December 2024

United States	Office of Chemical Safety and

v/trM Environmental Protection Agency	Pollution Prevention

Indoor Air Exposure Assessment for Formaldehyde

CASRN 50-00-0

A

December 2024


-------
TABLE OF CONTENTS

EXECUTIVE SUMMARY	7

1	INTRODUCTION	9

1.1	Risk Evaluation Scope	9

1.1.1 Indoor Air Exposure Assessment Scope	10

1.2	Revisions between Draft and the Revised Assessment	12

1.3	Conceptual Exposure Model	13

2	APPROACH AND METHODOLOGY	16

2.1	Indoor Air Monitoring Data	17

2.1.1 Systematic Review Prioritization for Formaldehyde Data	17

2.2	CEM Model Development and Parameterization	19

2.2.1	Model Output Time Period	20

2.2.2	Scenario Selection	20

2.2.3	Chemical-Specific Input Parameters for CEM	21

2.3	IECCU Model Development and Parameterization	27

2.3.1	IECCU Model Description	27

2.3.2	Model Output Time Period	28

2.3.3	Article Characterization and COU Mapping	28

2.3.4	Model Inputs for Single Article Models	30

2.3.5	Model Inputs for Aggregate Models	32

3	RESULTS	34

3.1	Monitoring Data	34

3.1.1	Summary of Monitoring Data Results	35

3.1.2	American Healthy Homes Survey II (AHHS II)	36

3.1.3	New Homes	37

3.1.4	Trailer Studies	38

3.1.5	Commercial and Other Buildings	39

3.1.6	Japan National Study	40

3.1.7	Relative Contributions of Formaldehyde Sources in Residential Indoor Air	40

3.1.8	Comparing Indoor to Outdoor Air	41

3.2	Tier 1: CEM Modeling	43

3.2.1	CEM Modeling Results	43

3.2.1.1 Aggregate Exposure	44

3.2.2	CEM Modeling Discussion	44

3.2.2.1	Data Integration	45

3.2.2.2	Automobile Exposures	46

3.3	Tier 2: IECCU Modeling	46

3.3.1	IECCU Modeling Results	46

3.3.2	IECCU Modeling Discussion	49

3.3.2.1	Residential	49

3.3.2.1.1 Data Integration	49

3.3.2.2	Aggregate Exposure	52

4	INDOOR AIR EXPOSURE: WEIGHT OF SCIENTIFIC EVIDENCE	54

REFERENCES	56

Page 2 of 81


-------
APPENDICES	63

Appendix A SYSTEMATIC REVIEW PRIORITIZATION FOR FORMALDEHYDE DATA. 63

A,1 Formaldehyde Data Needs	63

A.2 Boolean Search Terms	64

A.3 Indoor Air Data Extracted from Systematic Review for Consideration into the Indoor Air

Exposure Assessment	65

A.3.1 Indoor Air (ng/m3) - Vapor/Gas Fraction	65

A.3.2 Indoor Air (ppb) - Vapor/Gas Fraction	67

Appendix B AHHS II SUMMARY OF DATA COLLECTION METHODOLOGY	69

Appendix C FORMALDEHYDE RESIDENTIAL INDOOR AIR HALF-LIFE	70

C,	1 General Formaldehyde Dissipation Curve	70

Appendix D CEM SUPPLEMENTAL ANALYSES	72

D.	1 CEM Screening Combustion Assessment	72

I). 1.1 Methods	72

D. 1.2 Comparison of Formaldehyde Indoor Air Estimates from CEM Modeling of TSCA

COUs Relative to Homes with and without Reported Combustion Sources, According to

AHHS II	72

D,2 CEM Screening Composite Wood Article Assessment	76

D.2.1 Methods	76

D.2.2 Results	77

D.2.3 Conclusion	78

Appendix E SUMMARY OF WEIGHT OF SCIENTIFIC EVIDENCE	80

LIST OF TABLES

Table 2-1. Formaldehyde Indoor Air Conditions of Use and Relevant Exposure Scenarios	21

Table 2-2. Formaldehyde Emission Rates by TSCA Condition of Use (COU)	24

Table 2-3. Formaldehyde Emissions for Materials with the potential for large applications	27

Table 2-4. IECCU Modeling Inputs for Single Articles with Large Surface Area Representing TSCA

COUs	30

Table 2-5. IECCU Modeling Inputs for Aggregate Scenarios with Multiple Articles Representing a

Single TSCA COU	33

Table 3-1. Indoor Air Monitoring Concentrations for Formaldehyde	34

Table 3-2. Formaldehyde Monitored in Commercial Buildings in the United States	35

Table 3-3. Range and Weighted Quantiles of AHHS II Residential Indoor Air Formaldehyde

Concentrations (|ig/m3)	36

Table 3-4. Estimated Chronic Average Daily Formaldehyde Indoor Air Concentrations (According to

CEM)	43

Table 3-5. 15-Minute Peak, Mean 3-Month (Intermediate) and Mean 1-Year (Chronic) Formaldehyde
Concentrations (|ig/m3) in Indoor Air for Single Representative Article and Aggregate
Model Scenarios	47

LIST OF FIGURES

Figure 1-1. Risk Evaluation Document Summary Map	10

Page 3 of 81


-------
Figure 1-2. Formaldehyde Conceptual Model for Consumer Activities and Uses in Indoor Air:

Consumer Exposures and Hazards	14

Figure 2-1. Schematic of the Approach Used to Identify and Extract TSCA COU-Specific Data Pertinent

to the Formaldehyde Exposure Assessment	18

Figure 3-1. Histogram of Formaldehyde Indoor Air Sampling Results from AHHS II with Statistical

Weights Applied	37

Figure 3-2. Monitoring Formaldehyde Concentrations in Indoor Compared to Outdoor Settings	42

Figure 3-3. CEM Estimated 1-Year Average Daily Formaldehyde Indoor Air Concentrations	44

Figure 3-4. 15-Minute Peak Concentrations (|ig/m3) of Formaldehyde in Indoor Air for TSCA COU

Representative Article and Aggregate Models	48

Figure 3-5. Formaldehyde Concentrations in Indoor Air (|ig/m3) for TSCA COU Representative Article

and Aggregate Scenarios over the Course of ~1 Year	49

Figure 3-6. Figure from He et al. (2019) Displaying Formaldehyde Concentration Decay Curves from

Finished Wood Article Specimens Over Time, Using Test Chamber	52

LIST OF APPENDIX TABLES

TableApx A-l. Summary of Peer-Reviewed Literature that Measured FDH (ng/m3) Levels in the

Vapor/Gas Fraction of Indoor Air	66

Table Apx A-2. Summary of Peer-Reviewed Literature that Measured FDH (ppb) Levels in the

Vapor/Gas Fraction of Indoor Air	68

Table Apx B-l. Summary of Environmental Sampling and Analytical Method	69

Table Apx D-l. Estimated Concentrations in mg/m3 from Emission Standards in ppm	77

Table Apx D-2. Estimating Emission Rates from Product Specific Concentrations	77

Table Apx D-3. CEM Estimated Average Daily Concentration Over 1 Year	78

Table Apx E-l. Weight of Scientific Evidence Conclusions for the Indoor Air Exposure Assessments 81

LIST OF APPENDIX FIGURES

FigureApx A-l. Concentrations of FDH (ng/m3) in the Vapor/Gas Fraction of Indoor Air from 1998 to

2017	65

Figure Apx A-2. Concentrations of FDH (ppb) in the Vapor/Gas Fraction of Indoor Air in General

Population (Background) Locations from 2012 to 2013	67

Figure Apx C-l. General Formaldehyde Dissipation in a Residence	71

FigureApx D-l. Comparing the Relative Concentrations of Homes with and Without Sources of

Combustion for Formaldehyde in AHHS II	74

Figure Apx D-2. Comparison of AHHS II Monitoring to Modeling Estimates of Indoor Air

Concentrations	75

Page 4 of 81


-------
Key Points: Indoor Air Exposure Assessment for Formaldehyde

The indoor air environment includes commercial settings, new homes, mobile homes, and
automobiles—all of which are a major source of formaldehyde exposure to humans. Exposure to
formaldehyde in these settings is the result of ubiquitous use of formaldehyde in the manufacturing
of consumer products and articles (e.g., rubber mats, plastic chairs, hardwood floors). Formaldehyde
might also be present in the indoor environment due to the use of fireplaces, gas stoves, additional
combustion sources, human breath, and other sources. The number of potential sources of
formaldehyde to consider makes an indoor air exposure assessment of formaldehyde highly complex.

According to the U.S. Department of Housing and Urban Development (HUD) (OuanTech. 2021)
indoor air concentrations of formaldehyde range from 0.27 to 124 |ig/m3 (3.5-hour time-weighted
average [TWA]). Monitoring data are expected to represent concentrations of formaldehyde from all
sources of formaldehyde within the sampled space. In addition, the majority of the available
monitoring data are expected to represent typical formaldehyde concentrations in indoor air;
however, some datasets may capture peak exposures.

Indoor air monitoring data were used in this assessment to characterize known formaldehyde indoor
air exposures. However, these monitoring data do not differentiate between Toxic Substances
Control Act (TSCA) and other sources of formaldehyde, such as gas stoves. Therefore, EPA utilized
modeling tools to assess TSCA conditions of use (COU) exposures for formaldehyde in indoor air.
EPA used the Consumer Exposure Model (CEM) as a tier 1 modeling tool, which CEM was not used
to estimate peak exposures for the draft risk evaluation. Based on the available data, the Agency
assumed formaldehyde indoor air exposures would be primarily from long-term emissions. It is
unclear how often people purchase or replace articles, remodel their homes or article-specific lag-
times between the manufacture and installation of an article into an indoor environment. For a
comprehensive consideration of all potentially relevant durations of exposure to formaldehyde and to
address uncertainties associated with CEM's potential overestimation of long-term exposures, EPA
revised its assessment using the Indoor Environmental Concentrations in Buildings with Conditioned
and Unconditioned Zones (IECCU) as a tier 2 modeling tool to characterize 15-minute peak, 3-
month average and 1-year average formaldehyde residential indoor air concentrations. The IECCU
modeling considered article-specific lag-times between the manufacture and installation of an article
into an indoor environment by foregoing the use of initial weight fractions reported in article safety
data sheets (SDSs) and relying on the emission factors reported from chamber studies of finished
articles.

EPA identified four COUs under TSCA as significant formaldehyde contributors to commercial,
automotive, and residential indoor air environments. Formaldehyde exposures from individual
articles for these TSCA COUs were estimated using two different standard exposure models and
product-specific emissions data. Using CEM, estimated modeled average daily concentration over 1
year ranged from 4.01 [j,g/m3to 423.47 [j,g/m3. Using IECCU, the estimated formaldehyde 15-minute
peak concentrations ranged from 0.007 to 142 |ig/m3 and 1-year annual averages ranged from 0.0003
to 6.1 |ig/m3 in the applicable room(s) from IIECCU modeling. The highest exposures are expected
to result from laminate flooring and pressed wood furniture, using either modeling tool.

EPA also assessed aggregate exposure scenarios (new construction, new decor, and laminate with
background) and evaluated available formaldehyde indoor air monitoring data from within
residential homes, automotives, offices, and other buildings. These aggregate exposures are expected
to vary according to conditions of use, product specifications, activity patterns, purchase habits, etc.

Page 5 of 81


-------
Estimated formaldehyde 15-minute peak concentrations ranged from 3 to 160 |ig/m3 and 1-year
annual averages ranged from 0.1 to 6.9 |ig/m3 for aggregate exposures in the applicable room(s) of
use based on IECCU modeling.

EPA also considered other sources of formaldehyde including fireplaces, air cleaning devices,
cooking, candles, ethanol fireplaces and incense. Examination of indoor air monitoring data suggest
that in some spaces, combustion may substantially contribute to indoor air formaldehyde
concentrations, but typically, combustion does not substantially result in higher formaldehyde
concentrations in indoor air. Concentrations of formaldehyde in homes with combustion sources,
reported by residents at the time of indoor monitoring surveys, are not different than homes without
combustion sources or government buildings where combustion is not expected.

There is uncertainty in the precise estimates from both CEM and IECCU due to model limitations—
especially for long-term concentrations. However, when used together, EPA has high confidence in
the potential range of formaldehyde concentrations from TSCA COUs. This conclusion is supported
by available monitoring data. While the exposure durations and scenarios do not perfectly align,
consideration of all of the available data suggests model estimated and measured formaldehyde are in
reasonable agreement. Specifically, model estimated formaldehyde concentrations were within the
same order of magnitude as measured concentration data (e.g., HUD's American Healthy Homes
Survey II). In addition, the estimated aggregate exposures fall within the range of available
monitoring data. This suggests that TSCA COUs are contributors to real-world concentrations of
formaldehyde in indoor air. Formaldehyde concentrations are expected to be highest for newly
constructed residences with formaldehyde-based materials (including laminate flooring) and when
new formaldehyde-based articles are added to a residence (including furniture covers).

