
logs, or simply put, -2*log{ L(A)/L(B)} = -2*( log{L(A)} - log{L(B)} ) =
2*log{L(B)} - 2*log{L(A)}.  The likelihood values given by BMDS are in
fact the log likelihoods, therefore this becomes a subtraction problem. 
This value can then in turn be compared to a Chi-Square random variable
with a specified number of degrees of freedom.

Each log likelihood value has an associated number of degrees of
freedom.  The number of degrees of freedom for the Chi-Square test
statistic is merely the difference between the two model degrees of
freedom.  In the mini-example above, suppose L(A) has 5 degrees of
freedom, and L(B) has 8.  In this case, the Chi-Square value you would
compare this to would be a Chi-Square with 8 - 5 = 3 degrees of freedom.

In the A vs B example, what is exactly being tested?  In terms of
hypotheses, it would be:

H0:  A models the data as well as B

	H1:  B models the data better than A

Keeping these tests in mind, suppose 2*log{L(B)} - 2*log{L(A)} = 4.89
based on 3 degrees of freedom.  Also, suppose the rejection criterion is
a Chi-Square probability of less than 0.10.  Looking on a Chi-Square
table, 4.89 has a p-value somewhere between 0.10 and 0.25.  In this
case, H0 would not be rejected, and it would seem to be appropriate to
model the data using Model A.

The BMDS software provides three or four default tests, depending on the
variance model the user has specified (constant variance model, or a
non-constant variance model where the variance is a function of the
mean, namely, 

(i2=((i(

BMDS assumes the rejection criterion is a Chi-Square probability of less
than 0.10 for all of the tests; however p values are presented so that
the user is free to use any rejection criteria.  Each test in each model
will be discussed in some detail below. 

Test 1:  Tests the hypothesis that response and variance don't differ
among dose levels.  If this test is not rejected, there may not be a
dose-response relationship, although it is possible for some data sets
with a slightly significant trend to not reject this test.  This model
implies no differences in the mean or in the variance at each dose
level, and thus, there would be no adverse effect as dosage is
increased.  If this test is rejected, then modeling the data is
appropriate, and the user should consider the tests below.

Test 2:  Tests the hypothesis that variances are homogeneous.  Recall
that the goal is to simplify the model.  If this test is not rejected,
the simpler constant variance model may be appropriate.  If this test is
rejected, the user may want to run a non-constant variance model, or if
the non-constant variance model was run, then the user should look at
the second test 3 below to make further decisions.

Test 3 (Test 4 is a test of the variance model):  Tests the hypothesis
that the model for the mean fits the data.  If this test is not
rejected, the user has support for the selected model.  If this test is
rejected, the user may want to try a different model.

Test 4 (Non-constant variance model):  Tests the hypothesis that the
variances are adequately modeled.  Here, the test is to see whether or
not the variance model, (i2=((i(, is an appropriate assumption.  Again,
the purpose is to reduce the parameter space, and by modeling the
variances as a function of the mean (which also intuitively makes sense
that variance may have some dependence on the mean value) we achieve
some reduction.  If this test is not rejected, it may be appropriate to
conclude that the true variances have the form above.  If this test is
rejected, BMDS has no further way to model variance.

Visual fit, particularly in the low-dose region, was assessed for models
that had an acceptable global goodness of fit.  Acceptable global
goodness of fit was either a p-value > 0.1, or a perfect fit when there
were no degrees of freedom for a statistical test of fit.  Local fit was
evaluated visually on the graphic output, by comparing the observed and
estimated results at each data point.

Goodness-of-fit statistics are not designed to compare different models,
particularly if the different models have different numbers of
parameters.  Within a family of models, adding parameters generally
improves the fit.  BMDS reports the Akaike Information Criterion (AIC)
to aid in comparing the fit of different models.  The AIC is defined as
–2L+2p, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and p is the number of model parameters
estimated.  When comparing the fit of two or more models to a single
data set, the model with the lesser AIC was considered to provide a
superior fit. 

Selected Output from BMDS

 ==================================================================== 

   	  Polynomial Model. Revision: 2.2  Date: 9/12/2002 

  	  Input Data File: C:\BMDS\DATA\NPB-EST.(d)  

  	  Gnuplot Plotting File:  C:\BMDS\DATA\NPB-EST.plt

 							Wed Jun 29 17:09:10 2005

 ==================================================================== 

 BMDS MODEL RUN 

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

   The form of the response function is: 

   Y[dose] = beta_0 + beta_1*dose + beta_2*dose^2 + ...

   Dependent variable = MEAN

   Independent variable = dose

   Signs of the polynomial coefficients are not restricted

   The variance is to be modeled as Var(i) = alpha*mean(i)^rho

   Total number of dose groups = 5

   Total number of records with missing values = 0

   Maximum number of iterations = 250

   Relative Function Convergence has been set to: 1e-008

   Parameter Convergence has been set to: 1e-008

                  Default Initial Parameter Values  

                          alpha =     0.811333

                            rho =            0

                         beta_0 =      3.98617

                         beta_1 =  -0.00198177

                                 Parameter Estimates

                                                         95.0% Wald
Confidence Interval

       Variable         Estimate        Std. Err.     Lower Conf. Limit 
 Upper Conf. Limit

          alpha          73.6006          69.1595            -61.9495   
         209.151

            rho         -3.98394         0.778685            -5.51013   
        -2.45774

         beta_0           3.9907        0.0841431             3.82579   
         4.15562

         beta_1      -0.00199298      0.000310492         -0.00260154   
     -0.00138443

           Asymptotic Correlation Matrix of Parameter Estimates

                  alpha          rho       beta_0       beta_1

     alpha            1        -0.99      0.00051       0.0061

       rho        -0.99            1      -0.0012       -0.005

    beta_0      0.00051      -0.0012            1        -0.65

    beta_1       0.0061       -0.005        -0.65            1

     Table of Data and Estimated Values of Interest

 Dose       N    Obs Mean    Obs Std Dev   Est Mean   Est Std Dev  
Chi^2

Res.

------     ---   --------    -----------   --------   -----------  
----------

    0    25       3.96        0.539         3.99        0.545        
-0.282

  100    25       3.84        0.624         3.79        0.603         
0.403

  250    25       3.52        0.653         3.49         0.71         
0.194

  500    25       2.88         1.17         2.99        0.965        
-0.592

  750    25       2.56         1.26          2.5         1.39         
0.231

 Model Descriptions for likelihoods calculated

 Model A1:        Yij = Mu(i) + e(ij)

           Var{e(ij)} = Sigma^2

 Model A2:        Yij = Mu(i) + e(ij)

           Var{e(ij)} = Sigma(i)^2

 Model A3:        Yij = Mu(i) + e(ij)

           Var{e(ij)} = alpha*(Mu(i))^rho

 Model  R:         Yi = Mu + e(i)

            Var{e(i)} = Sigma^2

                       Likelihoods of Interest

            Model      Log(likelihood)   DF        AIC

             A1          -46.881357       6     105.762714

             A2          -31.698420      10      83.396841

             A3          -32.027970       7      78.055939

           fitted        -33.432720       4      74.865440

              R          -67.085741       2     138.171481

                   Explanation of Tests  

 Test 1:  Does response and/or variances differ among Dose

levels? 

          (A2 vs. R)

 Test 2:  Are Variances Homogeneous? (A1 vs A2)

 Test 3:  Are variances adequately modeled? (A2 vs. A3)

 Test 4:  Does the Model for the Mean Fit? (A3 vs. fitted)

                     Tests of Interest    

   Test    -2*log(Likelihood Ratio)  Test df        p-value    

   Test 1              70.7746          8          <.0001

   Test 2              30.3659          4          <.0001

   Test 3             0.659098          3          0.8828

   Test 4               2.8095          3          0.4219

The p-value for Test 1 is less than .05.  There appears to be a

difference between response and/or variances among the dose levels

It seems appropriate to model the data

The p-value for Test 2 is less than .05.  A non-homogeneous variance
model appears to be appropriate

The p-value for Test 3 is greater than .05.  The modeled variance
appears 

 to be appropriate here

The p-value for Test 4 is greater than .05.  The model chosen seems to
adequately describe the data

 

 Benchmark Dose Computation

Specified effect =           0.1

Risk Type        =     Relative risk 

Confidence level =          0.95

             BMD =       200.238

            BMDL =       162.391

 

BMDL computation failed for one or more point on the BMDL curve. 

 

			The BMDL curve will not be plotted

References

Birgfeld E. 2004.  Conversation with Sally Darney and Ralph Cooper, May
27, 2004.

CERHR.  2002.  NTP-CERHR Expert Panel Report on the Reproductive and
Developmental Toxicity of 1-Bromopropane.  National Toxicology Program. 
NTP-CERHR-1-BP-02.  March 2002.

EPA. 2001.  Help Manual for Benchmark Dose Software.  Office of Research
and Development, Environmental Protection Agency.  Washington, DC.  EPA
600/R-00/014F.

EPA.  2000a.  Benchmark Dose Software.  National Center for
Environmental Assessment with technical support from the National Health
& Environmental Effects Research Laboratory.  

www.epa.gov.ncea/bmds.htm

EPA. 2000b.  Benchmark Dose Technical Guidance Document. 
EPA/630/R-00/001.  External Review Draft.  Risk Assessment Form,
Washington, DC.  October 2000.

EPA.  1996.  Guidelines for Reproductive Toxicity Risk Assessment.  U.S.
Environmental Protection Agency, Risk Assessment Forum, Washington, DC,
630/R_96/009, 1996. 

ICF Consulting. 2004.  nPB Expert Panel Response Final Document. 
Delivered to the U.S. Environmental Protection Agency on December 13,
2004.  

ICF Consulting. 2002.  Acceptable Industrial Exposure Limit for n-Propyl
Bromide.  Delivered to the U.S. Environmental Protection Agency on May
3, 2002.  

Ichihara et al. 2004a.  A Survey on Exposure Level, Health Status, and
Biomarkers in Workers Exposed to 1-Bromopropane.  Am. J. Ind. Med.
45:63–75.  Wiley-Liss, Inc. 2004.  

Ichihara et al. 2004b.  Neurological Abnormalities in Workers of
1-Bromopropane Factory, Environmental Health Perspectives published by
the National Institute of Environmental Health Sciences, National
Institutes of Health, U.S. Department of Health and Human Services, June
2004.

Ichihara et al. 2002 Neurological disorders in three workers exposed to
1-bromopropane.  Journal of Occupational Health 44:1-7 (2002).

Ichihara G., Jong X., Onizuka J., et al. 1999.  Histopathological
changes of nervous system and reproductive organ and blood biochemical
findings in rats exposed to 1-bromopropane.  Abstracts of the 72nd
Annual Meeting of Japan Society for Occupational Health.  May 1999. 
Tokyo.

IZW. 2005.  Institute for Zoo and Wildlife Research. “Comprehensive
Science: Non-invasive Monitoring of Hormones.  Accessed on the World
Wide Web on May 31, 2005 at   HYPERLINK
"http://www.izw-berlin.de/en/research/fg4/index.html?reproduktionsmonito
ring.html~rechts" \o
"http://www.izw-berlin.de/en/research/fg4/index.html?reproduktionsmonito
ring.html~rechts" 
http://www.izw-berlin.de/en/research/fg4/index.html?reproduktionsmonitor
ing.html~rechts .  

