EPA/600/R-01/093
October 2001
Capstone Report
on the Development of a Standard Test Method
for VOC Emissions from Interior Latex and
Alkyd Paints
by
John C. S. Chang
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
National Risk Management Laboratory
Air Pollution Prevention and Control Division
Indoor Environment Management Branch
Research Triangle Park, NC 27711
-------
Abstract
This document provides a detailed report on the small-chamber test method developed by
EPA/NRMRL for characterizing volatile organic compound (VOC) emissions from interior latex
and alkyd paints. Current knowledge about VOC, including hazardous air pollutant, emissions
from interior paints generated by tests based on this method are presented. Experimental data
were analyzed to demonstrate the usefulness of the method and test results in terms of emission
characterization, material selection, exposure assessment, and emission reduction by product
reformulation. The conclusions drawn from the experimental results were used as input to
develop a standard practice to be adopted by the American Society of Testing and Materials
(ASTM). The draft standard practice is presented in Appendix A.
11
-------
Table of Contents
Abstract ii
List of Tables vi
List of Figures viii
Acknowledgments x
Executive Summary and Conclusions xi
Chapter 1 Introduction 1
Chapter 2 Literature Survey 5
Paint Definition and Composition 5
Paint Manufacture and Use Statistics 7
Paint Emissions Exposure and Health Effects 9
Small Chamber Emissions Testing Procedures 13
Chapter 3 Standardized Test Methods for Characterizing Organic Compounds Emitted from
Paint Using Small Environmental Chambers 16
Test Facilities 16
Construction of the Small Chamber 17
Air Mixing 19
Surface Velocity 19
Clean Air Generation System 19
Temperature Control 20
Humidity Control 20
Lighting Control 20
Environmental Measurement and Control Systems 21
Automatic Systems 21
Sample Collection and Analysis 22
Principal Components of the "Standard Practice" for Testing Paint Emissions 25
Storing and Handling Paint Prior to Analysis 26
Analyzing Paint in Bulk (As a Liquid) 27
Selecting and Preparing a Suitable Paint Substrate 28
Applying Paint to a Substrate to Create a Test Specimen 29
Operating the Small Chamber 31
Sampling the Specimen's Gaseous Emissions 31
Using Instruments to Measure Chemicals Present in the Emissions Sample ... 34
Analyzing the Results of the Analytical Instruments 34
Reporting the Experimental Results 38
Conducting Quality Assurance/Quality Control 40
in
-------
Table of Contents (Continued)
Chapter 4 Characterization of Emissions of Volatile Organic Compounds from
Interior Alkyd Paint 42
Experimental Work 43
Bulk Product Analysis by GC/MS 44
Small Chamber Emission Test Methods 44
Sampling and Analysis Methods 46
Results and Discussion 48
VOC Content Determined by GC/MS 48
Emission Tests - Mass Balance Calculations 50
Emission Test Results 55
Conclusions 65
Chapter 5 Methyl Ethyl Ketoxime Emissions from Alkyd Paint 68
Experimental Work 69
Results and Discussion 70
MEKO Contents 70
Chamber Emission Data 71
MEKO Emission Model 74
Indoor Air Quality Impact Assessment 75
Exposure Reduction Assessment 78
Conclusions 79
Chapter 6 New Findings About Aldehyde Emissions from Alkyd Paint 81
Experimental Work 82
Results and Discussion 84
Bulk Analysis for Hexanal 84
Chamber Emission Data 85
Hexanal Formation Mechanism 91
Hexanal Emission Model 92
Indoor Air Quality Simulation 96
Conclusions 96
Chapter 7 Substrate Effects on VOC Emissions from a Latex Paint 99
Experimental Work 100
Results and Discussion 102
Paint Composition 102
Substrate Effects on VOC Emissions 104
iv
-------
Table of Contents (Continued)
Substrate Effects on Composition of VOC Emissions 108
Emission Models 108
Emission Mechanisms 110
Long-Term Emission Data 113
Conclusions 113
Chapter 8 Experimental Work to Evaluate Low-VOC Paints 116
Experimental Work 117
Results and Discussion 119
Bulk Analysis 119
Emissions 121
Performance Evaluation 126
Conclusions 127
Chapter 9 Experimental Work to Characterize and Reduce Formaldehyde Emissions from
Low-VOC Paint 129
Experimental Procedure 130
Experimental Data 131
Emission Model 133
Data Analysis 136
Source Investigation 140
Biocide Replacement 142
Conclusions 143
Chapter 10 References 144
Appendix A A Proposed Standard Practice for Testing and Sampling of Volatile Organic
Compounds (Including Carbonyl Compounds) Emitted from Paint Using Small
Environmental Chambers A-l
-------
List of Tables
2-1 Wall Paint Production by Type from 1997 to 2000 8
3-1 Quality Control Objectives 40
4-1 Concentrations of the Predominant VOCs in the Liquid Primer and Three Interior Alkyd
Enamel Paints 49
4-2 Matrix of Tests Performed in the Program 51
4-3 Percent of the Applied VOC Mass Recovered in Emissions During 2-week Small
Chamber Tests with Alkyd Primer and Paints 53
5-1 Recovery of MEKO 71
5-2 Estimated Values of Model Parameter and Goodness-of-Fit to the Chamber
Concentration Data 75
6-1 Information About the Primer and Three Test Paints 83
6-2 The Total Emissions for the Three Most Abundant Aldehydes (in mg/g paint) 86
6-3 Hexanal Concentration in Test Chamber for Alkyd Paint A-l 87
6-4 Hexanal Concentration in Test Chamber for Alkyd Paint A-2 88
6-5 Hexanal Concentration in Test Chamber for Alkyd Paint A-3 89
6-6 Estimated Model Parameters for Hexanal Formation from the Three Alkyd
Paints Tested 95
7-1 Determination of Volatile Organic Compounds in the Latex Paint
by ASTM Methods (ASTM, 1989) 103
7-2 Total and Individual VOCs Determined by GC Analysis Concentration Units (mg/g) . 103
7-3 Comparison of Peak Concentrations Measured In the Environmental Chambers
(in mg/m3) 104
7-4 Weight Percentage of VOC in the Latex Paint Emitted in the First 336-Hour Testing
Period 108
7-5 Summary of Parameters of the Double Exponential Model for VOC Emissions from
Painted Gypsum Board Ill
7-6 Summary of Parameters of the Double Exponential Model for VOC Emissions from
Painted Stainless Steel (R20 = 0) Ill
8-1 The Low-VOC Latex Paints Tested 118
8-2 ASTM Methods Used for Performance Testing 120
8-3 Results of Bulk Analyses (mg/g) 121
VI
-------
List of Tables (Continued)
8-4 Results of Performance Testing 126
9-1 Estimated Values of Parameters of Equation (9-4) 138
9-2 Comparison of Quantitative Measures of Goodness of Model with ASTM Criteria
(ASTM, 1995) 139
9-3 Estimated Amount of Formaldehyde in the Paint Applied 139
9-4 Calculated Formaldehyde Content in Three Paints Tested 142
vn
-------
List of Figures
3-1 Small chamber paint testing facility 18
4-1 Decane emissions from alkyd primer and Paint A-l for first 20 hours after application to
glass, gypsum board, or pine board substrates 56
4-2 Effect of primer on short-term decane emissions from Paint A-l applied to pine board;
data presented for first 20 hours after paint application 59
4-3 Comparison of TVOC emissions from three alkyd paints for first 20 hours after
application to a pine board previously coated with primer 61
4-4 Effect of wet film thickness on short-term decane emissions from Paint A-l during first
20 hours following application to pine board previously coated with primer 63
4-5 Effect of air exchange rate on short-term decane emissions during small chamber tests
with Paint A-l applied to pine board previously coated with primer 64
4-6 Long-term hexanal emissions from three paints applied to pine board coated 48 hours
earlier with alkyd primer 66
5-1 Chamber concentrations resulting from the MEKO emissions from Paint A-l 72
5-2 Chamber concentrations resulting from the MEKO emissions from Paint A-2 72
5-3 Chamber concentrations resulting from the MEKO emission from Paint A-3 73
5-4 Cumulative MEKO emissions for Paint A-2 (calculated from chamber concentration
data) 73
5-5 Comparison of the predicted test house MEKO concentrations with the suggested
indoor exposure thresholds 77
5-6 Predicted test house MEKO concentrations at high (3.0 h"1) air exchange rate 79
6-1 Hexanal emission from Paint A-l and modeling results 90
6-2 Hexanal emission from Paint A-2 and modeling results 90
6-3 Hexanal emission from Paint A-3 and modeling results 91
6-4 Predicted hexanal concentration in a typical house after alkyd paint application 97
7-1 Effect of substrate on ethylene glycol emissions 105
7-2 Effect of substrate on propylene glycol emissions 105
7-3 Effects of substrate on 2-(2-butoxyethoxy)ethanol emissions 106
7-4 Effect of substrate on TPM emissions 106
7-5 VOC composition in chamber air with stainless steel substrate
(relative abundance = weight percentage in TOC) 109
7-6 VOC composition in chamber air with stainless steel substrate
(relative abundance = weight percentage in TVOC) 109
7-7 Long-term VOC emissions from painted gypsum board 114
8-1 Comparison of TVOC emission profile of paint L-2 with that of paint 0 (a conventional
latex paint). The method detection limit for TVOC was 0.02 mg/m3 122
Vlll
-------
List of Figures (Continued)
8-2 Aldehyde emission profiles of paint L-l. The method detection limits was 0.0007 mg/m3
for formaldehyde and acetaldehyde, and 0.0014 mg/m3 for benzaldehyde 124
8-3 Aldehyde emission profiles of paint L-3. The method detection limit was 0.0007 mg/m3
for formaldehyde, acetaldehyde, and propanal, and 0.0014 mg/m3 for benzaldehyde. . 125
9-1 Comparison of chamber data with model predictions. Note that model predications for
paints with no biocide and with a different biocide are not differentiable 132
9-2 Schematic of the first-order decay in-series model 134
9-3 Fraction of formaldehyde remaining in the paint 141
IX
-------
Acknowledgments
The contents of this report are the results of a research program performed between 1995 and
1999. The author wishes to acknowledge the work of Mark Bero, Roy Fortmann, Huei-Chen
Lao, Angelita Ng, and Nancy Roache of ARCADIS Geraghty & Miller, Inc. for collecting part
of the experimental data. The author also wishes to acknowledge the assistance of Ray Merrill
and David Berol of Eastern Research Group on literature survey, information compilation, and
computer graphics.
-------
Executive Summary and Conclusions
Americans spend about 90% of their time indoors, where concentrations of pollutants are
often much higher than they are outdoors. It is not surprising, therefore, that risk assessment and
risk management studies have shown that indoor environmental pollution poses significant risks
to human health.
The U.S. Environmental Protection Agency (EPA) has evaluated a number of indoor
materials and products as potential sources of indoor air pollution under the Indoor Air Source
Characterization Project (IASCP). Interior architectural coatings, especially alkyd and latex
paints, were identified as potentially high-risk indoor sources by the Source Ranking Database
developed under the IASCP. EPA conducted a literature survey and found that there was a lack
of reliable and consistent paint emission data for developing and evaluating risk management
options. Further investigation showed that a standardized test method needed to be developed so
that testing laboratories, researchers, and paint manufacturers could generate and report emission
data that were complete, consistent, and comparable.
Between 1995 and 1999, EPA's National Risk Management Research Laboratory
(NRMRL) conducted a paint emission characterization research program. The program was
devoted to developing, verifying, and demonstrating a small chamber test method for the
measurement of volatile organic compound (VOC) and hazardous air pollutant (HAP) emissions
from alkyd and latex paints. The test method has been documented and submitted to the
American Society for Testing and Materials (ASTM) for adoption as a standard practice.
This report summarizes the resulting test method, presents new findings, and describes
the key results generated by NRMRL as it assessed emissions from alkyd and latex paints. The
report is divided into four parts. After introducing the study and providing background
information about existing literature on the subject paint emissions testing, the report describes
the developed standard test method for characterizing organic compounds emitted from paint. It
also describes the results of NRMRL's tests on alkyd and latex paints.
Standardized Test Method
The standardized test method addresses the following key issues:
Storing and handling paint samples prior to analysis
Analyzing paint in bulk (as a liquid)
• Selecting and preparing a paint substrate for testing
• Applying paint to a substrate to create a test specimen
Establishing and controlling test conditions
xi
-------
Sampling the VOC emissions from the painted specimen
Analyzing the samples with chemical instruments
Calculating emission rates/factors using experimental data
Conducting quality assurance/quality control
The core experimental apparatus employed by the standardized test method is a device
called a Small Environmental Test Chamber ("small chamber" for short). A test chamber is a
hollow box that may range in size from a few liters to 5 m3. The chamber used at NRMRL is 53
L (0.053 m3) in volume. Chambers with volumes greater than 5 m3 are defined as "large"—they
may reach the scale of an entire room. The small chamber, on the other hand, is an apparatus
suited to the spatial and financial constraints of a typical laboratory environment. It is also more
convenient to operate than a large chamber. An environmental chamber test facility designed and
operated to determine organic emission rates from paints should contain the following: test
chambers, clean air generation system, monitoring and control systems, sample collection and
analysis equipment, and standards generation and calibration systems. The purpose of these
components is to provide a controlled environment for conducting emissions testing that can
reflect common indoor air conditions.
The standardized test method includes a series of procedures and guidelines for preparing
a painted test specimen. Procedures for handling and storing the paint to be tested were
established to guard against the possibility of evaporative losses, stratification, and property
changes. A modified version of EPA Method 311 (40 CFR, 1996) was adopted for the bulk
analysis of paints, to facilitate the experimental design of the emissions test and the selection of
sampling and analytical techniques. Instead of traditional test substrates such as glass, stainless
steel, and aluminum, common indoor materials such as gypsum board and wood are
recommended in the method for creating realistic and representative testing samples. Either a
roller or a brush should be used to apply the paint to the substrate. A protocol was developed to
quantify the amount of the paint applied so that the emission data can be consistent and
comparable.
The "time zero" for the start of an emission test is established when the chamber door is
closed (immediately after placing the test specimen inside the chamber). The small chamber
should be operated to match the actual environmental conditions at which people paint the
interiors of houses. The standardized method guides investigators in setting up their sampling
protocols. The instructions help to ensure that investigators collect an adequate quantity of
chamber air samples on the appropriate sampling media. The method describes several kinds of
analytical instruments that can be used to determine the amounts and kinds of VOCs in the
collected sample. Data reduction techniques and an example of an emission model are included
in the method—it describes the mathematical procedures used to convert the analytical results
into emission rates and emission factors. In addition, the method provides guidelines for
reporting and quality assurance. These guidelines should help investigators compile their results
in a consistent and complete fashion that allows for comparison or repeat emissions testing of
similar or new architectural coatings.
xn
-------
Alkyd Paints
Alkyd paint continues to be used indoors because it has desirable properties such as
durability, gloss, gloss retention, and fast drying. NRMRL has employed the developed
standardized test method to conduct research that characterizes VOC emissions from alkyd paint.
NRMRL used the results of its paint emissions tests to develop source emission models. These
models, in turn, were used for the assessment of indoor exposure levels and risk management
options.
The first test series that NRMRL performed on alkyd paints was integrated into the
process of developing and validating its new standard practices for paint testing. The tests
involved one primer and three alkyd paints. Bulk analysis indicated that the alkyd primer and
two of the three paints tested contained more than 100 different VOCs, primarily straight-chain
alkanes, with decane and undecane being the predominant compounds. The third paint had more
branched alkanes. All four coatings contained low levels of aromatic compounds. The total VOC
content of the liquid paints ranged from 32% to 42%. Measurements of the total VOC levels in
the liquid coatings by gas chromatography/mass spectrometry (GC/MS) agreed well with
manufacturers' data.
Mass balance calculations were conducted to compare the bulk analysis results and
chamber emission data to evaluate the recovery. It was found that for total VOC, the majority
(greater than 80%) of the mass in the applied paint could be accounted for in the subsequent air
emissions. The data for the more abundant compounds (e.g., nonane, decane, and undecane) in
the paint suggest that there was a margin of error of ± 20% in measuring these recoveries.
Due to the relatively high VOC content and fast emission pattern, peak concentrations of
total VOC as high as 10,000 mg/m3 were measured during small chamber emissions tests with a
loading factor of 0.5 m2/m3 and an air exchange rate of 0.5 h"1. Over 90% of the VOCs were
emitted from the primer and paints during the first 10 hours following application.
A series of tests were performed to evaluate those factors that may affect emissions
following application of the coatings. It was found that the type of substrate (glass, wallboard, or
pine board) did not have a substantial effect on the emissions with respect to peak
concentrations, the emissions profile, or the mass of VOCs emitted from the paint. The
emissions from paint applied to bare pine board, a primed board, and a board previously painted
with the same paint were quite similar. There were differences among the emissions from the
three different paints, but the general patterns of these emissions were similar. The effect of
other variables, including film thickness, air velocity at the surface, and air exchange rate, were
consistent with theoretical predictions for gas-phase, mass-transfer-controlled emissions.
Results from the testing performed in this study are being used to develop computational
methods for estimating the emission rate of total VOCs from solvent-based coating products
used indoors. The database on total VOC emission from alkyd paint should also be useful for
others involved in model development and validation.
Xlll
-------
In addition to studying the effects of substrates and other environmental variables on total
VOC emissions, small environmental chamber tests were conducted to characterize the
emissions of a toxic chemical compound—methyl ethyl ketoxime (MEKO)—from three
different alkyd paints. The data resulting from these tests facilitated the development of a set of
risk management options for MEKO.
Methyl ethyl ketoxime, another name for 2-butanone oxime or ethyl methyl ketoxime
[CH3C(NOH)C2H5, CAS Registry No. 96-29-7], is often used by paint manufacturers as an
additive to interior alkyd paints (Weismantel, 1981; Turner, 1988). MEKO has been found to be
a moderate eye irritant (Krivanek, 1982). It was also the subject of a Section 4 test rule under the
Toxic Substances Control Act (Fed. Regist, 1986). A number of toxicological endpoints have
been evaluated by testing conducted under the test rule (Fed. Regist., 1989). MEKO
demonstrated carcinogenic activity in long-term inhalation studies, causing liver tumors in both
rats and mice.
MEKO acts as an anti-skinning agent (or anti-oxidant) that prevents oxidative drying or
skinning of the alkyd paint to improve stability in the can. Usually, the MEKO content in a paint
is less than 0.5% (Krivanek, 1982). Due to its relatively high volatility (its boiling point is only
152°C), the majority of the MEKO in the paint is expected to be released into the surrounding
indoor air after painting to allow the paint to dry properly on the painted surfaces. The effects of
MEKO emissions on indoor air quality (IAQ) and associated exposure risk depend on
characteristics such as emission rates and patterns.
Bulk analysis showed that the MEKO content in alkyd paints can be as high as several
mg/g. Material balance from the chamber tests indicated that the majority (greater than 68%) of
the MEKO in the paint applied was emitted into the air. MEKO emissions occurred almost
immediately after each alkyd paint was applied to a pine board. Due to the fast emission pattern,
more than 90% of the MEKO emitted was released within 10 hours after painting. The peak
concentrations of MEKO in chamber air correlated well with the MEKO content in the paint.
The chamber data were simulated by a first-order decay emission model that assumed
that the MEKO emissions were mostly gas-phase mass-transfer-controlled. The first-order decay
model was used as an input to the continuous-application source term of an IAQ model to
predict indoor MEKO concentrations during and after the application of an alkyd paint in a test
house. The predicted test house MEKO concentrations during and after the painting exceeded a
suggested indoor exposure limit of 0.1 mg/m3 for all three paints. The predicted MEKO
concentrations also exceeded the lower limit of a suggested sensory irritation range of 4 to 18
mg/m3 with two of the three paints tested. The elevated MEKO concentrations can last for more
than 10 h after the painting is finished. The model was also used to evaluate and demonstrate the
effectiveness of risk reduction options. These options involved selecting lower MEKO paints
and establishing higher ventilation levels during painting. The higher ventilation should be
maintained about 2 h after the painting is finished to avoid exposure to residual MEKO
emissions.
xiv
-------
In addition to total VOC and MEKO emissions, the unpleasant "after-odor" which can
persist for weeks after application of alkyd paint has been a cause of IAQ concerns. Three
different alkyd paints were tested in small environmental chambers to characterize the aldehyde
emissions. Emission data indicated that significant amounts of odorous aldehydes (mainly
hexanal) were emitted from alkyd paints during the air-drying period. Bulk analyses showed that
the alkyd paint itself contained no aldehydes. Mass balance calculations indicated that any
aldehydes emitted should have been produced after the paint was applied to a substrate. The
aldehydes emission patterns were consistent with the theory that the aldehydes were formed as
byproducts from spontaneous autoxidation of unsaturated fatty acids in the applied paint.
Chamber data showed that the major volatile byproducts generated by the drying of the alkyd
paints were hexanal, propanal, and pentanal. These results facilitated the development of an
exposure assessment model for hexanal emissions from drying alkyd paint.
The hexanal emission rate was simulated by a model that assumed that the autoxidation
process was controlled by a consecutive first-order reaction mechanism with an initial time lag.
The time lag reflects an induction period after painting during which little oxygen is taken up by
the alkyd coating. As the final byproduct of a series of consecutive first-order reactions, the
hexanal emission rate increases from zero to reach apeak and is followed by a slow decay. This
model was confirmed by chamber concentration data. The modeling results also showed that the
hexanal emissions were controlled mostly by the chemical reactions that formed intermediates
(i.e., the precursors to hexanal production).
An IAQ simulation that used the emission rate model indicated that the hexanal
emissions can result in prolonged (several days long) exposure risk to occupants. IAQ simulation
indicated that the hexanal concentration due to emissions from an alkyd paint in an indoor
application could exceed the reported odor threshold for about 120 hours. The occupant exposure
to aldehydes emitted from alkyd paint also could cause sensory irritation and other health
concerns.
Latex Paints
The majority (over 85%) of the interior architectural coatings used in the United States
are latex paints. Previous testing of latex paint emissions has focused on determining cumulative
mass emissions of VOCs. The purpose of previous testing was to assess the effect of these paints
on the ambient air and to determine how they contributed to photochemical smog (Brezinski,
1989). NRMRL's concern has been to estimate people's time-varying exposure to overall VOC
levels and to specific VOCs from indoor latex paints.
The first test series that NRMRL performed on latex paints was integrated into the
process of developing and validating its new standard practice for paint testing. NRMRL's small
chamber tests indicated that the organic emission patterns of latex paints are very different from
those of alkyd paints. Bulk analysis showed that the total VOC content of a commonly used latex
paint is usually in the range of 2% to 5%, which is considerably lower than that of alkyd paints
(32% to 42%). Instead of alkanes, alkenes, and aromatics, only several polar compounds such as
glycols, alcohols, and aldehydes were found in the latex paints.
xv
-------
The chamber test results showed significant differences between the emissions of the
same latex paint applied to two different substrates (a stainless steel plate and a gypsum board).
The amount of VOCs emitted from the painted stainless steel was 2 to 10 times greater than the
amount emitted from the painted gypsum board during the 2-week test period. After the first 2
weeks, over 90% of the VOCs were emitted from the paint on the stainless steel plate but less
than 20% had left the gypsum board. The dominant species in the VOCs emitted also changed
from ethylene glycol to 2,2,4-trimethyl-l,3-pentanediol monoisobutyrate when stainless steel
was replaced with gypsum board. Data analysis by a double-exponential model indicated that the
majority of the VOC emissions from the painted stainless steel could be simulated by an
evaporation-like phenomenon with fast VOC emissions controlled by gas-phase mass transfer.
On the other hand, only a small fraction of the VOCs emitted from the painted gypsum board
appeared to be controlled by the evaporation-like drying process. The majority of the VOCs
were emitted after the painted gypsum board surface was relatively dry. They were probably
dominated by a slow, solid-phase-diffusion-controlled mass transfer process. Long-term
experimental data indicated that it may take as long as 3.5 years for all the VOCs to be released
from the paint applied to the gypsum board.
The small chamber test results demonstrate that, when the objective of a test is to provide
emissions data that are relevant to understanding a paint's emissions behavior in typical indoor
environments, one should use "real" substrates such as wood and gypsum board instead of
"ideal" substrates such as glass, aluminum, or stainless steel. Proper choice of substrate is
therefore crucial for exposure and/or risk assessment studies involving indoor latex paints.
NRMRL also used the small chamber test method to evaluate a relatively new type of
interior architectural coating, the so called "low-VOC" latex paint. Low-VOC paint has been
used as a substitute for conventional latex paints to avoid indoor air pollution. Low-VOC latex
paints are promoted for use in occupied hospitals, extended care facilities, nursing homes,
medical facilities, schools, hotels, offices, and homes where extended evacuation of an entire
building section for painting would be particularly difficult or undesirable.
Four commercially available low-VOC latex paints were evaluated as substitutes for
conventional latex paints. They were evaluated by assessing both their emission characteristics
and their performance as interior wall coatings. Bulk analysis indicated that the VOC contents of
the four paints (which ranged from 0.01% to 0.3%) were considerably lower than those of
conventional latex paints (3% to 5%). EPA Method 24 (40 CFR, 1994) for determining VOC
content (commonly used by paint manufacturers) is not accurate enough to quantify the VOC
contents of low-VOC latex paints for quality control and product ranking purposes. Other
methods such as EPA Method 311 are more suitable, especially when individual VOC content
data are needed.
The fact that "low-VOC" paint had relatively low VOC emissions was confirmed by
small chamber emission tests. However, the experimental data also indicated that three of the
four low-VOC latex paints tested either had some inferior coating properties or emitted
hazardous air pollutants. Significant emissions of several aldehydes (especially formaldehyde,
which is a HAP) were detected in emissions from two of the four paints. ASTM methods were
xvi
-------
used to evaluate the paints' coating performance including hiding power, scrub resistance,
washability, drying time, and yellowing. The results indicated that one of the four low-VOC
paints tested showed performance equivalent or superior to that of a conventional latex paint
used as control. It was concluded that low-VOC latex paint can be a viable option to replace
conventional latex paints for prevention of indoor air pollution. However, certain paints
marketed as "low-VOC" may still emit significant quantities of air pollutants, including HAPs.
In addition, some of these paints may not have performance characteristics matching those of
conventional latex paints.
Due to the use pattern of low-VOC paints proposed by their manufacturers (i.e., partial
occupancy during painting and immediate re-occupation after painting), the intimate exposure of
sensitive occupants to the low-VOC latex paint emissions (especially to HAPs such as
formaldehyde) is of special concern. Long-term environmental chamber tests were performed to
characterize the formaldehyde emission profiles of a low-VOC latex paint. The formaldehyde
emissions resulted in a sharp increase of formaldehyde concentrations within the chamber, rising
to a peak followed by transition to a long-term slow decay. Environmental chamber data
indicated that formaldehyde emissions from a low-VOC latex paint can cause very high (several
ppm) peak concentrations in the chamber air. When the paint was applied to gypsum board, the
formaldehyde emissions decayed very slowly after the initial peak, and the emission lasted for
more than a month. The results of these tests allowed for the development of exposure
assessment emissions models to facilitate pollution prevention efforts to reduce the amount of
formaldehyde released by low-VOC paints.
A semi-empirical first-order decay in-series model was developed to interpret the
chamber data. The model characterized the formaldehyde emissions from the paint in three
stages: an initial "puff of instant release, a fast decay, and a final stage of slow decay controlled
by a solid-phase diffusion process that can last for more than a month. The semi-empirical model
was used to estimate the amount of formaldehyde emitted or remaining in the paint. It also
predicted the initial peak concentration of formaldehyde and the time necessary for the
formaldehyde to become depleted from paint. Once the activity patterns of building occupants
were defined, the model was used for exposure risk assessment.
Additional small chamber tests were performed to investigate the major sources of
formaldehyde in the paint. Through comparing emission patterns and modeling outcomes of
different paint formulations, a biocide used to preserve one of the paints was identified as a
major source of the formaldehyde emissions. Chamber test results also demonstrated that paint
reformulation by replacing the preservative with a different biocide for the particular paint tested
resulted in an approximately 55% reduction of formaldehyde emissions. However, since other
sources (e.g., additives and binders) of formaldehyde are present in the paint, biocide
replacement can reduce only the long-term emissions. Short-term generation of high
concentrations of formaldehyde remains a problem. Additional research is needed to identify
other potential sources of formaldehyde to completely eliminate formaldehyde emissions from
low-VOC paints.
xvn
-------
Overall Conclusions
A standard test method was developed to characterize the VOC, including HAP,
emissions from interior architectural coatings. The advantages of the developed method and the
usefulness of the experimental data it can generate were demonstrated by extensive tests focused
on two types of commercially available and commonly used interior architectural coatings: latex
and alkyd paints. The experimental data generated by this test method can be used to estimate
emission rates, to compare emissions from different products, to predict a paint's effects on IAQ
and exposure levels, and to evaluate the effectiveness of risk management options. The test
method can also be used as a pollution prevention tool to assist paint manufacturers in reducing
or eliminating VOC emissions from their products.
xvin
-------
i a p ter
Introduction
The purpose of the report is to present the results of the U.S. Environmental Protection
Agency (EPA) National Risk Management Research Laboratory (NRMRL) Indoor Source
Characterization Research Program between 1995 and 1999. The focus of the research program
during this period was on emissions of volatile organic compounds (VOCs), including hazardous
air pollutants (HAPs), from architectural coatings. Among architectural coatings, the focus of
research was mainly on interior paints. Experimental work was conducted at EPA's Small
Chamber Source Characterization Facilities by the Indoor Environment Management Branch of
the Air Pollution Prevention and Control Division.
Americans spend about 90% of their time indoors. Concentrations of pollutants are often
much higher indoors than outdoors. Risk assessment and risk management studies (U.S. EPA
1987, 1990) have found that indoor environmental pollution is among the greatest risks to human
health and have advised EPA to address this problem. In response to these studies, EPA
established the Indoor Source Characterization Research Program and Experimental Facilities in
the late 1980s. Initial research was conducted to characterize organic emission profiles of
common indoor materials and products to facilitate the identification of high-risk indoor sources.
In the early 1990s, EPA's Office of Pollution Prevention and Toxics (OPPT) developed
an Indoor Air Source Ranking Database (SRD) as part of the Indoor Air Source Characterization
Project (Cinalli et al., 1993). The objectives were to categorize product classes and to score and
rank product classes based on relative risks. The SRD scores could be used to assign priorities to
product classes for further data development and risk screening. Based on formulation data,
emission profiles, use patterns, and health hazard data, architectural coatings, mainly interior
paints, were ranked as one of the top priority, high-risk indoor air pollution sources.
1
-------
Between 1995 and 1999, NRMRL's paint emission characterization research program
was devoted to the development, verification, and demonstration of a standard test method for
measurement of organic emissions from the two types of commercially available, most
commonly used indoor architectural coatings - latex and alkyd paints. A literature survey
indicated that there was a lack of reliable and consistent paint emission data needed to develop
and evaluate effective risk management options to address exposure risks and health concerns.
Further investigation showed that a standard test method was required for testing laboratories,
researchers, and paint manufacturers to generate and report emission data that were complete,
consistent, and comparable.
The research program produced a standard method for testing organic emissions from
paints and six technical papers published in peer-reviewed journals (Chang et al., 1997; Chang et
al., 1998; Chang and Guo, 1998; Fortmann et al., 1998; Chang et al., 1999; and Chang et al.,
2001). A Standard Practice (Appendix A), based on the standard test method, was developed and
submitted to the American Society for Testing and Materials (ASTM) for adoption as the
national standard. In 2000, the research program won an Agencywide Pollution Prevention
Leadership Award (James Craig Award) for identifying a major source of formaldehyde
emissions and successfully preventing air pollution by those formaldehyde emissions with paint
reformulation.
This report is divided into 10 chapters and an appendix. Chapters 1 and 2 provide
background information:
• Chapter 1, Introduction, summarizes the paint emission research program and the
scope, purpose, and organization of this report.
• Chapter 2, Literature Survey, provides background information from selected
literature about interior paint consumption data, latex and alkyd paint properties,
exposure and health risk concerns, and the need for a standard test protocol.
Chapter 3 focuses on testing methods:
-------
Chapter 3, Standardized Test Methods, presents the developed standard test
method for characterizing organic compounds emitted from paint using small
environmental chambers. The coded Standard Practice is provided in the
Appendix. This chapter provides a detailed description of the test facilities and
the rationale of critical experimental procedures.
Chapters 4, 5, and 6 focus on alkyd paint:
Chapter 4, Characterization of Emissions of Volatile Organic Compounds from
Interior Alkyd Paint provides experimental data to demonstrate the usefulness of
the small chamber test method. Results are presented for a series of tests
performed to evaluate factors that may affect organic emissions from interior
alkyd paints.
Chapter 5, Methyl Ethyl Ketoxime (MEKO) Emissions from Alkyd Paint, shows
how the standard test method can be used to measure the emission rate of atoxic
chemical compound, MEKO, from three different alkyd paints. The chamber data
were used to establish an emission model which predicted indoor MEKO
concentrations during and after painting. The effectiveness of risk management
options, including selection of lower MEKO paints and higher ventilation during
painting, was also evaluated by the model.
Chapter 6, New Findings About Aldehyde Emissions from Alkyd Paint describes
the use of a small chamber test method to characterize the emissions of odorous
aldehydes that were produced by autoxidation reactions during the curing (drying)
process. This chapter also demonstrates how to employ the chamber data to
simulate indoor air quality impact and evaluate exposure risk.