EPA has high confidence in the conclusions of this indoor air assessment, including that exposures to
formaldehyde occur as the result of TSCA COUs on a short- and long-term basis in indoor air
environments. However, the precise concentrations on a long-term basis are uncertain and are
expected to be highly variable.

Page 6 of 81


-------
EXECUTIVE SUMMARY

Formaldehyde is a chemical ingredient in many articles such as furniture, flooring, cabinets, to name just
a few, which can be significant contributors to its peak and long-term indoor air concentrations.

EPA used a combination of modeling and monitoring data to characterize formaldehyde indoor air
exposures while considering the relative contributions of TSCA COUs to the indoor air environment.
The Agency assessed indoor air exposures for four COUs expected to be significant sources of
formaldehyde in indoor air. Monitoring data were further used to characterize indoor air exposures and
provide context for estimated concentrations.

In the Draft Indoor Air Exposure Assessment for Formaldehyde, EPA used the CEM as a tier 1
modeling tool. However, CEM was not used to estimate peak exposures for the draft assessment and due
to uncertainties with CEM's potential overestimation of long-term exposures, along with an inability to
consider first-order exponential decay for articles in the long-term based on the E5 emission condition.
This revised assessment also utilizes IECCU as a tier 2 modeling tool to characterize 15-minute peak, 3-
month average, and 1-year average formaldehyde residential indoor air concentrations. Specific
consumer article categories were modeled by incorporating relevant low, median, and high emission
factors and corresponding surface areas expected in a room according to reasonably available
information. In addition, some COUs were aggregated to estimate indoor air concentrations from a home
renovation and a decor update.

According to CEM, the highest modeled average daily concentration over 1 year is for building wood
articles (423.47 (j,g/m3). The lowest concentrations were due to furniture seat covers (4.01 (J,g/m3) and
clothing (5.19 (j,g/m3). In contrast, IECCU-estimated 1-year concentrations ranging from 0.00003 to 6.9
[j,g/m3 across all COUs and scenarios. The COU contributing to the highest residential indoor air
formaldehyde concentrations was laminate flooring. Modeled 15-minute peak concentrations ranged
from 0.00009 to 142 [j,g/m3. In addition to individual COUs, IECCU was used to aggregate COUs and
estimate potential indoor air concentration. High-end aggregate scenarios for indoor air {i.e., new
construction and new decor scenarios) produced the highest peak modeled indoor air concentrations
(160 |ig/m3).

It should be noted that the CEM modeling for the formaldehyde does not consider the lag time between
when articles may be manufactured and installed indoors since it relies on the initial formaldehyde
concentration from a finished article, as reported by relevant SDSs, as the initial indoor air concentration
to which an individual is exposed. This results in a conservative emission rate and exposure estimate via
modeling because it is unlikely that a newly manufactured product will be instantly installed in a home
and individuals will immediately be exposed to those higher concentrations early in their exposure
period. It is more likely that the actual emissions from any given product will be the rate following some
period of storage time where initial off-gassing at those high-rates would have occurred prior to being
installed in a residence. The IECCU modeling considered this lead time by foregoing the use of initial
weight fractions reported in article SDSs and relying on the emission factors reported from chamber
studies of finished articles.

The 1-year results for IECCU are significantly lower than those from CEM and are likely an
underestimate. Available data suggest a biphasic emission profile (rapid emission of formaldehyde when
the product is new followed by a much slower emission of formaldehyde) for laminated wood products
that is not captured in the modeling results. This biphasic emission profile may also occur for other urea-
formaldehyde based products; however, data are not available to confirm. As such, CEM was used in
conjunction with IECCU to characterize 1-year average indoor air concentrations.

Page 7 of 81


-------
Monitoring data were incorporated into this assessment to characterize the indoor air concentrations
most people are exposed to in their homes. EPA used data acquired through systematic review and
publicly available to characterize indoor air concentrations of formaldehyde. These monitoring data do
not differentiate between TSCA COU and other sources of formaldehyde like cigarette smoke or gas
stoves. Monitoring data also may not capture peak concentrations or fluctuations in indoor air
concentrations of formaldehyde. Indoor air monitoring data can, however, provide insight to long-term
and aggregate exposures to formaldehyde. In addition to data captured from homes, EPA considered
monitoring data from various environments including schools, government buildings and mobile homes.

Monitoring data from the American Healthy Homes Survey II, a comprehensive survey of American
homes, ranged from 0.27 to 124.2 |ig/m3. Additional monitoring data identified through systematic
review demonstrates the potential for much higher concentrations (up to 4,500 |ig/m3) in the indoor
environment when construction materials with high emission factors are used. Some of these higher
concentrations are anticipated to be addressed by enactment and implementation of the composite wood
standards for formaldehyde emissions, but comprehensive monitoring data allowing for evaluation of
the impact of these standards are not yet available.

Indoor air monitoring data were an integral aspect of the formaldehyde indoor air assessment as it was
used to provide the best available data on real world formaldehyde indoor air concentrations. Although it
should also be noted that the formaldehyde indoor air monitoring data, even if recent, may not represent
future potential exposures from laminate wood articles in homes due to implementation of the composite
wood standards. This may be an artifact of how monitoring data cannot fully reflect how and when
formaldehyde-emitting materials—including imported articles from places with varying wood
standards—are installed. Similarly, monitoring data cannot explain how frequently these materials are
replaced. Lastly, the monitoring data may not reflect changes in energy efficiency home improvements
that reduce ventilation (e.g., leaks). Considering limitations in the monitoring data, it is reasonable to
rely on modeled concentrations according to TSCA COUs.

Based on consideration of the weight of scientific evidence, EPA has high confidence in the overall
findings for this indoor air exposure assessment.

Page 8 of 81


-------
1 INTRODUCTION

Formaldehyde is a naturally occurring chemical produced during combustion and the decomposition of
organic matter, and as a normal part of metabolism in humans as well as many other organisms.
Formaldehyde is also commercially used and manufactured extensively in construction, furniture
manufacturing, consumer products and articles. As such, formaldehyde is ubiquitous in indoor and
outdoor environments. Formaldehyde is a gas that is distributed in aqueous solution as formalin or in a
solid as paraformaldehyde.

Formaldehyde is a high priority chemical undergoing the Toxic Substances Control Act (TSCA) risk
evaluation process for existing chemicals following passage of the Frank R. Lautenberg Chemical Safety
for the 21st Century Act in 2016. It is concurrently undergoing a risk assessment under the Federal
Insecticide, Fungicide, Rodenticide Act (FIFRA). This document presents a formaldehyde indoor air
exposure assessment of TSCA COUs, as defined by TSCA sections 3(2) (defining "chemical
substance") and 3(4) (defining "conditions of use"). This TSCA-specific assessment serves to support
risk management needs by EPA's Office of Pollution Prevention and Toxics (OPPT) and is one of many
documents included within the Risk Evaluation for Formaldehyde.

1.1 Risk Evaluation Scope

The TSCA risk evaluation of formaldehyde comprises several human health and environmental
assessment modules and two risk assessment documents—the environmental risk assessment and the
human health risk assessment. A basic diagram showing the layout of these modular assessments and
their relationships is provided in Figure 1-1. This indoor air exposure assessment is shaded blue. In some
cases, individual assessments were completed jointly under TSCA and FIFRA. These modules are
shown in dark gray.

Page 9 of 81


-------
Figure 1-1. Risk Evaluation Document Summary Map

1.1.1 Indoor Air Exposure Assessment Scope	

Prior publications by regulatory and regulatory support bodies indicate that the indoor air environment is
a significant source of formaldehyde exposure (IPCS. 2002; ATSDR. 1999). EPA considered all
reasonably available data regarding TSCA conditions of use (COUs), including consumer products1 and
articles2 with high emissions and surface areas in the indoor environment. While the emission data
compiled for all articles initially modeled with CEM were published in 2009 or prior, for the revised
indoor air exposure assessment, EPA identified emission factors for composite wood materials that were

1	In the context of this TSCA risk evaluation, products are generally consumable liquids, aerosols, or semi-solids that are
used a given number of times before they are exhausted (EPA. 2019a). This is consistent with the regulatory definition of
"product" at 40 CFR 751.5.

2	In the context of this TSCA risk evaluation, articles are generally solids, polymers, metals, or woods, which are always
present within indoor environments for the duration of their useful life, which may be several years (EPA. 2019a). Articles
made of wood are often described as "wood products" commercially, but they are considered "articles" in this assessment.
This is consistent with the regulatory definition of "article" at 40 CFR 751.5). Similarly, a "composite wood product" as
defined at 40 CFR 770.3 is considered an "article" for purposes of this assessment.

Page 10 of 81


-------
published in 2019 (	) and included this information in the higher tier IECCU modeling. EPA

recognizes that while California established formaldehyde emission standards in 2010 by the California
Air Resources Board (CARB) which were intended to reduce formaldehyde emissions from pressed
wood articles in California, these standards did not apply nationally (	) Consequently,

Congress established formaldehyde emission standards for composite wood products (based on the
CARB formaldehyde emission standards for pressed wood) which began to go into effect on June 1,
2018, pursuant to the Formaldehyde Standards for Composite Wood Products regulations (40 CFR Part
770). which implement TSCA Title VI. No studies were identified that measured emission rates from
products that are known to be compliant with TSCA Title VI. Therefore, it is unknown if the currently
available emissions data reflect TSCA Title VI emission standard compliance. As such, emission factors
incorporated into the indoor models for composite wood materials (	) were estimated based

on assumed compliance with emission standards.

In 1982, the U.S. Consumer Product Safety Commission (CPSC) banned the sale of urea formaldehyde
foam insulation (UFFI) for use in residences and schools because of associated health concerns (47 FR
1662, January 13, 1982). However, this ban was reversed in 1983 (see Gulf S. Insulation v. United States
Consumer Prod. Safety Com., 701 F.2d 1137 (5th Cir. 1983)). During the public comment period for the
high priority designation of formaldehyde, the North American Insulation Manufacturers Association
submitted a comment that stated, "For those insulation products in which formaldehyde is a component
of the binder, the products are cured at high temperatures during the manufacturing process after the
binder has been applied, virtually eliminating the free formaldehyde content. Any free formaldehyde
released from the binder during heat cure is destroyed either during the cure process or by emissions
control equipment required by the MACT [maximum achievable control technology] standard....
Therefore, formaldehyde off-gassing from the majority of finished products is highly unlikely" (Docket
(I * « i \ 1 W >-OPPT-20.l'1 -01 I 0029). However, formaldehyde off-gassing has been reported from such
materials in the literature (Maddalena et at.. 2009). Thus, EPA considered the quantification of such
exposures from upholstery that are added to indoor air environments.

The Agency only quantified exposures from articles that are currently available on the consumer market.
Among other applications, formaldehyde is used for personal care products, embalming and taxidermy.
However, estimated exposures from these uses were not included in this indoor air exposure assessment
because these uses are excluded from the chemical substance definition under TSCA section 3(2)(B)(vi)
(pertaining to cosmetics as defined under the Federal Food, Drug, and Cosmetic Act) and (ii) (pertaining
to pesticides as defined under FIFRA), respectively.

Although formaldehyde is a combustion byproduct (AT	)), this indoor air exposure assessment

does not focus on byproduct or secondary formations of formaldehyde, but rather acknowledges
exposure to formaldehyde from these sources is possible. Specifically, generation of formaldehyde as a
combustion byproduct (e.g., cigarette smoking, fireplaces, wood stoves) may occur in indoor
environments and contribute to total indoor air concentrations of formaldehyde. When data are
available, combustion sources of formaldehyde are discussed in the monitoring data analysis.

Among the TSCA sources of formaldehyde contributing to indoor air exposure, wood articles are
expected to be the primary contributors (EPA. 2016) in addition to textiles and wallpaper due to the
relatively high emissions of formaldehyde and abundance in indoor environments (IPCS. 2002; AT SDR.
1999). Through a consideration of the aforementioned information, EPA modeled indoor air
formaldehyde concentrations for the following four conditions of use (TSCA COUs):

Page 11 of 81


-------
1.	Construction and building materials covering large surface areas, including wood articles;
Construction and building materials covering large surface areas, including paper articles; metal
articles; stone, plaster, cement, glass and ceramic articles;

2.	Fabric, textile, and leather products not covered elsewhere;

3.	Floor coverings; Foam seating and bedding products; Cleaning and furniture care products;
Furniture & furnishings including stone, plaster, cement, glass and ceramic articles; metal
articles; or rubber articles; and,

4.	Paper products; Plastic and rubber products; Toys, playground, and sporting equipment

EPA acknowledges that short-term {i.e., 15-minute peak), intermediate {i.e., 3-month), and long-term
{i.e., 1-year) formaldehyde indoor air inhalation exposures can occur via article uses {e.g., wood,
wallpaper, seat covers). Therefore, all three exposure durations are considered in this assessment.