National Research Council (NRC). 2001 Evaluating Chemical and Other
Agent Exposures for Reproductive and Developmental Toxicity, Commission
on Life Sciences (  HYPERLINK "http://www7.nationalacademies.org/dels/" 
CLS ), NATIONAL ACADEMY PRESS, Washington, D.C. 2001.

Sekiguchi S, Suda M, Zhai YL, Honma T. 2002.  Effects of 1-bromopropane,
2-bromopropane, and 1,2-dichloropropane on the estrous cycle and
ovulation in F344 rats.  Toxicol. Lett.  126(1):41-9.

Yamada T, Ichihara G, Wang H, et al. 2003.  Exposure to 1-bromopropane
causes ovarian dysfunction in rats.  Toxicol. Sci. 71:96-103.

WIL. 2001.  An Inhalation Two-Generation Reproductive Toxicity Study of
1-Bromopropane in Rats.  Conducted by Stump, D.G., at WIL Research
Laboratories, Inc., Sponsored by Brominated Solvents Consortium.  May
24, 2001.

Attachment   SEQ Attachment \* ALPHABETIC  B 

Derivation of a Reference Concentration for n-Propyl Bromide

Reference Concentration Derivation

EPA has critically evaluated the literature on n-propyl bromide (nPB)
for the purpose of evaluating its toxicity in animal models following
inhalation exposure.  From the available data, EPA has selected the most
relevant toxicity study to inform exposure limits for both occupational
and general exposures.  This attachment concerns the development of an
exposure limit for the general population.  The study selected to inform
this limit is a two-generation inhalation study in rats (WIL 2001).  EPA
derived an estimated reference concentration (RfC) for nPB as a
screening tool to assess risk to the general population including
sensitive individuals such as children and the elderly.

Recommended RfC: 	1 ppm 	

Basis:

Endpoints: 	A decrease in the number of estrous cycles in a 3-week
period prior to mating following 7 weeks of nPB exposure

Study: 	An Inhalation Two-Generation Reproductive Toxicity Study of
1-Bromopropane in Rats (WIL 2001)

Protocol:	Whole-body inhalation, 6 hours/day, 7 days/week for 70 days
prior to mating, during mating, gestation, lactation for two generations

Concentrations:		0, 100, 250, 500, or 750 ppm

BMDL:			162 ppm (mean number of estrous cycles in 3 weeks)

NOAEL:		100 ppm

LOAEL:		250 ppm (estrous cycle, sperm motility and hepatic effects)

BMDL [adj]:  		(162 ppm × 6 hours / 24 hours = 40 ppm)

BMDL [HEC]: 		40 ppm

Uncertainty/

Modifying Factors:	3 - animal to human extrapolation (pharmacodynamic
differences) 

10 - Sensitive individuals

RfC estrous cycle effects: 	1 ppm

An estimated RfC for nPB was derived using the results of the benchmark
dose modeling that was conducted on the data set from the WIL (2001)
study with nPB.  The summary results of the benchmark modeling are
presented in   REF _Ref127940825 \h  Table B.1 .  As shown in   REF
_Ref127940825 \h  Table B.1 , following the application of the selection
criteria, BMDLs ranging from 102 to 168 ppm were identified.  For a full
discussion on the development and selection of these BMDL values see  
REF _Ref127940782 \h  Attachment A .

Table B.  SEQ Table_B. \* ARABIC  1 : Comparison of BMD Modeling Results
and Final Decision

Model-Variance Model	AIC	P-Value	BMD	BMDL

Linear-Homogeneous	98	0.88	201	168

Linear-Heterogeneous1	75	0.42	200	162

Polynomial (2)-Homogeneous	100	0.75	178	109

Polynomial (2)-Heterogeneous	77	0.29	176	122

Polynomial (3)-Homogeneous	102	1.00	232	102

Polynomial (3)-Heterogeneous	76	0.64	234	146

Power-Homogeneous	102	0.43	201	168

Power-Heterogeneous	77	0.25	200	162

Hill-Homogeneous	104	0.84	233	103

Hill-Heterogeneous	77	0.41	249	143

1 Model that best fits the data; these results were used to derive the
AEL and estimated RfC (WIL 2001).

The estimated RfC was derived based on a decrease in the number of
estrous cycles measured in a 3-week period prior to mating following 7
weeks of nPB exposure in the WIL (2001) study.  This endpoint was chosen
because it is considered to be the most sensitive endpoint with the
lowest BMDL.  The BMDL of 162 ppm was selected because the
linear-heterogeneous model best fit the data and had the lowest AIC
value.  In the derivation of an estimated RfC for this endpoint, the
BMDL was adjusted for continuous exposure as discussed above, resulting
in a HEC of 40 ppm.

An uncertainty factor of 3 was applied for differences in
pharmacodynamics.  An uncertainty factor of 10 was applied for the
protection of sensitive individuals (intraspecies extrapolation). 
Uncertainty factors of up to a value of 10 are typically applied when
developing an RfC value unless there are data indicating that lower
values may be more appropriate.  The application of an UF of 10 for the
protection of sensitive individuals differs from the application of an
UF of 3 in the development of an AEL.  The occupational environment is
populated by individuals that represent healthy adults, excluding
children and the elderly.  The UF of 10 for the general population is
designed to be protective of the reproductive health of a broader range
of individuals.  Further, exposures in an occupational environment are
limited, whereas exposures are assumed to occur continuously in the
general population (for whom the RfC value is developed).  Therefore, an
overall uncertainty factor of 30 results (3 for differences in
pharmacodynamics and 10 for the protection of sensitive individuals). 
The application of the overall uncertainty factor to the HEC (40 ppm)
results in an estimated RfC of 1 ppm.

References

WIL. 2001.  An Inhalation Two-Generation Reproductive Toxicity Study of
1-Bromopropane in Rats.  Conducted by Stump D.G. at WIL Research
Laboratories, Inc., Sponsored by Brominated Solvents Consortium.  May
24, 2001.

Attachment   SEQ Attachment \* ALPHABETIC  C 

Evaluation of the Global Warming Potential for n-Propyl Bromide

Summary

Based on an analysis of a cooperative research effort by Atmospheric and
Environmental Research, Inc. (AER) and the Center for Chemical and
Environmental Physics at Aerodyne Research, Inc., and estimates on the
reaction products of n-propyl bromide (nPB), the global warming
potential (GWP) of nPB was estimated to be approximately 1.57.  However,
the value could be slightly lower—depending on the amount of rainout
removal of nPB reaction products containing carbon—or slightly higher,
due to radiative forcing of the reaction products.

Discussion of Results

The website for the Albemarle Corporation presents a GWP, Halocarbon GWP
(HGWP), atmospheric lifetime (ALT), and ozone depletion potential (ODP)
for nPB (  HYPERLINK "http://www.albemarle.com/abzregstatus.htm" 
http://www.albemarle.com/abzregstatus.htm ).  The calculations were
carried out in a cooperative effort by AER and the Center for Chemical
and Environmental Physics at Aerodyne Research, Inc (AER 1995).

The 100-year GWP, which is presented 0.31, is calculated relative to
CO2.  The HGWP, which is presented as 0.0001, is calculated relative to
CFC-11, which has a HGWP of 1.0.  GWP calculations were made using
different integration time horizons.  By the HGWP method, CFC-11 is ten
thousand times more detrimental as a global warming agent than nPB.  By
the GWP method, it is fourteen thousand times worse than nPB, and nPB is
only one tenth as detrimental as CO2.

The analysis by AER and Aerodyne, which was conducted in 1995, appears
to be the basis for the GWP of 0.31 for nPB.  Other companies’
websites suggest the same GWP and HGWP values, although, in some cases,
the GWP is listed as zero.

While the AER/Aerodyne analysis was conducted specifically for
Albemarle, both AER and Aerodyne are highly respected research groups. 
It can be assumed that Aerodyne measured the infrared absorption spectra
for nPB, and AER used their radiative transfer model to evaluate the
radiative forcing and GWPs.

While it is expected that these estimates are reasonably accurate, the
listed ALT of nPB on the Albemarle website is 11 days, which is smaller
than the published ALT of approximately 19 days (Wuebbles et al. 2001). 
Nonetheless, the evaluation for the direct GWP for nPB of 0.31 is
accepted within a factor of two.  The direct GWP for nPB, therefore, is
better approximated as 0.5.  A reevaluation of the direct GWP using the
longer ALT could be undertaken to confirm the approximated value, but is
probably not necessary.

However, this direct GWP evaluation does not consider the reaction
products of nPB, particularly the three carbon atoms that, except for
any rainout of the reaction products containing carbon, eventually will
become CO2.  Modeling has not been used to determine how the rainout
process removes carbon-containing products (focus in those studies was
on bromine reaching the stratosphere).

Ignoring any rainout of reaction products containing carbon, the
reaction products will contribute an additional (indirect) GWP due to
the production of CO2 equal to about 1.07.  This value could be slightly
larger due to the radiative forcing of reaction products before the
final production of CO2 or somewhat less depending on the amount of
rainout removal of nPB reaction products containing carbon.  The
indirect GWP of 1.07 is based on the following calculation:

1kg of nPB	1000g of nPB	1mol of nPB	3mol of CO2	44g of CO2	1kg of CO2	=
1.07kg CO2

	1kg of nPB	123g of nPB	1mol of nPB	1mol of CO2	1000g of CO2

	

Adding the direct GWP of 0.5 and the indirect GWP of 1.07, the 100-year
integration GWP for nPB is estimated at 1.57.

References

AER.  1995.  Estimates of the Atmospheric Lifetime, Global Warming
Potential and Ozone Depletion Potential of n-Propyl Bromide. 
Atmospheric and Environmental Research, Inc.  Independent study prepared
for Albemarle Corporation.

Wuebbles, D. J., K. O. Patten, M. T. Johnson, and R. Kotamarthi. 2001.
“The new methodology for Ozone Depletion Potentials of short-lived
compounds: n-propyl bromide as an example.” J. Geophys. Res., 106,
14551-14571.

Attachment   SEQ Attachment \* ALPHABETIC  D 

Occupational Exposure Analysis for n-Propyl Bromide When Used As an
Adhesive

Introduction

n-propyl bromide (nPB) is used in various solvent, aerosol, and adhesive
formulations as an alternative to ozone-depleting substances (ODS) such
as methyl chloroform.  However, nPB has been shown to exhibit toxicity
primarily to the central and peripheral nervous, reproductive, and
hematopoietic systems following inhalation exposure.  To date, the
potential for worker exposure to nPB across the range of facilities
where it can be used has not been adequately assessed.  This attachment
attempts to provide a brief assessment of these exposures.

In general, nPB usage in adhesive applications is highly emissive. 
Significant occupational exposure can occur when a spray gun is used to
aerosolize the adhesive.  Exposure levels depend on multiple factors
including the ventilation in the room, the size of the room, the amount
of nPB being used, and proximity to the spray gun.  The aerosol mist is
present in the work area throughout the production process.  Since
workers typically spend 8 hours per day applying adhesive, this
assessment focuses on long-term occupational exposures to nPB as a
result of inhalation.  Potential impacts from higher exposure levels,
such as those that could occur from an accident or spill, are not
assessed in this study. 

This occupational exposure analysis consists of both an occupational
exposure modeling exercise in Section   REF _Ref127941073 \r \h  2  and
an analysis of several studies containing actual monitoring data from
adhesive facilities in Section   REF _Ref127941094 \r \h  3 .  Section  
REF _Ref127941099 \r \h  4  lists references used to create this
analysis. 