Chapters 7, 8, and 9 focus on latex paint:
Chapter 7, Substrate Effects on VOC Emissions from a Latex Paint, discusses
how the small chamber test method was used to discover the significant effects of
two substrates — a stainless steel plate and a gypsum board — on the VOC
emission rates and patterns from a latex paint.
Chapter 8, Experimental Work to Evaluate Low-VOC Paints, reports the test
results of four commercially available low-VOC latex paints, evaluated as
substitutes for conventional latex paints.
Chapter 9, Experimental Work to Characterize and Reduce Formaldehyde
Emissions from Low-VOC Paint, illustrates how the Standard Practice was
employed to investigate the contributing sources of formaldehyde emissions from
a latex paint. This chapter also demonstrates that small chambers can be used as a
tool to conduct pollution prevention research to reduce formaldehyde emissions
by paint reformulation.
-------
and
• Chapter 10, References, lists the references cited in the report.
-------
Chapter L.
Literature Ourvey
This chapter summarizes scientific research pertaining to the measurement of paint
emissions and the evaluation of paints' effects on indoor air quality. It provides a context for
NRMRL's work on developing standard test practices and it illustrates the fact that these
practices met a pressing need within the community of indoor air scientists. The literature survey
begins by describing the content of paint; it then summarizes both the patterns of paint use in the
United States and the health problems that might result from that use. It concludes by describing
the state of small chamber paint emissions testing at the time when NRMRL began its research
into standardizing the practice.
r aint Definition and v-»o m p o s iti o n
Paint is a mechanical dispersion of pigments or powders with a liquid or solvent known
as a vehicle. The vehicle portion of the paint consists of a non-volatile portion and a volatile
portion. After application of a paint, the non-volatile portions of the vehicle (e.g., resin
polymers, film-formers, and binders) remain as a film on the coated surface, and the volatile
portion evaporates (U.S. EPA, 1997). In addition to pigments and vehicles, paints contain other
chemicals (called additives) that enhance their physical properties and make them easier to
apply. Wall paint vehicles can be alkyd (solvent/oil-based) or latex (water-based). A paint's
vehicle affects the ease with which it can be applied. Most paints are applied by brush, roller or
spray.
Solvents perform the following three functions:
They dissolve the chemicals that, when dried, form the coating's film.
-------
They dilute the paint solution or emulsion to achieve a proper solids content ratio
and viscosity.
• Their evaporation rate controls the paint's rate of drying.
Different types of solvents have varying characteristics:
Aliphatic hydrocarbon solvents are used in alkyd paints. These solvents are
petroleum fractions containing mostly aliphatic hydrocarbons with small amounts
of aromatic hydrocarbons. These solvents are commonly referred to as naphthas,
or mineral spirits. About 75% of the aliphatic hydrocarbons used in alkyd wall
paint are mineral spirits (U.S. EPA, 1997).
In the past, ethyiene giycoi was the organic primary solvent used in latex flat
paint. There is a current market trend towards replacing ethyiene glycol with
propylene glycol in both latex and alkyd paints. This substitution allows
manufacturers to develop formulations containing less organic solvent. These new
formulations, in turn, have lower and less hazardous organic compound
emissions.
Giycoi ethers and esters are fully soluble in water. They also aid the freeze-
thaw stability, coalescence, and wet-edge control in latex paints. They also make
it easier to spread paints on the surface to be coated. These compounds are
relatively volatile.
Aicohoi, ketone, and ester solvents are used in wall paints. Methyl ethyl ketone
and methyl isobutyl ketone are the most common compounds used (NPCA,
1992a).
es are aromatic hydrocarbons used in alkyd paints. Aromatics have been
largely phased out of wall paint use.
Other paint additives that can affect emissions of paint during application include
pigments, fillers, thickeners, plasticizers, surfactants, driers, stabilizers, and biocides. Any of
these additives can affect emissions through direct emissions or reaction with paint components
to produce volatile byproducts that evaporate into the air (U.S. EPA, 1997).
Interior latex paints are usually not considered to be in the same VOC category as alkyd
paints because they can be cleaned with water and are non-flammable. Nevertheless, latex paints
-------
contain volatile and semivolatile organic chemicals in the range of 3.5% to 9.5% (U.S. EPA,
1997). In general, latex paints are favored over alkyd paints indoors because they release much
less odor, cost less, dry faster, and are easier to clean up.
Alkyd paints are used indoors primarily for high-gloss applications or for situations in
which a particular surface or condition necessitates their use. For example, alkyd paint may be
needed when high moisture resistance is required to minimize mold growth. Alkyd paints are
used on a variety of substrates such as metal, plastic, wood, or glass.
r aint Manufacture and Use Otatistics
The total volume of alkyd and latex coatings used in the United States has remained
around 650 million gallons per year between 1998 and 2000 (U.S. Census, 2000, 2001). More
than 60% of this volume was paint applied to indoor surfaces (NPCA, 1992b; U.S. Census, 2000,
2001). Alkyd paint represents a small part (12%) of the total interior wall paint market. Latex
paint makes up over 87% of indoor wall paints produced in 2000.
Table 2-1 shows a breakdown of wall paint production between 1997 and 2000. Latex
flat paints comprised 98% of all flat wall paints. Latex semi-gloss paints comprise 90% of all
semi-gloss wall paints in 2000. For comparison purposes, in 1973, latex flat and latex semi-gloss
wall paints comprised 90% and 45% of the flat and semi-gloss wall paint market, respectively.
In 1981, those figures were 93% and 70%, respectively, and in 1992 those figures were 97% and
75%, respectively (U.S. Census, 2000, 2001).
-------
Table 2-1. Wall Paint Production by Type from 1997 to 2000
Interior lype (Quantity (thousands of gallons)
Alkyd
Latex
(solvent based)
Flat
Gloss
Semi-gloss
Primers and Sealers
Miscellaneous
(water based)
Flat
Semi-gloss
Primers
Miscellaneous
Stains/Sealers
Total Interior
2000
47,521
3,239
3,216
14,122
18,781
8,163
347,281
149,252
130,872
33,381
33,776
394,802
1999
51,181
3,322
3,409
12,355
19,503
12,592
339,225
150,931
125,002
31,581
31,711
390,406
1998
49,514
3,512
3,679
11,548
18,796
11,979
315,636
137,781
114,602
29,148
34,105
365,150
1997
52,432
3,727
6,067
11,803
18,111
12,724
335,035
149,115
120,635
26,131
39,154
387,467
Alkyd paint is composed primarily of organic solvents. Aliphatic and aromatic
hydrocarbon solvents make up about half the solvents in alkyd products. These solvents include
benzene, toluene, xylene, naphthas, and mineral spirits. The use of oxygenated solvents,
including ketones, alcohols, esters, glycol ethers, and glycols, is growing. Ketones (primarily
methyl ethyl ketone [MEK], methyl isobutyl ketone [MIBK], and acetone) comprise about 15 %
of the coatings solvents market (Markarian, 2000). Alcohols, such as ethanol, butanol, propanol,
and methanol, also hold about 15% of the market (Markarian, 2000).
While alkyd paint use is decreasing, emissions from alkyd paints still present a
significant potential health risk due to the types of volatile compounds that these emissions
contain. Per unit of volume, latex paints contain less hazardous volatile material than alkyd
-------
paints do. However, the volume of latex paints used nationwide each year is extremely large
(much larger than the volume of alkyd paints). Thus, widespread exposure to latex paint
emissions is still an important source of indoor air pollution. Even though paint composition and
consumption are moving away from alkyd paint and towards water-based (latex) paint, the
health effects from both types of paint are still under investigation.
Health effects assessments and environmental regulations are focused on the
hydrocarbons and hazardous air pollutants (HAPs) contained in both latex and alkyd paint. For
example, the hydrocarbons in paint are such significant contributors to photochemical smog
formation that California's South Coast Air Quality Management District (SCAQMD) developed
legislative standards for paint hydrocarbon content. These standards seek to eliminate 21.8 of the
estimated 59 million tons per day of hydrocarbon emission to the ambient air from paint in the
South Coast district (Sissell, 1999). At the federal level, the Architectural and Industrial
Maintenance (AIM) regulation took effect in 1999. It limits hydrocarbons in architectural
coatings and industrial maintenance applications (Markarian, 2000). Both hydrocarbon
reductions and HAP regulations are gradually becoming more stringent (Markarian, 2000).
r aint tmissions txposure and Health tffects
Indoor air exposure of painters and occupants to paint emissions is a health concern (U.S.
EPA, 1997). Conventional paints contain VOCs that vaporize, dispersing into the air we breathe.
Exposure to VOCs can result in irritation of the eyes, nose, and skin; respiratory problems;
headaches; nausea; and dizziness (Pennybaker, 1999; Wieslander et al., 1999). Many of the
VOC compounds in paint (such as benzene, formaldehyde, toluene, and xylene) are hazardous.
Some of these VOCs are carcinogens or neurotoxins (Pennybaker, 1999; U.S. EPA, 1993).
People are exposed to indoor air pollutants for 90% of the time. They frequently
complain about their health or well-being after moving into new or remodeled buildings.
However, it can be difficult to interpret the public health significance of these reports since no
federal or state air quality standards exist to limit the exposure of pollutants in non-industrial
-------
indoor air. Health threshold limits have not been set for long-term exposure to compounds
found in paint emissions. For this reason, the types of compounds and their emission
concentration are important to the long-term study of indoor health effects of wall paint
exposure.
In Germany, 2 months after a renovated building received occupants, a total VOC
concentration of 2,000-3,000 |_ig/m3 was measured. After 10 months, the concentration decreased
to 900-1,300 i-ig/m3 due to intense ventilation (Pitten et al., 2000). It was suggested that even this
lower concentration could pose serious threats to well being and health risk to vulnerable
occupants such as children with asthma and allergies. Since measurement methods for total
VOCs (TVOCs) are not standardized, different measurements of the same emissions can produce
significantly different values. Therefore, interpretation of these measurements should proceed
only with great caution.
Emissions from newly painted building interiors have been associated with respiratory
inflammation, asthma, and eye irritation. Many VOCs, such as terpenes and formaldehyde, are
often present in the indoor air of new buildings. However, it is difficult to determine which of
these chemicals are causing specific health effects because of the confounding effects of other
contemporaneous construction, such as new woodwork and floor coverings. Wieslander and
coauthors, from the Swedish Department of Occupational and Environmental Medicine, have
reported that these confounding effects prevented them from determining if the solvent-free
water-based paints used in renovating and reconstructing a building contributed to the health
effects identified in the building's occupants (Wieslander et al., 1999). NRMRL research
provides the standard procedures to characterize both the composition and concentration of paint
emissions over time. With validated models from NRMRL, health researchers can disentangle
the effects of paint emissions from other indoor air emission sources.
Asthma has been reported as a health concern associated with VOC emissions from
newly painted indoor surfaces (Wieslander et al., 1997). Wieslander measured exposure to
formaldehyde and VOCs at 62 dwellings. The relationships among exposures, asthma, and
10
-------
clinical signs were calculated by multiple linear or logistic regression, adjusting for possible
influence of age, gender, and tobacco smoking. The prevalence of asthma was elevated among
subjects with domestic exposure to newly painted surfaces, particularly newly painted wood
details and kitchen painting. A significantly increased prevalence of symptoms related to asthma
was also observed in relation to workplace exposure to newly painted surfaces. The indoor
concentrations of aliphatic compounds (C8-Cn), butanols, and 2,2,4-trimethyl 1,3-pentanediol
diisobutyrate (TPD also known as TXIB) were significantly elevated in newly painted dwellings.
The total indoor VOC concentration was about 100 i-ig/m3 higher in dwellings that had paint
coats less than a year old. A significant increase in formaldehyde concentration was observed in
dwellings with newly painted wood details. These results indicate that exposure to chemical
emissions from indoor paint is related to asthma and that some VOCs may cause inflammatory
reactions in the airways (Wieslander et al., 1997).
While many health professionals have focused on symptoms reported by painters and
occupants in new or renovated spaces, EPA has investigated the potential risks associated with
individual pollutants in the emissions from paint. Using paint emission test results and models
from NRMRL, EPA has found that:
Based on emission chamber work, the approximate 8-hr time weighted average
(TWA) for TVOC levels can exceed 2,000 mg/m3.
Xylene emissions from some paint samples approached the exposure levels at
which neurologic effects are often seen.
The high solvent exposures indicated by TVOC measurements from chamber
studies lead to concern about the possibility of chronic central nervous system
(CNS) effects in professional painters.
TVOC exposure levels indicate a high likelihood of complaints about indoor air
quality during and shortly after painting.
For MEKO, the risk posed by the Maximum Occupational Exposure (MOE)
during typical use presents concerns regarding developmental toxicity health
effects.
11
-------
The risk of cancer among consumers and professional painters from the inhalation
of MEKO during paint application and drying is a concern (U.S. EPA, 1997).
For latex paints, EPA determined that acute and chronic risks exist for professional
painters based on exposure to latex paint emissions. They found that:
TVOC levels are in the range that may result in complaints about indoor air
quality. All samples and test conditions in the chamber studies resulted in TVOC
levels exceeding 40 mg/m3 as an approximate 24-hr TWA.
Some latex paint is the source of formaldehyde and acetaldehyde. Neither the
exact concentrations of these chemicals nor the exact reason these chemicals are
emitted from latex paint had been well characterized by studies completed before
1993.
Acetaldehyde exposure presents a chronic health risk.
Acetaldehyde exposure also raises possible concern for cancer risk with risk
estimates around 10~4 based on current potency estimates.
Formaldehyde exposure raises a marginal concern for cancer risk to professional
painters. Risk estimates based on unit risk values using data from emissions tests
are around 10'6to 10'4.
Formaldehyde also presents an acute irritation concern for both consumers and
professional painters.
EPA recognizes that the paint industry has been in a state of flux and that it is lowering
the VOC content of paint—mainly due to existing regulations that aim to reduce ground-level
ozone levels. The marketplace has made a major switch from solvent- to water-based paint over
the last 30 years. In addition, the use of mercury as the biocide of choice in paints has been
discontinued, based partly on emission tests done by NRMRL (Tichenor et al., 1991). Industry is
aware of indoor air problems—for example, some companies have developed and marketed "no-
VOC" latex paints that have less odor. Although industry has made significant improvements to
address EPA's concerns, not all the factors in the risk analysis have been addressed, and further
research on emissions reduction and health effects is still necessary.
12
-------
Omall v-»hamber tmissions lestincj procedures
A number of studies have indicated the need for standard test practices for paint
emissions. Tichenor and Sparks describe how to manage exposure from indoor air pollutants in
residential and office environments. They find that intelligent source management requires
knowledge of the sources' emissions characteristics, including their chemical compositions,
emission rates, and emission decay rates. As an example, Tichenor describes several scenarios
that cover full-time occupancy, part-time occupancy, room ventilation, and room size. For each
of these scenarios, Tichenor predicts exposure of occupants to paint emissions. Tichenor
concludes that individual exposure to indoor air pollutants is affected by several parameters,
including source emission characteristics, source use, occupancy patterns, and ventilation
options. Standardized measurement of these parameters is necessary to choose source
management options that effectively decrease exposure (Tichenor and Sparks, 1996).
Saarela, in a review of papers describing the emissions from building construction
materials, shows that measurement and evaluation of emissions from indoor materials are more
meaningful when there are published guidelines for testing and policies for the use of emissions
data. According to Saarela, chamber test methods are becoming more important for material
classification, product development, and marketing. Therefore, it is important to develop and
standardize these measurement methods to ensure that emission measurements, regardless of
where in the world they are made, are comparable. There is a need to define all of the testing
parameters and analytical procedures in detail so as to generate reliable emissions test data.
These data, in turn, are required for developing selection guidelines for building material
(Saarela and Sandell, 1991).
Levin first reported on the use of emissions testing for product selection during the
design of a large office building (Levin, 1987a, 1987b, 1987c). In 1988, neither environmental
chamber testing nor any other product evaluation method provided more than limited data. Even
where test data were available, differing methodologies, the testing of non-comparable
specimens, or changes in product formulations decreased the data's value for product selection
13
-------
purposes. Levin points out that environmental chamber test results can be used by manufacturers
during product development to determine the potential magnitude, composition, and health
effects of exposure to their products' emissions. Therefore, Levin found that a standard protocol
was needed to guide the emissions testing used for product development, pollution prevention,
and exposure assessment programs. Levin presents the basic requirements for a well-designed
small environmental chamber, including the pros and cons for several design and monitoring
options. He also presents the major obstacles to adequate product documentation and the
comparability of test results (Levin, 1989a and b).
In 1991 Levin reviewed the protocols, applications, and problems associated with using
environmental chambers to determine VOC emissions from indoor air pollutant sources. His
review covered the efforts of NRMRL to develop a small chamber test method that would
standardize test procedures. It also covered the consensus American Society for Testing and
Materials (ASTM) standard D-5116-90 that NRMRL developed in cooperation with ASTM.
Levin compared the small chamber method development work performed by various groups to
the large-chamber and whole-house approaches used to characterize emissions from a variety of
home construction and decoration materials. Standards development efforts for low- or reduced-
VOC products are burdened by the large differences in testing requirements that exist within this
diverse group of products. In his review he concludes that the lack of standardized test protocols
makes data interpretation and comparison of test results difficult for most product types (Levin,
1991).
Chemical reactions further complicate the procedures for using small chambers to
evaluate paint emissions. Many chemical and physical transformations or reactions can occur,
both while the paint is stored in the container and after it is applied to a surface. In-can
preservatives and freeze/thaw protectants are incorporated into many paint formulations for
quality control purposes. So-called "natural" paints based on casein have been known to
deteriorate in the can, resulting in a "spoiled milk" odor after application. Homogeneous and
heterogeneous reactions occur after the paint is applied to building interior surfaces. Once there,
the change in the composition of the paint due to the evaporation of the volatile contents and the
14
-------
increased concentration of the non-volatile components creates a new chemical balance among
the remaining constituents with potential for chemical reactions. One example of this is the large
burst in emission of hexanal from latex paints 24 hours after application, as seen in the research
described in detail later in this report (Chang and Guo, 1998b).
15
-------
Chapters
Standardized Test Methods for Characterizing Organic Compounds
Emitted from Paint Using Small Environmental Chambers
A set of experimental methods was developed for testing the emissions properties of
household indoor paints. Those experimental methods have been documented and submitted to
ASTM for adoption as a standard practice. "A Proposed Standard Practice for Testing and
Sampling of Volatile Organic Compounds (Including Carbonyl Compounds) Emitted from Paint
Using Small Environmental Chambers" is included in the appendix of this report. This chapter
describes the principal elements of the standard practice and provides detailed information about
the background and the rationale of those methods. The goal of the standard practice is to
generate quantitative and comparable emissions data. These data, in turn, maybe used for:
Selecting paints for specific purposes
Establishing paint specifications
Conducting quality control checks of paint
Developing new low-emissions paints
Supporting any future paint labeling programs
Conducting indoor air quality analyses
Conducting exposure risk assessments
Test Facilities
The purpose of this section is to provide an overview of the test facilities used for
evaluating paint emissions. The design principles and major functions of critical components of
the test facilities are also discussed. The small environmental chamber system employed by
NRMRL is used as an example. Currently there are no commercially available chamber systems,
but investigators can design their own test facilities by following the same principles to suit their
16
-------
own needs. More technical discussions on small environmental chamber systems can be found in
ASTMD5116-97.
The core experimental apparatus with which NRMRL conducts its paint emissions
testing is a device called a Small Environmental Test Chamber ("small chamber" for short). A
test chamber is a hollow box that may range in size from a few liters to 5 cubic meters. The
chamber used at NRMRL is 53 L (0.053 m3) in volume. Chambers with volumes greater than 5
m3 are defined as "large"—they may reach the scale of an entire room. The small chamber, on
the other hand, is an apparatus suited to the spatial and financial constraints of a typical
laboratory environment. It is also more convenient to operate than a large chamber.
An environmental chamber test facility, designed and operated to determine organic
emission rates from paints, should contain the following: test chambers, clean air generation
system, monitoring and control systems, sample collection and analysis equipment, and
standards generation and calibration systems. Figure 3-1 illustrates a system with two test
chambers.
Construction of the Small Chamber
Small environmental test chambers should have non-adsorbent, chemically inert, smooth
interior surfaces so as not to adsorb or react with compounds of interest. Care must be taken in
their construction to avoid the use of caulks and adhesives that emit or adsorb VOCs. As an
example, electropolished stainless steel and glass may be used for interior surfaces. The chamber
must have an access door with air-tight, non-adsorbent seals. The chambers must be fitted with
inlet and outlet ports for air flow. Ports for temperature and humidity probes may also required.
Ports for sample collection are needed only if the sampling is not conducted in the outlet air.
17
-------
OO
PUMP
H
COMPRESSOR
MASS
FLOW
CONTROLLER
DRYER
SORBENT
TRAP
O O
SAMPLING
MANIFOLD
0
O
0-
SAMPLING
MAN FOLD
CATALYTIC
OXIDiZERS
*• Mixing Fan
TEST CHAMBER
source
Mixing Fan
TEST CHAMBER
source
^^H
I CU BATOR
MANIFOLD
dry
FT
wet dJ ,
Mass Flow Controllers
wet
ENVIRONMENTAL
CHAMBER COMPUTER
(for control and monitoring
of relative humidity,
temperature, and flow)
WATER FILLED
IMPINGERS
CONTROLLED
TEMPERATURE
WATER BATH
Figure 3-1. Small chamber paint testing facility.
-------
Air Mixing
The chamber and its air-moving components need to be designed to ensure good mixing
of the incoming air with the chamber air. While contaminant concentration gradients are expected
to exist in the chamber, particularly in the boundary layer near the emissions source, the mixing
issue concerns only the uniformity of the concentration in the bulk air inside the chamber. The
use of mixing fans and multi-port inlets and outlets are two techniques that have been used
successfully to ensure adequate mixing of air in the chamber.
Surface Velocity
The air's velocity near the surface of the material being tested can affect the mass transfer
rate. Thus, sources with evaporative emissions (e.g., wet paints) are tested under typical indoor
velocities (for example, 5-10 cm/s). A small fan can be used to achieve such velocities. A diffuser
should be used to eliminate the calm spot downstream of the fan hub. Velocity measurements can
be made with hot wire or hot film anemometers.
Clean Air Generation System
Clean air must be generated and delivered to the chambers to provide adequate
background for conducting tests. A typical clean air system might use an oilless compressor
drawing in ambient air followed by removal of moisture (for example, using a membrane dryer)
and trace organics (for example, by catalytic oxidation units). Other options include gas cylinders
or charcoal-filtered outdoor or laboratory air. If granular media (for example, charcoal) are used
for control of organics, a particle filter should be used downstream to remove particulate matter.
Calculations should be performed to determine the amount of air flow required before a decision
is reached on the supply system. For most sources to be tested, extremely clean air is needed.
Inlet concentrations should not exceed 2 i-ig/m3 for any single compound or 10 |_ig/m3 for the sum
of all VOCs. The purity of the air should be verified by routine analysis of background air
samples from a clean chamber.
19
-------
Temperature Control
Temperature can be controlled by placing the test chambers in incubator cabinets or other
controllable constant-temperature environments. The temperature of the inlet air can be controlled
by using conditioning coils.
Humidity Control
The humidity of the chamber air is controlled by adding deionized water or high-
performance liquid chromatography (HPLC) grade distilled water to the air stream. Injection by
syringe pumps followed by heating to vaporize the water can achieve desired humidity levels.
Other types of pumps (for example, HPLC) might also provide sufficient accuracy. The chamber
air can be humidified by bubbling a portion of theairstream through deionized water at a
controlled temperature (for example, in a water bath). The saturated air is then mixed with dry air
to achieve the desired humidity. Steam humidification can also be used. Coiled lines inside the
constant-temperature environment can be used to equilibrate the inlet temperature before delivery
to the test chambers.
Lighting Control
Small chambers are normally operated without lights. If the effect of lighting on emissions
is to be determined, appropriate interior illumination should be provided. If the heat generated by
lighting is a concern, a transparent chamber lid with a light source placed outside the chamber
may be an option.
20
-------
Environmental Measurement and Control Systems
Measurement and control are required for air flow, temperature, and humidity. Air flow
can be automatically monitored and controlled by electronic mass flow controllers, or
manualflow control (for example, a needle valve or orifice plate) and measurement (for example,
a bubble meter orrotameter) can be used. Temperature can be measured automatically via
thermocouples or thermistors; manual dial or stem thermometers can also be used. Humidity
control depends on the humidification system employed. If liquid injection is used, water flow is
controlled by the pump setting. Control of humidity by saturated air requires temperature
control of the water and flow control of the saturated air stream. Humidity can be measured by
several types of sensors, including dew point detectors and thin-film capacitors.
Temperature and humidity sensors should be located inside the chamber at least 5 cm from the
inside wall and near the midpoint between the air inlet and outlet ports.
Automatic Systems
Microcomputer-based measurement and control systems can be used to set air flow rates
and monitor temperature, relative humidity, and air flow during the experiments. Analog signals
from temperature, relative humidity, and flow sensors are converted to digital data that can be
stored by a microcomputer-based system, then processed to engineering units using appropriate
calibration factors. In this way, chamber environmental data can be continuously monitored, then
compiled and reduced for archival storage or display with minimal operator effort. Automatic
systems are also capable of certain control functions. Digital signals can be output to control
valves or converted to analog signals and sent out as setpoint signals to mass flow controllers. A
graphics overlay program can be used to show current setpoints and measured values on a system
schematic displayed on the microcomputer's monitor.
21
-------
Sample Collection and Analysis
Indoor sources of organic emissions such as paints vary widely in both the strength of
their emissions and the type and number of compounds emitted. Differences in emissions rates of
several orders of magnitude among sources are not unusual. To characterize organic emissions
fully, the sample collection/analysis system must be capable of quantitative collection and
analysis of volatile, semi-volatile, polar, and non-polar compounds. Any small chamber sampling
and analysis technique or strategy developed must consider the emission characteristics of the
specific source being evaluated. The design and operation of sample collection and analysis
systems must be appropriate for the organic compounds (and their concentrations) being
sampled. Such systems generally include sampling devices, sample collectors, and analytical
instruments. The remainder of this section discusses the alternatives available for small chamber
sampling and analysis of organic emissions.
Sampling Devices
The exhaust flow (for example, chamber outlet) is normally used as the sampling point,
although separate sampling ports in the chamber can be used. A multiport sampling manifold can
provide flexibility for duplicate samples. A mixing chamber between the test chamber
and the manifold can be used to permit addition and mixing of internal standard gases with the
chamber air stream (note that the effective chamber volume should be the sum of the test chamber
and the mixing chamber or manifold). Sampling ports with septums are needed if syringe
sampling is to be conducted. The sampling system should be constructed of inert material (for
example, glass, stainless steel), and the system should be maintained at the same temperature as
the test chambers. The exhaust from the sampling system should be ducted into a fume hood,
ensuring that any hazardous chemicals emitted by the test materials are isolated from the
laboratory environment.
Samples can be drawn into gas-tight syringes, gas chromatograph (GC) sampling loops,
evacuated canisters, or through sorbent cartridges using sampling pumps. Gas-tight syringes and
22
-------
closed loops are frequently used when chamber concentrations are high and sample volumes must
be small to prevent overloading the analytical instrument. Larger volume samples can be pulled
through sorbent cartridges using sampling pumps. Flow rate can be controlled by an electronic
mass flow controller (MFC) or other means. The sampling flow rate should be regulated to
prevent instabilities in the chamber system flow. Generally, this will require that the sampling
flow rate be limited to <50 % of the chamber flow rate. Valves and a vacuum gauge may be
incorporated into the system to permit verification of system integrity before samples are drawn.
The entire system can be connected to a programmable electronic timer to permit unattended
sample collection.
Sample Collectors
Selection of appropriate sample collectors will depend on factors such as boiling point,
polarity, and concentration ranges of the compounds of interest, as well as the amount of water
vapor in the sample airstream. No single sample collection, concentration, and delivery system
will be adequate for all analytes of interest, and the user must understand the limitations of any
system used to characterize source emissions. If the sample is collected by way of syringe or
closed-loop sampling, it is injected directly into an instrument for analysis. Collection in a
sampling bag (for example, Tedlar) or vessel (for example, glass, stainless steel) allows larger
samples to be collected and subsequently concentrated and analyzed. For many small chamber
evaluations of indoor paints, the fact that some of the compounds of interest are present only at
low concentrations requires the concentration of large-volume samples and collection on an
appropriate adsorbent medium. Several sorbent materials are available for use, singly or in
combination, including activated carbon (charcoal), glass beads, Ambersorb, Tenax
(polyphenylene oxide), graphitized carbon, and XAD-2. The selection of the sorbent (or sorbent
combination) depends on the compound(s) to be collected. XAD-2 resin can be used to collect
compounds considered to be semi- or non-volatile (i.e., boiling points above 18CTC). If sorbent
collection is used, the laboratory must be equipped with appropriate storage capabilities. Air-tight
glass tubes or chemically inert bags are both appropriate. Flushing the storage containers with
23
-------
high purity nitrogen prior to use will help ensure their cleanliness. Samples should be stored in a
freezer at -20 °C.
When sorbents are used for sample collection, desorption and concentration are necessary.
For example, a clamshell oven can be used to thermally desorb sorbent (e.g., Tenax) cartridges
with the vapors fed to the concentrator column of a purge and trap concentrator that thermally
desorbs the organic compounds. Another example is solvent extraction and dilution of samples
collected by charcoal cartridges. If possible, sorbent samples destined for thermal desorption
should be processed and analyzed within 48 h of collection. Samples requiring solvent extraction
should be processed within 7 to 14 days and analyzed within 40 days of extraction.
Organic Analysis Instrumentation
Several analytical instruments are available for determining the concentration of the
organics sampled from the chamber: GCs are the most commonly used. GCs have a wide variety
of columns available for separating organic compounds. Gas chromatography separates materials
based on the differences between their vapor pressures. Compounds "distill" from a GC column
based on how readily they remain in the vapor phase. Capillary columns are generally preferred
because they allow more opportunity for compounds to vaporize, condense, and separate from a
mixture before exiting the column to the detector. Several detectors can be used depending on the
purpose of the test and the compounds of interest. Mass spectrometers (MS) are the most versatile
and can be used in the scan mode to identify unknown compounds. When used in the scan mode,
a conventional MS has a sensitivity of about 10~9 g. An ion trap may have a sensitivity
approaching 10"12g in the scan mode. If conventional MS is being used to analyze known
compounds, it can be operated in the selective ion mode where its sensitivity increases to 10~12 g.
MS can be made even more sensitive via negative chemical ionization. Flame ionization detectors
(FIDs) are also widely used. They respond to a wide variety of organic compounds and have a
sensitivity of 10"11 g. Electron capture detectors (BCD) are used for analyzing electronegative
compounds (for example, halogenated organics) and have a sensitivity of 10~13 g.
24
-------
Some compounds are not easily measured with GCs; for example, low molecular weight
aldehydes. Analysis of these compounds requires other instrumentation. High Performance
Liquid Chromatography (HPLC) is one technique used to analyze compounds that are very polar,
thermally unstable, or of a high molecular weight. HPLC separations are based on the differential
solubility of compounds in a sample. Generally the compounds are derivatized before injection.
Once a sample is injected into the HPLC column, compounds dissolve into the solvent pumped
through the column. Solubility can be increased by automatically changing the solvent mixture.
A variety of packed solid-phase columns are available for separation of different classes of
compounds by HPLC. HPLC can be combined with any of several detector types including
ultraviolet (UV), visible (VIS), and infrared (IR) spectrometers. Much like gas chromatography,
data collected from HPLC analysis using these optical detectors consist of the specific retention
time of a compound (which indicates its identity) and the specific detector response (which
indicates the quantity of compound in the sample). HPLC can also be coupled with a mass
spectrometer (MS) to generate coordinated retention time, concentration, and molecular
fingerprint information.
Standards Generation and System Calibration
Calibration gas may be added to the test chamber or sampling manifold from permeation
ovens, gas cylinders, or dilution bottles. Calibration (or tracer) gas is added through the test
chamber in tests to determine chamber mixing, check for leaks, or to evaluate chamber "sink"
effects (i.e., the interactions between pollutants and chamber surfaces). Internal standards for
analysis quality control maybe added at the head of the sampling system. The internal standard
should not be added to the chamber due to the potential for adsorption on the material being
tested. Quality control can also be achieved by spiked samples.
Principal Components of the "Standard Practice" for Testing Paint Emissions
The purpose of this section is to provide a detailed description of the experimental
procedures included in the standard practice developed by NRMRL for measuring the emission
25
-------
rates of various organic compounds from indoor paints. The practice is applicable to both alkyd
and latex paints and primers. The following laboratory procedures are covered in the practice:
Storing and handling paint prior to analysis
Analyzing paint in bulk (as a liquid)
Selecting and preparing a suitable paint substrate
Applying paint to a substrate to create a test specimen
Operating the small chamber
Sampling the specimen's gaseous emissions
Using instruments to measure chemicals present in the emissions sample
Analyzing the results of the analytical instruments
Reporting the experimental results
Conducting quality assurance/quality control
Storing and Handling Paint Prior to Analysis
The purpose of this procedure is to ensure that the origin and history of each paint tested
are clearly documented and that the properties of each paint remain constant throughout the
testing process.
To ensure that paint is properly documented, NRMRL recommends that, upon acquiring a
new batch of paint, the investigator should record pertinent receipt information on a label
attached to the paint. The receipt information includes:
The date of acquisition
The source of the paint
The manufacturer and lot number
Repeatedly opening a large can of paint and extracting small samples from it allows
organic compounds to escape from the paint over time. To guard against the possibility of the
paint substantially changing in composition while it is being stored for analysis, the investigator
should open the can of paint only once and then immediately divide the paint into aliquots. Each
of these aliquots should be intended for a single application of paint to a sample. Although this
26
-------
procedure does not altogether eliminate the problem of uncontrolled vapor loss, it minimizes it
and standardizes it across all samples.