1.2 Revisions between Draft and the Revised Assessment

Substantial updates have been incorporated into this assessment. The most substantial change is the use
of a second EPA model to better characterize indoor air concentrations of formaldehyde. The Draft
Indoor Air Exposure Assessment for Formaldehyde relied on the CEM to estimate 365-day average
formaldehyde concentrations from articles that may be contributing to long-term indoor air
concentrations. This model is commonly used by EPA to estimate exposure to chemicals in consumer
products and articles for TSCA conditions of use. Since articles are the subject of the formaldehyde
indoor air exposure assessment, EPA considered using CEM's E6 model that is used to estimate
emissions from articles placed in an indoor environment. To improve the specificity of the CEM model,
EPA also considered replacing default emission rates with article-specific emission rates from the
literature. However, this is a key limitation of the model. Therefore, EPA used an alternate emissions
model (E5) that would allow for the incorporation of article-specific emission rates. Although, the E5
model is typically used to estimate emissions from products (not articles) placed in an indoor
environment (	). The subtle differences between products and articles may lead to a

mismatch in the exposure profile. Products are generally liquids, aerosols, or semi-solids that release
formaldehyde only when they are used whereas articles are solids, polymers, metals, or woods that may
continuously release formaldehyde for an extended period (EPA. 2019a). COUs modeled with CEM
were based on products with two key assumptions:

1.	how often an individual may be in a room, and

2.	a constant rate of emission.

Articles are more likely to have a first order decay rate of emissions and the E5 model does not account
for this. Thus, concentrations of formaldehyde in the indoor environment may be overestimated using
CEM.

In this revised assessment, EPA used the Simulation Program for Estimating Chemical Emissions from
Sources and Related Changes to IECCU to estimate short-term {i.e., 15-minute peak), intermediate {i.e.,
3-month), and long-term {i.e., 1-year) concentrations. This model is better parameterized for volatile
organic carbons like formaldehyde. It provides exposure decay curves allowing for better
characterization of exposure concentrations over time {i.e., after an article is introduced to the home).
However, available data suggest IECCU may underestimate long-term exposure concentrations. As
such, modeled concentrations for both CEM and IECCU are presented in the results of this assessment
to characterize the potential range of formaldehyde concentrations in indoor air.

In addition to this updated modeling, this technical support document for the risk evaluation further
characterizes formaldehyde concentrations in trailer homes, athletic fields with tire crumb surfaces, and

Page 12 of 81


-------
government buildings. Furthermore, feedback and resources from data submissions to the docket
(Docket ID: EPA-HQ-OPPT-2023-0613) were incorporated throughout this assessment.

1.3 Conceptual Exposure Model

Formaldehyde is a colorless, flammable gas at room temperature and has a strong odor. As described in
the Chemistry, Fate, and Transport Assessment for Formaldehyde (EPA. 2024a). formaldehyde is
subject to several transformation processes (e.g., photolysis) in outdoor air but is not expected to be
subject to transformation and degradation processes in the indoor environment (Salthammer et at..
2010). Thus, formaldehyde is expected to persist in indoor environments.

Formaldehyde concentrations in the indoor environment are driven by emission rates from articles and
subsequent dissipation. The most common mechanism for dissipation is by mechanical removal (i.e.,
ventilation). In addition, sorption of formaldehyde to surfaces can occur but is not expected to be
significant. Updates to insulation in American homes built after 1990 has generally improved
temperature control and energy efficiency, but these updates have also led to reduced dissipation of
formaldehyde due to decreased indoor-outdoor air exchange.

Depending on the article, formaldehyde emissions may last multiple years but are expected to decrease
over time and follow a first-order exponential process (EPA. 2016). For example, according to chamber
studies of formaldehyde emissions from pressed wood articles over time, emissions half-life for such
articles ranged from 1.5 to 2 years. Furthermore, with an emissions half-life of 1.5 years, the emission
rate of formaldehyde from pressed wood articles after 10 years was estimated to be approximately 1
percent of the initial emission rate (EPA. 2016). Such emissions and the half-lives may vary according
to formulations, chemical or article-specific properties, article thickness, article surface area, usage
patterns, and environmental conditions (e.g., humidity, sun exposure).

EPA considered reasonably available information, including (1) physical and chemical properties of
formaldehyde based on its specific forms in relevant products and articles, (2) public comments received
on the draft scope document, and (3) public comments and peer review on the draft risk evaluation for
formaldehyde in finalizing the relevant exposure pathways, exposure routes, and hazards. Figure 1-2 is a
graphical depiction of the actual or predicted relationships of a subset of TSCA COUs, exposure
pathways, exposure routes, hazards, and exposed groups throughout the consumer life cycle of
formaldehyde in indoor air.

It is important to note that the COUs assessed as part of this indoor air exposure assessment, including
an assessment of these COUs based on their expected persistence and relatively high emissions of
formaldehyde per room of use, were also assessed in the Consumer Exposure Assessment for
Formaldehyde under different exposure scenarios (e.g., hobbyist installs laminate flooring in living
room for 4 hours). Unlike the Consumer Exposure Assessment, the Indoor Air Exposure Assessment
especially focuses on longer (or less intermittent) durations of exposure expected in indoor air
environments. Furthermore, while the Consumer Exposure Assessment focuses on the installation (i.e.,
by hobbyists) and intermittent use of certain articles, the Indoor Air Exposure Assessment focuses on the
relative contributions of all relevant articles added to an indoor air environment using screening and
higher tier modeling approaches and facilitates the consideration of aggregate exposures for the general
population in indoor environments.

Page 13 of 81


-------
CONSUMER ACTIVITIES &

USES

EXPOSURE
PATHWAY

EXPOSURE
ROUTE

EXPOSED
GROUP

HAZARDS

Figure 1-2. Formaldehyde Conceptual Model for Consumer Activities and Uses in Indoor Air: Consumer Exposures and Hazards

Page 14 of 81


-------
A person who installs new laminate flooring or has it installed in their home may be exposed to peak
concentrations of formaldehyde when it is first installed, followed by progressively and significantly
lower concentrations of formaldehyde from off gassing to air over the span of the first year following
installation, especially within the first two to four months. Similarly, uses of consumer articles such as
wood, furniture seat covers, and wallpaper contribute to indoor air concentrations that lead to exposures
to formaldehyde. As such, this indoor air exposure assessment provides 15-minute peak, 3-month, and
1-year formaldehyde exposure concentrations resulting from TSCA COUs.

It should be noted that the CEM modeling does not consider the lag time between when articles may be
manufactured and installed indoors since it relies on the initial formaldehyde concentration from a
finished article, as reported by relevant safety data sheets (SDSs), as the initial indoor air concentration
to which an individual is exposed. This results in a conservative emission rate and exposure estimate via
modeling since it is unlikely a newly manufactured product will be instantly installed in a home and
individuals will immediately be exposed to those higher concentrations early in their exposure period. It
is more likely that the actual emissions from any given product will be the rate following some period of
storage time where initial off-gassing at those high-rates would have occurred prior to being installed in
a residence. The IECCU modeling considered this lead time by foregoing the use of initial weight
fractions reported in article SDSs and relying on the emission factors reported from chamber studies of
finished articles.

In addition, this assessment considers available indoor monitoring data to fully characterize total
formaldehyde concentrations across a range of indoor environments.

Page 15 of 81


-------
2 APPROACH AND METHODOLOGY

EPA used the CEM to estimate indoor air concentrations from articles in the Draft Indoor Air Exposure
Assessment. The Agency considered using CEM's E6 model which is normally used to estimate
emissions from articles placed in an indoor environment, since it articles were the subject of the indoor
air exposure assessment. EPA also wanted to parametrize CEM's E6 by incorporating article-specific
emission rates into the model. However, this is one key limitation of CEM's E6 model. The Agency
used CEM's E5 model, as an alternate, which is normally used to estimate emissions from products
placed in an indoor air environment. This facilitated the incorporation of article-specific emission rates
in the CEM modeled indoor air estimates.

The subtle difference between products and articles may lead to a mismatch in the exposure profile.
Products are generally liquids, aerosols, or semi-solids that release formaldehyde only when they are
used whereas articles are solids, polymers, metals, or woods that may continuously release
formaldehyde for an extended period. COUs modeled with CEM were based on articles with two key
assumptions:

1.	how often an individual may be in a room, and

2.	a constant rate of emission.

Previous studies have demonstrated that formaldehyde emissions from newly purchased or installed
articles generally exhibit an initial period of high emissions, followed by a rapid, non-linear decline in
the emission rate (Beckett et at.. 2022; June and Mahmoud. 2022; He et at.. 2019; Liu et at.. 2015). The
decline in emissions is often modeled using a decay rate described by an exponential or power function.
Modeling emissions at a constant rate over long durations (as was done via CEM) tends to over-estimate
indoor air concentrations as a result.

To address uncertainties with the CEM long-term modeling and to generate peak estimates, EPA
modeled indoor air concentrations using the higher tier IECCU model. IECCU is peer reviewed (EPA.
2019b) and used in TSCA risk evaluation including the Final Risk Evaluation for 1 -Bromopropane
(	320) to model chemical exposures. Modeling formaldehyde exposures with high emissions and

large surface areas in IECCU is expected to have several benefits. First, IECCU allows for the use of an
emissions model that includes applicable exponential decay rates from the literature (Beckett et at..
2022; June and Mahmoud. 2022; He et at.. 2019; Liu et at.. 2015) to better predict the long term
formaldehyde indoor air contributions from TSCA articles. The model may also estimate exposures over
longer periods of time as compared to the CEM E5 condition and is expected to provide relatively
improved estimates of long-term exposure. The IECCU modeling also improves estimation of
formaldehyde indoor air concentrations by considering lead time {i.e., the time it takes to procure
materials used to manufacture the articles of interest and the delivery of the finished article to a home).
This was done by considering emission rates from finished articles. In addition, IECCU models can be
configured to include multiple items with individual emission rates in the same space. This allows for
the generation of aggregate models, as suggested by the Science Advisory Committee on Chemicals
(SACC) during the peer review of the Draft Indoor Air Exposure Assessment, for scenarios in which
consumers may bring several articles belonging to the same TSCA COU into the home simultaneously
{e.g., newly built homes, decor change scenarios). A brief overview of the IECCU model is provided in
Section 2.3.

Using IECCU, EPA estimated 15-minute peak, 3-month, and 1-year average daily indoor air
concentrations of formaldehyde from individual TSCA COUs and aggregate scenarios. EPA also
considered measured formaldehyde residential, commercial, and automobile indoor air concentration
data from indoor air monitoring studies. Given the complexities of the exposure assessment of

Page 16 of 81


-------
formaldehyde in indoor air, multiple lines of evidence were considered to understand the indoor air
concentrations of formaldehyde resulting from formaldehyde TSCA COUs. These different data sources
are described in the following sections. Monitoring results are presented in Section 3.1, while modeling
results are presented in Section 3.2 for CEM and Section 3.3 for IECCU.

EPA considered the "best available science" under TSCA 26(h) and "reasonably available information"
under TSCA 26(k) to assess formaldehyde indoor air concentrations. This was accomplished by
applying EPA's systematic review process per the Draft Systematic Review Protocol Supporting TSCA
Risk Evaluations for Chemical Substances, Version 1.0: A Generic TSCA Systematic Review Protocol
with Chemical-Specific Methodologies (also called "draft systematic review protocol") (EPA, 202 1 h) to
identify literature and data of most relevance to the indoor air pathway and modeling formaldehyde
exposures from articles of most concern in residential air. Such articles (e.g., wood articles) have a
relatively large indoor surface area and significant formaldehyde emissions relative to other articles.
EPA used the CEM as a screening tool and IECCU to refine its final assessment of formaldehyde in
residential indoor air. Although, together these two models may provide a potential range of
formaldehyde concentrations in indoor air given the uncertainties of both when calculating long-term
estimates. CEM is expected to provide the highest concentrations while IECCU is expected to provide
the lowest concentrations for TSCA COUs. Though, it should be noted that IECCU modeling cannot be
performed for automobile cabins - this is a limitation of that model. Thus, CEM modeling results are the
only estimates provided for the TSCA COUs relevant to automobile cabins and is considered best
available.

2.1 Indoor Air Monitoring Data

Monitoring data were deemed as an integral aspect of the formaldehyde indoor air assessment. Below,
EPA presents its approach to acquiring such data using a fit-for-purpose systematic review process.