Occupational Exposure Modeling

Introduction

A box model was used to estimate concentrations of nPB that might be
present in the air of several hypothetical adhesive application
facilities.  In general, the setting in which nPB is applied varies
considerably depending on the size of the operation and the type of
application.  Adhesives are typically applied in either 1) an open-top
workbench spray area with side panels and some minor local ventilation,
or 2) an open room with no mist containment (i.e., supplemental
ventilation systems are assumed not to be present) (Swanson et al.
2002). 

The rest of this section is organized as follows: Section   REF
_Ref127941108 \r \h  2.2  presents the inputs used for the box model;
Section   REF _Ref127941110 \r \h  2.3  discusses the modeling approach;
Section   REF _Ref127941111 \r \h  2.4  presents and evaluates the
results of the analysis.

Model Inputs

The exposure assessment was conducted with the objective of determining
potential occupational exposure concentrations to nPB for a variety of
scenarios.  The model inputs were based on information provided in the
report prepared for EPA’s Design for the Environment Program,
“Alternative Adhesives Technologies: Foam Furniture and Bedding
Industries” (Swanson et al. 2002).  This report was prepared using
data gathered through facility site visits, publicly available chemical
data, and input from industry experts.

This exposure modeling analysis does not consider exposure reductions
associated with the use of personal protective equipment (PPE). 
Although certain types of protective equipment, such as safety glasses,
aprons, and caps, were described as being used at certain facilities
during the site visits conducted for the Swanson report (2002), these
types of safety equipment do not significantly reduce inhalation
exposures.  Additionally, it was found that no employees were wearing
gloves during the site visits, indicating that this is not a typical
practice in the adhesive industry.  While other types of protective
equipment, including self-contained breathing apparatus (SCBA), could
potentially reduce inhalation exposure, these or other similar devices
were not in use at the facilities under investigation and therefore are
not considered in this analysis.  

The following four emissions scenarios were considered in this analysis.
 These scenarios are presented in   REF _Ref127941227 \h  Table D.1 .

Emissions from a facility with average ventilation and average adhesive
use (S1);

Emissions from a facility with average ventilation and high adhesive use
(S2);

Emissions from a facility with poor ventilation and average adhesive use
(S3); and

Emissions from a facility with poor ventilation and high adhesive use
(S4).  

Table D.  SEQ Table_D. \* ARABIC  1 :  Scenarios Evaluated for nPB
8-Hour Exposure Concentration Analysis

Ventilation	Average Adhesive Use	High Adhesive Use

Average	Scenario 1 (S1)	Scenario 2 (S2)

Poor	Scenario 3 (S3)	Scenario 4 (S4)



The general characteristics used to model the four simulated facilities
for the bonding of furniture and mattresses are presented below:

Average Ventilation and Average Adhesive Use Facility (S1)

Average size facility: 65.7 meters by 65.7 meters;

Average height facility: 7 meters;

Typical lot size: 100 meters by 100 meters (~2.5 acres);

Urban setting;

Average-use adhesive mass emissions rate: 74 grams per minute;

Facility operates 2000 hours per year (40 hours per week, 50 weeks per
year);

Average breathing height: 1.8 meters; and

Air flow rate of 807 cubic meters per minute (28,500 cubic feet per
minute).

Average Ventilation and High Adhesive Use Facility (S2)

Same as Scenario 1 except for the adhesive use; and

High-use adhesive mass emissions rate: 739 grams per minute. 

Poor Ventilation and Average Adhesive Use Facility (S3)

Three-story (30 feet) house, with a footprint of 25 feet by 40 feet;

Urban setting;

Average-use adhesive mass emission rate: 74 grams per minute;

Facility operates 2000 hours per year (40 hour per week, 50 weeks per
year);

Average breathing height of 1.8 meters; and

Air flow rate of 192 cubic meters per minute (6780 cubic feet per
minute). 

Poor Ventilation and High Adhesive Use Facility (S4)

Same as Scenario 3 except for adhesive application rate; and

High-use adhesive mass emission rate: 739 grams per minute.

Scenario 4, although not representative of most facilities, is expected
to yield the highest exposure concentrations.  These concentrations are
useful for assessing the high-end exposure risk.  Comparing Scenarios 1
and 2 and Scenarios 3 and 4 will show how the quantity of adhesive used
affects exposure.  Comparing Scenarios 4 and 2 will show how ventilation
affects exposure.  Comparing Scenarios 4 and 1 will show the high and
low range of inhalation exposures and risks associated with variations
in adhesive use and facility parameters. 

Modeling Approach

The exposure point concentration for occupational inhalation exposure is
determined by calculating the workplace air chemical concentration.  The
indoor air concentration for nPB was estimated using the following
equation (EPA 1991):

Ca	=	Yv x 1000	mg	x	24.45



	g





AT x k

MW

Where:	

Ca	=	Concentration of the chemical in air (ppm)

Yv	=	Mass emission rate of volatile compound released (g/s)

AT	=	Air flow rate (m3/s)

k	=	Dimensionless room ventilation mixing coefficient (assumed as 0.5,
EPA 1991)

MW	=	Molecular weight (123.01 g/mole for nPB)



Note that 24.45 is a factor used to convert from mg/m3 to ppm.

This box model approach has been widely used for many years to estimate
probable exposures of workers to hazardous airborne materials and has
been described in detail by the National Institute for Occupational
Safety and Health (NIOSH 1999).

The mixing factor or turnover rate, k, accounts for the slow and
incomplete mixing of ventilation air with room air (i.e., the number of
times one volume of air in the room is replaced over a period of time—
typically one hour) and can be used to calculate an air flow rate if the
size of the room is known.  A value of 1 would represent complete mixing
throughout the room and a very small value of k would approach the
direct inhalation of the aerosolized nPB.  Note that this model assumes
that the entire area around the actual emissive source (i.e., around the
sprayer) contains the same nPB concentration.

The model implicitly assumes that chemicals are completely volatilized
into the air.  The assumption is appropriate for nPB application given
the volatility of nPB and the method of application (spray gun).

Other model assumptions include:

The incoming air is contaminant free;

The volume of air in the room is exchanged at a constant rate
(steady-state conditions); and

The worker is continuously exposed to the same steady-state
concentration of nPB during the 8-hour work day.

Results

The results of the analysis are outlined in   REF _Ref127941317 \h 
Table D.2 .  The exposure concentrations from each of the scenarios
modeled exceed the recommended 8-hour AEL of 17 ppm.  The exposure
concentration for Scenario 4 represents an upper range for the high-end
exposure.

Table D.  SEQ Table_D. \* ARABIC  2 : Eight-Hour Exposure Concentration
to nPB

Ventilation	Average Adhesive Use	High Adhesive Use

Average Ventilation	Scenario 1: 60.3 ppm	Scenario 2: 603 ppm

Poor Ventilation	Scenario 3: 253 ppm	Scenario 4: 2,533 ppm



The modeling and monitoring results suggest that in the absence of
supplemental ventilation systems (i.e., facilities with average or poor
ventilation), workers could be exposed to nPB at levels significantly
higher than the proposed AEL of 17 ppm.  NIOSH data from an actual
adhesive facility indicate that installation of additional ventilation
systems can reduce worker exposure.  Specifically, mean 8-hour TWA
employee exposure was reduced from 169 ppm to 19 ppm, or by almost 89
percent, after installation of new spray booths (NIOSH 2002a).  Despite
the drastic reduction in exposure caused by the ventilation systems, the
mean exposure to nPB still exceeded the AEL. 

Occupational Exposure Study Evaluations

Summary

Monitoring is sometimes performed by facilities to determine if, and to
what extent, a worker is exposed.  When available, monitoring data give
the best indication of the potential for exposure to workers.  It is
important to note, however, that the data reflect the workplace
conditions in which the monitoring occurred and are not representative
of conditions at all facilities.

Monitoring data for nPB exposure were available from three adhesive
application facilities.  These facilities are Custom Products, Inc., in
Morrisville, NC; STN Cushion Company in Thomasville, NC; and Marx
Industries in Sawmills, NC (NIOSH 2002a, 2002b, 2003).  The exposure
data were obtained in the immediate vicinity of the actual spray
application for long-term exposures.  Sections   REF _Ref127941364 \r \h
 3.2 ,   REF _Ref127941365 \r \h  3.3 , and   REF _Ref127941367 \r \h 
3.4  below describe the data obtained at each facility.  Section   REF
_Ref127941396 \r \h  3.5  summarizes the conclusions that can be drawn
about this data.  Note that throughout the rest of this section, data
for those workers actively spraying the nPB-containing adhesives are
presented in shaded cells, data for those workers exposed to
concentrations of nPB over the AEL are presented in bold text and data
for those workers exposed to nPB near the AEL are presented in italics. 

STN Cushion Company

STN Cushion Company assembles sofa cushions from several pieces of foam
for various furniture companies.  The foam pieces are glued together
using an nPB-containing adhesive applied with an air spray gun and then
hand-pressed together to set the bond.  The adhesive is sprayed in the
Fabrication room in spray stations with slotted local exhaust
ventilation (LEV) hoods.  There are also two rooms adjacent to the
Fabrication room—the Saw room and the Poly room—but adhesive is not
sprayed within these rooms.

Long-term exposure testing was done for employees working 8-hour shifts
at two different times.  The first sampling was done in November 2000. 
After this sampling was done, the spray tables were enclosed and turned
into spray booths and sampling was done again in July/August 2002.   
REF _Ref127941428 \h  Table D.3  below summarizes the results from the
air sampling before and after the spray stations were enclosed.

Table D.  SEQ Table_D. \* ARABIC  3 : nPB 8-Hour TWA Exposure
Concentration Results at STN Cushion Companya

Prior to Spray Station Enclosure	After Spray Station Enclosure

Job Title	8-hour 

(ppm)	Sample Location	8-hour

 (ppm)b

Sprayer	41.3	Station 2	15.4

Sprayer	143.0	Station 3	18.0

Sprayer	74.7	Station 4	19.1

Sprayer	29.4	Station 5	18.2

Sprayer	73.4	Station 6	27.9

Sprayer	48.6	Station 7	16.4

Sprayer	75.8	Station 8	15.2

Sprayer	78.3	Station 9	17.0

Sprayer	51.3	Station 10	17.0

Part Time	34.7	Station 11	25.8

Part Time	32.3	Station 12	20.2

Part Time	41.3



Floater	6.3



Floater	14.1



Average for Active Sprayers	68.4	Average for Active Sprayers	19.1

a Data in shaded cells for those workers actively spraying nPB.  Bold
values are above the AEL; italicized values are near the AEL.

b Average of 3 days of testing

As shown in   REF _Ref127941428 \h  Table D.3 , nPB exposures exceeded
the recommended AEL of 17 ppm in all but seven cases.  Prior to the
enclosure of the spray stations, only the exposures of the floaters
(i.e., workers moving throughout the facility during the day) were below
the recommended AEL.  After the enclosure of the spray stations, three
of the 11 samples were just below the AEL and two were at the AEL.  The
other six samples were still above the AEL.

Custom Products

Custom Products manufactures seat cushions for the commercial aircraft
industry, gluing together several pieces of foam and covering them with
a flame retardant and fabric covering.  An nPB-containing spray adhesive
is used to glue the foam pieces together and to seal the flame retardant
covering to the cushion.  The production process is divided into the
following four departments:

The Saw department, where the foam is cut;

The Assembly department, where the adhesive is sprayed by the Sprayers,
and the pieces are pressed together by-hand by the Assemblers;

The Sew department, where the flame retardant covering is produced; and

The Covers department, where Sprayers apply the adhesive to attach the
flame retardant covering.