It is essential to ensure that the paint within a new can is homogenous before it is divided
into aliquots. Over time, unshaken paint may separate into layers with different chemical
compositions. Before opening a new paint can and drawing aliquots, the investigator should
homogenize the paint; e.g., using a commercial paint shaker. Stainless steel mixing balls may be
added to each aliquot to aid in the re-homogenization of the sample immediately before use.
NRMRL investigators store paint and primer aliquots in clean, amber glass vials. The caps
of these vials are lined with Teflon®. The vials, once filled, are stored at room temperature, away
from light. The investigators take care to use them before the expiration of the shelf-life specified
by the paint manufacturer. These procedures guard against the deterioration of the paint while it is
in storage at the laboratory.
Analyzing Paint in Bulk (As a Liquid)
As a step preliminary to the design and implementation of an emissions test, the bulk
composition of the paint in question may be analyzed. NRMRL has adopted a slightly modified
version of Method 311 (1996) as its protocol for bulk analysis of paints. NRMRL chose this
method because it is a standard EPA method, it has been developed to be flexible, and it provides
consistent, comparable, and reproducible results with known performance for a wide variety of
materials. NRMRL investigators perform this analysis on an aliquot of every paint investigated in
its small chamber research program. By performing a bulk organic compound analysis of every
candidate paint, the research program is assured of an initial chemical composition analysis that is
consistent, realistic, and comparable to other paint materials tested in the program. The purpose
of this analysis is to ascertain the identities and relative abundance of the chemicals likely to be
emitted from the paint into the air. The concentrations of VOCs in the bulk paint can also be used
with an emissions source model to predict VOC concentrations in the small chamber during
testing. All of this information is useful in the design of the emissions test because it gives the
27
-------
investigators choices of the kinds of sampling and analytical techniques that can be used. Proper
sampling techniques, for example, vary depending upon the kinds and concentrations of
chemicals one expects to be present.
Investigators proceed with this method by diluting an aliquot of paint with an appropriate
solvent and centrifuging the mixture to remove solid matter. Potentially suitable solvents for this
step include dimethylformamide, methylene chloride (for alkyd paints), or acetone or acetonitrile
(for latex paints). The investigator may need to perform several initial tests to ascertain which
solvent most effectively dilutes the paint. The interested reader may refer directly to Method 311
(1996) for more specific details on procedures and quality control checks for analyzing the
supernatant.
A portion of the supernatant is injected into a GC/MS. The investigator's choice of GC
column and operating procedure depends on his or her initial assumptions about what kind of
chemicals are present in the supernatant. One general rule serves as a foundation for these
assumptions and is particularly noteworthy. It is that alkyd primers and paints primarily contain
relatively non-polar aromatic and aliphatic VOCs, while latex paints contain a significant amount
of polar compounds (e.g., glycols, glycol ethers, and alcohols).
Selecting and Preparing a Suitable Paint Substrate
The purpose of NRMRL's procedures for selecting and preparing paint substrates is to
provide a realistic, representative, and adequate substrate for painting. Traditional substrates, such
as glass, stainless steel, and aluminum plates, are not realistic and representative for interior
paints. NRMRL suggests that investigators use common indoor materials such as gypsum board
and wood. A substrate should be sufficiently large to generate adequate loading, it should not be
contaminated, and its edges should be sealed (with sodium silicate or Teflon® tape) to minimize
edge effects.
28
-------
The investigator should procure substrate material and cut it to an appropriate size (which
depends on the loading rate desired in the small chamber). For example, a 16 by 16 cm square of
substrate (a 0.0256 m2 square) would provide a loading of 0.48 m2/m3 for a small chamber with a
volume of 0.053 m3. The surface of the substrate must be smooth to facilitate the application of a
smooth layer of paint—when necessary, an appropriate abrasive should be used to produce a
smooth surface before the paint is applied.
The sides of the substrate (which are neither painted nor counted as part of the substrate's
area) should be sealed to prevent them from allowing the escape of organic vapors. At NRMRL,
this is accomplished by coating them with a solution of sodium silicate.
Until the substrates are ready to be used, they must be protected from exposure to ambient
organic vapors which might contaminate them and bias emissions tests. Two ways of doing this
are: (1) storing the substrate in an area with very low levels of VOCs, or if this is not possible, (2)
storing the substrate in an airtight container. For at least 24 hours prior to use, the substrates
should be conditioned in the small chamber at exactly the temperature and humidity at which the
sampling will be conducted. The purpose of this process is to make sure that the substrate is
already "conditioned" (in steady state with its environment) when it is painted.
Applying Paint to a Substrate to Create a Test Specimen
One of the key objectives of NRMRL's paint testing practice is to apply paint in a
controlled fashion that is realistic, quantifiable, and comparable. The first part of this objective is
to apply paint in a way that is realistic—that is, in a manner similar to how it would be applied by
actual house painters. In addition, the results of the emissions tests should be analyzed with
reference to a quantitative measurement of the mass of paint applied. Finally, in order for
different tests to be comparable, they must consistently involve approximately the same mass and
thickness of paint.
29
-------
With these goals in mind, NRMRL recommends using a roller or a brush under most
circumstances and a slit applicator for perfectly smooth surfaces such as glass. The slit applicator
will not leave a smooth coat on rougher surfaces such as gypsum board or wood because it cannot
compensate for valleys and ridges in the substrate. NRMRL did not find other applicators to be
particularly useful. Electronic applicators, for example, are expensive and they do not produce a
film that is sufficiently smooth and reproducible. Another applicator device, the spray gun, is
impractical given the small size of the test specimen, the waste from overspray, and the loss of
VOCs during spraying.
An investigator should begin the paint application process by determining the mass of
paint to be applied, using product data listed on the paint container label. One can calculate the
mass of paint needed using the spreading rate (ff/gal.), the density of the paint, and the area to be
painted:
A(DP)
Mp= -—^- (3-1)
SR
Where:
Mp = target mass of paint to be applied (g),
A = area of the substrate (cm2),
Dp = density of the paint (g/mL), and
SR = spreading rate, generally listed on container or data sheet as ft2/gal. (cm2/mL).
The appropriate mass of paint can then be weighed out into a tray. At this point, the
uncoated substrate should also be weighed. The paint may then be applied to the substrate using
the roller (or alternate applicator). Once application is complete, the amount of paint applied is
determined either by measuring the increase in the substrate's weight or the decrease in the roller
and tray's weight. The final weight of the paint applied should be within 10% of the target
application amount originally calculated.
30
-------
Operating the Small Chamber
The purpose of this phase of NRMRL's standard practice is to provide a controlled
environment for conducting emissions testing that can reflect common indoor air conditions.
Therefore, NRMRL recommends that the small chamber should be set to maintain an
environment within the following operating ranges:
Temperature: 20 to 25°C
Air Exchange Rate: 0.3 to 2 h"1
AirSpeed: 5tolOcm/s
Loading: 0.3 to 1.5 m/m
The relative humidity of inlet air should be maintained at 50% unless defined otherwise.
These operating ranges were chosen to match the actual conditions under which interior house
painting is likely to take place.
Prior to each use, the environmental test chamber, the sampling manifold, and all other
internal hardware should be cleaned to remove any chemicals which may have adhered over the
course of previous use. NRMRL uses alkaline detergent and deionized water as its cleaning
agents. The environmental conditions used for testing should be maintained for 24 hours prior to
the beginning of the test.
Sampling the Specimen's Gaseous Emissions
This part of the chamber operation and sample collection is critical to a successful
emissions evaluation. With new or unusual coating materials, some trial and error
experimentation may be necessary to develop a method that ensures complete and accurate
collection of the paint emissions.
31
-------
Sampling Protocol
As a test progresses, the experimenter may need to periodically adjust the airflow
controllers for the sorbent tubes and the length of sampling. At any given time, an important goal
of sampling is to deliver a precisely measured volume of air to the sampling media that conveys
an optimal mass of VOCs. An optimal mass of VOCs is one that is small enough to avoid
overloading the sorbent yet large enough to be detected by the analytical instrument.
During the early stages of a test, VOC emissions are often so high that they quickly
overwhelm sampling media. These conditions necessitate the collection of small sample volumes,
which can be difficult to measure precisely. The precision of the measurement of small sample
volumes should be maximized by adjusting the two settings that determine it: rate of airflow and
period of sampling. These two settings must multiply to produce the required sample
volume—particular settings should be chosen so as to maximize the overall precision of the
measured sample volume. Typical small-sample collection times range from 5 to 10 minutes. In
some cases, the investigator may choose to sample through two sorbent cartridges in series. A
separate analysis of the second sorbent cartridge will reveal whether or not overloading (i.e.,
"breakthrough") has occurred in the first cartridge.
Higher airflow rates and longer sample collection periods may be required in later stages
of the test as the concentrations of VOCs decrease in the chamber air. Again, the objective of the
selection of sampling time and flow rate is to collect sufficient organic material to be within the
sensitivity range of the analysis procedure yet avoid exceeding the sampling media capacity. The
total sampling flow rate must be less than 50% of the airflow from the chamber (e.g., less than
220 mL/min for a 53 L chamber with an air exchange rate of 0.5 h"1).
Air samples should be collected at predefined intervals during the test. The duration of the
test and the frequency of sample collection will vary, since they are determined by the objectives
of the test. A high frequency of sampling will be required if the objective of the test is to develop
or evaluate source emission models. A lower frequency of sampling may be appropriate for other
32
-------
test objectives. If resources are available, NRMRL's recommended sampling times for collection
of data for model development during a 2-week test with alkyd paint are at 0.25, 0.5, 1, 1.5, 2, 4,
6, 8, 10, 12, 24, 48, 72, 96, 144, 192, 264, and 336 hours following the start of the test. NRMRL's
recommended sampling times during a 2-week test with latex paint are at 2, 4, 8, 12, 24, 48, 72,
96, 144, 192, 264, and 336 hours following the start of the test. After sampling, one should
carefully seal sample tubes and record sampling flow rates and durations on them.
Sampling Media
During the first 10 to 20 hours after application of an alkyd primer or paint, air samples
for determination of VOCs may need to be collected on charcoal cartridges. These cartridges, if
used, are subsequently extracted with organic solvent. The use of charcoal cartridges may be
necessary because the very high concentrations of the VOCs in initial emissions can preclude the
use of Tenax and thermal desorption methods (due to the potential for breakthrough on the
sorbent media and overloading of the analytical instrument). Charcoal cartridge extracts can be
diluted if the emission sample concentration is too high for the analysis equipment. After the
initial drying period, the concentrations of VOCs will decrease to levels amenable to collection on
Tenax.
In addition to charcoal and Tenax sampling, air samples may be collected on silica gel for
the measurement of certain carbonyl-bearing compounds. A fourth sampling method (which
allows collection of aldehydes) uses silica gel coated with acidified 2,4-dintrophenylhydrazine
(DNPH). Samples collected on sorbent media or DNPH-silica gel cartridges should be analyzed
as soon as possible and no later than 30 days after collection. If they must be stored, they should
be refrigerated at lower than 4 °C.
33
-------
Using Instruments to Measure Chemicals Present in the Emissions Sample
The overall goal of the analysis of an emissions sample is to determine the amount and
kind of VOCs that the sample contains. Examples of the sampling and analysis methods followed
by NRMRL for monitoring paint emissions include:
GC/MS is used for compound identification. GC/MS is applicable to samples
collected on charcoal, Tenax, and silica gel. Compound identification is based on
chromatographic retention time and molecular fragmentation patterns from the
unknown compounds. Compound identification is confirmed by comparison with
the National Institute of Standards and Technology (NIST) mass spectral library
data base. The advantage of GC/MS is its ability to use an established data base of
mass spectra for a variety of organic compounds.
GC/FID is used for quantification of non-polar compounds such as alkanes,
alkenes, and aromatic compounds. Like GC/MS, GC/FID can be used to quantify
compounds collected on charcoal, Tenax, and silica gel. Compound identification
is confirmed using relative chromatographic retention times determined by
GC/MS. Compound quantification is based on calibration using at least five
standard concentrations covering the quantification range of the instrument.
GC/FID has the advantages of being less expensive than GC/MS and able to
quantify compounds over a wider concentration range.
HPLC/UV is used for identification and quantification of polar compounds such as
ketones and aldehyde that GC/FID cannot determine easily. HPLC is applicable to
samples collected on dinitrophenylhydrazene (DNPH). Compound quantification
is based on calibration with nine carbonyl compounds: formaldehyde,
acetaldehyde, propanal, benzaldehyde, pentanal, m-tolualdehyde, methyl isobutyl
ketone, hexanal, and heptanal. Additional carbonyl compounds can be analyzed
with this technique if the appropriate standards are available for calibration.
Analyzing the Results of the Analytical Instruments
One of the purposes of analyzing the results of the sampling instruments is to convert the
analytical results (chamber concentrations) into emission rates (ERs) and emission factors (EFs).
The two technical terms commonly used to describe the rate of emissions from indoor
materials, ER and EF, are related as follows:
34
-------
ER = A(EF) (3-2)
where:
ER = emission rate (mg h"1),
A = source area (m2), and
EF = emission factor (mg m"2 h"1).
Calculating VOC Emission Rates and Emission Factors
Once the chamber concentration data are obtained, the emission factor can be calculated
by three methods:
• Direct calculation from individual data points (Equation 3-3)
Direct calculation from the time/concentration profile (Equations 3-4 and 3-5)
• Using an explicit chamber model (Equation 3-6)
Selection of the most suitable method or methods depends on several factors, such as the
type of source, data quality, and sampling frequency.
If the emissions rate is nearly constant and the chamber has reached steady state, the
emission factor can be calculated from a single data point:
(3'3)
where:
Cs = steady state chamber concentration (mg m"3),
N = air change rate (h"1), and
L = loading factor (m-1).
35
-------
Note that this method may have significant error if the emission rate is not constant and/or
the chamber has not reached steady state.
Where the emission rate is not constant but there are enough chamber concentration vs.
time data points (for example, 10 or more) and the data are relatively smooth, a time-dependent
emission factor profile can be obtained directly from the concentration data:
AC,
'- + NCt
At. ' (3-4)
EP(t)=—!- l '
where:
= emission factor at time ^ (mg h"1),
C, = chamber concentration at time t, (mg m"3), and
AQ/Atj = the slope of the time/concentration curve at time t; (mg m"3 h"1).
36
-------
The slope is approximated by the average of the slopes of two adjacent intervals:
AC/Af,=
'r'i-1 tt+rtt (3-5)
Thus, if there are n + 1 data points for concentration, n - 1 emission factor values may be
obtained by this method. Such calculations can be easily carried out in an electronic worksheet.
Before making the calculations, make sure that replicate samples are replaced by the average
values to avoid dividing by zero.
The results from this method (Equations 3-4 and 3-5) are independent of any source
emission models. Additional benefits of these direct calculations include the fact that the results
can be used to check the validity of a chosen model, and that they can help select the most
appropriate model for further data analysis (see the following section on Using Emissions
Models). Note that differential methods such as this have the potential for high levels of
uncertainty. If there are sufficient data points but the random error is significant, a data smoothing
process can be considered before using this method.
Using Emissions Models
Another technique that NRMRL used to analyze 1he sampling results is to relate those
data to models of paint emission. Those mathematical models are useful for the purpose of
predicting (i.e., interpolating and extrapolating) emission rates, calculating their effect on indoor
air quality (when used as input to an indoor air quality model), and conducting exposure risk
assessment. When existing models, usually of an empirical nature, are not adequate, NRMRL
develops new mathematical simulations (as illustrated in Chapters 5 to 9) of the time-dependent
concentration data.
37
-------
The simplest model that is very useful for simulating short-term paint emissions assumes
that the chamber is an ideal continuously stirred tank reactor (CSTR) and that the change in
emission factor can be approximated by a first-order decay, as shown below:
EF=(EFQ)e~kt (3-6)
where:
EF0 = initial emissions factor, mg m"2 h"1, and
k = first-order decay rate constant, h"1.
The corresponding chamber model is:
dCldt= L(EFQ)e ~kt- NC (3-7)
Which has the following solution under the condition that C=0 when t=0 and N^k:
C= - (3-8)
N-k
Equation 3-8 is the model to be used to fit the chamber concentration data using non-
linear regression techniques. Using a curve-fitting program implemented on a computer requires
the user to provide initial values for the parameters to be estimated (that is, EF0 and k).
Reporting the Experimental Results
The purpose of guidelines for reporting is to ensure consistent and complete compilation
of results so that other investigators can compare or repeat emissions testing on similar or new
coating materials. NRMRL recommends the use the following template for the purpose of
ensuring that the experimental conditions and results are documented clearly and completely:
38
-------
Test objectives. This section provides a clear description of the purpose of the
testing program. This section should also clearly state the data quality objectives
and the limiting assumptions made during the experimental design and testing
phases of the research.
Facilities and equipment. Here, the report gives a description of the test
chambers, clean air systems, environmental measurement and control techniques,
sample collection (including sorbents if used), analytical instrumentation, and
standards generation and calibration.
Experimental design. This section describes the test conditions, including
temperature, humidity, air exchange rate, and material loading. It includes a test
matrix if appropriate.
Sample descriptions. This section provides a complete description of the
sample(s) tested, including type of material/product, size or amount of material
tested, product history, brand name (if appropriate), and sample selection process.
For wet samples, one describes the sample substrate. The section also includes
information about sample conditioning, including duration and environmental
conditions.
Experimental procedures. Here, the experimental procedures used during the
testing are described. The section includes details of the sampling and analysis
techniques. For wet samples, information about application method is provided
here.
Data analysis. This section shows the methods, including appropriate models and
equations, used to analyze the chamber data to produce emission factors.
Results. This section provides emission factors for each type of sample tested and
for each environmental condition evaluated. Emission factors can be provided for
individual organic compounds or for all organic compounds, or both. For sources
with variable emission rates, the section provides appropriate rate constants.
Discussion and Conclusions. Here, one discusses the relevance of the findings
and provides conclusions. For example, one might describe the effect of
temperature or air exchange rate on the emission factor.
Quality Assurance/Quality Control. This section describes the data quality
objectives and discusses the test's adherence to acceptance criteria. These are
described for both environmental variables and the chemical results. The results of
duplicate sampling and replicate analysis are provided here, as well as the outcome
of any audits.
39
-------
Conducting Quality Assurance/Quality Control
One should prepare a quality assurance and quality control (QA/QC) plan to specify the
quality requirements of the measured and reported data obtained during testing The plan should
include all facets of the measurement program from sample receipt to final review and issuance
of reports. ASTM Guide D 5116 describes QA/QC activities applicable to testing emissions from
paint.
NRMRL's QA/QC plan is based on established data quality objectives. These objectives
are listed in Table 3-1.
Table 3-1. Quality Control Objectives
Parameter Precision Accuracy Completeness
Temp erature
Relative Humidity
Air Flow Rate
Substrate Area
Sample Weighf
Organic Concentration
Emissio n Rate
±0.5 °C
± 5.0%
± 1.0%
± 1.0%
± 10.0%
± 20% RSDb
± 20% RSD
±0.5°C
± 10.0%
± 2.0 %
Not Applicable
Not Applicable
Not Applicable
Not Applicable
>90%
>90%
>90%
>90%
>90%
>90%
>90%
For wet samples
RSD=relative standard deviation=(s/m) * 100 %, where s=estimate of the standard deviation and m=mean.
To determine the accuracy of chamber test conditions and operating parameters
(temperature, air flow rate, and relative humidity), measuring devices should be calibrated with
NIST-traceable primary sources. Accuracy should be established for organic concentrations by
analyzing spiked samples, and a data quality objective should be established based on the data
needs of the project.
The chamber's background concentrations of VOCs, formaldehyde, and other carbonyl
compounds should be measured at the start of each test without the substrate in the chamber.
40
-------
NRMRL's QA/QC guidelines specify that the empty chamber background should be below the
following criteria:
For total VOCs, 10 i-ig/m3 or one-sixth of the lowest concentration to be measured,
whichever is lower.
For individual VOCs, 2 |_ig/m3 or one-sixth of the lowest concentration to be
measured, whichever is lower.
Field blanks, consisting of sorbent tubes and silica gel cartridges that are not used for
sampling, should be analyzed to verify that they have not been contaminated during handling and
storage.
The precision of the sampling and analysis should be determined by collecting multiple
samples. Duplicate samples should be collected concurrently at selected time periods during the
test. The number of duplicates for each type of sampling media should represent approximately
10% of the number of samples collected. Due to the limited volume of air available, duplicate
samples can be collected for only one type of media during a sampling period, especially when
low VOC concentrations dictate high sample volumes. Results from duplicate samples provide a
measure of bias between samples. Replicate analysis of duplicate samples (when possible)
provides a measure of precision for the analysis procedures used in the investigation.
41
-------
• hap te r
Characterization of tmissions of Volatile Urganic Compounds froi
Interior /\l k y d r a i n t
To demonstrate the usefulness of the small environmental chamber test method, this
chapter describes NRMRL's technical approach to characterizing emissions from interior alkyd
paints. It also provides results from NRMRL's bulk analysis of liquid paints and its measurement
of VOC emissions following application. Experimental data are presented for a series of tests
performed to evaluate factors that may affect emissions of VOCs from paints. Finally, the
chapter describes the use of a mass balance approach to evaluate emissions test results.
Building materials are recognized as important sources of indoor air contaminants
(Levin, 1989a). Although some building materials are relatively minor sources, paint may
represent a significant source of indoor air contaminants because of the volume of paint used and
the frequency of re-application during the life of a building. Despite a trend toward increased use
of water-based architectural coatings, alkyd paints are still used for some applications. EPA has
evaluated wall paint as a source of indoor air pollution under the Indoor Air Source
Characterization Project (IASCP) (Cinalli et al., 1993). Alkyd paint, which usually contains
greater than 30% by weight of organic solvents, has been identified in the Source Ranking
Database (SRD) developed under the IASCP as a potentially significant source of indoor air
pollution.
Emissions of VOCs have been determined for a number of different coatings. Wolkoff
cited 13 references on emissions of VOCs from lacquer, paint, and varnish in his review of
sources and emissions of VOCs (Wolkoff, 1995). Tichenor reported major organic compounds in
a variety of building materials, including polyurethane finish, latex paint, and wood stain
(Tichenor, 1987). Products with petroleum-based solvents contained alkanes (nonane, decane,
42
-------
undecane), substituted alkanes, and aromatics. Emissions data have been reported for wood stain,
varnish, latex paints, and alkyd paints (Tichenor and Guo, 1991; Howard et al., 1997; Krebs et al.,
1995; Fortmann et al., 1998). Data from these studies have been used to develop methods and
models for predicting VOC emissions from coatings used indoors (Sparks et al., 1996, Guo et al.,
1996a).
Alkyd paint was selected for testing because of the potential exposure of painters and
building occupants to high VOC concentrations following application. Alkyd paint continues to
be used indoors because of desirable properties such as durability, gloss, gloss retention, and fast
drying. The volatile portion of alkyd paints consists of aliphatic and aromatic hydrocarbons that
serve as the solvents. The solvents function to dissolve the film-former, reduce the solution or
emulsion to proper solids content and viscosity, and control the rate of film formation by their
evaporation rate. The solvents are generally straight-chain petroleum fractions and may contain
some aromatics.
txperi mental Work
The technical approach used in this project for characterizing emissions from alkyd paints
involved both the analyses of the liquid coatings and measurement of the emissions following
application to a substrate. The concentrations of the predominant VOCs in the liquid product are
not generally reported as part of emissions characterization projects. However, use of this
approach allows the researcher to evaluate the emissions test results using a mass balance
approach, as described in the section titled Emission Tests - Mass Balance Calculations. The
liquid product was analyzed using GC/MS to identify the VOCs in the paint that were most likely
to be emitted following application. The GC/MS measurement results were used to develop the
final design of the small chamber emissions tests, including selection of sampling methods, and to
develop the test protocols (e.g., sampling frequency and sample volume).
43
-------
Bulk Product Analysis by GC/MS
One alkyd primer and three alkyd enamel paints were purchased at local retail outlets for
the research program. All four products were commercially available and typical of products that
would be purchased by homeowners for application to walls or woodwork in residences. The
liquid products were analyzed by GC/MS to identify and quantify VOCs in the liquid product.
The extraction and analysis method was based on EPA Method 311 (40 CFR 1996), but the
analysis was not limited to quantification of only the HAPs regulated under the Clean Air Act.
The method was used to quantify predominant VOCs in the product with a boiling point of
approximately 35 to 250 °C. The coatings were extracted by diluting 1 g of paint with 10 mL of
methylene chloride. The sample was shaken for several minutes, then centrifuged for 5 minutes
to remove the solids. The supernatant was analyzed by GC/MS. The compounds were identified
by matching spectra using the NKT mass spectra library. Following identification of the
predominant peaks and calibration of the GC/MS, triplicate aliquots of each coating were
analyzed to quantify the concentration of 15 to 20 of the predominant VOCs. Blanks (solvent)
and spiked controls were analyzed as part of the QA program.
Small Chamber Emission Test Methods
Small chamber emission tests were performed to measure the emission rates of the
selected VOCs from the primer and paints following application to test substrates. (The basic
procedures of small chamber testing are described in Chapter 3 of this document.) The emissions
chamber operating conditions for the 15 tests conducted for this project were:
Temperature: 23 ± 0.5 °C ,
Relative humidity: 50 ± 5% (chamber inlet air),
Air exchange rate: 0.5 ± 0.05 h"1, for a nominal ventilation rate of 0.442 L/min, and
Air speed: 10 cm/s (nominal) at 1 cm above the surface of the test substrate.
The substrates for the tests were glass, new gypsum board, or white pine board. The
gypsum board and pine board were cut to a size of 16 x 16 cm for a total area of 0.0256 m2, which
gave a loading factor of approximately 0.5 m2/m3 in the 53 L chamber. (The same loading was
44
-------
used with glass.) The edges of these test specimens were sealed with sodium silicate. The
bottoms of the substrates were not sealed; the substrate was placed on the floor of the chamber
during the test. The substrates were conditioned in a chamber at 23 °C and 50% relative humidity
(RH) for at least 24 hours prior to application of the primer. The primer and paint were applied to
the top surface of the gypsum board or pine board substrates with paint rollers purchased at a
local retail outlet. The paint roller method has been demonstrated to provide reproducible film
applications on realistic substrates and it allows scale-up with the same application method for
tests in large chambers or full-scale test rooms or houses. The average mass applied was
2.8±0.3g/256 cm2 (89 |im wet film thickness) for the primer and 2.3±0.3 g/256 cm2 (68 |im wet
film thickness) for the paint. The application rates were slightly lower than the manufacturer's
recommended rates for wet film thickness of 100 |j,m for the primer and 76 |j,m for the paint, but
were realistic based on the application method and visual observation. Paint was applied to the
glass with a slit applicator. The slit applicator does not work well with the substrates (gypsum
and pine board) the edges of which are sealed with sodium silicate because of the rough edge.
Standardized methods were followed for gravimetric determination of the mass applied to the
substrate by weighing the paint container, roller, paint tray, and test substrate before and after
application. The test specimen was placed on the floor of the chamber for testing.
For most tests, the protocol involved application of the primer followed by application of
the paint after the primer was dry. During the first three tests of the project, the drying time for
the primer was only 1 hour, after which the test specimen was removed from the chamber and the
paint applied. In subsequent tests, the drying time was extended to 48 hours to obtain additional
information on the emissions from the primer alone and to simplify the modeling. Comparison of
the recovery of VOCs during the first three tests with recovery during the other 11 tests did not
indicate an effect of drying time on the recovery of VOCs during the tests. Most tests were 2
weeks in duration and involved application of primer and paint to the pine board substrate.
45
-------
Sampling and Analysis Methods
During small chamber emissions tests with alkyd paint, air samples were collected from
the chamber outlet on four types of sampling media. All air sampling was performed with
sorbents to collect integrated samples that were subsequently analyzed by GC/MS or HPLC.
During the first 10 to 20 hours after application of an alkyd primer or paint, air samples for
determination of VOCs were collected on activated charcoal cartridges. This was necessary
because the very high concentrations of the VOCs in the emissions precluded use of Tenax and
thermal desorption methods due to the potential for breakthrough on the sorbent media and
overloading of the GC column during analysis. After the initial drying period, the concentrations
of VOCs decreased to levels amenable to collection on Tenax and analysis by thermal
desorption/GC/MS.
In addition to the charcoal and Tenax samples, air samples were collected on silica gel for
the determination of methyl ethyl ketoxime (MEKO, 2-butanone oxime), a chemical of concern in
alkyd paint for health effects, including developmental toxicity, blood effects, and cancer risk
(EPA, 1997). Initial testing demonstrated that MEKO could not be quantitatively recovered from
air samples collected on charcoal, necessitating use of the alternative sampling method. The
fourth sampling method was collection of aldehydes on silica gel coated with acidified 2,4-
dintrophenylhydrazine (DNPH).
Samples were collected at a relatively high frequency in order to obtain sufficient
resolution of the emission profile and to obtain the best estimate of the peak concentration of
VOCs in the emissions for the purposes of model development. Each test with primer and paint
involved collection of 30 charcoal samples, 21 Tenax samples, 37 DNPH-silica gel samples, and
12 silica gel samples for MEKO during the 2-week test period.
46
-------
Charcoal Sorbent Samples
The sampling and analysis method for collection of air samples on charcoal sorbents was
based on ASTM Standard Practice D3686 (ASTM, 1996) and NIOSH Method 1500
(NIOSH, 1994a). Samples were collected on commercial sorbent tubes 6 mm O.D. x 70 mm long
containing two sections (50 mg/100 mg) of coconut shell charcoal. Air sample volumes of 0.5 to
3.0 L were collected at a flow rate of 200 cmVmin. After collection of air samples, the contents
of both sections were combined and extracted with 1% 2-propanol in carbon disulfide. A
surrogate (d10 Xylene) standard was added to determine the extraction efficiency. Analysis was
performed with a Varian Star 3400CX GC/Varian Saturn Ion Trap MS system. Quantification
was performed using an internal standard method. Performance of the instrument was verified by
analyzing daily calibration check samples prior to starting analysis of the samples.
Silica Gel Sorbent Samples for MEKO
Air samples were collected on silica gel cartridges, 8 mm OD x 70 mm long, consisting of
two sections (150 mg and 75 mg) of activated silica gel (20/40 mesh). The sampling and analysis
method was a modification of NIOSH Method 2010 (NIOSH, 1994b). The samples were
extracted with methanol and analyzed by GC/MS.
Tenax Sorbent Method
Air samples were collected on Tenax for quantification of VOCs after the concentrations
in the chamber decreased to levels that permitted use of the thermal desorption method. The
Tenax tubes used in this project were 203 mm long x 6 mm O.D. containing 250 mg of 60/80
mesh Tenax TA. Sample volumes of 0.5 to 8.0 L were collected at flow rates of 50 to
200 cmVmin. The samples were analyzed by thermal desorption/GC/MS. The method was based
on the EPA TO-1 method (Winnberry et al., 1988).
47
-------
DNPH-Silica Gel Samples
Air samples were collected on commercially available silica gel coated with acidified
DNPH. Sample volumes of 2 to 30 L were collected at flow rates of 200 to 400 crrrYmin.
Analysis was performed byHPLC. The sampling and analysis method was based on EPA
Method TO-11 (Winberry et al., 1988). The HPLC was calibrated for nine carbonyl compounds:
formaldehyde, acetaldehyde, propanal, benzaldehyde, pentanal, m-tolualdehyde, methyl isobutyl
ketone, hexanal, and heptanal.
rxesuilts and LJ i s c LJ s s i o n
One interior-grade alkyd primer and three interior semi-gloss enamel paints were used to
characterize emissions from alkyd paint. All four coatings were commercially available products
purchased at local retail outlets and were typical of products used in residences. The substrates,
glass, wallboard (gypsum board), and white pine board, were also purchased at local retail outlets.
VOC Content Determined by GC/MS
GC/MS analysis was performed to identify and quantify the predominant VOCs in the
coatings (Table 4-1). Decane, nonane, and octane were the three most abundant VOCs in the
primer. Decane, at a concentration of 30.7 mg/g, constituted approximately 10% of the TVOC
content of the primer. Undecane, o-ethyltoluene, decane, and dodecane were the most abundant
VOCs in Paint A-l, the paint used in most tests. The four compounds constituted 20% of the
TVOC of Paint A-l. Paint A-2, with undecane, decane, o-ethyltoluene, and dodecane as the
predominant compounds, was very similar to Paint A-l. Paint A-3 differed from Paints A-l and
A-2. The concentrations of the straight chain C-8 to C-12 alkanes were substantially lower in
Paint A-3, and this paint contained more branched alkanes. A number of aromatics were present
in the products, but at lower concentrations. MEKO was not detected in the primer, but was
present in all three paints at concentrations ranging from 0.92 to 2.93 mg/g. Hexanal, although
48
-------
I able 4~ I . v-*oncentrations of th e rre d o m i na nt \f vJv-* s in the Liquid rrii
Interior /\lkyd tnamel Paints
i n d I h re
ID
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
v^ompound
TVOC
toluene
octane
methyl ethyl ketoxime
ethylbenzene
m,p-xylene
nonane
o-xylene
propylcyclohexane
isopropylbenzene
n-propylbenzene
p-ethyltohene
1,3,5-trimethylbenzene
decane
branched decane a
branched decane b
o-ethyltoluene
1,2,4- trimethylbenzene
1,2,3- trimethylbenzene
2-methyldecane
trans-dec alin
undecane
branched undecane a
branched undecane b
branched undecane c
branched undecane d
branched undecane e
branched undecane f
dodecane
IT rimer
316
0.35
15.6
BDLb
0.29
1.39
18.4
0.23
4.06
BDL
0.03
0.21
0.02
30.7
c
~
~
0.14
~
2.19
2.28
6.68
~
~
~
~
~
~
0.063
Paint A-l
347
0.27
0.08
0.92
1.10
4.91
3.49
0.73
2.22
BDL
0.22
0.63
0.28
13.2
~
~
15.1
0.89
0.29
3.74
3.96
31.2
~
~
~
~
~
~
10.5
Paint A-2
350
0.83
0.62
2.93
1.54
5.48
7.55
1.45
5.00
0.18
0.42
1.54
0.62
23.4
~
~
27.6
1.01
0.41
BDL
4.54
32.6
~
~
~
~
~
~
8.57
Paint A-3
421
0.06
0.06
1.34
2.05
6.92
0.79
1.36
BDL
0.08
0.07
0.30
0.16
4.89
11.3
0.04
7.26
0.44
0.22
BDL
BDL
7.87
12.0
8.47
13.6
13.0
11.7
11.2
1.13
a Mean concentration of analyses of triplicate aliquots
b BDL = below method detection limit
0 Not detected
detected in the emissions from the primer and paints, was not detected in the liquid primer or the
three paints above the method detection limit of 0.03 mg/g.