2.1.1 Systematic Review Prioritization for Formaldehyde Data

Summary of the Fit-for-Purpose Systematic Review (SR) Approach for Exposure Discipline
OPPT refined the Draft Risk Evaluation for Formaldehyde by prioritizing high-quality, fit-for-purpose
data that is critical for the formaldehyde exposure analyses while meeting timeline requirements. A
targeted approach was implemented to the systematic review of exposure studies for formaldehyde to
address key data needs for the formaldehyde exposure assessment.

As of March 17, 2023, there were a total of 1,137 exposure studies; of which 1,029 studies had
completed initial reviews (i.e., primary evaluations performed by the contractor) and 388 studies had
quality control (QC) assessments completed by EPA staff. A total of 135 had data evaluation issues
pending resolution. Generally, after exposure studies undergo initial review and QC, data relevant to the
TSCA risk evaluation are extracted. Of all exposure studies, only about 30 percent were available for
data extraction with a due date of June 30, 2023. To meet deadlines and improve the quality and
relevance of formaldehyde data incorporated into the relevant exposure assessments, the formaldehyde
systematic review approach had to be improved to be more efficient and fit-for-purpose.

Prioritization Methodology

The data needs highlighted in Appendix A.l, according to exposure study type, emphasize the inhalation
pathway. Studies were prioritized if they contained indoor air concentrations and emission rate data that
were product-, article-, and COU-specific. Data were extracted from studies with an overall high rating
based on the draft systematic review protocol (	21b)—assuming that such studies would be

distinctly supportive to the formaldehyde exposure assessment.

Page 17 of 81


-------
To identify the most relevant studies to the formaldehyde exposure assessment, the Formaldehyde
Assessment Team performed a title and abstract screening (TiAB) using over 130 key words (see
Appendix A.2) determined to be associated with formaldehyde COUs and indoor air parameters of
interest, using a list of all existing formaldehyde exposure studies that provide integral or supplemental
information that is relevant to the population, exposure, comparator, and outcome of interest (PECO)
and have primary data. A Boolean search criterion was applied, generally separating keywords by
COU/product or article synonym using an or followed by an and with the air/emission criteria. For
example: (("paint" OR "vinyl wallpaper" OR "fiber glass" OR "fiberglass" OR "latex paint" OR "glue"
OR "adhesive") AND ("air" OR "indoor air" OR "ambient air" OR "air pollution" OR "air release" OR
"emission*" OR "emission rate*" OR "emission flux" OR "flux" OR "inhalation" OR "atmosphere" OR
"fume*" OR "fugitive" OR "gas*" OR "release*" OR "air release*"). Effectively, this creates a scenario
where the Agency identified a paper with a product term such as "adhesive" in its title or abstract, but
only when they appeared with an air/emission term.

Of 1,137 studies, approximately 290 were relevant to the exposure assessment of formaldehyde based on
the mentioned criteria. Of the 290 relevant studies, 185 had outstanding QCs that have been completed.
In addition, 41 articles out of the 290 prioritized studies were rated high according to the Exposure
discipline data evaluation metrics and proceeded through data extraction for incorporation into the
exposure assessment as needed. A visual representation of the formaldehyde exposure SR prioritization
scheme is included in Figure 2-1.

Figure 2-1. Schematic of the Approach Used to Identify and Extract TSCA COU-Specific Data
Pertinent to the Formaldehyde Exposure Assessment

Page 18 of 81


-------
EPA considered and incorporated several indoor air monitoring studies into this assessment (see Section
3.1). Among the presented monitoring studies, the Agency focused its review and analysis efforts on a
nationally representative formaldehyde indoor air monitoring data from the American Healthy Homes
Survey II (AHHS II) as it is the most current and first nationally representative residential indoor air
study of formaldehyde (OuanTech. 2021). See Appendix B for a detailed description of the AHHS11
data collection methodology. In addition, the SR-identified monitoring sources were supplemented by a
review of previous exposure and risk assessments, along with literature and data acquired through
backward searches, and those submitted by stakeholders including article chamber studies from industry.

In addition, the term "background" indoor air concentration has often been used in reference to indoor
air chemical assessments. This term may be generally used to define the typical measured concentrations
of a chemical in media {i.e., indoor air). It can also be used to describe the naturally occuring
concentration of a chemical. For the formaldehyde indoor air exposure assessment, this term is being
used to address the prior definition, "typical" formaldehyde indoor air concentrations, and to describe
aggregate indoor concentration of formaldehyde assumed to be composed of combinations of TSCA
sources and others that vary across indoor environments.

Interpreting available indoor air monitoring data is difficult because the ancillary data needed to
understand the data are not always available or is overwhelming in some cases for formaldehyde. This
results in assessment uncertainties associated with what the measured data represent. Ancillary
information needed to fully interpret available monitoring data include

•	Variability in sample locations (multiple or unknown formaldehyde sources, volume of samples,
varying ventilation rates, etc.);

•	Temporal variability (sample duration, lead time, collection time compared to article
introduction or installation, new homes/products, seasons, etc.);

•	Activity pattern variability across demographic categories {e.g., stay at home/work from home
individual vs. an office worker);

•	Variability in article-specific properties; and

•	Variability in consumer purchasing, use or installation of products and articles as appropriate.

These data are typically not available for the data considered in this assessment. In addition, the source
of formaldehyde resulting in the measured concentration values cannot be linked to TSCA COUs.

Generally, indoor air monitoring data available for consideration in this assessment represent typical
formaldehyde concentrations from all sources that people may be exposed to. It should be noted that
"concentrations" and "exposures" are used interchangeably in this assessment because the
concentrations of formaldehyde in indoor air are assumed to be equivalent to the exposure concentration
inhaled by people in an indoor environment. EPA used modeling tools to estimate the contribution to
formaldehyde indoor air concentrations from certain relevant TSCA COUs. The following sections
describe each modeling tool and how it was applied for this exposure assessment.

2.2 CEM Model Development and Parameterization

Formaldehyde indoor air concentrations from TSCA COUs were first estimated using CEM. Because
CEM is computationally efficient and estimates air concentrations from individual articles, it was chosen
to develop screening level estimates for this indoor air exposure assessment. CEM is a longstanding
model used by OPPT in several previous TSCA new and existing chemical risk evaluations to model
consumer and bystander exposures from products and articles. The model has been updated based on
feedback on both the performance and ease of use of the tool through beta testing and peer review (EPA.
2019aY

Page 19 of 81


-------
The scenarios, chemicals, and defaults currently included in CEM are based on available data and
professional judgment and allow the use of all parts of the model based on built-in default inputs. At any
time, defaults, chemicals, or use scenarios can be deleted, added, or refined based on newly available
information. In addition, generic (blank) scenarios are available that can be populated with user-defined
inputs (e.g., article-specific emission rates) (EPA. 2019a).

CEM retains 6 existing models from EPA's Exposure and Fate Assessment Screening Tool (E-FAST)
model and adds 15 additional models, including 6 emission models and 3 inhalation models. All CEM
models are used to estimate chemical concentrations in exposure media, including indoor air, airborne
particles, settled dust, etc. (EPA. 2019a).

A supplemental quantitative assessment, presented in Appendix D.l, compares formaldehyde indoor air
concentrations in homes with and without reported combustion sources from AHHS II monitoring
estimates vs. CEM TSCA COU modeling estimates. In addition, a supplemental quantitative assessment
of wood articles is presented in Appendix D.2 to consider the potential impact of the Title VI emission
standards for composite wood materials. This supplemental analysis allowed EPA to get an initial
projection of formaldehyde concentration reductions for hardwood floors, assuming compliance with the
Title VI rule 40 CFR part 770 (EPA. 2016) and using CEM as a screening tool. As described in
Appendix D.2, it relied on key assumptions and formulas used to determine indoor air formaldehyde
concentrations and emission rates derived from the Title VI emission standards for composite wood
materials (	). This approach was applied to EPA's assessment of wood articles in indoor air

using IECCU as a more refined modeling tool (see Section 2.3).

2.2.1	Model Output Time Period

While intermittent or peak article uses contribute to indoor concentrations of formaldehyde (including
peak concentrations), indoor air monitoring studies suggest long-term exposure to formaldehyde.
Therefore, to assess potential 1-year exposures from TCSA COUs in indoor air, through the CEM
screening assessment, EPA estimated 1-year average concentrations in automobiles and homes based on
TSCA COUs.

2.2.2	Scenario Selection

Certain materials are known to be significant emitters (e.g., wood articles) of formaldehyde compared to
others (	, Matthews et at.. 1984; Pickrell et at.. 1983). In addition, certain materials from

which formaldehyde may be emitted cover larger surfaces (e.g., wallpaper) compared to others (EPA.
2019a). After considering surface areas and rates of emissions from all formaldehyde TSCA COUs
based on the literature and professional judgement, EPA identified four COUs along with eight
scenarios deemed to be significant contributors to formaldehyde indoor air exposures. The application of
this criteria led to a focus on all relevant consumer articles and the omission of all consumer products as
defined in Section 1.1.1. Generally, consumer articles tend to be associated with more persistent
exposures over time and over relatively larger spaces, compared to consumer products, even when not
actively being used (EPA. 2019a).

Although the CEM exposure assessment focused on individual COUs, generally, an individual may be
exposed to a chemical such as formaldehyde through multiple use scenarios—including using
formaldehyde-emitting seat covers on automobile car seats vs. using formaldehyde-emitting seat covers
on a living room sofa. In these examples, the surface area of the article, size of room of use, and
interzonal ventilation rate are important inputs that have a major impact on the CEM-modeled indoor air
concentration and assumed exposure for the user of the formaldehyde-emitting article. A description of

Page 20 of 81


-------
the room of use and predefined exposure scenarios in CEM along with default parameters that can be
adjusted by the user (e.g., article-specific emission rates) can be found in the CEM user guide (EPA.
2019a). In certain cases, the modeler uses a generic product or article scenario if the scenario of interest
does not fit CEM's pre-built options. For this screening assessment, EPA utilized pre-built CEM
scenarios to model article-specific formaldehyde indoor air concentrations. The Agency assumed that
the formaldehyde indoor air exposure would occur according to the activity patterns of the individual.
The individual was assumed to be someone who goes to work or school for most of the day and spends 1
hour in a vehicle, 2 hours in a living room, and 10 hours in a bedroom every day (EPA. 2019a).

The four indoor air TSCA COUs and the relevant scenarios are described in Table 2-1.

Table 2-1. Formaldehyde Indoor Air Conditions of Use and Relevant Exposure Scenarios

Condition(s) of Use

CEM Exposure Scenarios

Construction and building materials covering large surface
areas, including wood articles; Construction and building
materials covering large surface areas, including paper
articles; metal articles; stone, plaster, cement, glass and
ceramic articles

Building/Construction Materials - Wood Articles:
Hardwood Floors (residential)

Fabric, textile, and leather products not covered elsewhere

Seat Covers (automobile)

Furniture Seat Covers (residential)

Fabrics: Clothing (residential)

Floor coverings; Foam seating and bedding products;
Cleaning and furniture care products; Furniture &
furnishings including stone, plaster, cement, glass and
ceramic articles; metal articles; or rubber articles

Furniture & Furnishings - Wood Articles:
Furniture (residential)

Paper products; Plastic and rubber products; Toys,
playground, and sporting equipment

Paper-Based Wallpaper (residential)

As shown in Table 3-1, some COUs may contain multiple exposure scenarios. In such cases, EPA
provided EPA selected a "representative" scenario (i.e., article) per COU according to the highest
estimated concentration. However, uncertainties with this approach include that the single identified
representative exposure scenario per COU, although based on the highest estimated concentration, may
not necessarily be the most common; and that an individual may be exposed to formaldehyde through
multiple articles that belong to the same COU all at once. In other words, this approach assumes that an
individual is only exposed to one of the potentially many articles that belong to the same COU at any
given time.

2.2.3 Chemical-Specific Input Parameters for CEM

EPA modeled indoor air concentrations by identifying

1.	The chemical of interest with a name (i.e., formaldehyde) and physical chemical properties (e.g.,
vapor pressure).

2.	The emission models of interest.

a. In this assessment, article-specific information used the E5 model (for products placed in
room) instead of E6 (for articles placed in room), even though the COUs identified are
articles. Under the E6 emission model, CEM estimates an initial rate of emissions based
on the initial concentration of the chemical in the solid article followed by a decline in

Page 21 of 81


-------
emission rate according to a first order exponential decay coefficient. However, EPA was
not able to obtain data for free formaldehyde content in the relevant article materials, and
it was determined that the use of an emission estimate based on gross formaldehyde
content would produce a significant overestimation of formaldehyde concentration in air
that could not be readily verified given the available data. Given the high level of
uncertainty in this approach, the E5 emission model was chosen instead. The E5 model
assumes that emissions occur at a constant rate until the chemical content is depleted; and
the emission rate is input directly by the user rather than estimated, allowing for the use
of real-world observations. While it was understood that the use of a constant rate
emission model would not yield ideal modeling conditions, and would potentially
overestimate exposures, the uncertainties arising from these conditions were well
understood and expected to be less impactful than issues arising from the use of the E6
model.