Due to the proximity of all departments, all workers are exposed to nPB,
although the Assemblers and Sprayers are more directly exposed while the
workers in the Saw and Sew departments are only indirectly exposed.

Long-term exposure testing was done for employees working 8-hour shifts
at two different times.  The first sampling was done in November 1998. 
After this sampling was done, the spray booths were replaced with new
and improved spray booths and sampling was done again in November 2000. 
  REF _Ref127941496 \h  Table D.4  below summarizes the results from the
air sampling before and after the spray stations were replaced.

Table D.  SEQ Table_D. \* ARABIC  4 : Summary of nPB 8-Hour TWA Exposure
Concentration Results at Custom Productsa

 Sample Set	Before Spray Booth Replacement	After Spray Booth
Replacement

	Number of Samples	Concentration (ppm)	Number of Samples	Concentration
(ppm)



Mean 	Min.	Max.

Mean	Min.	Max.

All Exposures 	69	168.9	60.0	381.2	30	19.0	1.2	58.0

   Assembly	36	169.8	60.0	250.7	11b	18.8	6.1	32.0

          Sprayers	15	193.0	115.3	250.7	3	21.7	14.9	32.0

          Assemblers	20	154.7	60.0	234.9	7	19.5	9.9	31.8

   Covers	21	197.0	117.3	381.2	12c	29.2	2.8	58.0

   Saw	12	117.1	85.1	159.2	6	1.8	1.6	2.0

   Sew	N.A.	N.A.	N.A.	N.A.	1	1.2	N.A.d	N.A.d

Average for Active Sprayers	36	195.0	14	29.5 e

N.A. Not applicable

a Data in shaded cells for those workers actively spraying nPB.  Bold
values are above the AEL.

b Data set includes exposure to the foreman in the Assembly department,
which was measured at 6.1 ppm.

c Data set includes exposure to the foreman in the Covers department,
which was measured at 2.8 ppm.

d Only one exposure measurement was taken in the Sew department.

e This number excludes the exposure data of the foreman in the Covers
department, which measured at 2.8 ppm.

As shown in   REF _Ref127941496 \h  Table D.4 , nPB exposures exceeded
the recommended AEL of 17 ppm in almost all cases, and exposures to
those employees actively spraying the adhesive all exceeded the AEL
prior to spray booth replacement.  After the spray booths were replaced,
only three out of 14 exposures to those employees spraying the adhesive
were below the AEL.  The other employees who had exposures less than the
AEL were the foremen in all departments, the operators in the Saw
department, and two Assemblers in the Assembly department.

Marx Industries

Marx Industries manufactures cushions by gluing pieces of foam together
with an nPB-containing adhesive applied using a siphon-cup feed,
compressed air spray gun; the pieces are then pressed together by hand. 
The adhesive is applied on two manufacturing lines—the Springs Line
and the Glue Line.  Both lines are equipped with spray tables and
exhaust fans.

Long-term exposure testing was done for employees working 8-hour shifts
at two different times.  The first sampling was done in November 1999. 
A second, more comprehensive, sample set was taken in January 2001.   
REF _Ref127941547 \h  Table D.5  below summarizes the results from both
sample sets.

Table D.  SEQ Table_D. \* ARABIC  5 : Summary of nPB 8-Hour TWA Exposure
Concentration Results at Marx Industriesa

First Sample Set (November 1999)	Second Sample Set (January 2001)

Job Title/ Sample Type	Location	Concentration (ppm)	Job Title/

Sample Type	Location	Concentration (ppm)

Adhesive Sprayer	Glue Line	105.9	Springs area	Springs Line 	10.7

Adhesive Sprayer	Glue Line	89.2	Supervisor 	Springs Line 	21.4

Adhesive Sprayer	Glue Line	77.3	Adhesive sprayer 	Springs Line 	215.8

Adhesive Sprayer	Glue Line	131.4	Adhesive sprayer 	Springs and Glue Line
77.7

Adhesive Sprayer	Glue Line	115.0	Adhesive sprayer 	Springs Line 	280.5

Adhesive Sprayer	Glue Line	66.3	Adhesive sprayer 	Springs Line 	70.6

Adhesive Sprayer	Glue Line	57.7	Adhesive sprayer 	Springs and Glue Line
173.2

Doffer	Glue Line	51.8	Adhesive sprayer 	Glue Line 	115.2

Supervisor/Set-up	Glue Line	18.1	Adhesive sprayer 	Glue Line 	44.5

Adhesive Sprayer	Springs Line	86.1	Adhesive sprayer 	Glue Line 	38.0

Adhesive Sprayer	Springs Line	160.0	Supervisor 	Glue Line 	7.2

Adhesive Sprayer	Springs Line	121.0	Baler 	Glue Line 	10.5

Adhesive Sprayer	Springs Line	253.9	Baler 	Glue Line 	19.9

Adhesive Sprayer	Springs Line	123.1	 	 	 

Foam Set-up	Springs Line	38.0	 	 	 

Doffer	Springs Line	45.9



	Average for Active Sprayers	115.6	 Average for Active Sprayers	 126.9

a Data in shaded cells for those workers actively spraying nPB.  Bold
values are above the AEL. 

As the table indicates, all but three of the exposures exceeded the AEL,
with the majority of the exposures being more than five times greater
than the recommended AEL.  The exposures that were below the AEL were to
employees that were not actively spraying the nPB-containing adhesive.

Conclusion

When comparing the monitoring data to the modeled concentrations for the
four hypothetical situations discussed previously, Custom Products
(prior to spray station replacement) and Marx Industries had exposure
concentrations that are most comparable with the estimated values of
Scenario 3, with average adhesive use and poor ventilation.  STN
Cushion Company had exposure data that were comparable to the estimated
values for Scenario 1, with average adhesive use and average
ventilation.  After spray station improvements, both STN Cushion
Company and Custom Products had exposure concentrations that were, in
general, lower than those of Scenario 1 but still exceeded the AEL in
most cases.  In each scenario, the majority of nPB exposures in the
adhesives industry did not meet the 17 ppm exposure limit.

The spray station improvements did result in significant reductions in
worker exposure.  For example, at the STN Cushion Company, there was a
72 percent reduction in worker exposure, and at Custom Products, there
was an 89 percent reduction in worker exposure.  Nevertheless, the
8-hour TWA exposures still exceeded the AEL for the majority of the
workers.  While actions to further reduce exposure are possible—such
as a decrease in unused space in spray booths to promote
ventilation—it is not clear whether these improvements will result in
the significant reduction needed to meet the 17 ppm AEL.  Furthermore,
it is not clear whether facilities would be able to reduce nPB exposure
for all employees, especially the most active sprayers.

References

ACGIH, 2004.  American Conference of Governmental Industrial Hygienists
Industrial Ventilation: A Manual of Recommended Practice, 23rd edition,
July 1999.

EPA. 1991.  Building Air Quality, Revised, EPA-402/F-91-102, EPA Office
of Indoor Air.  December 1991. 

NIOSH.  2003.  NIOSH Health Hazard Evaluation Report: HETA
#99-0260-2906; Marx Industries, Inc.; Sawmills, NC.  National Institute
for Occupational Safety and Health.  June 2003.  Available online at  
HYPERLINK "http://www.cdc.gov/niosh/hhe/reports/pdfs/1999-0260-2906.pdf"
 http://www.cdc.gov/niosh/hhe/reports/pdfs/1999-0260-2906.pdf . 

NIOSH.  2002a.  NIOSH Health Hazard Evaluation Report: HETA #
98-0153-2883; Custom Products, Inc.; Mooresville, NC.  National
Institute for Occupational Safety and Health.  November 2002.  Available
online at   HYPERLINK
"http://www.cdc.gov/niosh/hhe/reports/pdfs/1998-0153-2883.pdf" 
http://www.cdc.gov/niosh/hhe/reports/pdfs/1998-0153-2883.pdf . 

NIOSH.  2002b.  NIOSH Health Hazard Evaluation Report: HETA
#2000-0410-2891; STN Cushion Company; Thomasville, NC.  National
Institute for Occupational Safety and Health.  August 2002.  Available
online at   HYPERLINK
"http://www.cdc.gov/niosh/hhe/reports/pdfs/2000-0410-2891.pdf" 
http://www.cdc.gov/niosh/hhe/reports/pdfs/2000-0410-2891.pdf . 

Swanson, M.B., J.R. Geibig, and K.E. Kelly.  2002.  Alternative
Adhesives Technologies: Foam Furniture and Bedding Industries, Final
Draft.  Volume 2: Risk Screening and Comparison.  Chapter 4: Exposure
Assessment.  Produced by the University of Tennessee Center for Clean
Products and Clean Technologies under a grant from EPA’s Design for
the Environment Branch, Office of Pollution and Prevention and Toxics. 
June 2002.  Available online at   HYPERLINK
"http://eerc.ra.utk.edu/ccpct/aap1.html" 
http://eerc.ra.utk.edu/ccpct/aap1.html .

Attachment   SEQ Attachment \* ALPHABETIC  E 

Occupational Exposure Analysis for n-Propyl Bromide When Used As an
Aerosol Solvent

Occupational Exposure Modeling

 Introduction

n-propyl bromide (nPB) has been shown to exhibit toxicity primarily to
the central and peripheral nervous, reproductive, and hematopoietic
systems following inhalation exposure.  To date, the potential for
worker exposure to nPB across the range of facilities where it can be
used has not been assessed in great detail.  Some research has been
conducted to measure nPB exposure in bench top and assembly line
applications.  However, aerosol solvents are frequently used in in-situ
applications where the solvent is applied intermittently to the object
that is still in use (e.g., an electric control panel on a machine),
resulting in exposure of the worker to solvent vapor.  This attachment
attempts to provide a brief assessment of these potential exposures and
evaluates them based on the recommended acceptable exposure limit (AEL)
for nPB.

Aerosol nPB usage is emissive.  Significant occupational exposure can
occur when aerosol solvents are used to clean or treat a piece of
equipment.  Exposure levels depend on multiple factors that include room
ventilation, room size, the amount of nPB being used, and worker
proximity to the spray apparatus.  Therefore, workers applying nPB
solvent in this type of application can be exposed to varying
concentrations.  This assessment attempts to quantify nPB exposures
resulting from the most likely in-situ use scenarios.

The rest of this section is organized as follows: Section   REF
_Ref127941608 \r \h  1.2  presents the approach to modeling the exposure
of aerosol solvent use in the workplace; Section   REF _Ref127941612 \r
\h  1.3  presents and evaluates the results of the analysis.

Approach

The approach for assessing occupational exposure to nPB was to model the
use of an aerosol solvent.  The analysis is based on a simple box-model
approach that draws assumptions from spray tests performed by aerosol
solvent manufacturing companies.  In this case, the box-model approach
examines an area surrounding the face of the exposed worker and
determines exposure based on the velocity of ventilation present.  As
the velocity varies, the volume of moving air surrounding the worker
also changes, which in turn changes the level of exposure estimated by
the box model.  In this section the assumptions and application
scenarios used to create the exposure model are described.  These
application scenarios provide exposure estimates of very high
concentrations because intense use and low ventilation were assumed.

Assumptions

Assumptions regarding usage, area of exposure, ventilation, and the
constituents of the solvent formulation are described below. 