49
-------
The TVOC concentration, calculated by integration of all peaks in the chromatogram
between toluene and tetradecane and using the average response factor for toluene, was 316 mg/g
for the primer and ranged from 347 to 421 mg/g for the paints. Total volatile matter content was
also determined for the primer and Paint A-l by EPA Method 24, a gravimetric method (40 CFR
1994). The results were 333 mg/g for the primer and 331 mg/g for Paint A-l, results consistent
with the manufacturer's reported volatile content of 332 mg/g and 328 mg/g forthe primer and
Paint A-l, respectively. The TVOC concentrations estimated for the primer (316 mg/g) and Paint
A-l (347 mg/g) based on the GC/MS analyses were remarkably similar to both the Method 24
results and the manufacturer's data considering the potential magnitude of the errors associated
with the integration method and the use of toluene (only) as the response factor for TVOC.
Emission Tests - Mass Balance Calculations
A series of 14 small chamber emissions tests were performed to characterize emissions
from alkyd paint. The tests are summarized in Table 4-2, and include evaluation of the effect of
substrate, primer, paint, previous coat, film thickness, air exchange rate, and air speed at the
substrate surface.
For each test, the total mass of the individual VOCs and TVOC emitted during the 2-week
duration tests was calculated from the concentrations in the air samples. These results were used
with the data from the analysis of the liquid product by GC/MS to calculate amass balance for
each test. The mass balance can be used to evaluate the test results to determine if they are
reasonable, and can also be used to compare results from tests performed under different
conditions (e.g., air exchange rate) or with different test parameters (e.g., film thickness,
substrate). The mass balance was calculated as the total amount (in milligrams) of the TVOC or
individual VOCs emitted during the test versus the amount (in milligrams) of the TVOC or
50
-------
Table 4-2. Matrix
its Per
f o r m e d in the
Pr
o g r a m
Test No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
O u b str a te
Glass
Gypsum
Pine
Glass
Gypsum
Pine
Pine
Pine
Pine -Test 3
Pine
Pine
Pine
Pine
Pine
r r i me r
Yesa
Yesa
Yesa
Yes
Yes
Yes
Yes
None
None
Yes
Yes
Yes
Yes
Yes
Paint
A-l
A-l
A-l
None
None
A-l
A-l
A-l
A-l
A-2
A-3
A-l
A-l
A-l
Notes
Replicate - Test 6
Apply second coat
Velocity- 3cm/s
AERb=1.12
Heavier
application
1 est for
Effect of!
Substrate
Substrate
Substrate
Substrate
Substrate
Substrate
Precision
Primer
Two coats
Paint
Paint
Air veloc ity
AER
Thickness
a Primer dried for 1 hour before application of paint in Tests 1-3; dried for 48 hours in all
other tests
b Air exchange rate, h"1
individual VOCs applied. The total mass of primer or paint applied was determined
gravimetrically at the time of application. Concentration data for TVOC and individual VOCs
measured in the paint formulation by GC/MS were used with the mass applied to calculate the
mass of each VOC applied to the substrate. The mass of TVOC and individual VOCs emitted
was calculated as:
Amount Emitted (mg) = AC(Q)
where:
Ac
Q
the area under the time/concentration curve, mg»h/m3; and
the chamber air exchange flow rate, m3/h.
51
-------
The mass balance, reported as the percent of the applied VOC or TVOC recovered in the
emissions, was calculated for the entire test and included the emissions from both the primer and
paint. Emissions were measured over a 2-week test period. The results of the mass balance
calculations are summarized in Table 4-3. The average percent TVOC recovered in the 14 tests
was 111% ± 20%. The median recovered was 103%. The recovery of TVOC ranged from 84%
to 146%, but was 84% to 110% in nine of those tests. The reason for the high recoveries in five
of the tests could not be determined.
There are no criteria for determining what is an acceptable mass balance. The error in the
mass balance calculation is a function of the analytical errors associated with analysis of both the
liquid product and the air samples. Because more than 90% of the VOCs are emitted during the
first 10 hours, the data for the samples collected on the charcoal sorbents have the greatest effect
on the mass balance. Analyses of duplicate samples collected on charcoal during each test gave
an average %RSD ranging from 2% to 19%. The median %RSD was below 10% for all but 1 of
the 20 VOCs. ASTM Standard Practice D3687 for analysis of VOCs collected with the charcoal
tube method states that a relative precision of ± 15% can be expected for the method (ASTM
D3687, 1996).
TVOC concentrations were estimated by integration of all peaks in the chromatogram in
the retention time window between toluene and tetradecane and use of only the toluene response
factor for quantification. Recoveries of 85% to 115% were achieved.
The percentage recoveries for individual VOCs were highly variable. As shown in
Table 4-3, the median percentage recovered ranged from 30% for toluene to 190% for
1,3,5-trimethylbenzene. Eleven of the 20 VOCs had median recoveries between 74% and 120%.
The average percentage recovery was between 80% and 127% for 12 of the 20 VOCs. Data are
presented in Table 4-3 for individual VOCs in ascending order for the median percent recovered
(last column) in order to evaluate which compounds fall into an "acceptable" range for recovery.
52
-------
I able T- ~ O • Percent of th e /\ p p I i e d V \J (^ IVI ass rxecovered in t m issions Durin
wee k O m all v-*h a m b er I e s t s with /\l ky d rr i m er and Paints
/O of Applied IVlass Ixecovered
L>om pound
TVOC
toluene
octane
n-propylbenzene
m,p-xylene
ethylbenzene
nonane
o-xylene
methyl ethyl ketoxime
propy 1 -cy cl ohexane
decane
trans-decahydranaphthalene
undecane
dodecane
2-methyldecane
p-ethyltoluene
1 ,2,4-trimethylbenzene
1,3,5-trimethylbenzene
o-ethyltolueneb
1,2,3 -trimethylbenzene0
isopropylbenzened
Amt."
16
4
19
9
11
3
13
14
7
1
8
2
6
10
15
12
17
5
18
—
1V1 i n i m u m
84
2
17
8
27
48
55
51
63
71
83
92
80
86
36
65
88
38
—
—
—
1V1 a x i m u m
146
138
121
385
202
191
112
209
105
455
126
172
136
500
314
345
1712
2069
—
—
—
r\ver age
111
35
50
100
80
87
82
115
84
127
101
115
113
146
127
138
346
448
—
—
—
Std.
Dev.
20
36
28
107
42
47
14
55
13
93
11
21
16
101
68
71
470
615
—
—
—
Median
10
30
43
61
70
74
79
83
89
96
100
105
111
114
119
120
160
190
—
—
—
a Amt. = Relative amount in primer/Paint A-l system listed in order with 1 being the most
abundant VOC
b Mass balance not calculated due to analytical problems with this compound
0 Most air concentrations below practical quantification limit; mass balance not calculated
d Present only in Paints A-2 and A-3; mass balance not calculated
The column, labeled "Amt." is the relative abundance of the VOC in the primer/Paint A-l coating
system. Decane, listed as "1," is the most abundant VOC in the coating, followed by undecane (2),
and nonane (3). The trimethylbenzenes were present in the coatings at much lower
concentrations (Table 4-1). The mass balances were not calculated for 1,2,3-trimethylbenzene
because the concentrations in most air samples were below the practical quantification limit
(PQL) of the
53
-------
method in many of the tests. Due to analytical problems with o-ethyltoluene, the mass balances
were not calculated.
Recoveries of VOCs in the emissions were low for the more volatile compounds (e.g.,
toluene, octane, ethylbenzene, m,p-xy\ene). The low recovery of the more volatile compounds is
likely to be an artifact of the small chamber emissions test method which requires application of
the coating to the substrate outside of the chamber. Substantial losses of the most volatile
compounds in the paint may occur during the 3 to 4 minutes required to prepare and weigh the
test specimen.
The mass balances were very good for the most abundant compounds in the primer and
Paint A-l system. For decane, the most abundant compound, the average and median recoveries
were near 100% and the range for the 14 tests was from 83% to 126%. The results were also very
good for undecane and dodecane. The one test in which the mass recovery of dodecane was
500% was a test with primer only. Because the primer contained a low concentration of
dodecane, concentrations of dodecane in the air samples were low. Two of the 14 tests involved
application of only the primer to glass or a pine board. Because of the low concentrations of
aromatic compounds in the primer, the mass balances were poor. The maximum mass recoveries
for n-propylbenzene, p-ethyltoluene, 1,3,5-trimethylbenzene, and 1,2,4-trimethylbenzene
occurred in the tests with the primer only.
The data presented in Table 4-3 demonstrate that the error of the mass balance can be
significant if calculations are made for VOCs present at low concentrations in the coatings.
However, the data suggest that, for the more abundant compounds, an error in the mass balance of
less than ± 20% is probably reasonable. This assumes that all of the VOCs are emitted from the
test specimen and that the duration of the small chamber emission test is sufficient to collect most
of the emissions. The mass balance data are useful for evaluating the performance of the test
method for solvent-based liquid coatings when there are no substrate effects. The data maybe
more difficult to interpret for some products, including water-based products that emit less
volatile and more polar compounds, because the compounds may be emitted over longtime
periods (weeks to months) and substrate effects may affect the emission rates. However,
experience with a number of coatings suggests that the mass balance approach provides a
valuable "reality check" of the data and is useful for evaluating emission test results.
54
-------
Emission Test Results
Effect of Substrate
Previous research has demonstrated that there can be a substantial effect of the substrate
on emissions of VOCs from coatings. Geherig et al. (1993) reported that compounds with polar
oxygen-containing functional groups showed reduced emissions from a low-VOC paint applied to
gypsum board or wall paper as compared to applications to glass plate. They did not observe any
significant effect for less polar alkanes and aromatics. Guo et al. reported similar results for latex
paint applied to gypsum board (Guo et al., 1996b). They observed a significant effect of the
gypsum board substrate on emissions of ethylene glycol and propylene glycol from latex paint.
Therefore, the first tests for this project were performed with the primer and one paint (Paint A-l)
applied to glass, gypsum board, or a white pine board to assess substrate effects.
There were no substantial differences in the emissions of TVOCs or any of the individual
VOCs measured during tests with the primer and Paint A-l applied to glass, gypsum board, or
pine board. Emissions of decane, the most abundant VOC in the primer/Paint A-l coating
system, are depicted in Figure 4-1 and are representative of the emissions profiles for TVOCs and
individual VOCs. Decane emissions were not substantially different for the three substrates. The
highest concentrations of decane measured from the application of the primer were 1060, 926,
and 1450 mg/m3 for glass, gypsum board, and pine board, respectively. Following application of
the paint, the highest concentrations of decane were 370, 394, and 569 mg/m3 for the three
substrates. There were slight differences in the mass applied during the three tests. If the peak
concentrations of decane were normalized for the amount of primer and paint applied, the
difference between the
55
-------
1600
00
£
o
• 1—1
i-H
o
o
U
Glass
Gypsum
1200-
Paint applied
800
400
0
Pine
0
8 12
Elapsed Time (h)
16
20
Figure T-- I . LJecane e m i s s tort s fro m a I ky d pr i m er and raint r\~ I f o r first
gypsum board, or pine board substrates.
hours after application to glass,
-------
glass and gypsum board was less than 4%, but the peak concentration of decane in the test with
the pine board was nearly 40% higher than in the test with the glass. The reason for the
difference in peak concentrations was not determined. But, despite the differences in the peak
concentrations, the concentrations of decane in the emissions after 4 hours were very similar for
the three substrates. Air samples collected from the chamber with only the pine board, prior to
application of the primer and paint, did not show detectable levels of decane emissions from the
board.
The high cost of performing small chamber emissions tests precludes extensive replication
of tests. Therefore, the statistical significance of differences between the three tests cannot be
determined. Comparing peak concentrations of the emissions is not the best approach for
assessing the significance of the differences between tests. Although air samples are collected
frequently, they cannot be collected continuously, and the peak concentration maybe missed. As
discussed in the previous section, the mass balance approach is useful for evaluating chamber test
data. For the tests with the three different substrates, the percent of the applied TVOC recovered
in the emissions was 84%, 97%, and 133% for glass, gypsum board, and pine board, respectively.
Recoveries greater than 100% occur due to substantial background emissions from the substrate
or the inaccuracy of the sampling and analysis methods. Background air samples collected from
the test chambers prior to the tests with gypsum board and pine board demonstrated that
background emissions from the substrates were very low, generally below the method detection
limit. Therefore, the variation in recovery, ranging from 84% to 133% in these tests, is likely
related to the errors associated with estimating TVOC in the product and in the emissions, as
discussed in the previous section. For decane, the most abundant VOC in the primer/paint A-l
coating system, the recoveries were 83%, 99%, and 126% of the applied amount in the three tests.
Because it was the predominant VOC, the variation in the decane measurements would have a
substantial effect on the calculated recoveries of TVOCs.
In spite of the low recoveries of TVOCs from the glass and high recovery from the pine
board, the differences in emissions between the substrates were not substantial when compared to
the differences observed in previous tests with latex paint (Kreb s et al., 1995). In that study,
57
-------
emissions from latex paint applied to gypsum board were compared to the same paint applied to a
stainless steel plate. During a 2-week period following application of the paint, only 20% of the
TVOCs were recovered from the gypsum board substrate compared to 97% from the stainless
steel substrate. Nine percent of the ethylene glycol applied to the gypsum board was recovered
compared to 103% from the stainless steel.
Effect of Primer or a Previous Paint Coating
The effect of the primer on emissions of VOCs was determined by applying Paint A-l to a
pine board without application of the primer and comparing the results to a standard test
involving application of the primer followed 48 hours later by application of Paint A-l. Results
of the comparison for decane emissions are depicted in Figure 4-2. The peak concentrations and
the rate of emissions during the first 20 hours were not sub stantially different in the two tests. In
the test without the primer, the recovery in the air emissions was 102% for decane and 101% of
the applied TVOC. This compared to recoveries of 110% for decane and 104% for TVOCs in the
test with the primer.
A similar test was performed to determine if emissions from Paint A-l differed due to
application over a previous coat of the same paint. The results were similar to that of the primer.
The previous coat did not affect the peak concentrations or the rate of emissions. The total mass
recovered was 104 % for Paint A-l applied over the primer and 102% for Paint A-l applied to a
pine board coated with primer and Paint A-l 3 months prior to application of the second coat of
Paint A-l.
Comparison of Three Paints
Emissions from three alkyd paints purchased from major U.S. manufacturers were
compared during the study. The predominant VOCs (straight chain alkanes) and the TVOC
58
-------
i
In
i
s
o
Lj
G
a
U
300
250
200
150
100
50
No Primer
With Prim er
0
8 12
Elapsed Time (h)
16
20
Figure T-~^. tffect of primer on short~term decane emissions from r aint r\~ I applied to pin
data presented for first L\J hours after paint application
-------
concentration were similar in the liquid Paints A-l and A-2 (Table 4-1). Paint A-3 had a different
formulation; the TVOC concentration was higher and the predominant VOCs in the coating were
branched alkanes. A major difference between the three paints was the concentration of methyl
ethyl ketoxime (MEKO), which ranged from 0.92 to 2.93 mg/g. Figure 4-3 depicts the emissions
of TVOC for the three paints (the emissions from the primer applied 48 hours prior are not
depicted in the figure). The TVOC emission profile was similar for Paints A-2 and A-3. The
emissions from Paint A-l were somewhat lower during the initial 4 hours following application,
but approximately 10% less paint was applied to the substrate. There was little difference in the
total mass of emissions recovered during the 2-week tests. TVOC recoveries were 104%, 110%,
and 94% for Paints A-l, A-2, and A-3, respectively. For decane, the recoveries were 110%,
119%, and 100%.
The emission profile depicted in Figure 4-3 was typical for the tests with the alkyd paints.
The peak concentration of TVOC occurred within an hour after application. Calculated
cumulative emissions showed that approximately 90% of the TVOCs were emitted during the
first 10 hours following application of either the primer or paint. Within 100 hours after
application, concentrations of VOCs in the test chamber dropped by 3 orders of magnitude.
Emissions profiles for individual VOCs were similar for the alkanes, branched alkanes,
and the aromatics. For all compounds, including methyl ethyl ketoxime, but excluding the
aldehydes, peak concentrations occurred within 2 hours following application of the primer or
paint and the concentrations then decreased rapidly (e.g, Figure 4-3).
Effect of Film Thickness
The effect of film thickness on emissions has not been studied in detail (Wolkoff, 1995).
Clausen reported that, for a waterborne paint, the initial emission rate was not affected by film
thickness and that the first-order decay constant was inversely proportional to film thickness
(Clausen, 1993). The effect of film thickness on VOC emissions from the solvent-based
60
-------
8000
Paint A-1 ^PamtA-2 ^PamtA-3
0
4
8
12
16
20
Figure T-~O. v-*omparison of I VvJv-* emissions from three alkyd paints for first £\J hours after
application to a pine board previously co ate d w ith p r i m er.
-------
Paint A-l was evaluated by comparing an application with a paint film thickness of 59 |j,m with a
film thickness of Paint A-l of 82 |j,m, both on previously primed pine boards. The effect of wet
film thickness on emissions of decaneis depicted in Figure 4-4. With the increased film
thickness, the peak concentration of individual and total VOCs was higher and the peak
concentration occurred later. The total emissions were higher, consistent with the increased
amount of VOC mass applied.
Effect of Air Exchange Rate
The emission rates of VOCs from wet coatings, which may be controlled by evaporation,
have been shown to be affected by the air exchange rate (Tichenor, 1991; Tichenor, 1995;
Wolkoff et al., 1993). For evaporative emissions, increased air exchange rates (i.e., increased
ventilation) can be expected to result in lower peak VOC concentrations in the test chamber and
earlier occurrence of the peak concentration. Results of the test with Paint A-l in tests with air
exchange rates of 0.5 and 1.0 hr"1 were consistent with these predictions, as depicted in Figure 4-5
for decane.
Effect of Air Velocity at the Surface
To evaluate the effect of air velocity at the surface on emissions, one test was conducted at
0.5 air change per hour (ACH) with a mixing fan in the chamber. With the fan, the air speed at
the surface is approximately 10 cm/s. In a second test, the chamber was operated at 0.5 ACH
without a mixing fan. Measurements have shown that, without the mixing fan, the air speed at
the surface is less than 3 cm/s, the minimum speed that can be measured with the hot-wire
anemometer. Based on theoretical predictions for gas-phase mass-transfer-controlled emissions
(Tichenor et al, 1991), the peak VOC concentrations should be lower and occur later at lower air
speeds. Results of the tests were consistent with theoretical predictions. In the chamber test with
the fan operating, the peak concentration of decane from the primer was 814 mg/m3 at 1.29 hours
62
-------
C
O
CJ
G
O
U
400
59
82
300
200
100
0
0
4
8 12
Elapsed Time (h)
16
20
Figure T-~T-. tff ect of wet f i I m thickness on short~ter m d ec a n e e m i s s i o n s fro m r a i nt r\~ I during first
L-\J hours f o 11 o wi n g application to pine board previously coated with p r i m er.
-------
wo
.g
G
o
"i
Js
c
OJ
U
c
o
U
0
0
8 12
Elapsed Time (h)
16
20
Figure T-~O. tffect of air exchange rate on short~term decane emissions during small chamber tests
with r aint r\~ I applied to pine board previously coated with primer.
-------
after application, and the peak concentration of decane from the paint was 245 mg/m3 at
1.36 hours after application. But in the test without the fan, the peak decane concentrations were
523 mg/m3 at 2.44 hours after primer application and 164 mg/m3 at 2.23 hours after paint
application.
Emissions of Aldehydes
In addition to the solvent VOCs, aldehydes—formaldehyde, acetaldehyde, propanal,
pentanal, and hexanal—were detected in the emissions from the primer and paint. The
concentrations of formaldehyde were low and generally below the practical quantification limit of
the method. The predominant aldehyde in the emissions from the alkyd paints was hexanal, a
compound that was not detected in the liquid primer or paint by GC/MS analysis. Background
samples collected from chambers containing the unpainted pine boards had hexanal
concentrations ranging from non-detectable to 0.02 mg/m3, which was at least 2 orders of
magnitude lower than the peak hexanal concentrations during the tests. The hexanal was
apparently formed by oxidation (Hancock et al., 1989a and 1989b). An example of the hexanal
emissions for the primer and three paints is presented in Figure 4-6. The following observations
were made during the project: (1) the emissions of hexanal were lower from the primer than from
the paints, (2) peak concentrations differed, but the total mass emitted during the 2-week test
periods was similar for the three paints, and (3) unlike the VOCs, hexanal emissions were not
significant until about 10 hours after paint application and peaked at about 20 to 24 hours for
Paint A-l. Figure 4-6 is included in this chapter for illustrative purposes only; results of the
aldehyde emissions are presented in detail in Chapter 6.
v-*o nc I ui !
The alkyd primer and two of the three paints tested in this project contained primarily
straight-chain alkanes, with decane and undecane being the predominant compounds. Paint A-3
contained more branched alkanes. All four coatings contained low levels of aromatic compounds.
65
-------
CUD
„§
C
_o
s
G
u
G
O
U
12
10
8
6
0
0
40
80 120
Elapsed Time (h)
160
200
Figure T- ~ O • Long ~te r m h e x a n a I e m i s s i o n s fro m three paints applied to pine board coa te d
earlier with alkyd primer.
-------
The three paints contained methyl ethyl ketoxime, but the primer did not. The TVOC content of
the liquid paints ranged from 32% to 42%. Measurements of the VOCs in the liquid coatings by
GC/MS agreed well with Method 24 measurements and manufacturers' data.
Small chamber emissions tests, conducted by coating glass, gypsum board, or pine board
with primer followed by application of the paint 1 hour later, demonstrated that the substrate did
not have a substantial effect on the peak VOC concentrations, the emission rates, or the total mass
of VOCs emitted. Over 90% of the VOCs were emitted within the first 10 hours following
application of the primer or paint. There were differences in emissions of individual VOCs from
the three paints, but the general patterns of the emissions were similar. The effects of other
variables, including air velocity, air exchange rate, and film thickness, were consistent with
theoretical predictions.
Mass balance calculations showed that, for the most abundant compounds, the entire mass
of VOCs applied to the substrate could be accounted for in the air samples of emissions collected
during the test; i.e., there was 100% recovery. The data for the most abundant compounds in the
paint suggest that errors of ± 20% can be expected for this type of product. The mass balance
approach was useful for evaluating differences between tests and for assessing the reasonableness
of the test results.
67
-------
• hap te r
IVIethyl tthyl ixetoxi m e t m issions fro m /\ I k y d r a i n t
The previous chapter outlined how NRMRL's test methods have been used to characterize
overall VOC emissions from alkyd paints. The goal of the present chapter is to provide an
example of how these methods can be used to monitor a single chemical of concern.
Methyl ethyl ketoxime (MEKO), another name for 2-butanone oxime or ethyl methyl
ketoxime [CH3C(NOH)C2H5, CAS Registry No. 96-29-7], is often used by paint manufacturers
as an additive to interior alkyd paints (Weismantel, 1981; Turner, 1988). MEKO acts as an anti-
skinning agent (or anti-oxidant) that prevents oxidative drying or skinning of the alkyd paint to
improve stability in the can. Usually, the MEKO content in a paint should be less than 0.5%
(Krivanek, 1982). Due to its relatively high volatility (152°C boiling point), the majority of the
MEKO in the paint is expected to be released into the surrounding indoor air after painting to
allow the paint to dry properly on the painted surfaces. Impacts of MEKO emissions on indoor air
quality and associated exposure risk are significantly affected by source characteristics such as
emission rates and patterns.
MEKO has been found to be a moderate eye irritant (Krivanek, 1982). It has also been the
subject of a Section 4 test rule under the Toxic Substances Control Act (Fed. Regist, 1986).
Testing conducted under the test rule (Fed. Regist., 1989) has evaluated a number of lexicological
endpoints for the chemical. MEKO demonstrated carcinogenic activity in long-term inhalation
studies, causing liver tumors in both rats and mice. MEKO did not cause gene or chromosome
mutations in standard genotoxicity studies. Nielsen et al. (1997) reviewed and evaluated
68
-------
published lexicological data on noncarcinogenic effects of MEKO. No data were found on effects
in humans, and inhalation studies concordant with effects in oral studies were not available.
Using a safety factor to extrapolate from animals to humans, Nielsen et al. (1997) proposed a
tentative health hazard indoor air exposure limit of 0.1 mg/m3 for MEKO and suggested a range
of sensory irritation thresholds of 4 to 18 mg/m3. Those indoor exposure threshold values were
recommended to facilitate the evaluation of building materials and indoor climate.
The obj ective of NRMRL's research was to characterize the MEKO emissions from alkyd
paints. Alkyd paint in general contains significant amount (e.g., 40%) of organic solvent. While
solvent emissions were discussed by other papers (Fortmann et al., 1998; Chang and Guo, 1998),
this research focuses on MEKO. Small environmental chambers (ASTM, 1995) were used to
measure the MEKO emissions from three commonly available alkyd paints applied to pine
boards. The chamber data were interpreted by a first-order decay model to simulate the time-
varying emission rates. This emission model was used as input to an indoor air quality (IAQ)
model to evaluate the effectiveness of exposure reduction options.
tx p er i m e nta I Work
Experiments were designed to generate MEKO emission data from a newly applied alkyd
paint as it dried for more than 24 hours under controlled experimental conditions. Tests were
conducted in the EPA's small chamber source characterization facilities consisting of
electropolished stainless steel chambers (Tichenor, 1989). The facilities allowed close control of
temperature, relative humidity, and air flow rate in the chambers. Small fans were used in the
chambers to provide a velocity near (1 cm above) the test surface of 5 - 10 cm/s which is typical
of indoor environments. The standard test conditions were:
Air exchange rate (N) 0.5 h"1
Temperature 23 °C
Inlet relative humidity 50%
Nominal wet paint film thickness 80 |im
Substrate specimen surface area (A) 0.0256 m2 (0.16 x 0.16 m)
Chamber volume (V) 0.053 m3
Loading (L = A/V) 0.48 m'1
69
-------
Three commonly used alkyd paints (designated at paints A-l, A-2, and A-3) were
selected, acquired, and dispensed as described in Chapter 6.
The primer and the paints were analyzed for MEKO contents by extracting a sample
(usually Igin 10 mL) of methylene chloride and analyzing the extract by GC/MS.
A pine board substrate was used in the experiments. It was prepared, loaded with paint,
and placed into the small chamber as described in Chapter 6. The MEKO concentration in the
clean air flowing into the chamber was virtually 0. Any MEKO emissions from the substrate
would be reflected by the detection of MEKO in the exit air from the chamber.
Air samples at the chamber exit were collected on silica gel cartridges, 8 mm OD x 70
mm long, consisting of two sections (150 and 75 mg) of activated silica gel (20/40 mesh). The
sampling and analysis method was a modification of NIOSH Method 2010 (NIOSH, 1994b). The
samples were extracted with methanol and analyzed by GC/MS.
rxesuilts and LJ i s c LJ s s i o n
MEKO Contents
Bulk analysis results showed that the primer had no detectable amount of MEKO, but all
three alkyd paints tested contained an appreciable amount of MEKO. The MEKO contents of
paints A-l, A-2, and A-3 were 0.96, 2.93, and 1.34 mg/g of paint, respectively. It is believed that
the MEKO contents in these three paints cover the typical range of alkyd paints available in the
marketplace.
70
-------
Chamber Emission Data
No MEKO was detected in the chamber air during the substrate conditioning and primer
drying periods. The data indicated that there were no MEKO emissions from either the pine board
or the primer, which confirmed the bulk analysis results that the primer contained no MEKO.
However, significant MEKO emissions were measured for all three alkyd paints tested. The
concentration profiles (time/concentration curves) of MEKO in the environmental chambers from
paints A-l, A-2, and A-3 are shown in Figures 5-1, 5-2, and 5-3, respectively. It is seen that
MEKO emissions occurred immediately after painting, with the chamber concentration peaking
within 1 hour. Using paint A-2 as an example, Figure 5-4 shows that more than 90% of the
MEKO emitted was released in less than 10 hours. The fast release of MEKO resulted in a
decrease of chamber concentration by more than 2 orders of magnitude in 24 hours.
Table 5-1 lists material balance results represented by the recovery (as the ratio between
MEKO emitted in the chambers and applied with the paint). The amount of MEKO applied to the
pine board was estimated by multiplying the MEKO content and the quantity of each paint
applied. The total amount of MEKO emitted from each paint was estimated by a procedure which
integrates the concentration profile from time 0 to 24 hours. Table 5-1 indicates that the recovery
ranged from 68% to 105%, which suggests that most of the MEKO in the paints was emitted.
Table 5-1. Recovery of MEKO
IXecove ry
MEKO Applied, MEKO Emitted', (Emitted/Applied),
m g / m
I 2 I 2
» / m m g / r
Paint A-l
Paint A-2
Paint A-3
64.4
241
110
43.5
252
97.8
68
105
89
Estimated by integration of the first 24 h chamber concentration data.
71
-------
15
12
Data
Model
e
o
••s
1
u
a
o
U
10 15
Elapsed Time (h)
20
25
Figure O~ I . v-* h a m ber concentrations resulting fro m the IVI t rxvJ e m i s s i o n s
from PaintA-1.
iUU "
i
^ 80
! 60-
1
£ 40 -
S
3 ^O =
n =
Data Model
" =
5 10 15
Elapsed Time (h)
20
25
Figure O~^. v-*ha m ber concentrations resulting fro m the IVI t rxvJ e m i s s i o n s fro m
Paint A-2.
-------
m,
6 30
a
o
o
a
o
o
20
10
Data
Model
10 15
Elapsed Time (h)
20
25
Figure O~O. v-*h a m b er
concentrations resulting fro m the
MEKO
e m i s s i o n f r o m
IUU "
90 .
at
i"™1 O U :::
ft
§ 70 =
at
ft
60 =
i r -------__ -:::::.::::::-- :':::::"
— -":.!:::--—
"" -";""::.. I
":
;
:
-•
1
1
12
16
20
24
Elapsed Time (h)
Figure O~T-. cumulative IVItlxvJ emissions for ra'trtt. r\~ £. ^calculated from c h <
concentration cJataj.
73
-------
MEKO Emission Model
Assuming the MEKO emissions are an evaporation-like process controlled by gas-phase
mass transfer, it was found that the chamber concentration data can be simulated by a first-order
decay model (Guo et al., 1990; Clausen, 1993).
EF^M0ke~M (5-1)
where
EF = MEKO emission factor (mg/m2/h),
M0 = initial amount of MEKO in the paint film (mg/m2),
k = first-order decay rate constant (h"1), and
t = time elapsed (h).
Assuming the chamber air was well-mixed, mass balance equations give the following
expression for the chamber air MEKO concentration, C:
LMok , ,f Mt^
C = - (e~kt - e~m) (5-2)
N- k
If one uses the total amount of MEKO emitted listed in Table 5-1 as the value of M^ for
each paint, there is only one unknown parameter, k, in Equation (5-2). The values of the unknown
parameter were estimated by the best fit of Equation (5-2) to the chamber concentration profiles
shown in Figures 5-1 to 5-3 using a non-linear regression method.
Table 5-2 shows the estimated values of k and normalized mean square error (NMSE)
which is a parameter recommended by a standard guide of the American Society for Testing and
Materials (ASTM, 1995) to reflect the goodness-of-fit of the model for each paint tested.
Comparing the value of k with the MEKO content in each paint, the data seem to suggest that
there is a linear relationship between them. The values of k shown in Table 5-2 are also within the
range expected for emissions controlled by the gas-phase mass transfer process (Sparks et al.,
1996). The goodness-of-fit of the model was represented by NMSE which is a measure of the
74
-------
I able O~^. tstimated Values of IVIodel Parameter and \Joodness~of~l~it to the
v-*h a m ber v-*oncentration LJata
Paint A-l
Paint A-2
Paint A-3
*, „'
3.83±0.59a
5.19±1.20a
4.17±1.48a
.[Normalized 1V1 e a n
Oquare terror
0.056
0.071
0.219
MEKO in Paint, mg/g
0.96
2.93
1.34
a Mean ± standard deviation
b A value of 0 indicates perfect agreement for all pairs of observed and predicted values. A value
near 0.2 indicates differences of about 50%.
magnitude of prediction error relative to the predicted and measured values (ASTM, 1995). The
ASTM standard guide suggests that, considering the potential consequences of measurement
uncertainties, aNMSE value of 0.25 or less can be taken as generally indicative of adequate
model performance (ASTM, 1995). Table 5-2 shows that the NMSE values for the three paints
tested are all less than 0.25, which indicates that the first-order decay model is adequate for
predicting MEKO emissions.