3.	The room of use (e.g., automobile, whole home, living room) and relevant environmental inputs
for the area in which the article is assumed to be placed (e.g., building volume, use environment
volume, air exchange rates)

4.	The weight fraction of the chemical in the product or article.

a. While CEM provides a default value, weight fraction is typically identified via a search
of SDSs from articles currently on the consumer market, as was done for this exposure
assessment.

5.	Product or article properties (e.g., surface area of article, frequency of use, duration of use,
emission rates)

a. While defaults are typically based on the Westat (1987) survey and EPA's Exposure
Factors Handbook (EPA. 2021a). among other sources, the modeler has the ability to edit
these parameters as they see fit. For modeling of formaldehyde concentrations,
formaldehyde emission rates from articles identified in the literature were incorporated.
Only emission rates for relevant TSCA COUs were utilized. Emission rates for COUs
other than the ones identified were irrelevant to this assessment. CEM cannot be used to
model smoking and other sources of formaldehyde resulting from combustion.

i.	Emission rates presented in Table 2-2 were extracted from literature identified
through a Google search for relevant references (Maddalema et at.. 2009; Kelly et
I.LYu and Crump. 1998; Matthews et at.. 19' I, I t krell et at.. 1984;
Pickrell et a ?). Emission rates were not initially part of systematic review
search terms at the time of the draft assessment because EPA expected weight
fractions to be sufficient indicators of formaldehyde content in articles modeled.
However, article-specific emission rates were determined to be influential input
parameters for article-specific formaldehyde indoor air modeling and have since
been added to the systematic review search terms that yielded new or updated
emission rates as presented in Table 2-3. Emission rates identified for CEM
screening assessment were commonly reported in ranges. Therefore, the midpoint
of such ranges was calculated for each product identified in the literature to
estimate the typical emissions of formaldehyde in a residential or automobile
indoor air environment.

ii.	To estimate the most common emission rates per COU category for comparison
with nationally representative indoor air monitoring data from the American
Healthy Home Survey II, a central tendency of emission rates was estimated using

Page 22 of 81


-------
an average of the median emission rates for all articles identified per COU
category. However, indoor air exposures for potentially exposed or susceptible
subpopulations (PESS) are expected to be sufficiently addressed via an aggregate
assessment of indoor air exposure from multiple TSCA COUs.

iii. During systematic review, the relevant population, exposure, comparator, and
outcomes (PECO) were identified to target only relevant sources of data for the
exposure and risk evaluation of formaldehyde. From the identified PECO-relevant
sources, emission rates were reported as ranges except for a few studies where a
single value was reported per product (Table 2-2). Although the list of identified
sources does not represent all relevant sources of data, these sources were deemed
to be sufficiently representative of article-specific emission rate data per the
systematic review process (EPA. 2021b). In instances where the emission rate
was reported in a source as less than a given value (e.g., <0.1), the emission rate
was assumed by EPA as the given value (i.e., 0.1). Non-detects (ND) were
assumed to be 0. The median of ranges was used to approximate the 50th
percentile or central tendency. The average of medians were calculated to
generate a central tendency across articles based on COU.

6. The activity pattern that identifies the start time of exposure during a product or article use day
and the general expected movement from room to room in a residence over time (EPA. 2019a).

Page 23 of 81


-------
Table 2-2. Formaldehyde Emission Rates by TSCA Condition of Use (CPU)

Condition(s) of Use

Exposure
Scenario

Identified
Product
Types in
Literature

Source
(HERO ID)

Reported
Emission
Rates, per
Surface
Area
(jig/mMir)"

Median Emission

Rates per
Identified Product,
within a COU, per
Surface Area
(jig/m2-hr)

Average
Emission
Flux per
Exposure
Scenario
(jig/m2-hr)

Expected
Room of Use

Expected
Surface
Area of
Article in
Room of
Use (m2)

Average
Emission
Rates, per
COU, and
Room of Use
(mg-hr)





Pressed wood
articles

(concentration
: 0.05 ppm)

(Matthews et
al.. 1984)

10*

10













Pressed wood
articles

(concentration
: 0.10 ppm)

(Matthews et
al.. 1984)

40*

40









Construction and
building materials
covering large surface
areas, including wood
articles; Construction
and building materials



Pressed wood
articles

(concentration
: 0.20 ppm)

(Matthews et
al.. 1984)

70*

70









Wood
Articles:

Pressed wood
articles

(concentration
: 0.40 ppm)

(Matthews et
al.. 1984)

120*

120

454.5

Residence -

27.87

12.67

covering large surface
areas, including paper
articles; metal articles;
stone, plaster, cement,
glass and ceramic
articles

Hardwood
Floors

Pressed wood
articles

(Pickrell et
al.. 1983)

ND

(assuming
"0")—1,500

750

Living Room



Bare urea-
formaldehyde
wood articles

(Va~ 3/4")

(Kellv et al..
1999)

8.6-l,580c

794.3













Coated urea-
formaldehyde
wood articles

(Kellv et al..
1999)

<2.7-460c

231.35













Bare phenol-
formaldehyde
wood articles

(Kellv et al..
1999)

4.1-9.2C

6.65













Particle board

(Pickrell et
al.. 1984)

1,500-2,167rf

1833.5









Page 24 of 81


-------
Condition(s) of Use

Exposure
Scenario

Identified
Product
Types in
Literature

Source
(HERO ID)

Reported
Emission
Rates, per
Surface
Area
(jig/mMir)"

Median Emission

Rates per
Identified Product,
within a COU, per
Surface Area
(jig/m2-hr)

Average
Emission
Flux per
Exposure
Scenario
(jig/m2-hr)

Expected
Room of Use

Expected
Surface
Area of
Article in
Room of
Use (m2)

Average
Emission
Rates, per
COU, and
Room of Use
(mg-hr)





Plywood

(Pickrell et
al.. 1984)

1,292-1,375rf

1333.5









Cabinet
(including end
cabinet)

(Maddalena
et al.. 2009)

5.21-419

212.105

Door
(including
cabinet door)

(Maddalena
et al.. 2009)

14.3-91.8

53.05

Fabric, textile, and
leather products not
covered elsewhere

Furniture
Covers, Car
Seat
Covers,
Tablecloths

Curtain

(Maddalena
et al.. 2009)

14.4-323

168.7

118.8

Automobile
(Furniture Seat
Covers)

1

0.12

Permanent
press fabric

(Kellv et al..
1999)

42-215c

128.5

Cushion

(Maddalena
et al.. 2009)

69.2-410

239.6

Residence -
Living Room
(Furniture Seat
Covers)

1

0.12

Carpet

(Maddalena
et al.. 2009)

42.4—57.6

50

Fabrics

(Pickrell et
al.. 1983)

ND

(assuming
"0")-14.58

7.29

Fabrics:
Clothing

New clothing

(Pickrell et
al.. 1983)

0.63-31.25

15.94

15.9

Residence -

Bedroom

(Clothing)

1.18

0.02

Floor coverings; Foam
seating and bedding
products; Cleaning and
furniture care products;
Furniture & furnishings
including stone, plaster,
cement, glass and
ceramic articles; metal
articles; or rubber
articles

Wood

Articles:

Furniture

Bed Deck

(Maddalena
et al.. 2009)

4.1—136

70.05

116.6

Residence -
Living Room

27.87

3.25

Bench/Seat
Bottom

(Maddalena
et al.. 2009)

33.3-293

163.15

Page 25 of 81


-------
Condition(s) of Use

Exposure
Scenario

Identified
Product
Types in
Literature

Source
(HERO ID)

Reported
Emission
Rates, per
Surface
Area
(jig/mMir)"

Median Emission

Rates per
Identified Product,
within a COU, per
Surface Area
(jig/m2-hr)

Average
Emission
Flux per
Exposure
Scenario
(jig/m2-hr)

Expected
Room of Use

Expected
Surface
Area of
Article in
Room of
Use (m2)

Average
Emission
Rates, per
COU, and
Room of Use
(mg-hr)

Paper products; Plastic
and rubber products;
Toys, playground, and
sporting equipment

Paper
products

Paper-based
wallpaper

(Kellv et al..
1999)

27

27

27.0

Residence-
Living Room
(Drywall area
used as
surrogate for
wallpaper area)

20

0.54

" Emission rates were reported as ranges with the exception of a few cases where a single value could be found per product. In instances where the emission rate was
reported in a source as less than a given value (e.g., <0.1), the emission rate was assumed by EPA as the given value (i.e., 0.1). Non-detects (ND) were assumed to be 0.
The median of ranges were taken to approximate a 50% percentile value, median or central tendency. The average of medians were calculated to generate a 50%
percentile or central tendency values across products per COU.

h At 23 °C, 50% relative humidity, CH20 ER data interpolated to fixed CH20 concentrations from 0.05-0.40 ppm. This portrays the range of ERs according to range of
concentrations.

"Emission rates represent typical conditions, defined as 70 °F, 50% relative humidity, and 1 air change per hour.

J Range indicates different test conditions in temperature and relative humidity.

Page 26 of 81


-------
Table 2-3. Formaldehyde Emissions for Materials with the potential for large applications

Material Reported

Source (Hero ID)

Reported Emissions
Factors (jig/m2-hr)

min

mid

max

Pressed Wood Articles

(FIND. 2019)

72.0

130.0

257.0

Carpet

(Kellv etal.. 1999)

0.33

0.64

2.7

Carpet

(Maddalena et al., 2009)

42.4

50

57.6

Carpet

(Yu and Cruma 1998)



28.2



Vinyl Flooring

(Maddalena et al., 2009)

1.69

7.795

13.9

Laminate Flooring (with exposed seams)

(Sheehan et al., 2017)

17.7

49.8

81.9

Laminate flooring (no exposed seams)

(Sheehan et al., 2017)

7.9

61.8

115.7

Laminate Flooring (with exposed seams)

(Chen et al., 2018)

14.0

135.5

160

Laminate flooring (no exposed seams)

(Chen et al., 2018)

5.0

7.5

10.0

Drapery and Upholstery

(Kellv etal.. 1999)

0.25

3.6

14.6

Miscellaneous Textiles

(Kellv etal.. 1999)

0.83

2.01

4.17

Fabric Curtains

(Maddalena et al., 2009)

4.97

168.7

323.0

Foam and Fiberfill Seat Cushions

(Maddalena et al., 2009)

30.4

165.15

410.0

Wallpaper

(Kellv etal.. 1999)



27.0



2.3 IECCU Model Development and Parameterization	

In addition to the improvements imparted using IECCU, EPA further improved exposure models for
consumer articles by developing low, medium, and high exposure scenarios to better account for
inherent variability in key parameters. Where possible, EPA used the minimum, arithmetic mean, and
maximum reported values for emission factors (|ig/m2-hr) in specific article materials to populate low,
medium, and high exposure scenarios. In instances where only one value was reported for a material, it
was used to populate all scenarios. Where appropriate, high, medium, and low surface areas were also
modeled for each type of article to capture variability in items consumers may place in a home. In
instances where reported emission factors might be relevant to a variety of consumer goods with large
surface areas (e.g., pressed wood articles and textiles), a single representative item for each relevant
TSCA COU was chosen for modeling. These values are shown in Table 2-3. A detailed description of
all data sources and modeling inputs is provided in Sections 2.3.4, and 2.3.5.

2,3.1 IECCU Model Description

IECCU Version 1.1 (EPA 2019b) was selected to model consumer inhalation exposures to
formaldehyde in solid articles. Formaldehyde emissions from solid articles were modeled with a single
exponential decay model. In this model, the emissions are modeled at an initial rate specified by the user
followed by a decay in rate as the chemical concentration in the article declines.

The calculated emission rates are then used in a deterministic, mass balance calculation of indoor air
concentrations. IECCU can be configured with one, two, or three zones within a building. To match the

Page 27 of 81


-------
conditions used in CEM, articles modeled in a whole home used a single zone configuration while
articles modeled in a single room used a two-zone configuration. There is no distinction made between
user and bystander, and no near-field option as in CEM product models. IECCU does have
preconfigured models for environmental inputs including building volumes, ventilation rates, and
interzonal air flows, but these were not used in this assessment. Instead, these conditions were matched
to conditions used in CEM for specific room or whole house scenarios to ensure that calculations were
consistent across models.