Usage

An upper limit for heavy industrial usage of solvent (nPB mixture) for
cleaning equipment is 1000 grams per day (Anon. 1998);

The 8-hour exposure scenario will consist of 1000 and 500 grams of
solvent used per work day; and

The 15-minute exposure scenario will consist of 500 and 250 grams of
solvent used.  It is assumed that 500 grams is the highest amount of
solvent likely to be used in any 15-minute time period;

Area of Exposure

The area of exposure is 18 inches in equidistant directions (3 feet x 3
feet = 9 square feet of area at the face); used to estimate ventilation
volume for a given face velocity.

Ventilation

No ventilation is present: 50 feet per minute over 9 square feet of area
at the face is equivalent to 450 cubic feet per minute or 12.7 cubic
meters per minute (EPA 1994);

Poor ventilation is present: 100 feet per minute over 9 square feet of
area at the face is equivalent to 900 cubic feet per minute or 25.4
cubic meters per minute; and

Moderate ventilation is present: 150 feet per minute (minimum
recommended face velocity) over 9 square feet of area at the face is
equivalent to 1350 cubic feet per minute or 38.2 cubic meters per minute
(ACGIH 1982).

Solvent Formulation:

The aerosol spray formula consists of 10 percent propellant (HFC-134a)
and 90 percent solvent   (nPB or mixtures of nPB and HFC-365mfc) in four
different, commercially available formulations.    REF _Ref127941671 \h 
Table E.1  provides formulation details.

Each of the solvent formulations is assumed to have the same relative
cleaning effectiveness, so that the same total amount of solvent is used
for each formulation.

Table E.  SEQ Table_E. \* ARABIC  1 : Solvent Formulation Details

Mix	Solvent Composition

	HFC-134a	HFC-365mfc	nPB

1	10%	0%	90%

2	10%	23%	68%

3	10%	45%	45%

4	10%	68%	23%



Application Scenarios

Each of the following ventilation scenarios was modeled for both
moderate and high solvent use.  Exposure concentrations for each
scenario are presented in the tables in Section   REF _Ref127941695 \r
\h  1.3 .

Ventilation Scenario 1: No Ventilation Present

Assumptions:

This scenario represents the use of solvent with no local ventilation,
or only typical local air currents, with a flow rate of 12.7 cubic
meters per minute (EPA 1994). 

Amount of solvent used is 1000 and 500 grams over a period of 8 hours.

All four solvent mixtures are evaluated under this scenario.

Eight-hour average application rate:

1,000 grams used in 8-hour period	= 0.035 grams per second

28,800 seconds (in 8 hours)

	   500 grams used in 8-hour period	= 0.017 grams per second

28,800 seconds (in 8 hours)

	

Ventilation Scenario 2: Poor Ventilation Present

Assumptions:

This scenario represents the use of solvent in multiple-use scenarios
with poor ventilation of 25.5 cubic meters per minute.

Amount of solvent used is 1000 and 500 grams over a period of 8 hours.

All four solvent mixtures are evaluated under this scenario.

Application rates: the application rates are identical to those in
Scenario 1.

Ventilation Scenario 3: Moderate Ventilation Present

Assumptions:

This scenario represents the use of solvent in multiple-use scenarios
with moderate ventilation of 51.0 cubic meters per minute.

Amount of solvent used is 1000 and 500 grams over a period of 8 hours.

All four solvent mixtures are evaluated under this scenario.

Application rates: the application rates are identical to those in
Scenario 1.

Modeling Approach

The exposure point concentration for occupational inhalation exposure is
determined by calculating the workplace air chemical concentration.  The
indoor air concentration for nPB was estimated using the following
equation (EPA 1991):

Ca	=	Yv x 1000	mg	x	24.45



	g





AT x k

MW

Where:

Ca	=	Concentration of the chemical in air (ppm)

Yv	=	Mass emission rate of volatile compound released (g/s)

AT	=	Air flow rate (m3/s)

k	=	Dimensionless room ventilation mixing coefficient (assumed as 0.5,
EPA 1991)

MW	=	Molecular weight (123.01 g/mole for nPB)



Note that 24.45 is a factor used to convert from mg/m3 to ppm.

The mixing factor, k, accounts for the slow and incomplete mixing of
ventilation air with room air (i.e., ventilation effectiveness is
reduced by poor dispersion characteristics within the room).  In this
case a small area around the user (the breathing zone) was evaluated, so
complete mixing is assumed and a default value of one is used.  The
model assumes that the entire area surrounding the solvent sprayer
contains the same nPB concentration.  The model also assumes that the
chemicals are completely volatilized into the air after release.  This
assumption is appropriate for nPB-based applications given the method of
application (aerosol canister).

This approach also assumes that the worker stays within the in-situ
application area for the entire 8-hour period.  It is very likely that
the worker will move on to other tasks within the plant/factory or even
leave the site once the cleaning process has been finished.  By assuming
the worker stays in the contaminated area the entire time, this analysis
is expected to overestimate worker exposure.

Results

The three ventilation scenarios were modeled for both moderate and high
solvent use.  The results of the analysis are outlined in   REF
_Ref127941820 \h  Table E.2 ,   REF _Ref127941835 \h  Table E.3 , and  
REF _Ref127941832 \h  Table E.4 .

Table E.  SEQ Table_E. \* ARABIC  2 : Eight-Hour TWA Exposure
Concentrations (ppm) with No Ventilationa

Mix	Solvent Formulation	Moderate Solvent Use	High Solvent Use

	HFC-134a	HFC-365mfc	nPB	HFC-134a	HFC-365mfc	nPB	HFC-134a	HFC-365mfc	nPB

1	10%	0%	90%	2.2	-	16.3	4.4	-	32.6

2	10%	23%	68%	2.2	3.4	12.2	4.4	6.8	24.4

3	10%	45%	45%	2.2	6.8	8.1	4.4	13.5	16.3

4	10%	68%	23%	2.2	10.2	4.1	4.4	20.3	8.1

a Bold values are above the AEL; italicized values are near the AEL.

The exposure concentrations under no ventilation for each of the solvent
use scenarios modeled are under the recommended 8-hour nPB AEL of 17
ppm, with the exception of the 90 and 68 percent nPB mixtures (Mix 1 &
2) with high solvent use.  Exposure to the 90 percent nPB mixture (Mix
1) at moderate solvent use, and the 45 percent nPB mixture (Mix 3) at
high solvent use are estimated to be 16.3 ppm, or at the limits of safe
exposure.  Therefore, the modeling results suggest that in-situ
occupational exposure to nPB could pose a risk to worker’s health
during moderate and high solvent use when no engineered ventilation is
present.

Table E.  SEQ Table_E. \* ARABIC  3 : Eight-Hour TWA Exposure
Concentrations (ppm) with Poor Ventilation

Mix	Solvent Formulation	Moderate Solvent Use	High Solvent Use

	HFC-134a	HFC-365mfc	nPB	HFC-134a	HFC-365mfc	nPB	HFC-134a	HFC-365mfc	nPB

1	10%	0%	90%	1.1	-	8.1	2.2	-	16.2

2	10%	23%	68%	1.1	1.7	6.1	2.2	3.4	12.2

3	10%	45%	45%	1.1	3.4	4.1	2.2	6.7	8.1

4	10%	68%	23%	1.1	5.1	2.0	2.2	10.1	4.1

a Italicized value near the AEL.

The exposure concentrations under poor ventilation for each of the
solvent use scenarios modeled are below the recommended 8-hour AEL of 17
ppm.  It should be noted that the 90 percent nPB mixture (Mix 1) at high
solvent use is modeled just under the recommended AEL.  Therefore, the
modeling results suggest that in-situ occupational exposure to nPB could
pose a risk to worker’s health during high solvent use when poor
ventilation is present.

Table E.  SEQ Table_E. \* ARABIC  4 : Eight-Hour Exposure Concentrations
(ppm) with Moderate Ventilation

Mix	Solvent Formulation	Moderate Solvent Use	High Solvent Use

	HFC-134a	HFC-365mfc	nPB	HFC-134a	HFC-365mfc	nPB	HFC-134a	HFC-365mfc	nPB

1	10%	0%	90%	0.5	-	4.1	1.1	-	8.1

2	10%	23%	68%	0.5	0.8	3.0	1.1	1.7	6.1

3	10%	45%	45%	0.5	1.7	2.0	1.1	3.4	4.1

4	10%	68%	23%	0.5	2.5	1.0	1.1	5.1	2.0



The exposure concentrations under moderate ventilation for each of the
solvent use scenarios modeled are all below the recommended 8-hour AEL
of 17 ppm.

Exposure modeling suggests that when using solvent mixtures with nPB
concentrations greater than or equal to 45 percent and thereby using
more than 450 g of nPB per day, ventilation is required to keep the
occupational exposure levels below the recommended AEL exposure limit. 
But, if adequate ventilation exists and care is taken to monitor and
reduce the frequency of highly concentrated use, occupational exposure
to nPB should not pose a threat to workers’ health.  It should be
noted that this analysis does not consider the use of aerosol solvents
within confined spaces such as elevator shafts or electronic control
rooms that may have little or no ambient air flow.

Occupational Exposure Study Evaluations  

Summary

In addition to the occupational exposure modeling, three occupational
exposure studies for aerosol solvent use were evaluated to assess
possible exposure issues for the use of nPB as an aerosol solvent.  The
studies are evaluated based on the following:

The likelihood that the studies represent a reasonable approximation of
workplace conditions and actual aerosol solvent utilization
characteristics for nPB; and

The likelihood that the proposed 8-hour AEL and excursion limit can be
met by end users based on the results of these studies.

The three occupational exposure studies reviewed here (two for nPB and
one for HFE) appear to represent typical application settings for
aerosol solvent use.  However, in the majority of the studies,
ventilation appears to be inadequate to provide an acceptable level of
occupational exposure for all workers.  While the exposure results in
the studies are often well above the AEL, it is likely that these
standards would be met with sufficient workspace ventilation, including
the use of locally ventilated spray booths or fume hoods and personal
protective equipment.

The following sections detail the evaluation of these criteria for the
three exposure studies received.

Tech Spray

This study is comprised of exposure monitoring data for the following
three different nPB use applications (Tech Spray 2003).

Test 1 was designed to simulate an automotive repair shop in which nPB
is used to clean and degrease auto parts in bench-top and booth
applications for short-term exposure scenarios.  Samples 1, 3, and 4
used 205 grams, 355 grams, and 178 grams of nPB within the 15-minute
period.  These were worst-case scenarios because nPB was sprayed
continuously for 15-minutes.  In an actual work scenario, workers would
spray for a few minutes at a time.  Furthermore, the quantity used in
the samples is considerably more than would have been released in a
minute of spraying nPB.

Test 2 simulates the cleaning of printed circuit boards for the repair
of computers and electrical systems in a non-vented booth for long-term
exposure scenarios.  The samples from Test 2 were taken for 8 hours and
samples were taken in the general work area of concern only during the
actual spray-cleaning period.  Approximately 20 units were cleaned for
two to ten seconds to remove general use residues, fluxes, and other
such contaminants during the 8-hour (480-minute) period, resulting in an
average of 24 minutes between cleaning times with a total actual clean
time of less than 200 seconds (3.33 minutes), and a total daily use of
around 310 grams of nPB in aerosol form.

Test 3 is the same as the Test 2 except that the booth is vented.  The
booth was vented from the back using a constant draw of 0.9 cubic meters
per second during the 8-hour test.