Indoor Air Quality Impact Assessment
The impact of alkyd paint MEKO emissions on indoor air quality (IAQ) was assessed by a
case study which assumed that an alkyd paint was used to paint the trim, cabinetry, inside doors,
and frames of an EPA IAQ test house. The test house is an unfurnished, single-story, wood-frame
house with a central heating and air-conditioning system. The house contains three bedrooms, a
kitchen, a living room, a dining area, a den, and two full baths. The detailed floor plan of the IAQ
test house was reported elsewhere (Chang and Guo, 1994). The volume of the house (not
including the attic and the crawl space) was estimated to be 305 m3. The total area to be painted
(trim, cabinetry, and inside doors and frames) was estimated to be 80 m2. The painting was
assumed to be performed in two periods (one in the morning and the other in the afternoon) in a
single day. The morning period started (t = 0) at 9 AM and lasted for 3 h. There was a 1 h lunch
75
-------
break between the morning and the afternoon periods. The afternoon period started at 1 PM and
lasted for another 3 h. The painting was conducted continuously and evenly throughout the two 3
h working periods.
To simplify the calculations, the indoor air in the whole house was assumed to be well-
mixed and can be represented by a one-compartment model. No sink effects were accounted for
due to the lack of sink data. To assess the impact of MEKO emissions on IAQ, a continuous-
application source model was used. The model assumed that one coat of the alkyd paint was
applied to the surfaces at a constant rate. The approach was to break up the application into many
differential areas, and each area began emitting once the paint was applied. Evans (1996)
established the mathematical models for three types of continuous-application emission sources.
The analytical solution of the one-compartment model with a first-order decay emission source is:
_ a(t-g)H(t-g)
V
a(t) =
-Nf
Nk k(k-N)~ N(k-N)
e
(5-4)
(5-5)
where
q = the paint application time of each painting period (h),
a(t) = time shift term in Equation (5-3),
H(t-q) = Heaviside operator (O'Neil, 1991),
V = volume of the test house (m3), and
Mtot = total amount of emittable MEKO in the applied paint during each painting
period (mg).
The MEKO concentrations in the test house with an air exchange rate of 0.5 h"1 during and
after the painting, as predicted by the model for the three alkyd paints, are shown in Figure 5-5. It
is seen that the MEKO concentration rises sharply right after the painting started and reaches a
76
-------
peak at the end of the morning period. The MEKO concentration declines slightly between the
4th and 5th h which reflects the 1 h lunch break. The MEKO concentration rises again after the
afternoon painting period started and reaches another peak at the end. After all the painting is
finished, the MEKO concentration decreases continuously.
Compared with the suggested indoor air exposure limit of 0.1 mg/m3, Figure 5-5 shows
that the test house MEKO concentration exceeded the limit for all three alkyd paints. The episode
of IAQ deterioration started at the beginning of the painting and lasted for more than 7 h after the
100
10
o
U
O
M
-8
a
0.1
0.01
Sensory irritation
threshold range
Heal til-based indoor air
exposure limit
0
5 10
Elapsed Time (h)
15
20
Figure O~O. comparison of the predicted test h o u s e IVItlxvJ concentrations with
the suggested indoor exposure thresholds.
painting. The test house MEKO concentration also exceeded the lower limit of the suggested
sensory irritation range when paints A-2 and A-3 were used.
77
-------
Exposure Reduction Assessment
To reduce the exposure potential, alkyd paint with lower MEKO content should be
selected. Figure 5-5 shows that the peak MEKO concentration was reduced by more than 80%
when paint A-l (with 0.96 mg of MEKO/g of paint) was used instead of paint A-2 (with 2.93 mg
of MEKO/g of paint). Figure 5-5 also shows that using the paint with lower MEKO content can
decrease the duration of concentration elevation after the painting.
The exposure potential to MEKO can also be reduced by increased ventilation. Figure 5-6
shows that, if the air exchange rate of the case previously discussed were increased to 3.0 h"1, the
test house MEKO concentration could be maintained below the sensory irritation threshold range
during and between the two painting periods (hour 0 to 7 in Figure 5-6) even when paint A-2 was
used. The test house MEKO concentration also decreased rapidly after the painting, and fell
below the suggested indoor air exposure limit of 0.1 mg/m3 within 2 h (between hours 7 and 9 in
Figure 5-6).
Since Equations (5-3) and (5-4) indicate that the indoor MEKO concentration increases
are proportional to the total amount of emittable MEKO in the applied paint (M), another
exposure reduction option can be the division of the whole paint work into several small jobs.
Only a small area is painted during each paint job, and plenty of air-out time (more than 12 h)
should be allowed between the jobs. For example, when paint A-l is used and the painted area is
78
-------
100
10
Paint A-2
u
o
-8
a
0.1
0.01
Sensory irritation
threshold range
Pamt A-3
Paint A-l
Health-based indoor
air exposure limit
0
15
Elapsed Time (h)
Figure Q-D. Predicted test hi o u s e ML.KO concentrations at high (o.C) h ) air
exchange rate.
<1.6 m2, the test house MEKO concentration can be maintained < 0.1 mg/m3 at an air exchange
rate of 3.0 h"1. The MEKO concentration in the test house can be further reduced if the source
room can be isolated from the rest of the house and high local ventilation (e.g., 3 h"1) is
implemented in the source room.
Conclusions
Bulk analysis showed that the MEKO content in alkyd paints can be as high as several
milligrams per gram. Material balance from the chamber tests indicated that the majority (more
79
-------
than 68%) of the MEKO in the paint applied was emitted. Due to the relatively fast emission
pattern, more than 90% of the MEKO emissions occurred within 10 hours after painting. The
alkyd paint MEKO emissions can be simulated by a first-order decay model. The model indicated
that the MEKO emission is mostly gas-phase mass transfer controlled.
The first-order decay model can be used as input to the continuous-application source
term of an IAQ model to predict indoor MEKO concentrations during and after the application of
an alkyd paint. The IAQ model indicated that MEKO emitted from alkyd paints can cause indoor
MEKO concentration to exceed suggested indoor exposure limits. The elevated MEKO
concentrations can last for more than 10 h after the painting is finished. The indoor MEKO
concentration can be reduced by selection of lower MEKO alkyd paint, implementation of higher
ventilation, and isolation of the source room with local ventilation. The higher ventilation should
be maintained about 2 h after the painting is finished to avoid exposure to residual MEKO
emissions.
80
-------
\^r h a p te r D
IN e w Findings /\boi_it /\ I d e h y d e tmissions from /\lkyd raint
Using the small chamber apparatus as the basis of their testing methods, NRMRL staff
have made significant discoveries about the behavior and emissions of alkyd paints in the actual
conditions in which they are used. For example, NRMRL found that some alkyd paints release
odorous aldehyde compounds that are generated as the paint dries in a film. Only by analyzing the
gaseous emissions of a test material covered with alkyd paint was NRMRL able to detect and
model the generation of such compounds.
The unpleasant "after-odor" which can persist for weeks after application of alkyd paint
has been the cause of customer complaints and indoor air quality 0AQ) concerns. Hancock et al.
(1989a and 1989b) investigated the byproducts produced duringthe drying of a number of oil-
modified alkyd resins. They found that a number of odorous aldehydes were produced from the
reactions between unsaturated fatty acids and oxygen during the drying process of an alkyd-based
paint film
The unsaturated fatty acids, mostly derived from naturally occurring vegetable oils, are
usually included in alkyd paints to facilitate the formation of a polymer film. After alkyd paint is
applied to a surface at room temperature, the unsaturated fatty acids in the paint undergo
spontaneous reactions with atmospheric oxygen (autoxidation reaction) to produce
hydroperoxides (Porter et al., 1980). Under the influence of suitable promoters (catalysts) such as
cobalt, lead, benzil, or methanesulphonic acid, the hydroperoxides decompose to give free
radicals. The free radicals generated are responsible for the subsequent cross-linking of the
polyester polymers (the alkyd resins) by hydrogen abstraction, dimerization, or some other
mechanism to form a paint film with desirable physical/chemical properties. However, the
hydroperoxides not only generate free radicals, but also decompose by fragmentation to give
81
-------
byproducts, mainly aldehydes with lesser amount of ketones and alcohols. These byproducts are
responsible for the characteristic acrid odor during the drying process of alkyd paint since many
of the aldehydes formed are volatile and have pungent odors.
The objective of theresearch work described in this chapter of the report is to characterize
aldehyde emissions from alkyd paints. Small environmental chambers (ASTM D5116, 1990)
were used to measure the aldehyde emissions from three commercially available alkyd paints
applied to pine boards. The major aldehyde species emitted were identified, and the emission
rates of the most abundant one, hexanal, were quantified. The chamber data were analyzed by a
consecutive first-order reaction model to identify the rate-controlling step. The model, finally,
was used to simulate IAQ for the prediction of hexanal concentrations.
txperimental Work
Experiments were designed to generate aldehyde emission data from a newly applied
alkyd paint as it dried for more than 100 hours under controlled experimental conditions. Tests
were conducted in EPA's small chamber source characterization facilities consisting of
electropolished stainless steel chambers (Tichenor, 1989). The facilities allowed close control of
temperature, relative humidity, and air flow rate in the chambers. Small fans were used in the
chambers to provide a velocity near (1 cm above) the test surface of 5 -10 cm/s which is typical of
indoor environments. The standard test conditions were:
Air exchange rate (N) 0.5 h"1
Temperature 23 °C
Inlet relative humidity 50%
Nominal wet paint film thickness 80 |im
Substrate specimen surface area 0.0256 m2 (0.16 x 0.16 m)
Chamber volume 0.053m3
Loading (L) 0.48 m2/m3
Three commonly used alkyd paints, designated as paints A-l, A-2, and A-3, manufactured
by three major U.S. paint companies were acquired for testing (Table 6-1). Those
82
-------
I able O~ I . Info r m ati o n /\b out the rri m er and I hree lest raints
JT r od u ct \-s od e I\ a m e
Paint Type
Color
Density (g/cm3)
Recommended Wet Film
Thickness
rri me r
~
white
1.33
4 mils
(102 |im)
Paint A-l
semi-gloss
seafoam
1.26
3 mils
(76 |im)
Paint A-2
gloss
no pigment
1.10
4 mils
(102 |im)
Paint A-3
gloss
no pigment
N/Aa
46.4-5 1.0m2/gal.
(74.2-81.4 |im)
a not available
alkyd paints were selected after consulting with paint suppliers on the popularity of medium-
priced paints. The purchased paint was mixed in cans and then split into 150 mL amber vials.
Each vial was used for only one test.
The paints were analyzed for hexanal (the most abundant aldehyde emitted) contents by
extracting a paint sample (usually 1 g) in 10 mL methylene chloride and analyzing the extract by
GC/MS.
A pine board purchased from a local lumber supplier was used as the substrate. The board
was cut into 16 x 16 cm sections and the exposed edges were sealed with sodium silicate solution
to minimize any adsorption of aldehydes. For each experiment, the pine board substrate was first
treated with a primer recommended by paint suppliers. Experimental data showed that the primer
(Table 6-1) emits far less (more than an order of magnitude) hexanal than the paints. After 48
hours of drying, an alkyd paint was applied to one side (the side to be exposed to chamber air) of
the substrate specimen using a roller. The mass applied was determined using a protocol
involving weighing of the substrate before and after the application. The painted samples were
placed in the chamber promptly (within 3 minutes after paint application). The test start time
(t=0) was established when the door to the chamber was closed. The chamber was flushed with
clean air (<5 |ig/m3 VOCs) before each test. A typical test lasted for about a week with the clean
air flow through the chamber continued at a controlled rate.
83
-------
Air samples in the chamber outlet were collected based on EPA Method IP-6A
(Winnberry et al., 1990) using 55 to 105 jim chromatographic-grade silica gel coated with
acidified 2,4- dinitrophenylhydrazine (DNPH). Sample volumes ranged from 2 to 30 L with
sample flow in the range of 200 to 400 cnrVmin controlled by mass flow controllers. Each sample
was extracted with 5 mL acetonitrile (ACN). Twenty-five jiL of extract was analyzed on a
Hewlett Packard 1090 high-performance liquid chromatograph (HPLC). Chromatography was
performed with a C-18 reverse-phase column (4.6 x 250 mm) using a gradient program of 45%
ACN to 30 minutes, then 75% ACN to 35 minutes, then 100% ACN to 41 minutes, and finished
with 45% ACN to 55 minutes.
The HPLC was calibrated for nine carbonyl compounds: formaldehyde, acetaldehyde,
propanal, benzaldehyde, pentanal, m-tolualdehyde, methyl isobutyl ketone, hexanal, and heptanal.
The nine targeted compounds were identified by comparison of their chromatographic retention
times with those of the derivatized standards. Quantification was performed using an external
standard method with a five-point calibration based on peak area of derivatized standards.
Performance of the instrument was verified each day by analysis of a quality control check
sample prior to starting the sample analysis.
rxesuilts and LJ i s c LJ s s i o n
Bulk Analysis for Hexanal
No detectable amount of hexanal was found in any of the three alkyd paints tested. The
detection limit of bulk analysis was determined by the analytical instrument, the sample size, and
the extraction/dilution procedures. Typically, the instrument detection limit for hexanal was about
3 ng. Taking into account the 1.0 |j,L injection and the 10 mL extraction of 1 g of paint, this
detection limit corresponds to 0.03 mg of hexanal per gram of paint. Therefore, the bulk analysis
results indicated that the hexanal content in any of the three alkyd paints tested was below
0.03 mg/g.
84
-------
Chamber Emission Data
Significant aldehyde emissions were detected by DNPH sampling of the chamber air for
all three alkyd paints tested. All nine carbonyl compounds that HPLC was calibrated with were
detected in the chamber air samples. The three most abundant carbonyl compounds detected from
each paint are listed in Table 6-2. The emission profiles (time/concentration curves) of the
common most-abundant aldehyde, hexanal, from paints A-l, A-2, and A-3 are presented in
Tables 6-3 to 6-5 and Figures 6-1 to 6-3. It is seen that, for each paint, the hexanal concentration
in the chamber remained low for the first few hours after painting. Later, the hexanal
concentration increased to reach a peak followed by a relatively slow decay which lasted for more
than 100 hours.
The total amount of hexanal emitted from each paint, as shown in Table 6-2, was
estimated by integrating the time/concentration curves from time 0 to 200 h. The amount of
hexanal emitted was greater than 2 mg/g which confirms that hexanal was formed after painting
since the bulk analysis did not detect any hexanal.
85
-------
I able D~^. I he I otal tmissions for the I hree IVIost /Abundant /Aldehydes
\ i n m g / g p a i n t)
Paint A-l Paint A-2 Paint A~3
Hexanal
Pentanal
Propanal
2.04
0.26
0.08
2.33
0.33
0.26
2.48
0.48
0.30
86
-------
I able D-O. Hexanal l_>o nee ntrati o n in I est Chamber for Alkyd Paint r\-\
relapsed (Concentration
Time (h) (mg/m3)
1.14 0.045
2.30 0.032
2.30 0.035
3.37 0.029
4.54 0.043
6.24 0.321
6.24 4.67
10.93 5.12
20.38 4.16
22.49 3.53
26.79 1.56
26.79 1.25
30.45 0.768
48.00 0.225
53.86 0.114
71.47 0.118
173.74 0.079
87
-------
I able D-4. Hexanal Concentration in I est Crhamber for Alkyd Paint
JZljla p se d
Time (h)
1.20
2.30
2.30
4.56
6.55
6.55
11.43
21.61
21.61
23.07
25.20
25.20
30.51
47.83
53.63
76.41
100.37
143.61
v^oncentration
(^ing/m )
0.777
0.975
0.930
0.570
0.785
1.02
3.44
4.46
2.59
3.29
2.85
2.62
1.92
1.45
1.10
0.896
0.590
0.328
88
-------
I able D~O. Hex a n aI v-*oncerTtration in I e s I v-*h a m b er f o r /\l ky d r a i nt /\~O
JZljla p se d
Time (h)
0.33
1.26
2.37
4.63
6.38
10.99
22.19
25.00
30.13
47.17
56.80
95.88
148.96
195.83
v^oncentration
(^ing/m )
0.051a
0.142
0.375
4.42
9.82
4.98
2.15
1.77
1.44
0.811
0.562
0.325
0.167
0.122
Below practical quantification limit, but above method detection limit.
89
-------
40 80 120
Elapsed Time (h)
160
200
Figure 6-1 . H
i i s s io n fro m r a i nt r\~ I a n
cJ m o cJ e I i n g results.
12
Dl
Data
odel
40 BO 120
Elapsed Time (h)
Figure 6-2. H
180
ex a n aI e m
i s s io n fro m r a i nt r\~ L- and m odeling resul
90
200
-------
12
9
I
Data Model
.2 6 I
1 I
£ 1
S 3 1
c J I
o I
O
oil
40 80 120 160 200
Elapsed Time (h)
Figure D~O. Flexanal e m i s s io n fro m Faint /\~O and m odeling resu Its .
Hexanal Formation Mechanism
Hancock et al. (1989a and 1989b) found hexanal as one of the major volatile byproducts
resulting from the autoxidation reaction of fatty acids. The autoxidation and hexanal formation
process can be accounted for by a chain reaction mechanism as illustrated by the following
fragmentation scheme involving the ester of a fatty acid, methyl linoleate.
-H-, isomerization
CH3(CH2)4CH=CHCH2CH=CH(CH2)7COOCH3
methyl linoleate O2, +H-
91
-------
-OH-
CH3(CH2)4CHCH=CHCH=CH(CH2)7COOCH3
OOH
a hydroperoxide
scission
CH3(CH2)4CHCH=CHCH=CH(CH2)7COOCH3
O
a free radical
CH3(CH2)4CHO + -CH=CHCH=CH(CH2)7COOCH3 (6-1)
hexanal
The above-mentioned hexanal formation mechanisms indicate that oxygen from the air is
needed for the autoxidation reactions. It is known that, after the alkyd paint is applied to a
substrate, there is an induction period before the alkyd coating begins to take up oxygen from the
air (Byrnes, 1996). This induction period caused a delay of hexanal production from the alkyd
paints tested in current experiments and was reflected by the low hexanal concentration in
chamber air for the first few hours after painting. The mechanisms also involve a series of
intermediate reactions, and hexanal is the one of the final byproducts. Chemical kinetic theories
(Charles, 1977) indicate that, for consecutive reactions, the formation rate of the end products
should always exhibit a pattern including a peak followed by a relatively slow decay. This pattern
of time/formation rate relationship was also reflected by the hexanal concentration profiles in the
chamber air shown in Figures 6-1 to 6-3.
Hexanal Emission Model
Assuming that the formation of hexanal can be described by a set of simplified
consecutive first-order reactions:
92
-------
k! k2
[A] —> [I] —
(6-2)
where
[A] = alkyd resins related to hexanal formation (mg/m2),
[I] = intermediates related to hexanal formation (mg/m2),
[H] = hexanal formed (mg/m2),
kj = first order reaction rate constant (h"1), and
k2 = first order reaction rate constant (h"1).
Assume that there i s a delay ta before the initiation of the formation reaction. The reaction
rates can then be expressed as
d[A]
dt
d[I\
- - k, [A] Ua(f) (6-3)
= (k, [A] - k2 [/]) Ua(t) (6-4)
a?
—--— = &2 [7] t/a(0 (6-5)
where Ua(t) represents the unit step function reflecting the time lag ta and its value is 1 if t > ta,
otherwise 0.
The initial conditions of Equations (6-3) to (6-5) are: [I] = [H] = 0 and [A] = A0 at t = 0.
Solving Equation (6-3) by integrating from an initial condition of t = 0 and [A] = A,, yields
[A] - A0e--" (6-6)
Substituting Equation (6-6) into (6-4) and solving (6-4) by Laplace transform yields
93
-------
Substituting Equation (6-7) into (6-5) and assuming that hexanal emission is controlled by the
chemical reactions and all the hexanal formed is emitted, which means that the emission factor
EF(t) is the formation rate for H, yields:
EF(f) = k2[I\= > ( e -*, _
or
EF(t) = EFQ(e~kl (t~'a) U^ - e~^(t~^ UJf}) (6-9)
where EFQ -
As the proposed source model for aldehyde emissions from alkyd paint, Equation (6-9)
has the following characteristics: 1) It gives an initial emission factor of zero when t < ta; 2) It
gives the maximum emission factor EFmax at t = ta + In^/kj) / (k2 - kj; and 3) After reaching the
maximum, the emission factor decays slowly.
The concentration profile of hexanal in a well-mixed chamber can be obtained by material
balance with Equation (6-9) as the source term:
^ = -NC + LEF(f) (6-10)
at
where
C = hexanal concentration (mg/m3),
N = air exchange rate (h"1), and
L = loading (m"1).
94
-------
Solving Equation (6-10) yields
- *! (t- g ua(t) _ -N(t- ta) ua(t) - *2 ('- Q ua(t) _ -N(t- ta) ua(t)
= L EF0 ( * - -— ^ - - « - -— ^ - ) (6-11)
N - k N - &
The parameter A0 in Equation (6-9) reflects the total hexanal emitted which can be
obtained by integrating the concentration/time curves shown in Figures 6-1 to -3. The estimated
values of A0 of the three alkyd paints tested are listed in Table 6-2. The values of the three
remaining parameters (kl3 k2, and delay time ta) can be estimated by the best fit of Equation (6-11)
to the time/concentration curve using a non-linear regression method. The estimated values of the
three parameters by fitting Equation (6-11) to the concentration profiles shown in Figures 6-1 to
6-3 for the three alkyd paints tested are listed in Table 6-6. It is seen that each paint has a
different set of values for kj and k2 which reflects the fact that the fatty acids and promoters in the
three paints are different from each other. Also the values of kj are smaller than those of k2 which
indicates that the formation rate of intermediates is much slower than that of hexanal. Therefore,
the hexanal emissions were mostly controlled by the chemical reaction step during which
intermediates were formed.
I able D~D. tstimated IVIodel Parameters for Hexanal Formation from the I hre
Alkyd Paints Tested
EF0 (mg/m2/h)a
k, (h-1)
k2 (h-1)
Uh)
Paint A-l
9.41
0.056±0.012
0.744±0.414
4.3±1.4
Paint A-2
4.8
0.024±0.002
0.617±0.332
5.0±1.0
Paint A -3
11.9
0.056±0.009
1.52±1.60
2.U0.78
a Estimated from integrating the time/concentration curves.
95
-------
Indoor Air Quality Simulation
Simple IAQ simulation was performed by using Equation (6-11) to estimate hexanal
concentration after indoor use of alkyd paint in an example house, assuming that alkyd paint A-2
is used to paint the kitchen and bathroom cabinets (with a total area of 10 m2 being painted) of a
residential house (with a volume of 300 m3). The indoor air of the house is well-mixed with an air
exchange rate of 0.5 h"1. The estimated values of model parameters for paint A-2 shown in
Table 6-6 were used for the simulation. Figure 6-4 shows the indoor air hexanal concentrations
predicted by the simulation results. Also shown is the hexanal odor threshold (4.5 ppb) reported
in the literature (ASTM, 1978). It is seen that the indoor hexanal level remained above the odor
threshold for about 120 hours.
Table 6-2 indicates that hexanal is not the only aldehyde emitted. Other aldehyde
emissions such as pentanal and propanal also have very low odor threshold values, 12 and 9 ppb,
respectively (ASTM, 1978). The combination of those aldehydes and indoor sink effects can
result in a strong and irritating odor that lasts for weeks.
v>»o nc I LJ <
Three different alkyd paints were tested in small environmental chambers to characterize
the aldehyde emissions. Emission data indicated that significant amounts of odorous aldehydes
(mainly hexanal) were emitted from alkyd paints during the air-drying period. Bulk analyses
showed that the alkyd paint itself contained no aldehydes. Reaction kinetics and mass balance
calculations indicate that the aldehydes emitted should be produced after the paint was applied to
a substrate. The aldehydes emission patterns are consistent with the theory that they were formed
as byproducts from spontaneous autoxidation reactions of unsaturated fatty acids (in the applied
paint) with atmospheric oxygen. Chamber data showed that the major volatile byproducts from
the three alkyd paints tested were hexanal, propanal, and pentanal.
96
-------
jQ
Q.
CL
E
_g
-I—*
ns
4— >
E
0)
O
O
O
I
60 =
:
50
40 =
30 =
20 .
.
•
10 =
0 -
••-^
c
=• i
i "V-
::! , : ^,.... Odor Threshold =4.5 ppb
i '•—":•;-...
-- -""":.:::::"i:::::-
) 40 80 120 160 20
Time (h)
Figure O-T-. predicted hiexanal concentration in a typical house after alkyd paint
application.
The hexanal formation and emission rates can be simulated by a consecutive first-order
reaction model with an initial time lag. The time lag reflects an induction period after painting
during which little oxygen was taken up by the alkyd coating. As the final byproduct of a series of
consecutive first-order reactions, the hexanal emission rate should increase from zero to reach a
peak followed by a slow decay as confirmed by chamber concentration data. The modeling
results also showed that the hexanal emissions were mostly controlled by the chemical reactions
to form intermediates as the precursors to hexanal production.
The slow-decay aldehyde emission pattern results in prolonged exposure by occupants.
IAQ simulation indicated that the hexanal concentration due to emissions from an alkyd paint in
an indoor application could exceed the reported odor threshold for about 120 hours. Prolonged
97
-------
occupant exposure to the aldehydes emitted from an alkyd paint could cause sensory irritation and
other health concerns.
98
-------
v_y h a p te r /
Oubstrate tffects on \f vJ v_x t m issions fro m a Latex raint
NRMRL's paint testing practices are applicable to indoor latex paints as well as alkyd
paints. This and the following two chapters present examples of how latex paint emissions have
been studied using the small environmental chamber. The present chapter describes how NRMRL
discovered the important effect that choice of substrate has on the chemical profile of VOC
emissions.
Interior architectural coatings can increase indoor air pollution due to emissions of VOCs
(WHO, 1989; Clausenet al., 1991). As of 1992, more than 85%of the interior coatings used in
the United States were latex paint (National Paint & Coatings Association, 1992a and b).
Methods of assessing latex paint emissions have been developed to determine cumulative mass
emissions of VOCs for purposes of assessing their impact on the ambient air, specifically for their
contributions to photochemical smog (Brezinski, 1989). In indoor environments, the concern is
directed to determining the time-varying exposure of occupants to TVOCs, as well as specific
VOCs.
Environmental test chambers have been used to measure the VOC emission profiles under
simulated indoor conditions. These emission profiles can be used as a basis for exposure risk
assessment (ASTM D5116, 1990). The tests involve applying the selected latex paint to a
substrate and monitoring the chamber VOC concentration changes resulting from emissions of
the painted substrate as a function of time. The test substrates used by most previous emission
tests include stainless steel (Clausen et al., 1991; and Clausen, 1993) and glass (Sheldon and
Naugle, 1994). Other substrates such as aluminum pans (Stromberg and Wind, 1968; Hansen,
1974) and glass plates (Rosen and Andersson, 1990; Sullivan, 1975) were also used to study the
solvent evaporation and drying mechanisms of latex paints. Those substrates have the advantages
99
-------
of being non-porous, non-absorbent, and easy to handle. However, latex paint is seldom applied
to these substrates in real indoor environments. The surfaces most commonly painted with latex
paint in the United States are walls and ceilings made of gypsum board. The gypsum board has a
porous surface and absorbs liquid; this characteristic can alter the behavior of applied latex paint
and change the VOC emission patterns.
To evaluate the substrate effects on time-varying VOC emissions, environmental chamber
tests were conducted to measure the emissions from a latex paint applied to two different
substrates — a stainless steel plate and a gypsum board — under identical experimental conditions
(Tichenor, 1995; Krebs et al., 1995). The objective of this research is to report the summarized
experimental data and to assess the substrate effects on latex paint VOC emissions.
txperimental Work
Experiments were designed to generate VOC concentration data from newly applied latex
paint as it dried for several hundred hours under controlled experimental conditions. Tests were
conducted in EPA's small chamber source characterization facilities consisting of electropolished
stainless steel chambers (Tichenor et al., 1990). The facilities allowed close control and
monitoring of temperature, relative humidity, and air flow rate in the chambers. Small fans were
used in the chambers to provide a velocity near (1 cm above) the test surface of 5 - 10 cm/s which
is typical of indoor environments. The standard testing conditions were:
Air exchange rate (N) 0.5 h"1
Temperature 23 °C
Inlet relative humidity 50%
Nominal wet paint film thickness 100 |j,m
Substrate specimen size (A) 0.0256 m2 (0.16 x 0.16 m)
Chamber volume (V) 0.053 m3
Chamber loading (L = A/V) °'48 m2/m'
The latex paint used in this study was selected after consulting with local paint suppliers
on the popularity of medium-priced paints. Sufficient paint from the same lot was purchased to
100
-------
conduct the small chamber tests. The purchased paint was mixed in cans and then split into
150 mL amber vials. Each vial was used for only one test.
The paint was analyzed for water content, VOCs, and density by ASTM Standard Test
Methods (ASTM, 1989) and other related methods. Because the ASTM methods did not give
information on the individual VOCs, these compounds were analyzed by extracting a paint
sample in acetone and analyzing the extract by GC/MS and GC/FID. These analyses allowed the
individual VOC compounds to be identified by MS and quantified by FIDs. Paint samples were
periodically analyzed by GC to assess the stability of the target compound concentrations.
The substrates evaluated were newly purchased bare stainless steel and gypsum board.
The gypsum board was paper-coated with a total thickness of 0.127 cm (0.5 in.) commonly used
for walls and ceilings in residential and commercial buildings. The edges of the gypsum board
samples were sealed with sodium silicate to minimize any adsorption of the VOCs. The latex
paint was applied to the substrate specimen by a roller. It was estimated that 4.2 and 3.6 g of the
paint were applied to the stainless steel and the gypsum board, respectively. The mass applied
was determined using a protocol involving weighing of the substrate, the paint, and the roller
before and after the application. The painted samples were placed in the chamber shortly (within
5 minutes) after the paint was applied. The test start (t=0) was established when the door to the
chamber was closed. The chamber was flushed with clean air (with less than 5 i-ig/m3 TVOC)
before each test. A typical test lasted for about 2 weeks with the clean air flow through the
chamber continued at a controlled rate. The test duration was prolonged to more than 11 months
for a long-term test.
VOCs in the chamber outlet were collected on Tenax sorbent tubes and analyzed by
GC/FID. The desorption unit was a combination of an Envirochem Multisorbent Desorber and
Unacon sample concentrator. The analytical column was a J&W 30 m x 0.53 mm (Megabore)
DB-Wax operated at 40°C for 5 minutes and ramped at 5°C per minute to 240°C. The lower
quantification limits (i.e., the lowest calibration levels) for the target compounds were from 26 to
37 ng on GC column, which is equivalent to a chamber concentration range of 0.0026 to 0.0037
101
-------
mg/m3 for 10-L samples. Of the VOC samples collected, 16% were duplicates. The average
relative standard deviation between duplicates was 5% to 17% for individual compounds and 9%
for TVOCs.
rxe s LJ It s and Discussion
Paint Composition
Table 7-1 shows the paint composition obtained by ASTM methods (ASTM, 1989). Table
7-2 shows five major VOCs (propylene glycol, ethylene glycol, 2-(2-butoxyethoxy)ethanol, 2,2,4-
trimethyl-l,3-pentanediol monoisobutyrate (TPM also known as Texanol), and diethylene glycol)
detected by GC analysis. The TVOC content in the bottom of each column in Table 7-2 was the
sum of the five VOC contents in the same column. Comparison between Tables 7-1 and 7-2
indicates that the TVOC contents determined by the ASTM methods are lower than those by GC
analysis. Since the ASTM TVOCs were determined by the relatively small difference between
total volatiles and water content, it is suspected that the results from the ASTM analysis are
biased for latex paint. Besides, the ASTM gravimetric method determining total volatiles
(including water) involves heating the diluted sample in an oven at 110±5°C for 60 min.
However, the boiling points (ranging from 187°C for propylene glycol to 244°C for TPM) for the
VOCs in the paint are much higher than the oven temperature. It is suspected that the evaporation
in the oven was incomplete at the specified temperature. Furthermore, at temperatures over
100°C, the paint polymerization is accelerated and may form a film over the residue which could
significantly decrease evaporation. The GC VOC data are believed to be more reliable than the
ASTM values and will be used for mass balance calculations.