2.3.2	Model Output Time Period

As with the CEM screening approach, IECCU was also used to generate 1-year average concentrations
of formaldehyde in indoor air due to article uses. To evaluate long term air concentrations relevant to
chronic exposure, the simulation duration selected was 10,000 hours (~1 year), and the number of data
points selected was 5,000. At the end of this duration, air concentrations in all models were not zero, but
small enough to be negligible when compared to typical indoor air concentrations of formaldehyde,
according to monitoring studies summarized in Section 3.1. Concentrations began approaching zero
after approximately three months and peaked within the first day of article uses. Therefore, simulations
were additionally generated for 3-month average and approximately 15-minute peak exposures. To
evaluate 3-month average concentrations relevant to intermediate exposures, the simulation duration
selected was 2,190 hours (-90 days) and the number of data points was 1,098. To evaluate peak air
concentrations relevant to short-term exposure, the simulation duration selected was 1,200 hours (50
days), and the number of data points selected was 5,000. The highest value reported in the output thus
represents the peak air concentration for a 14.4 min period.

2.3.3	Article Characterization and COU Mapping

Based on the results from the CEM screening analysis, EPA updated the list of reported emission factors
to focus on emissions from materials likely to be used as building materials or manufactured into
consumer articles with large surface areas for emissions. To ensure that emission factors were
representative of items which might enter a home, EPA only used emission rates reported for goods
purchased from a retailer or sampled from a finished home environment. Key updates included addition
of emission factors for carpet, laminate flooring, vinyl flooring and foam/fiber filled seat cushion
materials and the removal of emission factors related to clothing. See Table 2-3 for all emission factors.
Emission standards are based on the concentration of formaldehyde in the air around the product in a
controlled test environment, typically expressed in parts per million (ppm). This measurement is
conducted using standardized chamber testing methods, which allows for the estimation of emission
factors. Emission standards for pressed wood articles commonly used in building material and furniture
manufacturing contexts include 0.05 ppm from hardwood plywood, 0.09 ppm from particleboard, and
0.11 ppm from medium-density fiberboard as defined at	'70.3. Under standard test conditions

(25°C, 1 atm pressure, 67-L chamber operating at one air change per hour) these values can be
converted to area specific emission rates of 72 |ig/m2-hr for hardwood plywood, 130 |ig/m2-hr for
particleboard, and 257 |ig/m2-hr for medium-density fiberboard. These values were used to define a
range of values for emission factors in pressed wood articles likely to be relevant to household indoor air
as shown in Table 2-3.

The emission factors were grouped by material (pressed wood articles, carpet, vinyl flooring, laminate
flooring, textiles, and seat cushion materials) and a range of emissions from each material type was
defined as shown in Table 2-3. The high and low values are the highest and lowest values reported
across all data sources for a given material type. The average value is calculated by averaging the mean
or midrange values for a given material type across all data sources.

Page 28 of 81


-------
For each material type, exposure scenarios were developed based on representative articles with large
surface areas. For flooring materials, including carpet, vinyl flooring, and laminate flooring, the
representative scenario is newly installed flooring in a whole home or only part of the home. For
wallpaper, the representative article was assumed to be newly installed wallpaper in part of the home.
For both textiles and seat cushion materials, the representative article chosen was a new furniture set
consisting of a couch and loveseat. However, in the case of textiles it was also considered that the
material may also be used for furniture covers rather than as upholstery. While surface area and
emission rates would be the same for these items, they belong to separate TSCA COUs, and the material
must therefore be considered in both capacities. Similarly, for pressed wood articles, the materials were
considered both as building materials and furniture. As such, both kitchen cabinets and an entertainment
center were selected as representative articles for pressed wood materials.

In total, EPA mapped the reported emissions factors for materials likely to be used as building materials
or manufactured into consumer articles with large surface areas to eight consumer articles representing
four TSCA COUs. Surface areas were then estimated for each representative article as described below.

Flooring

To estimate surface areas for flooring materials (vinyl tile, laminate flooring, and carpet), it was
assumed that the material may be used in the whole home or only part of it. As such, surface area of
flooring materials was equal to 100, 50, and 25 percent of the total floor space in high, medium, and low
exposure scenarios, respectively. The value for whole house floor space (-205 m2) was back calculated
from the CEM house volume (492 m3) and an assumed ceiling height of 8 ft (2.4 m).

Furniture Set (Upholstered or with a Textile Cover)

A furniture set consisting of a couch and loveseat was used as the representative article for both textile
and furniture cushion articles. To estimate the total surface area for a furniture set, an informal survey
was conducted to identify common dimensions sold by various internet retailers. Based on this
information, it was determined that there was considerable variability in sizes available so small,
medium, and large estimates were developed. The low, medium, and high surfaces areas, respectively,
are based on open bottom (the bottom surface is not typically upholstered or covered) prisms measuring
60" x 30" x 25", 80" x 36" x 30", and 100" x 42" x 35" for a couch; 48" x 30" x 25", 60" x 36" x 30",
and 72" x 42" x 35" for a loveseat. EPA added the lowest values for couch and loveseat to estimate
exposures to smaller furniture in the low-end scenario, and similarly for the medium and high estimates.

Pressed Wood Furniture

To estimate the total surface area for an entertainment center, an informal survey was conducted to
identify common dimensions sold by various internet retailers. Based on this information, it was
determined that there was considerable variability in sizes available so small, medium, and large
estimates were developed. The low estimate is for a small unit with dimensions of 1.22 m x 1.22 m x
0.46 m and two internal shelves. The medium estimate is for a unit with dimensions of 1.83 m x 1.52 m
x 0.61 m and three internal shelves. The high estimate is for a large unit with dimensions of 2.44 m x
1.83 m x 0.76 m and four internal shelves.

Pressed Wood Cabinets

The default value for the kitchen cabinet surface area (59 nr) from the Formaldehyde Indoor Air Model
(FIAM) was used in low, medium, and high exposure scenarios (	). Because this value is a

central tendency value calculated from multiple floorplan schematics and has been previously peer
reviewed; it was considered reasonable for use without further consideration of variability.

Page 29 of 81


-------
Wallpaper

The surface area used in low, medium, and high exposure scenarios (100 m2) is the recommended value
for wallpaper surface area in a home provided in the Exposure Factors Handbook Table 9-13 (EPA.
2011). This source is considered the default for exposure related inputs for EPA risk assessments and
was thus considered a reliable estimate without further consideration of variability.

2.3.4 Model Inputs for Single Article Models

Key inputs for modeling air concentrations in IECCU using a first order exponential decay emission
model include initial emission rate (|ig/hr), first order exponential decay constant (hr '), ventilation rate
(m3/hr), interzone exchange rate (m3/hr), and zone 1 volume (m3). Values for zone 1 volume, interzonal
air exchange rate, and the ventilation rate were taken from CEM 3.2 default values (EPA. 2019a).

Values were chosen based on the expected area of article use or installation, which may be a single room
or whole home. For whole house models, a single zone model was used so interzone air exchange rates
were not needed. The initial emission rate was calculated as the product of the emission factor and
article surface area. Development of values for emission factors and surface area are described in detail
for each representative article in Section 2.3.3 and all parameters which vary between articles are
summarized in Table 2-4.

Table 2-4. IECCU Modeling Inputs for Single Articles with Large Surface Area Representing

TSCA COUs







Emission
Factor
Gig/m2-
hr)"

Article
Surface
Area
(m2)b

Initial
Emission

Rate
(jig/hr)b

Zone 1

Ventilation

Interzone
Air

COU(s)

Article

Level

Volume

(m3) b

Rate
(m3/hr) b

Exchange

Rate
(m3/hr) b





High

160.0

202

32,320









Laminate
Flooring

Med

50.1

101

5,060

492

221.4

N/A



Low

5.0

50.5

252







Construction and building
materials covering large surface



High

57.6

202

11,635







Carpet

Med

26.3

101

2,654

492

221.4

N/A

areas, including wood articles;
Construction and building
materials covering large surface
areas, including paper articles;



Low

0.3

50.5

17







Vinyl
Flooring

High

13.9

202

2,808







metal articles; stone, plaster,
cement, glass and ceramic articles

Med

7.8

101

787

492

221.4

N/A



Low

1.7

50

84.5









Pressed

High

257.0

59

15,163









Wood
Cabinets

Med

130.0

59

7,670

24

221.4

108.98



Low

72.0

59

4,248







Fabric, textile, and leather products
not covered elsewhere

Textile

High

323.0

17

5,491







Furniture

Med

58.1

12

697

50

221.4

108.98

Covers

Low

0.3

7.9

2







Floor coverings; Foam seating and

Textile
Furniture

High

323.0

17

5491







bedding products; Cleaning and
furniture care products; Furniture
& furnishings including stone.

Med

58.1

12

697

50

221.4

108.98

Components

Low

0.3

7.9

2







Page 30 of 81


-------
COU(s)

Article

Level

Emission
Factor
Gig/m2-
hr)"

Article
Surface
Area
(m2)b

Initial
Emission

Rate
(jig/hr)b

Zone 1
Volume

(m3) b

Ventilation
Rate
(m3/hr) b

Interzone

Air
Exchange

Rate
(m3/hr) b

plaster, cement, glass and ceramic
articles; metal articles; or rubber
articles

Foam

Furniture

Components

High

410.0

17

6,970

50

221.4

108.98

Med

165.2

12

1,982

Low

30.4

7.9

240.2

Pressed

Wood

Furniture

High

257.0

30.2

7761

50

221.4

108.98

Med

130.0

16.4

2,129

Low

72.0

7.46

537

Paper products; Plastic and rubber
products; Toys, playground, and
sporting equipment

Wallpaper

High

27.0

100

2,700

492

221.4

N/A

Med

27.0

100

2,700

Low

27.0

100

2,700

" See Table 2-3 for a list of sources used for article-specific emission factors.

b The article surface areas and buildins configurations were sourced from the EPA Exposure Factors Handbook (EPA.
2011). Formaldehyde Indoor Air Model (EPA. 2012). CEM User Guide (EPA. 2019a) and IECCU User Guide (EPA.
2019b).

The decay rate of formaldehyde emissions can be estimated by measuring the emission rates from
consumer goods over time in a controlled environment and fitting these measurements to a single
exponential decay model. Estimated decay rates for formaldehyde emissions from consumer goods were
obtained from two studies. The selected studies focus on consumer goods purchased from a retailer
(rather than immediately after manufacture) to ensure that decay rates are relevant to emissions for items
which could be purchased and installed in a home. In the first study, four types of furniture products
with solid wood, stained wood veneer, and pressed wood components were tested for 4000 hrs in a
controlled study chamber and emissions data were fitted to a first order exponential decay model (Liu et
al.. 2015). In the second study, three kinds of flooring and three kinds of wallpaper were tested for 400
hours in a controlled study chamber and decay rates were estimated by fitting emissions data to a first
order decay model (Jung and Mahmoud. 2022). In total, eleven values for single exponential decay rate
of emissions were reported from separate chamber experiments in these two studies. The reported values
ranged from 8.7xl0~4 to 4.18 xlCT3 hr_1, with an average reported value of 2.41 xl0~3 hr_1. The
relatively low level of variability in this parameter is notable given that these studies examined different
materials, used different test conditions, and calculated decay rates using different methods. As such, the
mean value for decay rate (2.41 x 10~3 hr"1) was chosen for use in all exposure scenarios.

For articles modeled in the same area of the home, environmental factors affecting air concentration
(e.g., volume, interzonal air flow, and ventilation rate) will be the same. In these cases, the only factor
which would produce a difference in air concentration between these items is the initial emission rate
(|ig/hr); higher initial emission rates will result in higher air concentrations. As such, when more than
one representative article belonging to the same TSCA COU would be modeled in the same area of the
home, it was not necessary to generate models for each article to identify the item which would result in
the highest air concentrations. In these cases, only the representative article with the highest initial
emission rate was modeled. As such, models for carpet, vinyl flooring, textile furniture components, and
foam furniture components were not run and will not be reported in individual model results. However,
some of these items were included in the aggregate scenarios described in Section 2.3.5.

Page 31 of 81


-------
2,3,5 Model Inputs for Aggregate Models

In addition to the model scenarios generated for individual articles representing TSCA COUs, EPA
generated two composite scenarios in which multiple new articles belonging to the same COU are
assumed to be installed in the home at the same time. The first scenario is the "decor change" scenario in
which it is assumed that a new couch, loveseat, and entertainment center are purchased and brought into
the home at the same time. For the upholstered furniture, both textile and foam furniture components are
assumed to emit formaldehyde.

The second scenario is the "new build" scenario in which the home has newly installed flooring and new
kitchen cabinets. Note that building materials which may emit formaldehyde are used widely in home
construction, but many of these items such as underlayment, insulation, and plywood subflooring are
generally covered by other materials which would serve as a barrier to emissions. In addition, these
materials are added earlier in the construction process and will have an air out period before construction
is complete. Flooring and cabinetry are generally added at the end of construction and surfaces can emit
freely to air; thus, EPA considers it reasonable that these materials may contribute to air concentrations
at the time of occupancy by the purchaser. In this scenario, the home is assumed to have mixed flooring
installed. This includes four carpeted bedrooms with the remainder of the home flooring made of
laminate material (except in the garage, which is assumed to be unfinished concrete). The surface area of
carpet was calculated using the default CEM room volume of 36 m3 per bedroom and an assumed
ceiling height of 8 ft (2.4 m). Garage floor area (-37.5 m2) was calculated using the default CEM room
volume 90 m3 for the garage and an assumed ceiling height of 8 ft (2.4 m). The area of laminate flooring
surface area was calculated as the difference between the previously described value for whole home
floor area and the combined values for carpeted areas and unfinished garage floor. Surface areas for
pressed wood cabinets were the same as described in single item scenarios. All materials used the same
value for single order decay rate previously described. Model inputs for aggregate scenarios are
summarized in Table 2-5.