Eleven samples were taken during the three tests to determine potential
operator exposure to nPB.  Seven out of the eleven tests were taken in
the breathing zone (collar) of the workers.  These samples are more
characteristic of the potential risk to the worker and are therefore
presented in the table.  The results are presented in   REF
_Ref127941926 \h  Table E.5 .

Table E.  SEQ Table_E. \* ARABIC  5 : Tech Spray Study - nPB Exposure
Resultsa

Test	Sampleb	Location	Ventilation (cfm)	Time (min)	Exposure (ppm)
Pass/Fail

AEL

(17 ppm)c

1	1	Operator in work area 	640	15	370	NA

1	3	Operator in work area  	640	15	1,100	NA

1	4	Operator at booth 	472	15	190	NA

1	5	Worker in adj. room 	NA	15	11	NA

2	6	Operator at booth 	0	480	32	Fails

2	11	Operator at booth 	0	480	13	Passes

3	9	Operator at booth	1907	480	5.5	Passes

a Bold value is above the recommended AEL of 17 ppm.

b Sample 2 was taken at the spraying operator’s wrist and Samples 7,
8, and 10 were taken above the substrate in the booth and therefore are
not representative of an inhalation dose.  Therefore, these samples were
removed from this analysis.

c Because samples 1, 3, 4, and 5 are only 15-minute exposures, they
cannot be directly compared to the AEL unless one assumes that no other
exposure occurred throughout the workday.

As can be seen in the above table, two of the three 8-hour samples met
the recommended 8-hour exposure limit for nPB.  The 15-minute exposures
ranged from 11 ppm to 1,100 ppm.  By dividing a 15-minute exposure by
two, it can be converted to a 30-minute TWA, assuming that 15 minutes is
the maximum amount of time that nPB would be continually sprayed and
inhaled within a 30-minute period and that no nPB would linger in the
area to be inhaled.  Furthermore, if it is assumed that the 15-minute
period was the only time the operator was exposed during the 8-hour day,
this 15-minute TWA exposure can be divided by 32 to convert it to an
8-hour TWA.  Sample 3, which was a 15-minute exposure of 1,100 ppm,
would exceed the 30-minute excursion limit and 8-hour AEL.  The 370 ppm
and 190 ppm exposures, if they were the only exposure in the entire
8-hour work day, would meet the 8-hour AEL but still exceed the
30-minute TWA excursion limit.  This is not surprising, considering the
low ventilation rate during these tests.  Only the 11-ppm exposure,
which was to a worker in an adjacent room, would meet both the 8-hour
AEL and the 30-minute excursion limit.

The high levels of exposure in the Tech Spray study raised several
issues when the ventilation characteristics of the three test scenarios
were evaluated.

The use of an aerosol solvent outside of a booth or hood would very
likely result in the high exposure results found in Test 1.  Also, the
airflow in a vented booth with 472 cubic feet per minute (cfm) would
seem inadequate for the use of this compound.  Given the dimensions of
the booth (40 inches by 40 inches by 20 inches) the estimated face
velocity is approximately 42.5 feet per minute (fpm).  It is recommended
that in most cases a face velocity of 100 fpm be used for substances of
a toxic nature (USIP 2003).  Also, the fact that the sample taken in an
adjacent room outside of the operator’s environment (Sample 5, 11 ppm)
was so high suggests a pressure problem between the two rooms, where the
adjacent room actually draws in contaminated air that should be directed
through the exhaust ventilation system.

In Test 2, the application takes place in an unventilated booth (also 40
inches by 40 inches by 20 inches).  This case would seem to provide the
lowest possible protection to the worker and possibly increase the risk
due to the accumulation of vapor in the non-vented booth.

Test 3 provides a glimpse of what an acceptable use scenario would be. 
The application again takes place in a booth, except this booth is
ventilated to provide a face velocity of 171 fpm.  The recorded worker
exposure is 5.5 ppm, well below the AEL of 17 ppm.

Company A

This study is from a confidential submitter and is based on a single nPB
use application (Anonymous 1998).  For clarity in this discussion, the
submitter of this report will be referred to as “Company A.”  The
testing was conducted in the maintenance shop of an unspecified facility
in which nPB is used as a cleaning agent in a bench-top application. 
The test substrate was a small electric motor of approximately 156
square centimeters in area.  Test subjects sprayed 250 grams of nPB at
four evenly-spaced two-minute periods throughout an 8-hour work day,
resulting in a total use of 1000 grams (approximately two cans) of nPB. 
Sixteen samples were taken during the three tests to determine employee
exposure to nPB.  Twelve out of the 16 tests were taken in the breathing
zone of the workers.  These samples are more characteristic of the
potential risk to the worker and are therefore presented in the table. 
The results are presented in   REF _Ref127941970 \h  Table E.6 .

Table E.  SEQ Table_E. \* ARABIC  6 : Company A Study - nPB Exposure
Resultsa

Sample	Employee or Location	Ventilation (cfm)b	Time (min)	Exposure (ppm)
Pass/Fail

AEL

(17 ppm)c

1	QC Technician 	300	480	15.1	Passes

2	Mechanic  	300	480	11.3	Passes

3	Mechanic  	300	480	17	Passes

4	Maintenance Supervisor 	300	480	30.2	Fails

9	Mechanic	300	15	57.5	NA

10	Maintenance Supervisor	300	15	254	NA

11	Mechanic	300	15	45.1	NA

12	Maintenance Supervisor	300	15	243	NA

13	Mechanic	300	15	151	NA

14	QC Technician  	300	15	134	NA

15	Mechanic	300	15	123	NA

16	QC Technician  	300	15	92.8	NA

a Bold value is above the recommended AEL of 17 ppm; italicized value is
at the AEL.

b Shop ventilation was decreased from the normal rate of 1510 cfm to 300
cfm to model a worst-case scenario.

c Because samples 9 through 16 are only 15-minute exposures, they cannot
be directly compared to the AEL unless one assumes that no other
exposure occurred during the workday.

As can be seen in the above table, three of the four samples met the AEL
for nPB exposure, with one of these samples being at the recommended
AEL.  The 15-minute exposures in samples 10 and 12 would meet the 8-hour
AEL but exceed the 30-minute excursion limit, assuming these were the
only exposures that the employees received in the day.  The rest of the
15-minute exposures would meet both the 8-hour AEL and 30-minute
excursion limit, assuming that these were the only exposures the
employee received in the 8-hour work day.  While many of the 15-minute
samples in this study exceed the 30-minute excursion limit, overall,
these samples are much lower than the values in the Tech Spray study.

The Company A exposure study produced results where several of the
employee exposures were above the AEL and excursion limit of nPB. 
However, this study was conducted where ventilation was set at 300 cfm
(a relatively low value).  It is noted in the study that under normal
operating conditions, the ventilation in the work area is 1510 cfm, or
five times greater than when the exposure monitoring took place.  Using
a proportional drop in nPB concentration, the recorded exposure values
can be reduced by 80 percent to provide an approximate estimate of
exposure under actual operating conditions.  In this scenario, all
samples would meet the AEL and excursion limit of nPB.  This study
provides evidence that under normal operating conditions for bench-top
cleaning, virtually all of the nPB exposure samples would be within
acceptable limits and again emphasizes the importance of adequate
ventilation when using this compound.

3M Corporation

This study is composed of exposure monitoring tests for nine different
aerosol contact cleaner formulations and three neat HFEs (3M 1997).  The
study is designed to simulate an electronics cleaning process where
circuit boards are treated with HFE-based aerosol solvents.  Twelve
samples were taken during simulated operation of the process to
determine operator exposure to HFE and other component chemicals
associated with solvent application in this process.  From these twelve
samples, the highest measured concentrations of the formulation
components are presented in   REF _Ref127942022 \h  Table E.7 .

Table E.  SEQ Table_E. \* ARABIC  7 : 3M Study - Aerosol Exposure
Results

Component	Highest Measured Concentration (ppm)	8-Hour Exposure Limit
(ppm)	Pass/Fail

AEL

HFE-7100	194.0	750	Passes

HFE-7200	148.0	200	Passes

Trans-1,2-dichloroethylene	76.3	200	Passes

HCFC-141b	150.0	500	Passes

HCFC-225ca/cb	79.7	100	Passes

Acetone	70.1	750	Passes

Pentane	56.5	600	Passes

HFC-134a	35.5	1000	Passes



The 3M Corporation exposure study produced results in which all worker
exposures were below the AEL limits for eight component compounds.  In
order to glean information about possible nPB exposure, the highest
measured concentration of 194 ppm for HFE-7100 was used as a surrogate. 
Differences in molecular weight, vapor pressure, and latent heat of
vaporization characteristics are assumed to be negligible for the sake
of this discussion.  This exposure estimate is several times greater
than the 17 ppm AEL for nPB.  However, this study was conducted under a
worst-case scenario in which mechanical ventilation was shut off.  It is
assumed that in this worst-case scenario, the workspace had an air-flow
of 48 cfm.  In order to approximate operating conditions with minimal
ventilation present, a conservative air-flow estimate of 500 cfm is
assumed.  This would reduce the measured chemical concentrations in the
workspace by a factor of 10, resulting in an estimated exposure value of
19 ppm for the highest measured concentration.  While this value is just
above the AEL, the rest of the exposures would be below the AEL. 
Therefore, it is assumed that under operating conditions with minimal
ventilation, worker exposures could be close to the AEL limits for nPB. 


Conclusion

While these exposure studies were conducted for a reasonable application
end use, the poor or worst-case ventilation present during nPB (or
surrogate) exposure produced results that are well above the AEL and
excursion limit for worker safety.  However, in each of the studies it
can be shown that when sufficient ventilation is present, it is possible
to meet occupational safety limits for the use of nPB.  The Tech Spray
study reports an 8-hour nPB occupational exposure of 5.5 ppm when used
in conjunction with a highly ventilated spray booth.  The remaining two
studies imply that nPB exposure standards would be met under most
operating conditions using artificial ventilation, albeit certain
exposure scenarios could slightly exceed the recommended AEL.  It should
also be noted that in addition to adequate ventilation, appropriate
worker personal protective equipment (PPE) and procedural best
management practices to reduce worker exposure are also critical.

The evaluation of these three aerosol exposure studies determined the
following:

These studies represent likely end use scenarios, and the exposure
measurements suggest that it is possible that sufficient ventilation
will not be provided to meet safety standards in all cases.

These studies also show that it is possible for end users to meet the
AEL and excursion limit with sufficient general workplace ventilation
and/or the use of locally-ventilated booths.  Users that previously used
other less-toxic solvents will most likely have to increase ventilation
rates or employ locally ventilated booths in order to meet the low AEL
exposure limits for nPB.

References

3M. 1997. 3M HFE Aerosol User Exposure. 3M Corporation.  May 22, 1997.

Anonymous. 1998.  Airborne Exposure Assessment of 1-Bromopropane. 

ACGIH. 1982.  Industrial Ventilation: A Manual of Recommended Practice,
17th Edition.  Chapter 4.  Cited in U.S. EPA.  1994.  SNAP Technical
Background Document:  Risk Screen on the Use of Substitutes for Class I
Ozone-Depleting Substances: Aerosols. 

EPA.  1994.  SNAP Technical Background Document:  Risk Screen on the Use
of Substitutes for Class I Ozone-Depleting Substances: Aerosols. 

EPA. 1991.  Building Air Quality, Revised, EPA-402/F-91-102, EPA/Office
of Indoor Air, December 1991.