102
-------
I a b le / - I . LJeter m i n ati on of Volatile vJr g a n ic v-*o m pounds in the Latex Paint lested
by
ASTM
(ASTM, 1989)
Oa m p le 1L/
Non-Volatiles
Total Volatiles (incl. water)
Water Contenf
Density (g/mL)
TVOC Content (mg/g)
1
57.23%
42.77%
39.91%
1.41
28.6
2
57.16%
42.84%
40.31%
1.42
25.3
3
57.14%
42.86%
39.93%
1.43
29.3
Mean
57.20%
42.8%
40.1%
1.42
28.0
ASTM method D4017 (Karl Fischer)
I able /~^. I ota I and Individual VvJv-*s Determined by \Jv-* /Analysis v-*o nee ntr ati <
Units (mg/g)
O a m p le 1L/
v^ompound
Propylene Glycol
Ethylene Glycol
Butoxyethoxyethanol
TPM
Diethylene Glycol
TVOCs
la
2.52
23.6
5.20
14.0
0.98
46.3
lb
2.46
25.8
5.15
14.0
0.56
48.0
4d
2.14
23.3
4.88
13.2
0.43
44.0
4e
2.16
23.3
4.68
12.6
0.37
43.1
IVlean
2.32
24.0
4.98
13.5
0.59
45.4
103
-------
Substrate Effects on VOC Emissions
Figures 7-1 to 7-4 show the chamber air VOC concentration profiles (concentration vs.
time curves) during the 2-week test period for the four major VOCs emitted from the latex paint
applied to the two substrates. A significant substrate effect on latex paint VOC emissions is
evidenced by comparing the VOC concentration profiles between stainless steel and gypsum
board tests. Figures 7-1 to 7-4 indicate that, for each VOC, the chamber concentrations resulting
from the gypsum board substrate were considerably lower than those from the stainless steel
substrate during the 2-week test period. Table 7-3 shows the comparison of peak VOC
concentrations measured from the two tests. When the latex paint was applied to the gypsum
board instead of the stainless steel plate, the peak concentrations of propylene glycol, ethylene
glycol, butoxyethoxyethanol, and TPM decreased by 81%, 90%, 56%, and 23%, respectively.
I a b I e /~O. v-*omparison of Peak v-*oncerTtrations Measured In the tnvironments
v-*h
a m b er s i n m g m
)
c
o m p o u n d
Ota inless
Oteel Plate
*Jy p s u m
Doa r d
Ethylene Glycol
Propylene Glycol
BEE
TPM
TVOCs
76.8
10.0
11.6
23.3
122
7.88
1.88
5.14
18.0
32.9
104
-------
iao
d
o
o
C
o
O
100
80
60
40
20
D SS Plate
0 Gyp, Board
— M odel
Figure 7-1 . Eff
100 200 300
El ap sed Tim e (h)
'ect of substrate on ethylene glycol emissions.
400
s
WJ
a
d
-------
o
'*J
d
(J
a
o
O
12
10
100 200
Elapsed Time (h)
" SS Plate
0 Gyp. Board
— Model
300
400
Figure / ~O. t
ff ect s of s
ubstrate on 4—-
~to utoxyeth oxyj eth a n o I e m is s i o n s .
d
O
d
o
O
30
25
20
1 5
1 0
0 100 200
Elapsed Tim e (h)
Figure /~T-. tffect of substrate on I r IVI emissions.
300
400
106
-------
The differences are also reflected by the amount of each compound emitted. The amount
of each VOC emitted from the latex paint in the 2-week test period (W) was estimated by
integrating the concentration curve (Colombo and De Bortoli, 1992; Guo et al., 1996b):
where
(7-1)
Q = air exchange flow rate (nrVh),
C = chamber concentration of a given compound (mg/m3),
t = time (h), and
n = total number of data points.
and the percent of each VOC emitted in the 2-week test period can be calculated as:
W
Percent emitted = « 100% (7-2)
Wo
where
W0 = The amount of VOCs in the latex paint applied (mg).
Table 7-4 shows that 89% to 100% of the VOCs were emitted when stainless steel was the
substrate. However, only 9% to 29% of the VOCs were emitted during the same period when
gypsum board was the substrate. Similar to the peak concentration reduction, the total emission
reduction was not uniform among the four major species. Ethylene glycol had the largest
reduction (from 100% to 9%) and TPM the smallest (from 89% to 29%). The other significant
difference in emission patterns is that the peak concentrations appeared earlier with gypsum
board than with stainless steel and this happened for all the individual compounds (see Figures 7-
1 to 7-4).
107
-------
Table 7-4. Weight Percentage of VOC in the Latex Paint Emitted in the First
336-Hour Testing Period (in %)
v^om pound
Ethylene Glycol
Propylene Glycol
BEE
TPM
TVOCs
Ot ainless
Oteel Plate
100
89
93
89
97
VJy p s u m
Jjoa r d
9
17
14
29
20
Substrate Effects on Composition of VOC Emissions
When the same paint was applied to the two substrates, the compositions of VOC
emissions were dramatically different. As shown in Figure 7-5, when stainless steel was the
substrate, ethylene glycol dominated the VOC emissions for the first 100 h, and TPM was the
dominant VOC after that. On the other hand, when the gypsum board was the substrate, the trend
was reversed. TPM became the dominant VOC emitted for the first 100 h, and ethylene glycol
dominated the VOC emissions after that (Figure 7-6).
Emission Models
Chang and Guo (1992b) analyzed wood stain emission data that look similar to current gypsum
board data, by using a double-exponential model:
EF(t) = Rl + R2 = Rloe-klt + R2Qe~^ (7-3)
108
-------
100
80
o
ctf
fl
ft
-------
where
EF(t) = Emission factor of a VOC (mg/m2/h),
R! = Phase 1 emission factor (mg/m2/h),
R2 = Phase 2 emission factor (mg/m2/h),
R10 = Phase 1 initial emission factor (mg/m2/h),
kj = Phase 1 emission rate decay constant (h"1),
R20 = Phase 2 initial emission factor (mg/m2/h), and
k2 = Phase 2 emission rate decay constant (h"1).
Integrating the chamber mass balance equation with the source term defined by
Equation (7-3) and assuming an initial concentration of zero gives:
C = L
N- ki N- k2
(7-4)
Based on the techniques suggested by Chang and Guo (1992b), the double exponential
model was used to analyze the chamber data using a non-linear regression curve fit routine
implemented on a microcomputer. It was found that Equation (7-4) provides an adequate
representation of the chamber concentration profiles resulting from gypsum board VOC
emissions (see Figures 7-1 to 7-4). The stainless steel emission data can also be described by the
double exponential model by assuming that the phase 2 emissions are negligible (R20 = 0).
Tables 7-5 and 7-6 list the values of parameters estimated by the non-linear regression fit.
Emission Mechanisms
Chang and Guo (1992b) indicated that the double exponential model can represent a two-
phase emission process. In this previous work, wood stain was applied to hardwood. The phase 1
emissions, Rl3 correspond to the period shortly after the stain was applied while it was still
relatively wet. During this phase, it appeared that VOC emissions were related to evaporation
processes, characterized by relatively fast emissions. These fast emissions caused a rapid
110
-------
I able / ~O. Oui m m ary of rara m eter s of the Double txponential IVIo del f o r V vJv-*
t m i s s i o n s from Painted vJypsum Doard
v^om pound
Ethylene
Glycol
Propylene
Glycol
BEE
TPM
RIO
(mg/m /h)
18.5
4.78
10.8
29.7
ki Rio/ki
(h-1) (mg/m2)
0.696 26.6
0.724 6.60
0.331 32.6
0.795 37.4
R20
(mg/m /h)
1.96
0.561
0.475
15.9
k2
0.00694
0.0115
0.00767
0.0317
(mg/m )
282
48.8
61.9
502
I able /~D. Oui m m ary of rara m eters of the LJouble txponential IVIo del f o r \f vJv-*
t m i s s i o n s fro m r a i nted Ota i n le s s Otee I \rx20 = /
C
o m p o u n d
\m g / m / hj
Rig/ki (mg/
Ethylene Glycol
Propylene Glycol
BEE
TPM
100
14.0
13.4
30.0
0.0235
0.0368
0.0166
0.0169
4255
380
807
1775
111
-------
depletion of organic compounds on the surface of the wood, resulting in a rapid rise and decline
of the chamber concentrations. A similar phenomenon was observed in the latex paint data for the
first 50 to 100 h, as shown in Figures 7-1 to 7-4.
The phase 2 emissions, R2, corresponded to the period when the wood stain applied was
relatively dry. There was evidence that the emissions were controlled by diffusion through a solid
(the wood), making the emission rates low, but the emission lasted for a long time. The latex
paint behaved in a similar fashion, as reflected by the small values of R20 and k2 (compared with
the corresponding R10 and k^ values) as shown in Table 7-5.
According to the model, R^/k] and R20/k2 represent the total quantities of organic species
emitted in phases 1 and 2, respectively. The values of those two parameters estimated from the
chamber data are also listed in Table 7-5. It is seen that for gypsum board, the VOCs emitted in
phase 2 far exceeded those in phase 1. It is likely that the majority of the latex paint VOCs
penetrated into the relatively porous gypsum board and became relatively dry shortly after
application. Only a small fraction of the paint VOCs stayed wet on the surface of the gypsum
board, resulting in the relatively short phase 1 emissions. The rest of the VOCs applied were
probably imbedded in the gypsum board and had to diffuse through the substrate to be emitted as
the phase 2 emissions.
However, the opposite was observed when the stainless steel plate was the substrate.
Almost all of the VOCs in the latex paint applied were emitted in phase 1. The phase 2 emissions
were so small that, for modeling purposes, the emission rate was negligible and the value of R20
was set at 0 (see Table 7-6). This is most likely due to the fact that the stainless steel was
impervious and all of the latex paint applied remained on the surface. As a result, it took longer
for the latex paint to dry, and the majority of the VOCs were emitted rather rapidly while the
paint was still relatively wet.
112
-------
Long-Term Emission Data
Figure 7-7 shows the temporal history of the composition of the VOC emissions from a
painted gypsum board over an 11-month period. As mentioned in the section above, entitled
"Substrate Effects on Composition of VOC Emissions," TPM had the highest initial emission
rate, but ethylene glycol exceeded TPM after about 100 h and continued to be the dominant VOC
emitted over the long term. Even after 11 months, an elhylene glycol concentration of 0.1 mg/m3
was still being measured in the chamber air. This approximates a steady state emission rate of
about 0.1 mg/m2/h. The levels of other VOCs were near the quantification limit of the sampling
and analytical methods used. Assuming a constant emission rate of 0.1 mg/m2/h, after 1 year
about 40% of the ethylene glycol will be emitted, and it will take as long as 3.5 years for all of the
ethylene glycol to be released.
Although the double-exponential model indicated that the phase 2 VOC emissions should
last for a long time, it failed to predict that the ethylene glycol emission can persist for more than
11 months. The discrepancy between the model predictions and the chamber data is probably due
to the empirical nature of the model which oversimplifies the emission process. A physical model
based on mass transfer fundamentals which take into account the characteristics of physical and
chemical processes is needed to fully represent the long-term behavior of latex paint VOC
emissions.
v>»o n c I LJ <
Environmental chamber tests showed significant differences in emission rates and patterns
between the VOCs released from the same latex paint applied to two different substrates—a
stainless steel plate and a gypsum board. After the first 2 weeks, over 90% of the VOCs were
emitted from the paint on the stainless steel plate, but less than 20% had left the gypsum board.
The dominant species in the VOCs emitted in the first 100 h also changed from
113
-------
100
5-1
•J-j
P!
-------
ethylene glycol to TPM when stainless steel was replaced with gypsum board. Data analysi s by a
double-exponential model indicated that the majority of the VOC emissions from the painted
stainless steel might be simulated by an evaporation-like phenomenon with fast VOC emissions
controlled by gas-phase mass transfer. On the other hand, only a small fraction of the VOCs
emitted appeared to be controlled by the evaporation-like drying process from the painted gypsum
board. The majority of the VOCs were emitted after the painted gypsum board surface was
relatively dry and were probably dominated by a slow, solid-phase, diffusion-controlled mass
transfer process. Long-term experimental data indicated that it may take as long as 3.5 years to
release all the VOCs in the paint applied to the gypsum board.
Therefore, when the objective of the test is to provide emissions data that are relevant to
understanding the emissions behavior in typical indoor environments, instead of "ideal"
substrates such as glass, aluminum, or stainless steel, "real" substrates such as wood and gypsum
board should be used to evaluate the time-varying VOC emissions and drying mechanisms of wet
products like latex paint.
115
-------
• hap te r
8
_ x p e r i
ntal Work to Evaluate Low-VOU Paint;
This chapter exemplifies the usefulness of NRMRL's paint testing practices for evaluating
new paint products. It deals with a new class of paints developed and marketed on the grounds
that they emit very low levels of VOCs as they dry. A standard paint testing practice provides an
avenue by which environmental and public health agencies can research the properties of new
products like these.
Conventional water-based latex paints use solvents and additives which contain volatile
organic compounds (VOCs) to provide the following functions: binder/coalescing agent, polymer
plasticizer, freeze/thaw stablizer, defoamer, and carriers for other additives such as colorants,
thickening agents, surfactants, and biocides (Turner, 1988). As the paint dries, these solvents and
additives evaporate and release VOCs (Clausen et al., 1991). When the paint is applied to interior
surfaces, the VOCs resulting from the drying process cause increased indoor VOC concentrations
and associated indoor air pollution problems. The major VOC emissions from conventional latex
paints include propylene glycol (PG), ethylene glycol (EG), 2-(2-butoxyethoxy)ethanol (BEE),
and 2,2,4-trimethyl-l,3-pentanediol monoisobutyrate (TPM a.k.a Texanol) (Chang et al., 1997).
The elevated indoor air VOC concentrations are often related to complaints of unpleasant odors
and sometimes irritant and allergic reactions by building occupants.
Considerable efforts have been made by the entire coatings industry to decrease levels of
VOCs in coatings and paints (Klein, 1993; Bjorseth andMalvik, 1995). In recent years, newly
developed interior architectural coatings, advertised as "low-VOC," "low-odor," or sometimes
116
-------
"no-VOC" latex paints, have been commercially available. Those interior coatings are labeled as
"VOC free" based on EPA Reference Method 24 (40 CFR, 1994). They are also recommended by
their manufacturers as an alternative to conventional paints to reduce indoor air pollution.
However, the relevance of replacing conventional latex paints with low-VOC coatings to reduce
indoor air pollution has not been studied. While the off-gassing from conventional paints is a
cause of indoor air pollution, no emission data from low-VOC paints are available. In addition,
the VOC-containing solvents and additives are normally used to improve several important
properties of conventional paints. Eliminating the VOCs in paint formulations may compromise
the performance of the paint and make it improper for certain uses.
To address those issues, NRMRL evaluated four commercially available low-VOC latex
paints. The purpose was to provide experimental data on the emission characteristics and the
coating performance. The results can be used to assess whether those paints are suitable
substitutes for conventional latex paints for preventing indoor air pollution and reducing VOC
exposure. The tests conducted included bulk analyses, chamber emission tests, and performance
evaluation.
Lx p er i m e nta
Wor,-
Four commercially available low-VOC latex paints (designated as paints L-l, L-2, L-3,
and L-4) and a reference paint (a commonly used conventional latex paint designated as paint 0)
were acquired for the evaluation. Table 8-1 lists the characteristics of the low-VOC latex paints
based on the manufacturers' specifications. EPA Method 24 was used to determine the total
volatile matter content, water content, and the total VOC content of each paint. The method is the
same as that used by manufacturers for labeling purposes and to support the claims about low-
VOC paints. The concentrations of individual VOCs in the bulk liquid paints were determined by
a GC/MS method adapted from EPA Method 311(40 CFR 1996). The paints were diluted with
either acetone or acetonitrile at a rate of 1 g of the paint with 10 mL of solvent. Acetone formed
an emulsion with some paints, requiring the use of acetonitrile. Samples of the diluted and
centrifuged paints were analyzed by injection onto the GC column.
117
-------
Table 8-1. The LowVOC Latex Paints Tested
Pa int
Color
Density, kg/L
Solid content, vol.%
Water content, wt.%
VOCs, g/L
Coverage, m2/L
L-l
Antique white
1.35
33.4
49.4
0
9.8-11.1
L-2
Antique white
1.33
38
45.4
1
9.8
L-3
White
1.21
NAa
40.7
0
12.3-19.7
L-4
White
1.25-1.31
26-32
55.4
0
9.8-11.1
a NA - Information not available from manufacturer-supplied data.
Emissions tests were conducted in the EPA's small chamber source characterization test
facility. Each paint was applied to the top side of a glass plate (16 cm by 16 cm) by a roller. Glass
plate was used as the substrate to maximize the VOC emissions to facilitate sampling and
chemical analysis. The painted glass plate was weighed to ensure the amount of the paint applied
was adequate before it was placed inside a 53-L electropolished stainless steel chamber with the
painted side exposed to the chamber air. Each test started at the moment when the chamber door
was closed. Each chamber was operated at 0.5 h"1 air exchange rate, 50% relative humidity, and
23°C. A small fan was running continuously to ensure good mixing of the air inside the chamber.
Air samples were collected at the exit of each chamber on Tenax tubes for analysis of VOCs and
dinitrophenylhydrazine(DNPH)-treated silica gel cartridges for determination of aldehydes. The
details of the test protocol and analytical procedures can be found in the literature (Chang et al.,
1997; Fortmann et al., 1998).
The physical performance of the paints was evaluated by testing each paint with seven
standardized methods adopted by the American Society of Testing and Materials (ASTM). The
seven methods were selected based on information in the ASTM Standard Guide for Testing
Latex Flat Wall Paints (ASTM 2931) and discussions with paint testing laboratories. The paint
performance evaluated included specular gloss (ASTM 523), hiding power (ASTM 2805), scrub
resistance (ASTM 2486), washability (ASTM 3450), sag resistance (ASTM 4400), drying time
118
-------
(ASTM 1640), and yellowness index (ASTM E3 13). A summary description of each method is
provided in Table 8-2. In addition to the four low-VOC paints, the reference paint (paint 0) was
also tested as a control.
rxesuilts and LJ i s c LJ s s i o n
Bulk Analysis
The VOCs identified by the GC/MS analyses included EG, PG, BEE, TPM, and
dipropyleneglycol. Table 8-3 shows that the VOC concentrations in the liquid paints were quite
low and some (expressed as BDL) were below the detection limit of 0.01 mg/g. Paint L-2 had the
highest TVOC concentration (defined as the sum of the five VOCs quantified) of 3.05 mg/g,
which is still more than 1 order of magnitude lower than that (45.39 mg/g) of paint 0 (a
conventional latex paint) as shown in Table 8-3.
Analyses based on Method 24 (40 CFR, 1994) were also performed to estimate the VOC
content of the four low-VOC latex paints. The results ranged from -1.7% to 0.8%, which
indicated that Method 24 is an inaccurate method for estimating the VOC content of coatings that
contain significant amounts of water and small amounts of VOC. Equation 8-1 is used by the
method to calculate VOC content:
- Ww (8-1)
where
Wvoc = VOC content estimated based on Method 24 (wt.%),
Wvol = total volatile content (wt.%), and
Ww = water content (wt.%).
119
-------
8-2. ASTM Meth<
U
s e d f o
r P
erf o r m ance I e s ti n g
ASTM
Method
Method description
D523 This test method identifies the specular gloss. The reading is made with a glossmeter at an
85° angle. The difference between 1.6 and 2 is not visible to the naked eye. The range
from 4 to 6 is much more noticeable, and the range from 9 to 11 shows significant light
reflecting off the surface for a flat paint.
D2805 This is an instrumental method to measure the coverage hiding power of the paint. The
paint is applied to a standard chart with a wet film thickness of 1.5 mils which represents
one coat of paint and a wet film thickness of 3 mils which represents two coats of paint.
The paint is allowed to air dry and measurements are made. The contrast ratio is reported.
Generally, a contrast ratio in the range from 0.90 to 0.95 indicates poor hidingpower, and
the range from 0.95 to 1.00 indicates good hiding power. Readings below 0.90 indicate
very poor hiding power.
D2486 This test method determines the resistance of latex flat wall paints to erosion caused by
scrubbing. The paint is applied to a standard chart and allowed to dry for 7 days. The chart
is then placed in a machine that scrubs the surface with a brush and an abrasive cleanser.
The reported number indicates how many cycles weie requited before wearing through the
dried paint film. The analytical laboratory that performed the tests indicated that the
average number of cycles for most paints is between250 and 500.
D3450 This test method covers the relative ease of removing soilant discolorations from the dried
film of an interior coating by washing with either an abrasive or non-abrasive cleaner. The
paint is applied to a standard chart and allowed to dry for 7 days. Then a dark oil stain is
applied to the paint and baked on the surface overnight. The next day the stain is washed
with a detergent and a sponge. The higher the percentage rating, the more of the stain that
was removed by the method. Performance of the paint is evaluated relative to other paints.
D4400 This method uses a multi-notched applicator to determine the sag resistance of aqueous and
non-aqueous liquid coatings at any level of sag resistance. The method used for the flat
latex paint has the value of 12 as the perfect Anti-Sag Index. Values between 9 and 12 are
considered good Anti-Sag Indexes. Any Index number lower than 7 is considered poor.
D1640 This method measures the time it takes for the coating to dry to touch in minutes (finger
test).
E313 The Yellowness Index method compares the differences in the whiteness of the initial dry
film before it is expo sed to sunlight to the whiteness after it has been exposed to sunlight.
The lower the initial number, the whiter the paint. The difference in the initial number and
after- exposure number indicates the effect of sunlight on the film. A negative number
indicates that the film has bleached in the sun, and a positive number indicates yellowing.
Differences less than ±0.2 are not detectable.
120
-------
I able O~O. rxesults of Dulk /Analyses
Taint
L-l
L-2
L-3
L-4
^a conventional
latex paintj
Ethyl ene
Glycol
24.00±1.20a 0.04±BDL 0.59±0.01 BDL
BDL
Propylene 2.32±0.19
Glycol
Dipropylene BDLb
Glycol
2-(2-butoxy- 4.98±0.25
ethoxy)ethanol
0.03±BDL 0.09±BDL 0.16±0.01 BDL
O.lliBDL 0.81±0.03 BDL
0.14±0.01
TPM
TVOCsc
13.50±0.68
45.39±2.32
O.lliBDL 1.51±0.01 0.01±BDL BDL
0.03±BDL 0.05±BDL BDL BDL
0.32±BDL 3.05±0.05 0.17±0.01 0.14±0.01
a Standard deviation
b Below the method detection limit estimated to be 0.01 mg/g of the paint
0 Total volatile organic compounds = sum of all the volatile organic compound concentrations
For low-VOC latex paints, Ww is usually in the range of 40 to 55%, and Wvoc could be
0.3% or less. Even if Wvol is absolutely accurate and precise, a 1% error of Ww can cause more
than 200% error of Wvoc. However, the within-laboratory analytical precision of Ww and Wv is
expected to be 2.9 and 1.5%, respectively (40 CFR, 1994). As a result, the Method 24 data of
Wvoc scattered a wide range and often showed negative numbers. Therefore, Method 24 is not an
adequate method to quantify the VOC content for those waterborne coatings.
Emissions
Very little VOC emissions were detected in the chamber air samples of the four low-VOC
paints. Among the four paints, paint L-2 had the highest emissions which reflects the highest
VOC content as indicated in Table 8-3. Figure 8-1 shows the TVOC concentration profile for
121
-------
140
to
to
6
'oi)
fl
o
i-H
*J
£
OJ
o
fl
o
U
Paint 0
(A conventional latex paint)
0
10
20
30
40
50
Time, h
Figure
O ~ I . v-* o m parison of I \f vJ v-* e m i s s i o n p ro file of paint \—-J— w ith that of paint U ^a conventional la te x p a i n t) .
T TV/Of"* n no / ^
he m ethod detection M m it f o r IV v^O wa s U.U£ m g/ m .
-------
emissions from paint L-2. Also shown is the TVOC profile of paint 0 (a conventional latex paint).
It is seen that the paint L-2 TVOC concentrations are at least 1 order of magnitude lower than that
of the conventional paint.
However, significant aldehyde emissions were detected from two low-VOC latex paints
(paints L-l and L-3 as illustrated in Figures 8-2 and 8-3, respectively). The aldehyde emissions
detected included formaldehyde, acetaldehyde, propanal, and benzaldehyde. Among them, the
most abundant aldehyde species was formaldehyde as shown in Figures 8-2 and 8-3. The
measured peak concentrations of formaldehyde were 3.15 mg/m3 for paint L-l and 5.53 mg/m3
for paint L-3. The peak concentrations of the second abundant aldehyde, acetaldehyde, were 0.11
and 0.30 mg/m3 for paints L-l and L-3, respectively. In addition to the high peak concentrations,
the formaldehyde emissions lasted for more than 2 days for both paints (Figures 8-2 and 8-3). The
formaldehyde concentrations in the chamber air samples taken 50 h after testing were 0.011 and
0.223 mg/m3 for paints L-l and L-3, respectively. The estimated mass of formaldehyde emitted
during the 50 h test period was 0.26 and 0.51 mg/g from paints L-l and L-3, respectively.
Formaldehyde is a listed hazardous air pollutant by the Clean Air Act Amendments
(CAAA, 1990). Literature data showed that formaldehyde is a primary upper respiratory tract
irritant and its odor is characterized as "pungent" (AIHA, 1989). Symptoms of eye, nose, and
throat irritation, such as tearing, running nose, and a burning sensation in these areas, are
relatively common with formaldehyde exposure. Formaldehyde is also classified as a probable
human carcinogen based on sufficient evidence in animal studies (Grindstaff et al., 1991).
Formaldehyde is listed as a California Proposition 65 carcinogen and as a Toxic Air Contaminant
by the California Air Resources Board (CARB) based on potential carcinogenicity.
The formaldehyde emissions from low-VOC latex paints are of special concern since
those paints are promoted as the "perfect choice" for use in occupied buildings during normal
business hours, without evacuating entire building sections. The newly painted room purportedly
123
-------
to
Pi
O
iH
-4-J
Pi
€>
U
P!
O
U
10
5
1
0.5
Formaldehyde
Acetaldehyde
—*—
Benz aldehyde
Time, h
Figure O~^. /Aldehyde emission profiles of paint l_~ I. I he method detection limits was U.UUU/ mg/m for
f o r m aide hyd e and acetaldehyde, and U.UU I 4 m g/ m for benzaldehyde.
-------
to
ca
n
be
pu
t
ba
in
us
e
al
m
OS
t
im
m
ed
iat
el
y
to
10
C
o
cd
u
u
C
O
U
o.i
o.oi
o
Formaldehyde Acetaldehyde
Benzaldehyde
10
20 30
Time, h
40
50
Figure O~O. /Aldehyde emission profiles of paint I_~O. I he method detection limit was
U.UUU / m g/ m for f o r m aldehyde, acetaldehyde, and p r o p a n a I, and U.UU 14 mg/ m
f o r benzaldeh yd e .
-------
maintain productivity. In addition, the low-VOC latex paints are promoted for use in occupied
hospitals, extended care facilities, nursing homes, medical facilities, schools, hotels, offices, and
homes where extended evacuation of the entire building section for painting is difficult and often
not desirable. However, occupants in those buildings frequently include those who are vulnerable
and susceptible. When there is indoor air pollution by a potentially harmful VOC, such as
formaldehyde released from the paint, the exposure potential maybe high since the occupants
may be located in the newly painted room and some of the normal precautions may be neglected.
Performance Evaluation
The results for performance testing are summarized in Table 8-4. It is seen that the
performance characteristics of the four low-VOC latex paints varied significantly.
I able O~T-. rxesults of Performance I esting
Paint
0
(a conventional
latex paint)
L-l
L-2
L-3
L-4
Specular gloss
Hiding power
Contrast ratio: 1.5 mils wet
3.0 mils wet
Scrub resistance (cycles)
Washability (%)
Sag resistance
Dry to touch (minutes)
Yellowing index
1.6
0.973
0.987
508
50
12
12
-0.01
2
0.968
0.987
254
41
12
14
-0.07
1.5
0.982
0.998
2000+
47
12
14
0.14
9.4
0.928
0.961
2000+
36
12
11
-0.79
4.7
0.966
0.982
49
50
12
15.5
0.03
126
-------
For specular gloss, the higher the value, the more light is reflecting off the surface.
Table 8-4 shows that the measured values ranged from 1.5 (not visible) to 9.4 (significant light
reflection) with paints 0 and L-2 at the low end. As to hiding power, the measured contrast ratios
are good for all of the paints except paint L-3 at 1.5 mils wet film thickness. For scrub resistance,
paints L-2 and L-3 exhibited the best performance which even exceeded that of the conventional
latex paint. On the other hand, paint L-4 exhibited extremely poor scrub resistance with the
number of cycles (49) far less than the typical range (250 to 500) of conventional paints. As to
washability, paints 0 and L-4 scored the highest, and paint L-2 was only slightly lower. Sag-
resistance was measured with a multi-notched applicator. A perfect score is 12, which was
achieved with all five paints. The dry to touch time ranged from 11 to 15.5 minutes, with the
fastest time exhibited by paint L-3 which had the lowest water content (Table 8-1). The inverse
relationship between dry to touch time and water content was also suggested by paint L-4, which
had the highest water content and exhibited the slowest time. The yellowness indices measured
were relatively small and would not be considered significant.
Based on the ASTM test results, paint L-2 appeared to perform better than the other three
low-VOC paints for most of the features assessed. Table 8-4 also shows that, when compared
with paint 0 (the conventional latex paint used as a reference), paint L-2 also provided equivalent
or superior performance on most of the categories evaluated.
wonclusjons
Based on the experimental data of the four paints tested, the following conclusions are
drawn:
One commercially available low-VOC latex paint (i.e., paint L-2) emitted
considerably less VOCs and performed equal to or better than a conventional latex
paint used as a control. However, the data also indicated that the rest of the low-
VOC latex paints tested either had some inferior properties or emitted hazardous
air pollutants.
127
-------
VOC content data based on Method 24 are not accurate enough to quantify the
VOC contents of low-VOC latex paints for quality control and product ranking
purposes. Other methods, such as EPAMethod 311, are more suitable, especially
when individual VOC content data are needed.
Chamber test data indicated that, despite having low total VOC content in the
paint, emissions of individual VOCs and hazardous air pollutants (e.g.,
formaldehyde and acetaldehyde) can still be significant. Due to the use pattern
(i.e., no full evacuation during painting and immediate re-occupation after
painting) suggested by the manufacturers, the intimate exposure of sensitive
occupants to those hazardous air pollutants are of special concern.
In addition to VOC content data, consumers need emission information and
performance evaluation results to assist in making purchasing decisions about low-
VOC latex paints.
128
-------
\^r h a p te r <3
t x p e r i m e n ta I W o r k to Characterize and rxeduce Formaldehyde
t m issions fro m l_ o w ~ V \J (^* r a i n t
One of the conclusions of the previous chapter was that some of the "low-VOC" latex
paints that were tested actually released significant quantities of formaldehyde as they dried. This
chapter provides an example of how small chamber testing methods can fit into an integrated
pollution prevention (P2) effort—in this case, an effort aimed at reducing formaldehyde
emissions from certain low-VOC latex paints.
Formaldehyde emissions have been detected from certain latex paints (Hansen et al.,
1987). The formaldehyde emissions from low-VOC latex paints are of special concern since those
paints are often used by sensitized and allergic consumers as an alternative to conventional latex
paints to avoid VOC exposure. Low-VOC latex paints are also promoted for use in occupied
buildings during normal business hours, without the necessity of evacuating entire sections of the
building. It is claimed that the newly painted room may be put back in use almost immediately to
maintain productivity. The cost of painting can also be reduced by eliminating premium pay for
night or weekend painting work. As a result, low-VOC latex paints are often sold for use in
occupied hospitals, extended care facilities, nursing homes, medical facilities, schools, hotels,
offices, and homes where extended evacuation of the entire building section for painting is
difficult and often not desirable.
Formaldehyde has been recognized as a hazardous air pollutant (CAAA, 1990), an
odorous respiratory tract irritant (AIHA, 1989), and a probable carcinogen (Grindstaff et al.,
1991). The World Health Organization guideline for indoor air formaldehyde concentration is
0.1 mg/m3 (WHO, 1987). The California Air Resources Board recommends for homes an "action
level" of 0.12 mg/m3 and a "target level" of 0.06 mg/m3 or lower (CARB, 1991). The U. S.
129
-------
Department of Housing and Urban Development recommends that indoor formaldehyde
concentrations from all sources should not exceed 0.5 mg/m3 (HUD, 1984). The California
Environmental Protection Agency suggests an acute 1-h exposure concentration of 0.17 mg/m3,
based on early symptoms of eye irritation (OEHHA, 1995). Those guidelines and
recommendations are established for the general population. However, occupants of buildings
painted with low-VOC latex paints often include those who are vulnerable and susceptible. In
case of indoor air pollution by formaldehyde released from the paint, the exposure and associated
health risk will be higher than average since the occupants are located in the newly painted room
and some of the normal precautions may be neglected.
To address those concerns, NRMRL evaluated four commercially available low-VOC
latex paints (Chang et al., 1999), as described in Chapter 8. The results indicated that, compared
with conventional latex paints, the VOC contents and emissions of those paints are considerably
lower. However, formaldehyde emissions were detected from two of the four paints tested.
NRMRL conducted additional studies with the low-VOC paint that had the highest formaldehyde
emissions. Long-term environmental chamber tests were performed to characterize the
formaldehyde emission profiles. A mathematical model was developed to interpret the chamber
data. Biocide was identified and confirmed as a major source of the formaldehyde in the paint.
The paint was reformulated as a pollution prevention effort to reduce formaldehyde emissions.
The objective of this research is to summarize the results from the additional studies and to
present the newly developed mathematical model for formaldehyde emissions.
tx p er i m e nta I procedure
Experiments were designed to generate formaldehyde emission data from the newly
applied low-VOC latex paint as it dried for more than 350 hours under controlled experimental
conditions. Tests were conducted in lEMB's small chamber source characterization facilities,
along the lines of the experimental conditions described in Chapter 7.
130
-------
Gypsum board purchased from a local supplier of building materials was used as the
substrate to provide a realistic emission profile. It was prepared, loaded with paint, and placed in
the small chamber as described in Chapter 7, Experimental Work section. A typical test lasted for
about 2 weeks with the clean air flow through the chamber continued at a controlled rate.