Page 32 of 81


-------
Table 2-5. IECCU Modeling Inputs for Aggregate Scenarios with Multiple Articles Representing a
Single TSCA CPU

COU

Scenario

Article

Level

Emission
Factor
(jig/m2-
hr)

Article
Surface
Area
(m2)

Initial
Emission
Rate
(jig/hr)

Zone 1
Volume

(m3)

Ventilation
Rate
(m3/hr)

Interzone

Air
Exchange
Rate
(m3/hr)

Floor coverings;
Foam seating
and bedding
products;
Cleaning and
furniture care
products;
Furniture &
furnishings
including stone,
plaster, cement,
glass and
ceramic articles;
metal articles;
or rubber
articles

Pressed

Wood

Furniture

High

257.0

30.2

7,761

Med

130.0

16.4

2,129

Low

72.0

7.46

537

Decor
Change

Indoor

Furniture

(Foam

Components)

High

410.0

17

6,970

Med

165.2

12

1,982

Low

30.4

7.9

240.2

50

221.4

108.98

High

323.0

17

5,491

Indoor
Furniture
(Textile
Components)

Med

58.1

12

697

Low

0.3

7.9

Construction
and building
materials
covering large
surface areas,
including wood
articles;
Construction
and building
materials
covering large
surface areas,
including paper
articles; metal
articles; stone,
plaster, cement,
glass and
ceramic articles

High

160

104.5

16,720

Laminate
Flooring

Med

50.1

104.5

5,236

Low

104.5

523

Pressed

Wood

Cabinets

High

257.0

59

15,163

Med

130.0

59

7,670

New Build

Low

72.0

59

4,248

492

221.4

N/A

High

57.6

60

3,456

Med

26.3

60

1,577

Carpet

Low

0.33

60

20

Page 33 of 81


-------
3 RESULTS

3.1 Monitoring Data

A summary of formaldehyde concentrations identified from residential (

Table 3-1) and non-residential (Table 3-2) indoor air monitoring studies was evaluated to understand
how people are exposed to formaldehyde in these settings. These monitoring data do not differentiate
among the various sources of formaldehyde whether they be from TSCA articles or other sources.
Additionally, these monitoring data are not expected to represent peak exposure concentrations since the
air monitoring often occurs some period of time after articles are installed in a home. These data are
more representative of long-term exposures to formaldehyde. In a few cases, however, monitoring
studies do capture peak concentrations such as those that sample new homes or trailers (e.g., unoccupied
Federal Emergency Management Agency [FEMA] trailers). These tend to be outliers in the monitoring
data and show substantially higher concentrations when compared to other monitoring data. It is
important to note that the available monitoring studies were conducted prior to the implementation of
the Title VI pressed wood product emission standards. Formaldehyde concentrations in all indoor air
settings may be reduced as these standards are implemented over time.

Table 3-1. Indoor Air Monitoring Concentrations for Formaldehyde

Reference

Monitoring Study Description

Formaldehyde Concentrations (jig/m3)

Central Value

Range/Percentiles

(ATSDR 2007)

96 unoccupied FEMA trailers
assessed during the summer of
2006

Mean: 1,280

Range: 12.28-4,500

American Healthy
Home Survey II
(OuanTech. 2021)

Nationally representative
sample of 689 U.S. homes of
various ages, types, conditions,
and climates

Mean: 23.2

Range (lower/upper 95%
tiles of mean): 21.4-24.9

(Board. 2004)

Portable and traditional
classrooms in 67 California
schools (Phase II study)

Arithmetic Mean:
18.42 (portable)
14.74 (traditional)

95th Percentile:
31.93 (portable)
27.02 (traditional)

(Gilbert et al.. 2005)

59 homes in Prince Edward
Island, Canada

Geometric Mean: 33.16

Range: 5.53-87.33

(Gilbert et al.. 2006)

96 homes in Quebec City,
Canada

Geometric Mean: 29.48

Range: 9.58-89.91

(Hodason et al.. 2004)

4 new relocatable classrooms

Unspecified Mean:
9.83 (indoor-outdoor)

Range: 4.91-14.74
(indoor-outdoor)

(Hodeson et al.. 2000)

New homes in eastern/SE U.S.:
4 new manufactured homes

7 new site-built homes

Geometric Mean:
41.76

44.22

Range:
25.79-57.73

17.2-71.24

(Liu et al.. 2006)

234 homes in Los Angeles
County, CA; Elizabeth, NI; and
Houston, TX

Median: 20.02

Range:

12.53-32.43
(5th-95th percentiles)

(LBNL. 2008)

4 FEMA camper trailers

Unspecified Mean:
568.67

Range: 330.39-924.85

Page 34 of 81


-------
Reference

Monitoring Study Description

Formaldehyde Concentrations (jig/m3)

Central Value

Range/Percentiles

(Murphv et al.. 2013)

Sample:

All structures (519)
Travel trailers (360)
Park models (90)
Mobile homes (69)

Geometric Mean:

94.57

99.49

54.04

70.01

Range:

3.68-724.65

3.68-724.65

3.68-196.52

13.51-393.03

(Offermann et al..
2008)

108 new SF homes in CA

Median: 38.2

Range: 4.67-143.33

(Sax et al.. 2004)

Inner-city homes:

NY City (46) - winter (W),
summer (S)

Median:

12.28 (W), 18.42 (S)

Range:

4.91-22.11 (W), 6.14-
50.36 (S)



Los Angeles (41) - winter (W),
fall (F)

18.42 (W), 14.74 (F)

7.37-55.27 (W), 7.37-
31.93 (F)

Table 3-2. Formaldehyde Monitored in Commercial Buildings in the United States

References

Monitoring Study Description

Formaldehyde
Concentrations
(iug/m3)

Descriptor

(Ceballos and Burr, 2012)

Office space indoor air monitoring for
formaldehyde in a commercial building

24.56

Average

(EPA. 2023 c)

Indoor air monitoring across 100
randomly selected U.S. commercial
buildings

3.68

5th Percentile

14.74

50th Percentile

30.71

95th Percentile

(Pase and Couch, 2014)

Indoor air U.S. government offices

<61.41

Maximum

(Lukcso et al., 2014)

12.28

Geometric
mean

56.50

Maximum

(Dodson et al., 2007)

Classrooms in U.S. school buildings

17.69

Median

An important consideration in these data is how formaldehyde may dissipate in homes. The most
prominent cause is home ventilation either through mechanical systems or through open windows. Due
to improved insulation in American homes built after 1990, formaldehyde may persist longer in newer
homes compared to older homes as a result of reduced indoor-outdoor air exchange (see Appendix C).
Sorption is not expected to be a key source of dissipation for homes. Also, the available monitoring data
may not reflect expected future indoor concentrations as energy efficiency and building materials
change over time. These factors are not readily presented in the monitoring data but do play an
important role in understanding why some homes may have higher formaldehyde concentrations than
others.

3.1.1 Summary of Monitoring Data Results

As previously mentioned, indoor air environments may have a variety of attributes that influence indoor
formaldehyde concentrations including, but not limited to, the age of the building, ventilation rates,
types, quantity, and age of articles. Indoor air concentrations of formaldehyde in the AHHS II study

Page 35 of 81


-------
ranged from 0.27 to 124 |ig/m3 (QuanTech. 2021).Table 3-1. Indoor Air Monitoring Concentrations for
Formaldehyde Outside of the United States, the average measured formaldehyde concentrations are
between 20 and 40 [j,g/m3 in European homes (ECHA. 2019) and 30 to 40 [j,g/m3 in Canadian homes
(Canada. 2005).

Unlike residential settings, most commercial settings are not expected to have sources of formaldehyde
attributable to combustion. A comparison of formaldehyde indoor air concentrations from both
residential and commercial settings and residential settings suggests similar concentration ranges of
formaldehyde. The similar ranges of concentrations across residential and commercial settings also
suggests that, while combustion sources may be notable contributors in some residential settings,
combustion is not a substantial contributor to typical indoor air concentrations that Americans may be
exposed across all indoor air environments.

3.1.2 American Healthy Homes Survey II (AHHS II)

The AHHS II formaldehyde residential indoor air monitoring survey is the most recent and relevant
high-quality American residential indoor dataset for formaldehyde. It is a well-designed study. See
Appendix B for a detailed summary of the AHHS II data collection methods. The AHHS II was the first
national study of formaldehyde concentrations in homes across the contiguous United States. The AHHS
II survey was sponsored by the U.S. Department of Housing and Urban Development (HUD) along with
EPA, and was conducted by QuanTech, Inc. (QuanTech. 2021) from March 2018 through June 2019.
The AHHS II measured household levels of lead, lead-based paint hazards, pesticides, formaldehyde,
and mold in American homes. Ninety-eight percent of the homes surveyed (689 out of 703 housing
units) were randomly selected in 78 cities and counties across 37 states. A sample was not collected in
14 housing units due to air sampling pump failure. The sampled homes represent both owned housing
units and rented units (QuanTech. 2021).

Summary statistics for the AHHS II measured formaldehyde concentrations in indoor air are presented
in Table 3-3 and a histogram of the data is provided as Figure 3-1. Indoor air concentrations of
formaldehyde in the AHHS II study ranged from 0.27 to 124 |ig/m3 (3.5-hour TWA). These samples
represent the indoor air concentration of formaldehyde in the most used room in the home. Statistical
weights reported in the AHHS II data are applied here to reduce sampling bias and provide a more
nationally representative distribution of monitored values (QuanTech. 2021).

Table 3-3. Range and Weighted Quantiles of AHHS II Residential Indoor Air Formaldehyde
Concentrations (ng/m3)				

Minimum

10th Percentile

Median

90th Percentile

Maximum

0.27

7.54

19.8

41.8

124.2

Page 36 of 81


-------
(/)

„	„

*~—'

CO

c
3

_c

CD

U)

"CD

c

'(f)

£

D
O

CO
o

X



M—

CO

o

-4—'

CO



-4—'

CD

C0

_Q

E

E

=3

z

o

i_
M—

	

"O

co

CD

c

>

o





CD

03

Q

z



20,000,000-

5,000,000-

0-

20	40	60	80	100

Indoor Air Formaldehyde Concentration (|jg/m3)

120

Figure 3-1. Histogram of Formaldehyde Indoor Air Sampling Results from AHHS II with
Statistical Weights Applied

Surprisingly, AHHS II researchers found that homes without tobacco smoke had significantly higher
average formaldehyde concentrations compared to homes with reported smokers (23.79 |ig/m3 vs 19.21
|ig/m3). In addition, homeowners were twice as likely to report smoking at least once per day compared
to renters; 26.2 vs. 12.5%, respectively (OuanTech. 2021). The AHHS II does not provide a reason for
these differences. However, these differences suggest that tobacco smoke is not a significant source of
formaldehyde in the surveyed homes.

Despite the utility of this robust dataset, the measured concentrations cannot be linked to a specific
TSCA COU or group of COUs. The concentrations represent aggregate formaldehyde concentrations in
homes across the United States and provide context on how much formaldehyde people are regularly
exposed to in their homes.

3.1.3 New Homes	

New homes tend to have significantly higher concentrations of formaldehyde due to improved insulation
and the incorporation of new articles that may still be off-gassing formaldehyde. Two studies were
identified that highlight these issues.

Hodgson et al. (2000) measured indoor air formaldehyde concentrations in newly manufactured homes
and new site-built homes. Formaldehyde concentrations ranged from 25.79 to 57.73 |ig/m3 among four
newly manufactured homes in the Eastern and Southeastern United States after 2 to 9.5 months. In
contrast, indoor air concentrations ranged from 17.20 to 71.24 |ig/m3 among seven new site-built homes
after only 1 to 2 months (Hodgson et al.. 2000). The difference in the range of these concentrations are
attributed to the construction type and the sampling period. Furthermore, several of the site-built homes
had relatively poor ventilation rates per the American Society of Heating, Refrigerating and Air-
Conditioning Engineers. All homes were in hot and humid climates that generally increase emissions of
formaldehyde. According to the authors, plywood flooring, latex paint, and sheet vinyl flooring were
major sources of formaldehyde (Hodgson et al.. 2000). All are TSCA COUs.