Tech Spray. 2003.  nPB and aerosol exposure.  Tech Spray, L.P.
Management.  July 22, 2003.

USIP. 2003.  Laboratory Fume Hoods.  University of the Sciences in
Philadelphia.

  HYPERLINK "http://www.usip.edu/safety/lmanual/LSfumehood.htm" 
http://www.usip.edu/safety/lmanual/LSfumehood.htm 

Attachment   SEQ Attachment \* ALPHABETIC  F 

General Population Exposure Assessment for n-Propyl Bromide

Introduction

n-propyl bromide (nPB) is used in various solvent, aerosol, and adhesive
formulations as an alternative to ozone-depleting substances (ODS) such
as methyl chloroform.  However, nPB has been shown to exhibit toxicity
to the liver, nervous system, and reproductive system following
inhalation exposure.  To date, the potential for significant exposure of
the general populations living near facilities that use nPB has not been
adequately assessed.  This attachment attempts to provide a brief
assessment of these exposures. 

This assessment focuses on the long-term general population exposure as
a result of routine adhesive application.  This end use was chosen
because adhesive use is more emissive than either the solvent cleaning
or the aerosol solvent end uses.  Thus, EPA expects that the greatest
nPB exposure to the general population will result from adhesive
operations. 

Approach

The exposure assessment was conducted with the objective of determining
the most reasonable foreseeable worst-case scenario for general
population exposure associated with the use of nPB.  The exposure
assessment uses a hypothetical adhesive application facility based on
data collected from actual facilities (EPA 2001).  The collected data
were used in this analysis to characterize the typical high-use adhesive
application facility and an average-use adhesive application in an urban
locale, assuming a row-house-type warehouse setting.

Two emission release scenarios were considered in the analysis:

Emissions released from the average-size adhesive application facility
for the bonding of furniture and mattresses (S1); and

Emissions released from the urban row-house-type warehouse setting in
which application of nPB is used to bond furniture and mattresses (S2).

Average Size Adhesive Application Facility (S1)

The general characteristics used in the modeling of the average
adhesive-application facility for the bonding of furniture and
mattresses are:

Average size facility: 65.7 meters by 65.7 meters;

Average height of the facility: 7 meters;

Typical lot size of 100 meters by 100 meters (~ 2.5 acres);

Urban setting;

Average high use adhesive emission rate of 12.3 grams per second;

Facility operates 2000 hours per year (40 hours per week, 50 weeks per
year);

Average breathing height of 1.8 meters; and

For the vented scenario the facility operates a ventilation system at
the ASHRAE recommended flow rate of 0.50 cubic feet per minute per
square foot of floor space (i.e., 1.3 air exchanges per hour) (ASHRAE
1999).  The stack was located on the roof, and a short 0.3 meters stack
with an inside diameter of 0.8 meters was assumed.  Estimates of
concentration included the effects of building downwash within cavity
region, for both the near- and far-wake regions.

Within S1, three types of releases were considered:

Emissions released as a fugitive (non-vented) area source upwards
through the roof of the facility;

Emissions released as a single point source (vented) upwards on the roof
of the facility; and

Emissions released as a fugitive (non-vented) source horizontally
through cracks, leaks, window ventilation and shaft ventilation (natural
ventilation).

Urban Row-house-type Warehouse Adhesive Application Operation (S2)

The general characteristics used in the modeling of the urban
row-house-type warehouse adhesive application operation are:

Three-story (30 feet) house, with a footprint of 25 feet by 40 feet;

Urban setting;

Average high use adhesive emission rate of 1.23 grams per second;

Facility operates 50 weeks per year, 8 hours per day, 5 days per week;

Average breathing height of 1.8 meters;

Nearest resident is adjacent to the house at a distance of 3 meters; and

For the vented scenario, the facility operates a ventilation system at
the ASHRAE recommended flow rate of 0.50 cubic feet per minute per
square foot of floor space (i.e., 1.3 air exchange per hour) (ASHRAE
1999).  The stack was located on the roof with a short 0.3 meter stack
with an inside diameter of 0.8 meters.  Estimates of concentration
included the effects of building downwash within cavity region, for both
the near- and far-wake regions.

Within S2, four types of releases were considered:

Emissions released as a fugitive (non-vented) area source upwards
through the roof of the warehouse;

Emissions released as a single point source (vented) upwards on the roof
of the warehouse;

Emissions released as a fugitive (non-vented) source horizontally
through cracks, leaks, window ventilation and shaft ventilation (natural
ventilation); and

Emissions released as a fugitive source horizontally using commercially
available fans capable of accomplishing at least 1.3 air exchanges per
hour.

Modeling System 

EPA’s SCREEN3 (EPA 1995) air dispersion model was used to assess
dispersion of emissions using the full range of meteorological
conditions (all stability classes and wind speeds) to estimate the
highest 1-hour concentration.  The SCREEN3 model is a “screening”
model used to assess the likely maximum-potential concentration from a
single source.  The technique used here is typically used to evaluate
air quality impacts of sources pursuant to the requirements of the Clean
Air Act, such as prevention of significant deterioration, new source
review, and existing sources of air pollutants, including air toxics. 
The approach applied here is the initial-phase approach used to
determine if either (1) the source clearly poses no air quality problem
or (2) the potential for an air quality problem exists.  If a potential
problem exists then a more refined analysis is necessary following
approaches discussed in the Guideline on Air Quality Models (Revised)
(40 CFR Parts 51 and 52).

EPA’s screening procedures (EPA 1992) for annual average concentration
require that estimates be developed by multiplying the maximum 1-hour
concentration determined by SCREEN3 by a factor of 0.10.  The
multiplying factors are based upon general experience with relationships
between maximum short-term concentrations and annual average
concentrations based on observations.  The factors account for the fact
that the meteorological conditions that produce the highest 1-hour
concentrations do not occur throughout the year (i.e., variations in
wind speed, direction, atmospheric stability).  The factors are intended
as a way to estimate the maximum annual concentrations from short-term
averages; a degree of conservatism is incorporated in the factors to
provide reasonable assurance that maximum annual values will not be
underestimated.  For a site-specific analysis using site-specific
meteorological data, the use of the long-term dispersion
model—Industrial Source Complex Long Term (ISCLT)—is preferred, but
unfortunately this approach cannot be used in this analysis because
site-specific meteorological data are not available.  The range of
factors that may be used with this model is 0.06 to 0.10.  Using a
factor of 0.10 in this risk screen incorporates an additional degree of
conservatism since this number represents a worst-case maximum
annual-average concentration assuming that aerodynamic downwash and/or
low stack height is a problem for most adhesive application facilities.

Additionally, the SCREEN3 model can provide estimated concentrations for
distances less than 100 meters (down to one meter as in other regulatory
models).  However EPA/OAQPS recommends caution when using results at
distances of less than 100 meters, as concentrations may be suspect.  A
recent model validation study (EPA 1998) for similar conditions (flat
terrain, non-buoyant, near-surface release, measurements at arcs
positioned 50 meters to 800 meters downwind) using the short-term
Prairie Grass tracer study have shown that the Industrial Source Complex
(ISC3) model predicts well within a factor of two for all downwind
distances.  The robust highest-concentration statistic shows how the
model predicts for the tail end of the distribution.  For the Prairie
Grass study, ISC3 shows an overall tendency to over predict by an
average of 50 percent for all downwind distances.  This gives further
support that the screening results are conservatively based and fairly
estimates the near source concentration.  

The certainty to which the ISC3 and SCREEN3 type model estimates are
considered accurate have been evaluated in past studies (EPA 1982, EPRI
1983) by comparison with observed concentrations.  The studies found
that: (1) these types of models are more reliable for estimating longer
time-averaged concentrations than for short-term concentrations; and (2)
the models are reasonably reliable in estimating the magnitude of the
highest concentrations occurring sometime, somewhere within an area. 
Errors in the highest estimated concentrations of 10 to 40 percent are
found to be typical, which is well within the accuracy of a factor of
two that has long been recognized for these models, as noted in the
Guideline on Air Quality Models (Revised) (40 CFR Parts 51 and 52).

Results

Results from the two scenarios are presented in   REF _Ref127942153 \h 
Table F.1  and   REF _Ref127942178 \h  Table F.2  with a comparison to
the estimated reference concentration (RfC) for nPB (See   REF
_Ref127942192 \h  Attachment B ).

Table F.  SEQ Table_F. \* ARABIC  1 : Scenario 1 - Maximum Annual
Average Air Concentrations in the Vicinity of an Average-Sized
High-Adhesive-Use Application Facility Using nPB

Distance (m)	Non-Vented

Vertical Release

Scenario (ppm)	Vented Vertical Release Scenario (ppm)	Non-Vented
Horizontal Release Scenario (ppm)	Reference

Concentration Level (ppm)



100	

0.06	

0.01	

0.08	

1



200	

0.04	

0.01	

0.04	

1



300	

0.03	

0.01	

0.03	

1



400	

0.01	

0.01	

0.01	

1



500	

0.01	

0.01	

0.01	

1



600	

0.01	

0.01	

0.00	

1



700	

0.00	

0.00	

0.00	

1



800	

0.00	

0.00	

0.00	

1



900	

0.00	

0.00	

0.00	

1



Table F.  SEQ Table_F. \* ARABIC  2 : Scenario 2 – Maximum Annual
Average Air Concentrations in the Vicinity of an Average Adhesive Using
Operation Located in an Urban Row-House-Type Warehouse Using nPB

Distance (m)	Vented Vertical Release Scenarioa (ppm)	Non-Vented Vertical
Release

Scenario (ppm)	Non-Vented Horizontal Release

(natural vent) Scenario (ppm)	Non-Vented

Horizontal

Release

(w/fans)

Scenario (ppm)	Reference Concentration Level (ppm)



3	

0.00	

0.00	

0.06	

0.24	

1



5	

0.00	

0.00	

0.06	

0.19	

1



10	

0.00	

0.00	

0.04	

0.13	

1



20	

0.00	

0.01	

0.03	

0.08	

1



30	

0.03	

0.01	

0.01	

0.05	

1



40	

0.03	

0.01	

0.01	

0.04	

1



50	

0.03	

0.01	

0.01	

0.04	

1



60	

0.01	

0.01	

0.01	

0.03	

1



70	

0.01	

0.01	

0.01	

0.03	

1



80	

0.01	

0.01	

0.00	

0.01	

1



90	

0.01	

0.01	

0.00	

0.01	

1



100	

0.01	

0.01	

0.00	

0.01	

1

a Maximum cavity region annual average concentration of 0.02 ppm
assuming half of all hours when facility is operating are conducive to
downwash conditions 

For the average-sized high-adhesive-use facility (S1), the vented
emission concentrations are lower than the non-vented emissions and are
well below the estimated RfC.  These values include building downwash. 
The additional momentum from the ventilation system appears to be of
sufficient magnitude to lower the concentrations compared to the
non-vented scenario.  All ventilation scenarios for S1 show no
concentration exceeding the estimated RfC.

For the urban row-house-type warehouse (S2), again none of the scenarios
exceed the estimated RfC, although it is worth noting that the
non-vented horizontal release with fans case shows relatively high
values compared to other S2 cases.  The two vertical release scenarios
show maximum concentrations occurring at a distance of 30 meters from
the house due to the height and upward momentum associated with the
release.

References

ASHRAE. 1999.  ASHRAE Standard Ventilation for Acceptable Indoor Air
Quality, ASHRAE Standard 62-1999, American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Atlanta, GA.