Air samples in the chamber outlet were collected based on EPA Method IP-6A (Winberry
et al., 1990) using 55 to 105 jim chromatographic-grade silica gel coated with acidified 2,4-
dinitrophenylhydrazine (DNPH). Sample volumes ranged from 2 to 30 L with sample flow in the
range of 200 to 400 cnrVmin controlled by mass flow controllers. Each sample was extracted with
5 mL acetonitrile (ACN). Twenty-five jiL of extract was analyzed on a Hewlett Packard 1090
high-performance liquid chromatograph (HPLC). Chromatography was performed with a C-18
reverse-phase column (4.6 x 250 mm) using a gradient program of 45% ACN to 30 minutes, then
75% ACN to 35 minutes, then 100% ACN to 41 minutes, and finished with 45% ACN to 55
minutes.
The HPLC was calibrated for formaldehyde. Quantification was performed using an
external standard method with a five-point calibration based on a peak area of derivatized
standards. Performance of the instrument was verified each day by analysis of a quality control
check sample prior to starting the sample analysis.
Lx p er i m e
ntal Date
The formaldehyde concentrations measured at the chamber exit are shown in Figure 9-1
(as squares). The first sample of the test, taken within 30 min after the start of the experiment, had
the highest concentration. The formaldehyde concentrations decreased rapidly and monotonically
by about 1 order of magnitude in the first 10 h of the test. Chang and Guo (1992a) suggested a
fast organic emission model to interpret such results. The model assumed that a portion of the
formaldehyde in the paint was emitted to the chamber air almost instantly (perhaps in a few
minutes) at the start of the experiment. As a result, there was a sharp increase of
131
-------
10
to
E
"o>
E,
c
o
"-^
CD
at
o
£Z
O
O
1 -
0.1 -
0.01
Original (O)
Nobiocide(N)
Different biocide (D)
Model (O)
Model (N)
-Model (Dl
100
200
Elapsed Time (h)
300
400
Figure c/~ I. v-*omparison of chamber data with model predictions. l\lote that model
pred icati o n s f o r paints with no biocide and with a d if fere nt biocide are not
d iff ere nti a b I e .
-------
formaldehyde concentration as soon as the painted gypsum board was placed inside the chamber.
The first "puff of formaldehyde emission created a relatively high initial chamber concentration,
which then decayed rapidly over the initial 10-h period.
As the experiment proceeded beyond 10 h, the formaldehyde concentration decrease
slowed considerably. After 100 h, the formaldehyde concentration in the chamber decayed at a
very slow pace and gave a long tail to the time/concentration curve. When plotted on a semi-
logarithmic scale, the tail section of the time/concentration curve became almost a straight line as
shown in Figure 9-1. Chang and Guo (1992b) characterized these emissions as process controlled
by the solid-phase diffusion of the organic species. The formaldehyde had to diffuse through the
relatively dry and dense paint embedded in the substrate. The internal diffusion process made the
emission rates low, but the emission lasted for a long time.
t m i s s i o n IVI o d e I
It was found that the formaldehyde emissions from the paint applied to the wallboard can
be represented by first-order decay in-series model as shown in Figure 9-2. The empirical model
assumed that the formaldehyde in the applied paint was distributed in two layers. The top layer
can be considered as the thin surface coating section of the paint. Formaldehyde was emitted from
the top layer directly into the chamber air. The formaldehyde emissions from the top layer were
assumed to be controlled by diffusion through the surface coating section. The bottom layer can
be seen as the paint beneath the surface and embedded in the substrate (wallboard). The
formaldehyde in the bottom layer had to migrate to the top layer before it was emitted. The
internal diffusion in the embedded paint is the controlling mechanism of the formaldehyde
migration from the bottom to the top layer.
133
-------
Chamber Air
km
Top Layer
M
Bottom Layer
B
Figure \j~ L- Oche m atic of the first~order d ec ay in~series m ode
134
-------
The model also assumed that the formaldehyde emissions can be divided into three stages.
First, there was an almost instant "puff-like release of formaldehyde immediately after the paint
was applied to the gypsum board. The "puff release caused an initial peak concentration of
formaldehyde, C0, which is equivalent to the total instant formaldehyde emission divided by the
chamber volume. In the second stage, the top layer formaldehyde was almost depleted, and the
chamber air formaldehyde concentration decreased by more than 1 order of magnitude. The third
stage formaldehyde emissions mainly came from the bottom layer and were controlled by the
slow diffusion process.
With the assumption that both the formaldehyde emission and migration can be described
by the first-order decay model (Clausen et al., 1991), the following mass balance equations can be
written to represent the formaldehyde mass transfer between the two layers and the chamber air:
dB
= -kbB (9-1)
dt
dM
dt
dC
= h B - km M 0-2)
dt
= - N C + L km M (9-3)
Initial conditions: C= C0, M=M0, and B=B0 at t=0.
where
B = the amount of the formaldehyde in the bottom layer (mg/m2),
kb = the first-order decay constant for formaldehyde migration from the bottom
layer (V),
M = the amount of formaldehyde in the top layer (mg/m2),
k^ = the first-order decay constant for formaldehyde emission from the top
layer (h"1), and
C = formaldehyde concentration in the chamber air (mg/m3).
135
-------
Equations (9-1) to (9-3) can be solved by Laplace Transform, and the analytical solution for C is:
CA ^— . A -—m i A „— kbt (n A\
= Ai e + Ai e + Ai e (9-4)
where
C0 = Ai + A2 + A3 (9-5)
Mo = - - - — - - - - (9-6)
L Km
A3 (N - h) (km - h)
J~*' A-771 &•&
In addition,
t, , ,
M = Mo e** * + (9-8)
Km- h
B= £oe-**f (9-9)
LJata /\n a lys i s
Of the six parameters in Equation (9-4), N is determined by the air exchange rate (0.5 h"1)
used for the chamber testing, and the other five need to be estimated. Difficulties were
encountered when a non-linear regression curve-fitting routine was used to estimate the values of
the five parameters. However, it was found that the five unknown parameters in Equation (9-4)
could be estimated by an exponential peeling procedure (Serber and Wild, 1989; M01have et al.,
1995) for a three-compartment analysis. The procedure was based on the assumption that
136
-------
N»km»kb>0. When this assumption is valid, the first and second terms on the right-hand side of
Equation (9-4) become negligible for t large, and Equation (9-4) is reduced to:
ln(C) = ln(A3) - kbt (9-io)
A plot of ln(C) against t for t large should give a straight line with intercept ln(A3) and slope -kb.
The peeling procedure was initiated by the estimation of the long-term parameters, A3 and
kb, using a least-square regression method to fit Equation (9-10) to the chamber data between 100
and 300 h shown in Figure 9-1. The estimated values of A3 and kb are listed in the second column
of Table 9-1. Subsequently, the same iterative regression method was used to estimate A2 and k^
by fitting Equation (9-11) to the chamber data between 10 and 300 h. (Note that the values of A3
and kb are already known.)
C= A2e~htf + Ase'*** (9-11)
Finally, Aj was estimated by setting N=0.5 and fitting Equation (9-4) to the entire set of
chamber data shown in Figure 9-1. The estimated values of Al3 A2, and k^ and their standard
deviations are listed in the second column of Table 9-1.
It is seen from Table 9-1 that the value of N (0.5) is considerably greater than that o
(0.0745) which is much greater than that of kb (0.0042). Therefore, the assumption of the
exponential peeling procedure, N»km»kb>0, was validated. Also, the values of k^ and kb are in
the same range (from 0.1 to 0.001 h"1) as estimated before for solid-phase diffusion-controlled
emissions (Chang and Guo, 1992b).
137
-------
I able b/~ I. tstimated Values of Parameters of tquation
A,,
A2,
A3,
km,
kb,
ITa r a m ete r
mg/m3
mg/m3
mg/m3
h'1
h'1
Origin al
paint
1.926±0.027
0.513±0.021
0.274±0.005
0.0745±0.0031
0.0042±0.0001
Witho ut
b ioc ide
2.563±0.069
0.592±0.067
0.129±0.018
0.0518±0.0062
0.0066±0.0009
With a different
b ioc ide
2.549±0.070
0.602±0.086
0.130±0.015
0.0547±0.0081
0.0066±0.0008
a expressed as mean ± standard deviation
The goodness-of-fit of the model to the experimental data was evaluated based on the
ASTM criteria (ASTM, 1995). The ASTM criteria (the fifth column of Table 9-2) include the
slope and intercept of the best-fit line between measured and predicted values, the correlation
coefficient between measured and predicted concentrations, the magnitude of prediction error
relative to the measured values represented by normalized mean square error (NMSE), and the
fractional bias.
The estimated values of the five goodness-of-fit criteria for the low-VOC latex paint
tested are listed in the second column of Table 9-2. It is clear from Table 9-2 that all five ASTM
criteria were met and the first-order decay in-series model is adequate for simulating
formaldehyde emissions from the low-VOC latex paint as shown in Figure 9-1.
When all six model parameters are known, the values of C0, M0, and B0 can be
calculated by Equations (9-5), (9-6), and (9-7), respectively. The calculated values are shown in
the second column of Table 9-3. The significance of the formaldehyde emissions from this low-
VOC paint is reflected by the peak chamber concentration, C0, of 2.71 mg/m3. This concentration
is considerably higher than any of the guidelines mentioned in the beginning of the chapter.
138
-------
I able b/~^. v-*omparison of vj LJ a ntitati ve Measures of \J o o d n e s s of IVIodel with /\O I IVI
Criteria (ASTM, 1995)
Slope
Intercept
Correlation coefficient
NMSEa
Fractional bias
Original
paint
0.998
0.000377
0.999
0.0015
-0.0024
Without
biocide
0.951
-0.0236
0.993
0.0333
-0.0083
With a different
biocide
0.998
-0.0108
0.996
0.0101
0.0136
ASTM
criteria
0.75 to 1.25
-0.25 to 0.25
>0.9
<0.25
<0.25
a Normalized mean square error
The total amount of formaldehyde per unit area of the applied paint can be estimated by
integrating Equation (9-4) and multiplying the result byN/L. On the other hand, the total amount
of formaldehyde can also be calculated as the sum of C0V/A, M0, and B0 based on the model
assumptions The term C0V/A is equivalent to the amount of formaldehyde in the initial "puff."
Table 9-3 indicates that the amount of formaldehyde emitted in the initial "puff," although
resulting in the high peak concentration, accounts for only 6.8% of the total formaldehyde
emitted. The majority (80.4%) of the formaldehyde emitted, according to the model, came from
the bottom layer (i.e., B0) via the slow solid-phase diffusion-controlled process.
I able y~O. tsti mated /\mouint of l~ o r m a I d e hy d e in the r a i nt/\p p I ie d
C0, mg/m3
M0, mg/m2
B0, mg/m2
Total (C0V/A+M0+B0), mg/m2
LJrigin al
paint
2.71
10.0
63.2
78.62
Without
b ioc ide
3.28
12.65
17.80
37.24
With a different
b ioc ide
3.28
12.70
17.53
37.02
139
-------
Since all the parameters are known, Equations (9-8), and (9-9) can be used to estimate the
amount of formaldehyde remaining in the paint which is equivalent to the sum of M and B. The
results are shown in Figure 9-3. It is seen that (at t=350 h) 15.4 mg/m2, or 19.6% of the total
emittable formaldehyde, was still remaining in the paint. The model predicted that it would take
about 1056 h for 99% of the formaldehyde in the paint to be emitted.
Oo LJ rce I n
vesticjation
The paint manufacturer was informed about the test results, and the paint formula was
reviewed. The paint manufacturer pointed out that the biocide used to preserve the paint might
contain as much as 5% formaldehyde, and it was suspected that the biocide could be a major
source of the formaldehyde. To confirm this hypothesis, the paint manufacturer provided a
sample that contained no biocide for chamber testing.
The chamber test data with no biocide are also shown in Figure 9-1 (as circles). It is seen
that the formaldehyde concentration profile followed the same pattern as the original one. But the
formaldehyde concentration decreased faster and the tail section was considerably lower than that
of the original paint.
The chamber concentration data with no biocide were analyzed by the first-order decay in-
series model and the exponential peeling procedure. The results are listed in the third columns of
Tables 9-1, 9-2, 9-3, and 9-4. Table 9-1 indicates that the assumption of the exponential peeling
procedure, N»km»kb>0, was still valid, which confirmed that the formaldehyde was emitted in
the same fashion as before. Table 9-2 shows that the first-order decay in-series model was
adequate to simulate the formaldehyde emissions from the paint without biocide as illustrated by
Figure 9-1. The bottom row of Table 9-4 shows that each g of the no-biocide paint contained
0.269 mg of total emittable formaldehyde which is about 56.4% less than that of the original.
140
-------
Table 9-3 indicates that the decreased formaldehyde content significantly affected only the
bottom-layer diffusion-controlled emissions (B0) which was reflected by the lower tail section of
the emission profile shown in Figure 9-1. Figure 9-3 shows that (at t=350 h) only 1.93 mg/m2, or
5.4% of the total emittable formaldehyde, was still remaining in the paint without biocide. The
model also predicted that it would take about 607 h for 99% of the formaldehyde to be released.
Compared with the original paint, the formaldehyde in the no-biocide paint was depleted
considerably faster.
1.0
0.8 -
0.6 -
o
'.jZj
CO
0.4 -
0.2 -
0.0
"Original
'No biocide
Different biocide
50 100 150 200 250
Elapsed Time (h)
300
350
400
Figure c/~O. Fraction of f o r m aldehyde re m aining in the paint.
141
-------
Table 9-4. Calculated For
m aldehyde
v-*ontent in I hree Paints leste
lest paint
Paint applied, g
Substrate area, m2
Total formaldehyde, mg/m2
Formaldehyde content of paint, mg/g
Origin al
paint
3.26
0.0256
78.62
0.617
Witho ut
b io c id e
3.54
0.0256
37.24
0.269
With a different
b ioc ide
3.43
0.0256
37.02
0.276
Therefore, the chamber data confirmed that the biocide was a major source of
formaldehyde in the paint. However, biocide was not the only source: other sources (e.g,
additives and binders) also contributed significantly (43.6%) to the formaldehyde emissions.
BD
iocide rxeplace m e nt
In a pollution prevention effort to reduce formaldehyde emissions through reformulation,
the paint manufacturer decided to replace the biocide with a new one. A sample of the
reformulated paint was evaluated by a small chamber test following the same procedure as
described in Chapter 7, Experimental Work section. The experimental results are also shown in
Figure 9-1 as triangles. It is seen that the formaldehyde emission profile basically coincided with
that of the paint without biocide. Data analysis (Tables 9-1 and 9-2) by the first-order decay in-
series model and the exponential peeling procedure indicated that each g of the reformulated paint
contained 0.276 mg of total emittable formaldehyde which is very similar to that (0.269 mg) of
the paint without a biocide (Table 9-4). Apparently, biocide replacement eliminated a major
source of formaldehyde in the paint and reduced the amount of formaldehyde emission by 55.3%.
When compared with the emissions of the original paint (Table 9-3), only the third-stage long-
term formaldehyde emissions were reduced by biocide replacement.
142
-------
wonGlusions
Environmental chamber data indicated that formaldehyde emissions from a low-VOC
latex paint can cause very high (several ppm) peak concentrations in the chamber air. When the
paint is applied to gypsum board, the formaldehyde emissions decay very slowly after the initial
peak, and the emission can last for more than a month. Certain biocides, when used as a
preservative for latex paints, can be a major source of formaldehyde emissions from the paint.
This particular source was verified by the significant reduction of formaldehyde emissions in the
absence of the biocide.
The formaldehyde emissions from a latex paint applied to gypsum board can be
characterized by a first-order decay in-series model. The semi-empirical model can be used to
estimate the amount of formaldehyde emitted or remaining in the paint. It can also predict the
initial peak concentration and the time of depletion. When the activity patterns of building
occupants are defined, the model can further be used for exposure risk assessment. Note that the
model assumed an initial, instant release of formaldehyde and can be applied only to similar
cases.
The biocide-contributed formaldehyde emissions can be eliminated by paint reformulation
via biocide replacement. However, when other sources (e.g., additives and binders) of
formaldehyde are present in the paint, biocide replacement can reduce only the long-term
emissions. Short-term exposure potential to high peak concentrations of formaldehyde still exists.
Additional research is needed to identify other potential sources of formaldehyde to completely
eliminate formaldehyde emissions from those paints.
143
-------
• ha pte r
1 0
R
ef e r e n c e s
40 CFR. (1996) Method 311—Analysis of hazardous air pollutants in paints and coatings by
direct injection into a gas chromatograph. Code of Federal Regulations, Title 40, Part 63,
Appendix A. Washington, DC.
40 CFR. (1994) Method 24—Determination of volatile matter content, water content, density,
volume solids and weight solids of surface coatings. Code of Federal Regulations, Title 40, Part
60, Chap. 1, Appendix A. Washington, DC.
AIHA. (1989) Odor thresholds for chemicals with established occupational health standards.
American Industrial Hygiene Association. Akron, OH.
ASTM E313 (1997) Standard practice for calculating yellowness and whiteness indices from
instrumentally measured color coordinates. Annual book of ASTM standards. 6(2). American
Society for Testing and Materials. Philadelphia, PA.
ASTM D523 (1997) Standard test method for specular gloss. Annual book of ASTM standards.
6(1). American Society for Testing and Materials. Philadelphia, PA.
ASTM D1640 (1997) Standard test method for drying curing, or film formation of organic
coatings at room temperature. Annual book of ASTM standards. 6(1). American Society for
Testing and Materials. Philadelphia, PA.
ASTM D2486 (1997) Standard test method for scrub resistance of interior latex flat wall paints.
Annual book of ASTM standards. 6(1). American Society for Testing andMaterials.
Philadelphia, PA.
ASTM D2805 (1997) Standard test method for hiding power of paints by reflectrometry. Annual
book of ASTM standards. 6(1). American Society for Testing and Materials. Philadelphia, PA.
ASTM D2931 (1997) Standard test method for testing latex flat wall paints. Annual book of
ASTM standards. 6(1). American Society for Testing and Materials. Philadelphia, PA.
ASTM D3450 (1997) Standard test method for washability properties of interior architectural
coatings. Annual book of ASTM standards. 6(1). American Society for Testing and Materials.
Philadelphia, PA.
144
-------
ASTM D4400 (1997) Standard test method for sag resistance using a multinotch applicator.
Annual book of ASTM standards. 6(1). American Society for Testing and Materials.
Philadelphia, PA.
ASTM D5116. (1997) Standard guide for small-scale environmental chamber determination of
organic emissions from indoor materials/products. Annual book of ASTM standards. 11(3).
American Society for Testing and Materials. Philadelphia, PA.
ASTM D 3686 (1996) Standard practice for sampling atmospheres to collect organic compound
vapors (activated charcoal tube adsorption method). Annual book of ASTM standards. 11(3).
American Society for Testing and Materials. Philadelphia, PA.
ASTM D 3687 (1996) Standard practice for analysis of organic compound vapors collected by
the activated charcoal tube adsorption method. Annual book of ASTM standards. 11(3).
American Society for Testing and Materials. Philadelphia, PA.
ASTM D5157-91 (1995) Standard guide for statistical evaluation of indoor air quality models.
1995 Annual book of ASTM standards. ll(3):484-487. American Society for Testing and
Materials, Philadelphia, PA.
ASTM D5116 (1990) Standard guide for small-scale environmental chamber determination of
organic emissions from indoor materials/products. Annual book of ASTM standards. 11(3):480-
491. American Society for Testing and Materials. Philadelphia, PA.
ASTM D 4017 (1990) Test method for water in paints and paint materials by Karl Fischer
method. Annual Book of ASTM standards. 6(1). American Society for Testing and Materials.
Philadelphia, PA.
ASTM. D3960 (1989) The ASTM standard practice for determining volatile organic compounds
(VOC) contents of paints and related coatings. American Society for Testing and Materials.
Philadelphia, PA.
ASTM. (1978) Compilation of odor and taste threshold values data. American Society for Testing
and Materials, Data Series DS 48A. Philadelphia, PA.
Bjorseth, O., and Malvik, B. (1995) Reduction in VOC emissions from water-based paints. In:
Proceedings of Healthy Buildings 1995, University of Milan and International Center for
Pesticide Safety. Milan, Italy. Vol. 2, pp. 881-886.
Brezinski, J. (1989) Manual on determination of volatile organic compounds in paints, inks, and
related coating products. ASTM Manual Series. MNL 4. American Society for Testing and
Materials, Philadelphia, PA.
Byrnes, G. (1996) Alkyds. Journal of Protective Coatings and Linings. December 1996. 73-87.
145
-------
CAAA (1990) Clean Air Act Amendments. Public Law 101-549. Title El—Hazardous Air
Pollutants. November 15, 1990. 104 STAT 2534.
CARB (1991) Indoor air quality guidelines #1: formaldehyde in the home. California Air
Resources Board, Research Division, Sacramento, CA.
Chang, J.C.S., Guo, Z., Fortmann, R., and Lao, H. (2001) Characterization and reduction of
fomaldehyde emissions from a low-VOC latex paint. Accepted for publication by Indoor Air.
Chang, J. C. S., Fortmann, R. C., Roache, N. F., and Lao, H. (1999) Evaluation of low-VOC latex
paints. Indoor Air 9:253-258.
Chang, J.C.S., Guo, Z.,and Sparks, L (1998) Exposure and emission evaluations of methyl ethyl
ketoxime (MEKO) in alkyd paints. Indoor Air 8:295-300.
Chang, J.C.S., and Guo, Z. (1998) Emissions of odorous aldehydes from alkyd paint.
Atmospheric Environment 32:3581-3586.
Chang, J.C.S., Tichenor, B.A., Guo, Z., and Krebs, K.A. (1997) Substrate effects on VOC
emissions from a latex paint. Indoor Air 7:241-247.
Chang, J.C.S., and Guo, Z. (1994) Modeling of alkane emissions from a wood stain. Indoor Air
4:35-39.
Chang, J. C. S., and Guo, Z. (1992a) Modeling of the fast organic emissions from a wood-
finishing product—floor wax. Atmospheric Environment 26A(13):2365-2370.
Chang, J. C. S., and Guo, Z. (1992b) Characterization of organic emissions from a wood finishing
product—wood stain. Indoor Air 2:146-153.
Charles, G. H., Jr. (1977) An introduction to chemical engineering kinetics and reactor design.
New York, NY. John Wiley & Sons, Inc.
Cinalli, C.A., Hohnston, P.K., Koontz, M.D., Girman, J.R., and Kennedy, P.W. (1993) Ranking
consumer/commercial products and materials based on their potential contribution to indoor air
pollution. In: Proceedings of the 6th International Conference on Indoor Air Quality and Climate.
Indoor Air '93. Helsinki, Finland. Vol. 2, pp 425-430.
Clausen, P. A., Wolkoff, P., and Host, E. (1991) Long term emission of volatile organic
compounds from waterborne paints. Indoor Air 4:562-576.
Clausen, P. A. (1993) Emissions of volatile and semivolatile organic compounds from waterborne
paints—the effect of the film thickness. Indoor Air 3:269-275.
146
-------
Colombo, A., and De Bortoli, M. (1992) Comparison of models used to estimate parameters of
organic emissions from materials tested in small environmental chambers. Indoor Air 2:40-57.
Evans, W.C. (1996) Development of continuous-application source term and analytical solutions
for one- and two-compartment systems. In: Tichenor, B.A., ed., Characterizing sources of indoor
air pollution and related sink effects. Philadelphia, PA. American Society for Testing and
Materials. ASTM STP 1287. pp. 279-293.
Federal Register (1989) 54(176):37799-37810.
Federal Register (1986) 51(220):41329-41331.
Fortmann, R., Roache, N., Chang, J.C.S., and Guo, Z. (1998) Characterization of emissions of
volatile organic compounds from interior alkyd paint. Journal of the Air and Waste Management
Association 48:931-940.
Geherig, R., Hill, M., Zellweger, C., and Hofer, P. (1993) VOC-emissions from wall paints - a
test chamber study. In: Proceedings of the 5th International Conference on Indoor Air Quality and
Climate. Indoor Air 1993. Helsinki, Finland. Vol. 2, pp. 431-436.
Grindstaff, G., Henry, M., Hernandez, O., Hogan, K., Lai, D., and Siegel-Scott, C. (1991)
Formaldehyde risk assessment update. U.S. EPA, Office of Toxic Substances. Washington, DC.
Guo, Z., Fortmann, R., Marfiak, S., Tichenor, B., Sparks, L., Chang, J., and Mason, M. (1996a)
Modeling the VOC emissions from interior latex paint applied to gypsum board. In: Proceedings
of the 7th International Conference on Indoor Air Quality and Climate. Nagoya, Japan. Indoor
Air'96, 1:987-992.
Guo, Z., Tichenor, B.A., Krebs, K.A., and Roache, N.F. (1996b) Considerations on revision of
emissions testing protocols. In: Characterizing sources of indoor air pollution and related sink
effects, ASTM STP 1287, Tichenor, B.A., ed., American Society for Testing and Materials.
Philadelphia, PA. pp. 225-236.
Guo, Z., Tichenor, B.A., Mason, M.A., and Plunket, C.M. (1990) The temperature dependence of
the emission of perchloroethylene from dry cleaned fabrics. Environmental Research 52:107-115.
Hancock, R.A., Leeves, N.J., and Nicks, P.F. (1989a) Studies in autoxidation, Part I—the volatile
byproducts resulting from the autoxidation of unsaturated fatty acid methyl esters. Progress in
Organic Coatings 17:321-336.
Hancock, R.A., Leeves, N.J., and Nicks, P.F. (1989b) Studies in autoxidation, Part II—an
evaluation of the byproduct formed by the autoxidation of fatty acid modified alkyd resins under
the influence of different promoters. Progress in Organic Coatings 17:337-347.
147
-------
Hansen, M. K., Larsen, M., and Cohr, K. (1987) Waterborne paints. Scand. J. Work Environ.
Health 13:473-485.
Hansen, C.M. (1974) The air drying of latex coatings. Ind. Eng. Chem. Prod. Res. Develop.
13(2):150-152.
Howard, E.M., McCrillis, R.C., Krebs, K.A., Fortmann, R., Lao, H.C., and Guo, Z. (1997) Indoor
emissions from conversion varnishes. In: Proceedings of the 1997 Engineering Solutions to
Indoor Air Quality Problems Symposium. Air & Waste Management Association, Research
Triangle Park, NC.
HUD (1984) Manufactured home construction and safety standards; final rule. Department of
Housing and Urban Development. 24CFR Part 3280. Federal Register, August 9, 1984.
49(155):31996-32013.
Klein, RJ. (1993) Formulating low-odor, low-VOC interior paints. Modern paint and coatings,
March, pp. 37-39.
Krebs, K., Lao, H., Fortmann, R., and Tichenor, B. (1995) Test methods for determining short
and long term VOC emissions from latex paint. Presented at: The International Symposium on
Engineering Solutions to Indoor Air Quality Problems. Air and Waste Management Association,
Research Triangle Park, NC.
Krivanek, N. (1982) The toxicity of paint pigments, resins, drying oils, and additives. In:
Englund, A., Ringen, K. & Mehlman, M.A., eds, Advances in modern environmental toxicology,
Vol. II Occupational health hazards of solvents. Princeton Scientific Publishing Co., Princeton,
NJ. pp. 1-42.
Levin, H. (1991) "Environmental Chamber Determinations of VOC Emissions from Indoor Air
Pollutant Sources: Overview of Applications, Protocols, and Issues" in Measurement of Toxic
and Related Air Pollutants, Air and Waste Management Association Special Publication VIP 21.
pp. 403-408.
Levin, H. (1989a) Building materials and indoor air. Problem buildings: building-associated
illness and the sick building syndrome. State of the Art Reviews—Occupational Medicine, Vol. 4,
No. 4. Hanley & Belfus, Inc., Philadelphia, PA.
Levin, H. (1989b) "Problem Buildings: Building Associated Illness and Sick Building Syndrome"
Occupational Medicine, Volume 4, Number 4, October-December 1989, Hanley & Belfus:
Philadelphia, PA.
Levin, H. (1987a) "Protocols to Improve Indoor Environmental Quality in New Construction."
Practical Control of Indoor Air Problems, Proceedings of IAQ '87, May 18-20, 1987 (Atlanta,
GA: American Society of Heating, Refrigerating and Air-Conditioning Engineers, 1987), pp. 157-
170.
148
-------
Levin, H. (1987b) "The Evaluation of Building Materials and Furnishings for a New Office
Building." Practical Control of Indoor Air Problems, Proceedings oflAQ '87, May 18-20, 1987
(Atlanta, GA: American Society of Heating, Refrigerating and Air-Conditioning Engineers,
1987), pp. 88-103.
Levin, H. (1987c) "A Procedure Used to Evaluate Building Materials and Furnishings for a Large
Office Building." Indoor Air '87, Proceedings of the 4th International Conference on Indoor Air
Quality and Climate ed. B. Seifert (Berlin: Institute for Water, Soil and Air Hygiene, 1987), I, 54-
58.
Markarian, J. (2000) Coatings Solvents Continue to Be Pressured, Chemical Market Reporter,
Schnell Publishing Co. Inc., New York, NY, October 16.
Method 311 (1995) Analysis of Hazardous Air Pollutant Compounds in Paints and Coatings by
Injection into a Gas Chromatograph; December 7, 1995; 40 CFR Part 63, Appendix A.
M01have, L., Dueholm, Z., and Jensen, L. K. (1995) Assessment of exposures and health risks
related to formaldehyde emissions from furniture: a case study. Indoor Air 5:104-119.
National Paint & Coatings Association (1992a) U.S. Paint Industry Database. Washington, DC.
National Paint & Coatings Association (1992b) U.S. Paint & Coatings Market Analysis -Industry
Data base (1992), National Paint & Coatings Association, Washington, DC.
http://www.paint.org/contact.htm.
Nielsen, G.D., Hansen, L.E., and Wolkoff, P. (1997) Chemical and biological evaluation of
building material emissions. II. Approaches for setting indoor air standards or guidelines for
chemicals. Indoor Air 7:17-32.
NIOSH (1994a) Manual of analytical methods. Method 1500, hydrocarbons. National Institute of
Occupational Safety and Health, Washington, DC.
NIOSH (1994b) Manual of analytical methods. Method 2010, aliphatic amines. National Institute
of Occupational Safety and Health, Washington, DC.
O'Neil, P.V. (1991) Advanced Engineering Mathematics, 3rd ed. Belmont, CA: Wadsworth pp.
236-240.
OEHHA (1995) Technical support document for the determination of acute toxicity exposure
levels for airborne toxicants (draft). Office of Environmental Health Hazard Evaluation,
California Environmental Protection Agency. Sacramento, CA.
Pennybaker, Mindy, "The Right Paint (nontoxic paints)"Mother Earth News, June 1999.
149
-------
Pitten, F.A., Bremer, J., and Kramer, A. (2000) "Air pollution by volatile organic compounds
(VOC) and health complaints" Dtsch. Med. Wochenschr. 125(18):545-50.
Porter, N.A., Weber, B.A., Weenen, H., and Khan, J.A. (1980) Autoxidation of polyunsaturated
lipids—factors controlling the stereochemistry of product hydroperoxides. Journal of the
American Chemical Society 102:5597-5601.
Rosen, G., and Andersson, I.M. (1990) Solvent emission from newly painted surfaces. Journal of
the Oil and Color Chemists Association 5:202-211.
Saarela, K., and Sandell, E. (1991) "Comparative Emission Studies of Flooring Materials with
Reference to Nordic Guidelines" in the Proceedings of IAQ'91 Healthy Buildings, Washington,
DC, September 4-8, 1991. pp. 262-265.
Serber, G. A., and Wild, C. J. (1989) Non-linear regression. Wiley Series in Probability and
Mathematical Series. Auckland, New Zealand: J. Wiley & Sons. pp.409-412.
Sheldon, L.S., and Naugle, D.F. (1994) Determination of test methods for interior architectural
coatings. Research Triangle Institute, REI/5522/042-02 FR. Research Triangle Park, NC.
Sissell, K. (1999) "California Clamps Down on Paints" Chemical WeekMay 26, 1999, pg. 47.
Sparks, L.E., Tichenor, B.A., Chang, J., and Guo, Z. (1996) Gas-phase mass transfer model for
predicting volatile organic compound (VOC) emission rates from indoor pollutant sources. Indoor
Air 6:31-40.
Stromberg, S.E., and Wind, G.J. (1968) Investigation of the evaporation patterns of mixed
solvents in water-thinnable resins. Journal of Paint Technology 40(525):459-467.
Sullivan, D.A. (1975) Water and solvent evaporation from latex and latex paint films. Journal of
Paint Technology 47(610):60-67.
Tichenor, B.A., and Sparks, L.E. (1996) "Managing Exposure to Indoor Air Pollutants in
Residential and Office Environments." Indoor Air 6: 259-270.
Tichenor, B. (1995) Evaluation of emissions from latex paint. Presented at: Low- and No-VOC
Coating Technologies, 2nd Biennial International Conference. U.S. Environmental Protection
Agency. Durham, NC.
Tichenor, B.A., Guo, Z., and Dorsey, J.A. (1991) Emission rates of mercury from latex paints. In:
IAQ 91 - Healthy Buildings, American Society of Heating, Refrigerating and Air-Conditioning
Engineers, Inc., Atlanta, GA, pp. 276-279.
Tichenor, B.A., and Guo, Z. (1991) The effect of ventilation on emission rates of wood finishing
materials. Environ. Intl. 17: 317-323.