EPA.  2001.  Chapter 4.  Exposure Assessment, Adhesive Technology
Partnership, EPA Toxic Office, Design for Environment.

EPA.  1998.  Model Evaluation Results for AERMOD, EPA web site
(http://www.epa.gov/scram001). 

EPA. 1995.  SCREEN3 Model User’s Guide, EPA-454/B-95-004, EPA/OAQPS,
Research Triangle Park, NC.  September 1995.

EPA. 1992.  Screening Procedures for Estimating the Air Quality Impact
of Stationary Sources, Revised, EPA-454/R-92-010, EPA/OAQPS, Research
Triangle Park, NC.  October, 1992.

EPA. 1982.  A Survey of Statistical Measures of Model Performance and
Accuracy for Several Air Quality Models, EPA-450/4-83-001, Research
Triangle Park, NC.

EPRI. 1983.  Overview, Results, and Conclusions for the EPRI Plume Model
Validation and Development Project: Plains Site, EPRI EA-3074.  Electric
Power Research Institute (EPRI).  SEQ CHAPTER \h \r 1 

 The purpose of an excursion limit is to ensure that when a worker is
exposed to a short-term high concentration during the performance of
normal procedure or maintenance operations, the worker stays within the
TWA exposure limit.  In a typical work environment for adhesives, nPB
will be used over the course of the entire workday.  While the
excursion limit is defined for a 30-minute period, this does not mean to
imply that it is acceptable for workers to be exposed to this level of
nPB for 30-minutes continuously.  If a worker were exposed to this
level of nPB for 30 minutes continuously, the worker would have to be
removed from the work area to stay within the AEL.  The excursion limit
should be seen as a safety guidance limit or marker put in place to warn
industrial hygienists that a sample measured over the excursion limit is
an indicator of a potential failing ventilation system or the need for
additional personal protective equipment.

 This same protocol was used to develop short-term occupational exposure
limits (EGLs) in EPA’s background document (1994).

 Short-term exposures are less relevant for adhesives than long-term,
8-hour exposures because workers applying adhesives are typically
exposed steadily during the entire course of the work day, rather than
for short periods of time.

 The purpose of an excursion limit is to ensure that when a worker is
exposed to a short-term high concentration during the performance of
normal procedure or maintenance operations, the worker stays within the
TWA exposure limit.  In a typical work environment, nPB will be used
for short periods of time.  While the excursion limit is defined for a
30-minute period, this does not mean to imply that it is acceptable for
workers to be exposed to this level of nPB for 30-minutes continuously.
 Studies where nPB is sprayed for 15 or 30 minutes continuously are
assuming a worst-case scenario and are not illustrative of a realistic
usage scenario.  If a worker were exposed to this level of nPB for 30
minutes continuously, the worker would have to be removed from the work
area to stay within the AEL.  The excursion limit should be seen as a
safety guidance limit or marker put in place to warn industrial
hygienists that a sample measured over the excursion limit is an
indicator of a potential failing ventilation system or the need for
additional personal protective equipment.

 This same protocol was used to develop short-term occupational exposure
limits (EGLs) in EPA’s background document (1994).

 Assuming that nPB was sprayed at most for 15-minutes at a time within a
30-minute period , the 15-minute exposures can be divided by two and
compared to the recommended 30-minute excursion limit.

 Although nPB used in aerosol solvent applications is an emissive use,
aerosol solvents are typically used intermittently and in relatively
small amounts.  This is in contrast to adhesive applications where nPB
is aerosolized to the ambient air in large amounts throughout a typical
workday.

 The purpose of an excursion limit is to ensure that when a worker is
exposed to a short-term high concentration during the performance of
normal procedure or maintenance operations, the worker stays within the
TWA exposure limit.  The excursion limit should be seen as a safety
guidance limit or marker put in place to warn industrial hygienists that
a sample measured over the excursion limit is an indicator of a
potential failing ventilation system or the need for additional personal
protective equipment.  This risk screen does not evaluate short-term
exposure data for solvent cleaning because solvent cleaning in equipment
typically exposes workers over longer periods during the day and because
short-term data were not available.

 This same protocol was used to develop short-term occupational exposure
limits (EGLs) in EPA’s background document (1994).

 Freeboard height is the distance between the vapor zone and the top of
the vapor degreaser.

 Sometimes referred to as a chiller, a freeboard refrigeration device is
a set of secondary coils mounted in the freeboard that promotes
condensation of the vaporized solvent.

 In general, drafts, which could be caused by ventilation, should be
eliminated in the area of the degreaser itself, as this could increase
the concentration of vapor in the air (MnTAP 2005).

 Although nPB use for solvent cleaning applications is an emissive use,
vaporized solvents used for cleaning are typically condensed for re-use
and therefore emissions are relatively low when compared to emission
from adhesive applications, where nPB is aerosolized to the ambient air
in large amounts throughout a typical workday.

 For the rest of this document, this endpoint is simply referred to as
“estrous cycles,” without the clarifier that they were measured
within a 3-week period.

 EPA 2000b

 Maximum likelihood is an estimate of a population parameter most likely
to have produced the sample observations (EPA 2000b).

 Degrees of Freedom is the difference between the number of data points
and the number of parameters in the model (EPA 2000b).

 The data on average yearly adhesive-use rates were
log-normally-distributed.  Following log-transformation, the average
value was taken as the average adhesive use rate—2,874 gallons per
year.  The use rate was then converted to a mass emission rate using the
specific gravity of nPB (1.353) and assuming a 2000-hour work year. 
(Swanson et al. 2002)

 The data on average yearly adhesive-use rates were
log-normally-distributed.  Following log-transformation, the 90th
percentile value was taken as the high adhesive use rate—28,736
gallons per year. The use rate was then converted to a mass emission
rate using the specific gravity of nPB (1.353) and assuming a 2000-hour
work year.  (Swanson et al. 2002)

 A number of facilities are known to have no mechanical ventilation, and
for these poor ventilation facilities an air turnover rate, k, of 0.5
hr-1 (ACGIH 2004) was used. Using the turnover rate of 0.5 hr-1 and an
average room size of 1,440 square meters, an airflow rate of 192 cubic
meters per minute was determined to reflect facilities with poor
ventilation. (Swanson et al. 2002)

 Five hundred grams of solvent is equivalent to roughly one can.

9

3

4

5

6

7

8

9

:

;

W

X

Y

Z

\

]

{

|

}

—

˜

™

š

›

œ

ž

Ÿ

»

¼

.¼

½

¾

À

Á

å

æ

ç

-

€

Ù



&

'

(

*

+

,

-

.

/

K

L

M

N

P

Q

]

^

_

y

z

$z

{

}

~



€

‚

ž

Ÿ

 

¡

£

¤

¶

·

¸

Ò

Ó

Ô

Ö

×

Ø

Ù

Ú

Û

÷

ø

ù

ú

ü

ý

j¸

j;

j¬

j/

j

,

ԀĤ␆ਁ&䘋␱㜁Ĥ␸䠁Ĥ

>

$

@

$

@

ԀĤ␆ਁ&䘋␱㜁Ĥ␸䠁Ĥ

ԀĤ␆ਁ&䘋␱㜁Ĥ␸䠁Ĥ

\

\

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

Öü³

㄀Ĥ

@



 

0

1

?

 

%

,

6

<

G

M

P

	?

@

\

^

w

x

‡

ˆ

—

˜

¦

§

·

¹

Î

Ï

Þ

ß

î

ï

ý

þ

	

&

'

5

6

C

D

’

“

®

°

¸

¹

Ñ

Ò

Ó

Ô

é

ñ

ò

÷

ø

ÿ

÷ñêñ÷ñ÷ñ÷ñ÷ñêñ÷ñ÷ñ÷ñ÷ñêñ÷ñ÷ñ÷ñ÷ñêñ÷ñ
÷ñ÷ñ÷ñêñ÷ñ÷ñ÷ñ÷ñáñáñÝ×Í×ÍÂÍ×»ñ»ñ»ñ»ñ
»ñ

s

}

ƒ

“

£

®

´

µ

Ê

Ô

Ú

ä

ê

ô

ú

 

#

,

2

;

A

B

é

ò

ø

᠁ Ā

␃؁Ĥ옍)

؀Ĥ옍)

옍)

옍)

„@

^„@

„@

÷^„@

„@

÷^„@

愀Ĥ

·

@

@

@

D	3

D	3

j

$

@

@

@

+

,

-

0

1

ú

@

ø

♛♜

⚪⚫

⛲⛳

✸✹

➇➈

⟁⟂

⠅⠆

⡋⡌

ø

ø

ꐔx␱㜁Ĥ␸䠁Ĥ

⨁؃萏̪ꐓð葞̪

D	3

ñÿX9

ñÿX9

ñÿX9

Š

Š

Š

Š

Š

Š

hJ

਀Ħ䘋

਀&䘋

䀁

਀&䘋

@

@

@

@

@

@

@

@

@

D

@

@

kdj

@

@

kdZ

@

@

@

@

@

@

@

@

@

The purpose of an excursion limit is to ensure that when a worker is
exposed to a short-term high concentration during the performance of
normal procedure or maintenance operations, the worker stays within the
TWA exposure limit.  In a typical work environment, nPB will be used
for short periods of time.  While the excursion limit is defined for a
30-minute period, this does not mean to imply that it is acceptable for
workers to be exposed to this level of nPB for 30-minutes continuously.
 Studies where nPB is sprayed for 15 or 30 minutes continuously are
assuming a worst-case scenario and are not illustrative of a realistic
usage scenario.  If a worker were exposed to this level of nPB for 30
minutes continuously, the worker would have to be removed from the work
area to stay within the AEL.  The excursion limit should be seen as a
safety guidance limit or marker put in place to warn industrial
hygienists that a sample measured over the excursion limit is an
indicator of a potential failing ventilation system or the need for
additional personal protective equipment.

 Face velocity is a measure of air-flow entering a fume hood or booth
through the worker access point.  In this case the opening is estimated
to be 40 inches by 40 inches, or 11.1 square feet.  This area divided
into 472 cfm yields a face velocity of 42.5 feet per minute.

 Work-space volume is given as 5,760 cubic feet.  Using an ACGIH air
turnover constant of 0.5 times per hour for no mechanical ventilation,
an estimate of 48 cubic feet per minute is calculated (5,760 / 2 / 60 =
48).

 Based on the given dimensions of the work space, 500 cfm of general
ventilation would equal about 5 air changes per hour (ACH).  This would
be a minimal amount of ventilation for a facility with significant
aerosol solvent usage.

 In solvent cleaning applications, the use of vapor degreasers generally
reduces emissions of nPB to very low levels in the ambient air. 
Although nPB use in aerosol solvent applications is an emissive use,
aerosol solvents are typically used intermittently and in relatively
small amounts.  This is in contrast to adhesive applications where nPB
is aerosolized to the ambient air in large amounts throughout a typical
workday. 

 The basic dispersion algorithms contained within SCREEN3 are identical
to ISC3; the ISC3 model allows the user more flexibility in inputting
meteorological data, multiple sources, and source locations.

 An RfC is an estimate of a continuous inhalation exposure to the
general public (including sensitive subgroups) that is likely to be
without an appreciable risk of adverse health effects during a lifetime.
 

February 17, 2006

  PAGE  16 

April 2006	  										             PAGE  73 

This risk screen does not contain Clean Air Act Confidential Business
Information (CBI) and, therefore, can be disclosed to the public.