150
-------
Tichenor, B., Sparks, L., White, J., and Jackson, M. (1990) Evaluating Sources of Indoor Air
Pollution, Journal of the Air and Waste Management Association 40(4):487-492.
Tichenor, B. A. (1989) Indoor air sources: Using small environmental test chambers to
characterize organic emissions from indoor materials and products. EPA-600/8-89-074 (NTIS
PB90-110131), U.S. Environmental Protection Agency, Air and Energy Engineering Research
Laboratory, Research Triangle Park, NC.
Tichenor, B.A. (1987) "Organic emission measurements via small chamber testing" In:
Proceedings of the 4th International Conference on Indoor Air Quality and Climate, Indoor Air
1987. Berlin, West Germany, Vol. 1. pp. 8-15.
Turner, G.P.A. (1988) Introduction to Paint Chemistry and Principles of Paint Technology, 3rd.
ed. New York, NY: Chapman and Hall. Chapters 10 and 12.
U.S. Department of Commerce (2000) MA325F(98)-1 "Paint and Allied Products - MA325F"
US Census Bureau Feb-00, U.S. Census Bureau, Suitland, MD,
http ://www. census.gov/cir/www/ma28f.html.
U.S. Department of Commerce (2001) MA325F(00)-1 "Paint and Allied Products - MA325F"
US Census Bureau Jul-01, U.S. Census Bureau, Suitland, MD,
http://www.census.gov/cir/www/ma28f.html.
U.S. EPA (1997) RM 1 Risk assessment of wall paints—indoor screening cluster. Office of
Pollution Prevention and Toxics, U.S. Environmental Protection Agency. Washington, DC., May.
U.S. EPA, (1993) Office of Pollution Prevention and Toxics, Regulatory, Chemical Engineering
Branch, Washington, DC. July 23, 1993 "Chemical Use Clusters Scoring Methodology."
U.S. EPA (1990) Reducing risk: setting priorities and strategies for environmental protection.
SAB-EC-90-021. U.S. Environmental Agency, Science Advisory Board, Washington, DC.
U.S. EPA (1987) Unfinished business: acomparative assessment of environmental problems.
U.S. Environmental Protection Agency, Office of Policy Analysis. Washington, DC.
Weismantel, G.E. (1981) Paint handbook. New York, NY: McGraw-Hill, Inc. Chapter 3.
Wieslander, G., Norback, D., Walinder, R., Erwall, C., and Vernge, P. (1999) "Inflamation
markers in nasal lavage, and nasal symptoms in relation to relocation to a newly painted building:
a longitudinal study."Int. Arch. Occup. Environ. Health 72(8):507-15.
Wieslander, G., Norback, D., Bjornsson, E., Janson, C., and Boman, G. (1997) "Asthma and the
indoor environment: significance of emissions of formaldehyde and volatile organic compounds
from newly painted indoor surfaces."Int. Arch. Occup. Environ. Health 69(2): 115-24.
151
-------
Winnberry, W.T.Jr., Forehand, L., Murphy, N.T., Ceroli, A., Phinney, B., and Evans, A. (1990)
Compendium of methods for the determination of air pollutants in indoor air. EPA-600/4-90/010
(NTIS PB90-200288), U.S. Environmental Protection Agency, Atmospheric Research and
Exposure Assessment Laboratory, Research Triangle Park, NC.
Winnberry, W.T., Murphy, N.T., and Riggan, R.M. (1988) Compendium of methods for the
determination of toxic organic compounds in ambient air. EPA/600-4-89/017 (NTIS PB90-
127374). U.S. Environmental Protection Agency, Atmospheric Research and Exposure
Assessment Laboratory, Research Triangle Park, NC.
Wolkoff, P., Clausen, P. A., Nielsen, P. A., and Gunnarsen, L. (1993) Documentation of field and
laboratory emission cell "FLEC"- identification of emissions processes from carpet, linoleum,
paint, and sealant by modeling. Indoor Air 3:291-297.
Wolkoff, P. (1995) Volatile organic compounds - sources, measurements, emissions, and the
impact on indoor air quality. Indoor Air, Supplement No. 3/95.
World Health Organization (1989) Indoor air quality: organic pollutants. Copenhagen, Denmark:
World Health Organization, Regional Office for Europe (EURO Reports and Studies, N. 111).
World Health Organization (1987) Air quality guidelines for Europe. World Health Organization
Regional Publications, European Series No. 23. Copenhagen, Denmark: World Health
Organization, Regional Office for Europe.
152
-------
/\ppendix /\
r\. Iroposed Otandard L ractice
for 1 e sting and Oamplingof Volatile vJrganic L^ompounds ^Including L^arbonyl V^-ompoundsJ
JlLmitted from r aint Using Omall JlLnvironmental V^-hambers
l.Se
ope
1.1 This practice provides procedures for preparing test samples of alkyd primer, alkyd paint,
latex primer, or latex paint applied to building materials such as gypsum wallboard, wood, or
engineered wood products and procedures for sampling volatile organic compounds (VOCs)
emitted from those test samples. Emissions are sampled from small environmental chambers
operated under controlled conditions.
1.2 This practice describes procedures for preparation of test specimens by application of
primer or paint to common building materials. Use of the procedures described in this practice
for tests with other application methods or substrates may affect the results and not meet the
criteria recommended in the practice.
1.3 This practice describes procedures for col lection of VOCs on sorbenttubes and carbonyl
compounds on silica gel treated with 2,4-dinitrophenylhydrazine (DNPH) that require analytical
methods for measurement of individual organic compound and total VOC concentrations. This
practice does not describe the detailed procedures of analytical methods, but refers to published
methods for these analyses.
1.4 This practice describes procedures for testing and sampling VOCs emitted from paint
under controlled conditions. The test conditions, when combined with analytical data, can be
used to calculate emission rates. This standard practice does not recommend a method for the
calculations.
1.5 Values stated in the International System of Units (SI) are to be regarded as the standard.
1.6 This standard does not purport to address all of the safety concerns, if any, associated
with its use. It is the responsibility of the user of the standard practice to establish practices for
appropriate safety and health, and to determine the applicability of regulatory limitations prior to
use.
LJ. Ixeference .Documents
2.1 ASTMStandards
D 16 Standard Terminology for Paint, Related Coatings, Materials, and Applications
D 355 Practice for Gas Chromatography Terms and Realtionships
D 1005 Standard Test Method for Measurements of Dry Film Thickness of Organic Coatings
D 1212 Standard Test Method for Measurement of Wet Film Thickness of Organic Coatings
Using Micrometers
A-l
-------
D 1356 Standard Terminology Relating to Sampling and Analysis of Atmospheres
D 1914 Practice for Conversion Units and Factors Relating to Sampling and Analysis of
Atmospheres
D 3686 Practice for Sampling Atmospheres to Collect Organic Compound Vapors (Activated
Charcoal Tube Adsorption Method)
D 3687 Practice for Analysis of Organic Compound Vapors Collected by the Activated
Charcoal Tube Adsorption Method
D 5116 Guide for Small Scale Environmental Chamber Determinations of Organic Emissions
from Indoor Materials/Products
D 5197 Test Method for Determination of Formaldehyde and Other Carbonyl Compounds in
Air (Active Sampler Methodology)
D 6196 Practice for Selection of Sorbents and Pumped Sampling/Thermal Desorption
Analyses Procedures for Volatile Organic Compounds in Air
D 6345 Standard Guide for Selection of Methods for Active, Integrative Sampling of Volatile
Organic Compounds in Air
E 355 Standard Practice for Gas Chromatography Terms and Relationships
E 741 Test Method for Determining Change in a Single Zone by Means of Tracer Gas
Dilution
2.2 Other Referenced Documents
Compendium of Methods for the Determination of Toxic Organic Compounds in Ambient
Air. Compendium Methods TO-15 and TO-17, EPA/625/R-96-010b (NTIS No. PB99-172355),
(website http://www.epa.gov/ttn/amtic/airtox.html). U.S. EPA, Center for Environmental
Research Information, Cincinnati, OH, January 1999.
O. 1 er minol
ogy
3.1 Definitions - For definitions and terms used in this standard practice, refer to Standard
Terminology D 1356, Standard TerminologyD 16, and Standard Practice E 355. For definitions
and terms related to test methods using small-scale environmental chambers, refer to Standard
Guide D 5116.
3.2 Definitions of Terms Specific to this Standard Practice:
3.2.1 air changes per hour (ACH) - the volume of clean air brought into the chamber in 1
hour divided by the chamber volume measured in identical volume units, normally expressed in
air changes per hour (h"1).
3.2.2 alkydpaint - also referred to as oil paint, it is a paint that contains drying oil or oil
varnish as the basic vehicle ingredient.
3.2.3 chamber loading (m2/m3) - the exposed surface area of the test specimen coated with
paint divided by the environmental test chamber volume.
A-2
-------
3.2.4 clean air - air that does not contain any individual VOC at a concentration in excess of 2
l-ig/m3 and does not contain greater than 10 |_ig/m3 for the sum of the VOCs measurable in the
sample. The air should be conditioned to remove particulates and ozone.
3.2.5 environmental enclosure - a temperature controlled enclosure of sufficient size to
contain the environmental test chamber(s) and allow adequate access to it to conduct the testing.
3.2.6 environmental test chamber - a chamber constructed of inert materials into which a
material can be placed and tested to determine the VOC emission rate under controlled
environmental conditions.
3.2.7 latex paint - a paint containing a stable aqueous dispersion of synthetic resin, produced
by emulsion polymerization, as the principal constituent of the binder. Modifying resins may be
present.
3.2.8 primer - the first of two or more coats of a paint.
3.2.9 spreading rate - the area covered by a unit volume of coating material, frequently
expressed as square feet per gallon. It may also be referred to as coverage or coverage rate on
paint container labels.
3.2.10 test specimen - a specimen of the paint applied to a substrate such as gypsum
wallboard, wood, or engineered wood products.
3.2.11 volatile organic compound (VOC) - an organic compound with saturation vapor
pressure greater than 10~2 kPa at 25°C
4. Oummary of Otandard Iractice
4.1 This practice describes procedures for testing and sampling emissions of VOCs, including
formaldehyde and other carbonyl compounds, from paint applied to building materials such as
gypsum wallboard, wood, and engineered wood products. Emissions tests are conducted using
small environmental chambers operated in a dynamic mode with continuous flow of humidified
VOC-free air through the chambers. The environmental chambers are operated at designated
conditions of airflow rate, temperature, and relative humidity. The VOCs in the emissions are
sampled by adsorption on an appropriate single, or multiple sorbent media that can be analyzed
by thermal desorption and combined gas chromatography/mass spectrometry (GC/MS) or
GC/flame ionization detection (GC/FID). Formaldehyde and other carbonyl compounds are
collected on silica gel coated with DNPH reagent that can be analyzed by high performance liquid
chromatography (HPLC).
4.2 This practice describes the procedures for handling and storage of paint, setup of the small
environmental test chambers, preparation of test specimens, chamber performance tests, sampling
and reporting.
J. Oignificance and Use
A-3
-------
5.1 Latex and alkyd paints are used as coatings for walls, wooden trim, and furnishings in
occupied buildings. Paint may be applied to large surface areas and may be applied repeatedly
during the lifetime of a building. VOCs are emitted from paint after application to surfaces.
5.2 There is a need for data on emissions from paint. The data can be used to compare
emissions from different products. The data may be used to assist manufacturers in reducing or
eliminating VOC emissions from their products. The data may be used to predict concentrations
of VOCs in a room or building when used with appropriate indoor air quality models.
5.3 Standard test practices and procedures are needed for the comparison of emissions data
from different laboratories.
D. Apparatus
6.1 This standard practice requires the use of an environmental chamber testing facility and
air sample collection systems.
6.2 Environmental Chamber Testing Facility, consisting of an environmental test chamber, a
controlled-temperature environmental enclosure, a system for supplying clean and conditioned air
to the chamber, and fittings and manifolds on the chamber outlet for collection of air samples.
All materials and components in contact with the test specimen or air prior to sample collection
should be chemically inert and accessible for cleaning. Suitable materials include stainless steel
and glass. All gaskets and flexible components should be made from chemically inert materials.
General guidance for design, construction, configuration, and validation of an environmental test
chamber facility is provided in ASTM Guide D 5116-97.
6.2.1 Environmental Test Chamber., constructed of inert materials of sufficient size to hold the
test specimen. Small environmental test chambers may range in size from a few liters to 5 m3.
Procedures recommended in this standard practice have been evaluated using environmental
chambers with a volume of 0.053 m3. This volume is used in the standard practice for discussion
and illustrative purposes. Chambers of different size and shape may be used if the standard
environmental test chamber conditions can be maintained and chamber performance can be
demonstrated. The chamber should be equipped with an opening large enough for loading the
test specimen and for cleaning the chamber. The chamber will be equipped with a port to supply
air to the chamber, an air outlet from the chamber, and ports for temperature and relative
humidity probes. The chamber may be equipped with a fan to promote mixing in the chamber
and to achieve the desired air velocity across the surface of the test specimen. The performance of
the chamber should be evaluated prior to use to determine air-tightness, surface adsorption
effects, air mixing, and air velocity at the surface of the substrate. Some of the chamber
performance (i.e., air mixing and air velocity) should be evaluated with an uncoated test substrate
in place. The chamber performance should be tested and demonstrated following the guidelines
presented in ASTM Guide D 5116-97.
6.2.2 Environmental Enclosure, of sufficient size to accommodate the environmental test
chamber and capable of maintaining the desired temperature within + 0.5 °C.
A-4
-------
6.2.3 Clean Air Supply System, capable of supplying a controlled flow of clean, humidified air
into the test chamber, as described in ASTM Guide D 5116-97. The system should incorporate
hardware for removing particles, ozone, and VOCs from the air supplied to the chamber(s).
Concentrations of VOCs and aldehydes measured at the chamber inlet should not exceed 2 i-ig/m3
for any single compound or 10 |_ig/m3 for the sum of all measurable VOCs in the sample. The
relative humidity (RH) of the air supplied to the chamber(s) should be controlled to the desired
set point + 5% RH. The flow rate of the air supplied to the chamber should be controlled to the
desired set point + 5% of the set-point airflow rate. Ideally, the chamber system will be designed
such that a positive pressurization of the chamber of approximately 10 Pa relative to the
environmental enclosure will be maintained and monitored at all times during the test.
6.2.4 Environmental Measurement System, consisting of hardware and software to measure
and record the temperature, RH, and airflow rates during operation of the test system. A system
for continuous recording of the data is recommended.
6.3 Air Sampling Systems, consisting of sorbent tubes and DNPH-silica gel cartridges, a
sampling manifold, vacuum pumps, and airflow controllers/meters. Airflow controllers should
control the airflow rate through the sampling system to within + 5% of the specified value. All
system components between the chamber and the sampling media should be constructed of
chemically inert materials.
6.3.1 A glass or stainless steel manifold should be connected to the outlet of the chamber for
collection of air samples. The manifold should be designed for collection of multiple samples
simultaneously. The exhaust from the manifold should be vented into a laboratory fume hood or
other appropriate exhaust device to prevent contamination of the air in the laboratory or
environmental enclosure.
6.3.2 Vacuum pumps should be used to draw air through the sorbent tubes. The required
airflow rate is a function of the type of sampler used, the size of the chamber, and the air change
rate. The total airflow rate through the samplers should not exceed 50% of the flow rate from the
chamber outlet. For collection of VOCs on sorbent tubes, the pump should be capable of
maintaining a constant flow in the range of 10 to 200 mL/min. For collection of air samples on
DNPH-silica gel cartridges, the pump should be capable of maintaining a constant flow in the
range of 100 to 500 mL/min.
6.3.3 For collection of VOCs during the emissions test, tubes containing single or multiple
sorbents may be used. The sorbents may be porous polymers or graphitized carbon blacks.
Select an appropriate single or multi-layered sorbent tube following the procedures in ASTM
Practice D 6196 and D6345, and EPA Method TO-17. Recommendations on the use of sorbent
tubes from manufacturers or suppliers should be followed in selecting the sampling airflow rate
and sampling period to avoid breakthrough of VOCs through the sorbent tube. The required air
sampling volume at each collection time point should be determined through consideration of the
safe sampling volume (SSV, see Practice D 6196) of the VOC with the lowest retention volume,
concentrations to be measured, and detection limits of the analytical method.
A-5
-------
6.3.4 For collection of VOCs during the first lOto 20 hours folio wing application of alkyd
primer or paint, charcoal sorbents (ASTM D3687) may be used due to the high concentrations of
VOCs in the chamber air.
6.3.5 For collection of formaldehyde and other carbonyl compounds, DNPH-silica gel
cartridges should be used following the ASTM Method D 5197.
6.3.6 An airflow meter/controller should be used to control and measure the airflow rate
during sample collection. The controller may consist of a precision flow control valve, a critical
orifice, or a mass flow controller. The measurement device may consist of soap film bubble
meter, calibrated high precision rotameter, or mass flow meter. A mass flow meter/controller is
recommended for use during sample collection. All flow measurements should be referenced to
standard temperature and pressure.
/. Irocedures for L aint Oelection, .Handling, and Otor
age
7.1 Procedures for selection of the paint to be tested are a function of the objectives of the
tests. Paint may be procured from clients, manufacturers, distributors, or retailers. Record
pertinent information upon receipt of the paint including date of acquisition, source of the paint,
manufacturer, container size, lot number, and other relevant information on the label. Obtain and
review the Material Safety Data Sheet (MSDS) for the paint. At least two containers of the same
lot number of paint should be procured (one for testing and one to archive).
7.2 Upon receipt of the paint, it should be split into storage vials for handling and testing. The
paint should be mixed in the original container on a paint shaker before the split.
7.3 Split the paint into aliquots. Special care should be taken to minimize the loss of volatile
compounds during the process. Paint containers should not be left open except when required for
transfer to storage vials. The size of the aliquot and storage vial is a function of the amount of
paint required for the test. Vials of 40 to 60 mL volumes hold sufficient paint for GC/MS
analyses of the liquid product or preparation of test specimens of 256 cm2 area for chamber tests.
Store paint in clean amber glass vials that can be sealed with caps that have Telfon® liners.
Clean vials with alkaline detergent, rinse thoroughly with deionized water, then dry before use.
Individual vials of paint are used for testing to minimize losses of volatile compounds during
handling and preparation of test specimens. Repeated opening of a large container of paint will
result in losses of VOCs.
7.4 Pour the mixed paint into the vials, filling the vial to near the top to minimize the volume
of headspace and loss of VOCs when the vial is opened. Two or more clean stainless steel balls
may be placed in the vial to aid in mixing prior to use of the sample. Prepare a sufficient number
of sample vials for all analyses and tests planned for the paint. Label the vials individually with a
sample code and the date of preparation.
7.5 Store the vials of the paint samples and the original containers of the paint in the dark at
room temperature.
A-6
-------
7.6 Use the paint sample within the manufacturer's specified shelf-life time. VOC
concentrations should be measured by GC/MS or GC/FID after extended periods of storage to
verify that they have not changed during storage. Compare the concentrations of the VOCs to
results from the original analyses performed immediately after the paint was obtained.
O. Irocedures for r reparation of lest Opecimens
8.1 Procure the substrate material that is appropriate for testing the emissions from the paint.
The substrate should not emit any of the compounds that are to be quantified in the emissions
from the paint or compounds that may interfere with quantification of the emissions from the
paint. Cut the substrate material to an appropriate size (e.g., 16 by 16 cm for a loading of 0.5
m2/m3 in a 0 .053 m3 chamber). Use an appropriate cutting device to obtain smooth edges. If
necessary, sand wood substrate to provide a smooth surface.
8.2 To minimize emissions from the cut edges, seal the edges. Techniques that can be used to
seal the edges include: (1) coating the edges with technical grade liquid sodium silicate, (2)
wrapping with aluminum foil, or (3) mounting the substrate in a stainless steel frame that covers
the exposed edges.
8.3 The substrate material should be properly stored to remain clean, dry, and structurally
sound.
C/. Irocedure for L aint Application
9.1 Prior to application of paint for emissions testing, one must perform practice applications
using the same type of applicator (e.g., a roller or brush) and substrate in order to develop a
technique that will produce uniform and repeatable applications. The VOC emission
characteristics will be influenced by the thickness of the paint applied, and care should be taken to
ensure uniform paint application. If desired, wet and dry film thickness of paint applied can be
determined by ASTM methods referenced in Section 2 to verify the uniformity of the application.
The application should also be performed as quickly as possible to minimize VOC losses during
preparation of the test specimen.
9.2 Paint Application Procedure
9.2.1 Determine the target mass of paint to be applied from product data sheets or the
spreading rate (ft2/gal) listed on the container label. Calculate the mass of paint needed using the
spreading rate (ff/gal) and the density of the paint. The density is generally listed on product data
sheets or the MSDS. The mass to be applied can be calculated as follows:
Where:
Mp = target mass (g) of paint to be applied,
SR = spreading rate (cm2/mL), generally listed on container or data sheet as fiVgal,
A-7
-------
Dp= density (g/mL) of the paint, and
A = area of the substrate (cm2).
As an example, for paint with a spreading rate of 400 ff/gal and a density of 1.4 g/mL, the
amount of paint to be applied to a 16 by 16 cm test substrate would be 3.65 g:
Mp =
2 gal O.Q92903*s2
1.4 g
16cm *16cm* =- (A-2)
9.2.2 Select a vial of paint for the test. Prepare the paint by vigorously shaking the vial for 5
minutes. Mixing must be done in the sealed vial to minimize loss of VOCs.
9.2.3 Wear cotton gloves during specimen preparation to prevent contamination by oils from
the skin.
9.2.4 Weigh the substrate.
9.2.5 Pour the paint from the vial into a small, clean tray of appropriate size for the roller or
brush being used for the application. Disposable trays can be made using clean aluminum foil.
For a 7.6-cm wide paint roller, a tray that is 10 cm (4 in.) wide, 15 cm (6 in.) long, and 2 cm (0.75
in.) deep is appropriate.
9.2.6 Place the paint applicator into the tray and wet the surface uniformly with the paint. If a
roller is used, apply the roller to the substrate starting at the center and roll outward to all edges
until the surface is covered. Then rotate the substrate 90 degrees and roll back and forth across
the entire surface in this direction. Use a similar application procedure if using a brush to obtain
a uniform coating on the substrate surface.
9.2.7 Immediately after application, weigh the substrate with the coating and record the
weight. Calculate the weight of coating applied by the difference in the weight of the substrate
before and after paint application. The final weight of paint applied should be within ± 10% of the
target application amount. If the weight of coating is greater than ± 10% of the target mass, adjust
the application procedure (e.g., by changing the number of strokes and/or the amount of the paint
in the brush) and try again with a new substrate until desirable and repeatable applications can be
achieved.
9.2.8 Each paint application process, including the weight measurement and calculation,
should be finished within 3 minutes.
9.2.9 If a suitable balance is not available to weigh the substrate before and after application
of the paint, an alternative method can be used involving weighing of the paint and applicator.
Prior to preparation of the test specimen, pour the paint in the aluminum tray. Determine the
combined weight of the tray, paint, and the applicator (roller or brush). Then apply the paint to
A-S
-------
the substrate. After the application is complete, weigh the paint tray and the applicator.
Determine the amount of paint applied by calculating the weight difference.
1U. v>hamber JlLmissions lest Irocedure
10.1 Preparation of the Environmental Chamber Testing System - Prior to each test, the
chambers should be cleaned and operating parameters set.
10.1.1 Clean the environmental test chamber and all internal hardware by wiping the interior
surfaces with an alkaline detergent, rinsing thoroughly with deionized water, and drying with
clean laboratory tissue. Solvent such as acetone or methanol and/or chamber heating can be used
when necessary. Use the same procedure to clean all hardware on the chamber outlet and the
sampling manifold.
10.1.2 Place the environmental test chamber in the environmental enclosure and set the
enclosure temperature to the desired level. Close the chamber. Assemble the chamber outlet and
manifold.
10.1.3 Adjust the water vapor concentration in the supply air to achieve the desired test
conditions (e.g., 50% RH at 23 °C). For water-based paints, %RH should be recorded throughout
the test because the drying paint will affect the water vapor concentration. Ideally, in those cases
the water vapor concentration in the outlet air (rather than the supply) should be controlled to
50% RH. Start airflow to the chamber at the rate required for the desired air change rate (typically
0.5 or 1.0 ACH). For a 53-L chamber the airflow rate will be 442 mUmin for an ACH of 0.5 hr1.
Measure the airflow rate at the inlet and the outlet of the chamber. If the inlet and outlet flow
rates differ by more than + 5%, check for leaks.
10.1.4 Operate the empty chamber at desired test conditions for 24 hours prior to use.
10.1.5 Collect background samples from the empty chamber at the end of the 24 hours.
Analyze the samples as soon as possible to ensure that the chamber background meets the criteria
in Section 14.3.2 and is therefore ready for use. If the background VOC concentrations are too
high, further cleaning or other measures may be required.
10.2 Emissions Test Procedure
10.2.1 Place the uncoated test substrate (and holder if used) into the chamber and condition
for 24 hours at the same temperature, RH, and ACH that will be used in the test.
10.2.2 At the end of the 24 hours, collect an air sample from the chamber outlet, using a
sorbent tube to measure VOC background concentrations from the environmental test chamber
loaded with the uncoated substrate. Collect an air sample from the chamber outlet on a DNPH-
silica gel cartridge to measure background concentrations of carbonyl compounds. Analyze the
chamber background samples as soon as possible after collection. Individual VOCs or carbonyl
compounds that are targeted for quantification in the emissions from the paint should not be
present in the samples at concentrations greater than the chamber background criteria in Section
A-9
-------
14.3.2. The emissions from the substrate should not include any compounds that may interfere
with quantification of the emissions from the paint.
10.2.3 After the substrate background levels have been determined to be acceptable, open the
chamber and remove the test substrate. Maintain airflow to the chamber and close it during
preparation of the test specimen to minimize contamination of the chamber with laboratory air.
10.2.4 Prepare the test specimen by applying the paint as described in Section 9. If the
application is acceptable, immediately place the test specimen into the chamber and close the
chamber. Record time that the chamber is closed as "Time Zero" for the start of the test. Total
preparation time from chamber opening, painting, weighing, to sealing of the coated test
specimen within the chamber should not exceed 5 minutes.
10.2.5 Operate the test chamber at desired conditions for a test period during which air
samples are routinely collected from the chamber outlet.
11. Oample (Collection
11.1 Throughout the emission test, air samples will be collected from the outlet of the
chamber. Samples will be collected on either charcoal media or thermal desorption sorbent tubes
for VOC analyses and on DNPH-silica gel cartridges for analyses of carbonyl compounds.
11.2 Set up the vacuum pump(s) and mass flow controllers/meters for collection of sorbent
tube and DNPH-silica gel cartridge samplers.
11.3 Set the flow controllers for the required airflow rates for the sorbent tube and DNPH-
silica gel samplers. The required airflow rates will be a function of the emission rates for
individual compounds and the volume of air sample required to obtain VOC mass on the sampler
that is within the calibration range of the analytical instrument. During early stages of a test, due
to the relatively high VOC concentrations, samples may need to be collected over periods of only
5 to 10 minutes at a low flow rate in order to obtain accurate sample volumes. Higher airflow
rates and longer sample collection periods will be required in later stages of the test as the
concentrations of VOCs decrease in the chamber air. The total sampling flow rate must be less
than 50% of the airflow from the chamber (e.g., less than 221 mL/min for a 53 L chamber
operated at 0.5 ACH). Set airflow rates with a sampling tube or cartridge connected to the
sampling line. Sample flow rates should not exceed manufacturer's recommendations.
11.4 Collect air samples for VOCs and carbonyl compounds at predefined intervals during the
test. The duration of the test and the frequency of sample collection will be determined by the
objectives of the test. A high frequency of sampling will be required if the objective of the test is
to develop or evaluate source emission models. A lower frequency of sampling may be
appropriate for other test objectives. If resources are available, the recommended sampling times
for collection of data for model development during a 2-week test with alkyd paint are at 0.25,
0.5, 1, 1.5, 2, 4, 6, 8, 10, 12, 24, 48, 72, 96, 144, 192, 264, and 336 hours following the start of the
test. Recommended sampling times during a 2-week test with latex paint are at 2, 4, 8, 12, 24, 48,
A-10
-------
72, 96, 144, 192, 264, and 336 hours following the start of the test. After sampling, carefully seal
sample tubes and record sampler flow rates and duration.
11.5 Collect duplicate air samples simultaneously at a subset of time periods. Use separate
flow controllers for each sampler. Adjust total flow to be less than 50% of the chamber outlet
airflow rate.
11.6 Store samples collected on sorbent media refrigerated at less than 4°C in a clean
environment. Analyze samples as soon as possible after collection. Samples should be analyzed
within 30 days as recommended in EPA methods referenced in Section 2. Analyses should be
performed following methods referenced in Section 2 or other published methods that produce
results that meet the project objectives for precision and bias.
11.7 Store samples collected on DNPH-silica gel cartridges refrigerated at less than 4°C.
Analyze the samples within 30 days. Analyses should be performed following methods
referenced in Section 2 or other published methods that produce results that meet the project
objectives for precision and bias.
\.L. v>hemical Analysis
12.1 For VOCs collected on sorbent mediafor analysis by thermal desorption/GC, analyze
samples following guidance in ASTM Practice D6196 and EPA Compendium Methods TO-15
and TO-17.
12.2 For VOCs collected on charcoal, follow procedures described in ASTM D3687 to extract
the analytes from the charcoal media with carbon sulfide and analyze the liquid extract by GC.
12.3 For carbonyl compounds collected on DNPH-silica gel cartridges, follow procedures
described in ASTM Method 5197. The Method involves extraction of the media with acetonitrile
and analysis of the extract byHPLC.
J.O. Ixeporting lest Ixesults
The report should include the following information:
13.1 Test objectives - Provide a clear description of the purpose of the test.
13.2 Equipment and Methods - Provide a description of the small chamber test facility
equipment, including chamber size and materials, environmental control and measurement
systems, sampling and analysis equipment, and description of the test methods and protocols.
13.3 Product identification - Provide the name, specific identifiers from the manufacturer and
a brief description of the product, its application, and history.
A-ll
-------
13.4 Test conditions - Provide data for the temperature, humidity, air change rate, dimensions
of the test specimen, empty chamber and substrate background concentrations, mean air velocity
above the surface of specimen, and mass applied.
13.5 Samples Collected - Provide a record of samples taken including sampling schedule,
flow rate, volume, frequency, and type.
13.6 Quality Assurance/Quality Control Results - Describe the data quality objectives and
quality control activities for the test. Provide results for quality control samples (replicates,
blanks, controls).
14. v^/uality Assurance and v^/uality v>ontrol
14.1 A quality assurance and quality control (QA/QC) plan should be prepared and
implemented to ensure the integrity of the measured and reported data obtained during the tests.
This plan should encompass all facets of the measurement program from sample receipt to final
review and issuance of reports. QA/QC activities applicable to tests to measure emissions from
paint are described in ASTM Guide D 5116.
14.2 Data Quality Objectives and Acceptance Criteria - The QA/QC plan should be based on
established data quality objectives and acceptance criteria Recommended data quality objectives
are listed in Table A-l.
14.3 Quality Assurance Activities and Quality Control Samples
14.3.1 Determine the accuracy of chamber test conditions and operating parameters
(temperature and RH) by calibration of measurement devices with National Institute of Standards
and Technology (NIST)-traceable primary sources. Determine the precision of the test conditions
by continuous recording of the parameters. Failure to meet acceptance criteria requires
immediate corrective action. Check instrument performance before and after each test by
comparison of measurements with a reference device.
14.3.2 The empty chamber background concentrations of VOCs should be measured at the
start of each test without the substrate in the chamber. The empty chamber background
concentration shall meet the following criteria: (1) less than 10 |J,g/m3 or 1/6 of the lowest
concentration to be measured, whichever is lower, for TVOC, and (2) less than 2 |j,g/m3or 1/6 of
the lowest concentration to be measured, whichever is lower, for individual VOCs.
14.3.3 Field blanks, consisting of sorbent tubes and DNPH silica gel cartridges that are not
used for sampling, should be analyzed to verify that the tubes and cartridges have not been
contaminated during handling and storage. The blanks should be stored and handled in the same
manner as the samples collected during the test. The number of blanks analyzed should represent
approximately 10% of the number of samples collected for each type of sampling media; a
minimum of 2 blanks for each type is recommended.
A-12
-------
14.3.4 Precision of the sampling and analysis methods should be determined by collection of
replicate samples. Duplicate samples should be collected concurrently at selected time periods
during the test. The number of duplicates for each type of sampling media should represent
approximately 10% of the number of samples collected. Due to the limited volume of air
available, duplicate samples can be collected for only one type of media during a sampling
period, especially when low VOC concentrations dictate high sample volumes. Each sampling
and analytical method used must have detection limits lower than the acceptable background
levels stipulated in Section 14.3.2.
14.3.5 The accuracy and precision of the small chamber test method can be measured prior to
start of a test program. The precision of the chamber test method, to include the combined
variability of test specimen preparation, chamber operation, sampling, and analysis, can be
measured by performing multiple chamber tests with the same source and substrate. Variation in
test data from a single chamber and among chambers can also be established by use of
standardized sources such as permeation tubes for determining organic concentrations and
calculated emission rates.
Table A-l: Recommended Data Quality Objectives
Parameter Precision Accuracy
Chamber temperature +0.5 °C + 0.5°C
Chamber relative humidity +5.0 % +5.0%
Airflow rate +5.0% +5.0 %
Amount of paint applied + 10 % + 10%
A-13
------- |