Offerman et al. (Offermann et al.. 2008) also measured formaldehyde concentrations in newly
constructed homes. In this study, 108 newly constructed homes were sampled in California (Offermann

Page 37 of 81


-------
et ai. 2008). The measured indoor air concentrations of formaldehyde ranged from 4.67 to 143.33
|ig/m3. Given that a primary focus of this study was on the effect of ventilation on indoor air
formaldehyde concentrations, the authors determined that because new single-family homes in
California are built relatively air-tight, and because the windows and doors were kept shut during the
duration of the study—the indoor-outdoor air exchange rates were generally low {i.e., 0.2 air exchanges
per hour). This resulted in significantly elevated indoor air concentrations of formaldehyde (Offermann
et ai. 2008).

3.1,4 Trailer Studies

Trailers may also have substantially higher concentrations of formaldehyde when constructed with
materials that significantly emit formaldehyde. Three monitoring studies were identified that provide
insight to how formaldehyde concentrations may vary among these housing units.

In the summer of 2006, as part of a health consultation, the Agency for Toxic Substances and Disease
Registry (ATSDR (2007)) was requested by the FEMA to assess formaldehyde indoor air concentrations
in 96 unoccupied residential trailers (ATSDR. 2007). Such trailers are often provided to victims as a
form of temporary housing following displacement from their home due to natural disasters. The air
samples were collected by the EPA. Formaldehyde concentrations were initially measured when the
trailers were unventilated, and again during a two-week intervention in which either air conditioning
was turned on or windows were open. Results showed that each intervention was able to lower
formaldehyde concentrations significantly. Without any intervention, the measured formaldehyde
concentrations ranged from 12.28 to 4,500 |ig/m3, with an average of 1,280 |ig/m3. The interventions
reduced formaldehyde concentrations, with open windows causing greater reductions (to an average of
110 |ig/m3) compared to air conditioning (to an average of 480 |ig/m3). The authors noted that while
concentrations were reduced during the air conditioning intervention, formaldehyde concentrations in
the unoccupied FEMA trailers remained at a range that may be associated with acute health symptoms in
certain individuals. They also stated that during both interventions, formaldehyde concentrations
remained above certain federal health-based guidelines. Furthermore, when windows were closed,
higher temperatures were associated with higher formaldehyde concentrations. In addition,
formaldehyde levels tended to differ across commercial brands of trailers. However, it is important to
note that this health consultation did not assess formaldehyde under normal use or living conditions.
Hence, the authors caution that these results should not be generalized to all FEMA trailers, and that the
findings do not predict health outcomes due to living in such trailers. Lastly, since the health
consultation did not assess human exposures, the authors noted that this analysis could not be used to
define levels of concern (ATSDR. 2007).

Lawrence Berkeley National Laboratory (LBNL) measured formaldehyde indoor air concentrations
within four FEMA camper trailers with concentrations ranging from 330.39 to 924.85 |ig/m3 (LBNL.
2008). According to LBNL, relatively high concentrations of formaldehyde measured in FEMA
temporary housing units are likely due to the very high composite wood surface area relative to room
volume in addition to low ventilation rates—specifically for low area-specific fresh air flow rates in
relation to the internal surface area in the assessed temporary housing units. Notably, the authors
indicated that results from this study were not representative of all FEMA temporary housing unit
conditions given only four such units were assessed. It is, however, representative of other temporary
housing unit indoor air conditions with similar materials and low air flow conditions (LBNL. 2008).

According to a study by Murphy et al, with a sample of 519 FEMA-supplied trailers, including travel
trailers, park models, and mobile homes, peak formaldehyde indoor air concentrations ranged from
196.52 to 724.65 |ig/m3 according to trailer type (Murphy et al.. 2013). The geometric mean

Page 38 of 81


-------
concentration of formaldehyde in such homes were higher than levels found in traditional homes, as also
presented in the (Murphy et al.. 2013) study. (Murphy et al.. 2013) noted that low air flow (caused by
closed windows more than by lack of adequate air conditioning) was a key reason for the relatively high
concentrations of formaldehyde found in trailers. Increased indoor air temperate and relative humidity
also correlated with increased formaldehyde concentrations in trailer indoor air. Although the authors
did not investigate the impact of material type and trailer material composition on measured
formaldehyde indoor air concentrations, all trailer brands had some trailers with formaldehyde
concentrations exceeding 123 ug/m3 (Murphy et al.. 2013).

It is important to note that these studies which show elevated indoor air formaldehyde concentrations,
especially compared to other settings, were conducted prior to the implementation of the Title VI
pressed wood product emission standards. Therefore, it is conceivable that formaldehyde concentrations
in these settings (and all indoor air settings) may be reduced as these standards are implemented over
time.

3.1,5 Commercial and Other Buildings	

Formaldehyde emissions in commercial and other buildings can expose office workers, students, and the
general population. These settings may have high formaldehyde concentrations due to off gassing from
building materials and other articles. Likewise, these emissions may be from multiple COUs (e.g.,
composite wood articles; coatings, paints, adhesives, sealants; formaldehyde-based furnishings; and
building materials). Twelve monitoring studies were identified that provide some context for
formaldehyde exposures in these settings.

Ceballos and Burr (2012) evaluated formaldehyde indoor air exposures in an office located in a two-
story commercial building. The office contained cubicles separated by fabric-covered dividers and most
of the office was carpeted. Over the 2-day sampling period inside the office, area concentrations
remained at 25 |ig/m3. Dodson et al. (2007) conducted personal breathing zone (PBZ) sampling of
teachers in primary and secondary schools, as well as office workers. The median of these personal
samples was 18 |ig/m3.

Additionally, EPA identified studies measuring formaldehyde exposure in office environments outside
of the United States. Han a zato et al. (2018) measured area concentrations in a newly constructed
commercial bank in Japan, and the formaldehyde concentrations in the samples ranged from 1.5 to 3.2
|ig/m3. Samples were collected in the lobby, office space, seminar room, and outdoor space. Another
study in Japan measured area formaldehyde concentrations across 17 office buildings with
concentrations ranging from 3.4 to 21 |ig/m3 in the winter and 12 to 45.2 |ig/m3 in the summer (Azuma
et al. 2017).

In Sweden, the PBZ of 79 participants across 8 office buildings was measured (Glas et al.. 2004). The
PBZ samples ranged from 2 to 18 |ig/m3, with an average of 9 |ig/m3. Another study measured the PBZ
of office workers in Sweden and Finland, with geometric means of 7.6 |ig/m3 and 8.1 |ig/m3,
respectively (Glas et al.. 2014). Dingle et al. (2000) measured the area concentrations of formaldehyde
across 18 conventional offices and 20 portable office buildings located on a university campus in
Australia. The concentrations in the conventional office buildings ranged from 12 to 90 |ig/m3, and the
concentrations in the portable office buildings ranged from 516 to 2,592 |ig/m3. The elevated
formaldehyde concentrations in the portable office buildings were believed to be from the particleboard
and plywood present in those buildings (Dingle et al.. 2000).

Page 39 of 81


-------
In general, higher formaldehyde indoor air concentrations had been reported in the past. EPA identified
two studies from office environments in the United States in the 1990s with area concentrations ranging
from less than 12 to 2,456 |ig/m3 (Hedge et a >; Kaiser and Sylvain. 1994). Additionally, EPA
identified three studies spanning multiple office buildings in Canada with individuals exposed to
formaldehyde (Haghighat and Donnini. 1999; Allaire et at.. 1997; Menzies et at.. 1996). Menzies et al.

5) measured formaldehyde concentrations in the air ranging from 15 to 59 |ig/m3 in two office
buildings. Between the other two studies, concentrations ranged from less than 2 to 2,590 |ig/m3, and
most of the office buildings were carpeted (Haghighat and Donnini. 1999; Allaire et al.. 1997). A study
conducted in 29 office buildings in northern Sweden measured air concentrations of formaldehyde
ranging from 11 to 59 |ig/m3 (Sundell et; 3).

3.1.6	Japan National Study	

In 1996, Japan's National Institute of Health Sciences administered the first national survey of
formaldehyde in approximately 230 homes with an arithmetic mean concentration of approximately
74.92 [j,g/m3. After repeating this monitoring study with 1,181 homes in 2005, the arithmetic mean of
formaldehyde across Japanese homes decreased to approximately 29.98 (.ig/m3 (Osawa and Havashi.
2007; Azuma et al.. 2005). As reported by the World Health Organization, this reduction in average
formaldehyde concentration in Japanese homes from 1996 to 2005 was likely due to an amendment of
the national building codes and, more specifically, a restriction of materials that emit formaldehyde in
interior finishing (WHO. 2010).

3.1.7	Relative Contributions of Formaldehyde Sources in Residential Indoor Air

Monitoring data from AHHS II suggests that concentrations of formaldehyde may range from 0.27 to
124.2 |ig/m3 for all homes, with 95 percent of homes having concentrations below 47 |ig/m3 (QuanTech.
2021). Those data include formaldehyde produced from both TSCA sources and other sources of
formaldehyde such as tobacco smoke or the use of fireplaces, gas-burning appliances, candles, incense
and air purifiers (QuanTech. 2021; Salthammer. 2019). Of these other sources, the most common and
significant contributors to indoor air are likely from fireplaces, gas burning appliances and air purifiers.
It should be emphasized that some of these sources {i.e., tobacco smoke) may not contain formaldehyde
but rather lead to the formation of formaldehyde during use.

For other sources of formaldehyde in indoor air, simulated 50th percentile room concentrations ranged
from 12.3 to 44.2 [j,g/m3 individually for wood combustion, air cleaning devices, cooking, candles, and
incense, and up to 152.2 (.ig/m3 for ethanol fireplaces (ECHA.; ). Air cleaning devices such as
photocatalytic air purifiers can produce formaldehyde from irradiation of air contaminants, leading to
increased indoor air concentrations of formaldehyde (Salthammer. 2019). Formaldehyde production
associated with cooking depends on many factors, including cooking temperature and type of oil and
variety of food being cooked. Select gas-oven cooking tests involving a variety of cooking parameters
resulted in formaldehyde concentrations ranging from 36.5 to 417.3 (.ig/m3 (Salthammer. 2019). Tobacco
smoke is also known to be a contributor to formaldehyde concentrations within all indoor air
environments (• • \ , >rman et al.. 1982). although according to the World Health Organization
tobacco smoke primarily increases formaldehyde concentrations in indoor air environments where the
rates of smoking are high with minimal ventilation (IPCS. 2002).

In addition, tire crumb-based surfaces may also be a source of formaldehyde exposure for some
populations including children or adults who play sports on fields made of such materials. Due to
increased concerns from parents, athletes, schools and communities regarding potential chemical
exposures from recycled tire crumb surfaces, in 2016 the Centers for Disease Control and
Prevention/Agency for Toxic Substances and Disease Registry (CDC/ATSDR) and the EPA, in

Page 40 of 81


-------
collaboration with the CPSC, launched a multi-agency effort via the Federal Research Action Plan on
Recycled Tire Crumb Used on Playing Fields and Playgrounds (FRAP) to investigate potential human
exposures to the substances associated with recycled tire crumb rubber used on synthetic turf fields
(EPA. 2019c). Researchers collected tire crumb rubber samples from nine tire recycling facilities, and
tire crumb rubber infill material from 40 synthetic indoor and outdoor turf fields located throughout the
United States. These samples were quantitatively analyzed for formaldehyde through chamber emission
testing at 25 °C and at 60 °C using high-performance liquid chromatography/ultraviolet spectroscopy.
At 25 °C, all formaldehyde measurements were below limits of quantification for synthetic field tire
crumb rubber infill. At 60 °C, the average formaldehyde emissions from the chamber studies were 16
ng/g/h for such materials. The average emissions of formaldehyde were 23 ng/g/h for indoor field based
materials and 12 ng/g/h for outdoor field based materials. The authors reported no statistical difference
among the measured emissions based on the period of installation (2004-2008, 2009-2012, and 2013-
2016) and that no specific chemical hazards from recycled tires in playground surfacing were identified
(EPA. 2019cY

Many of these other sources of formaldehyde represent temporary emission sources, which may affect
the overall impact on indoor air quality. Further, qualities such as the frequency and duration of use of
these temporary formaldehyde sources (e.g., using a fireplace), age of the indoor spaces and
formaldehyde-containing finishes and furnishings, and ventilation rate will impact the total
concentration of formaldehyde in indoor air and the relative contribution of TSC A and other sources to
indoor air.

Although there are some uncertainties in estimating indoor air concentrations of formaldehyde, EPA
generally expects that a larger number of formaldehyde sources will lead to higher concentrations of
formaldehyde in the indoor air (EPA. 2016; IPCS. 2002; AT SDK I • !')9; Girman < d s * 2). As
previously noted, there is insufficient data to quantify the relative contributions of the modeled TSCA
COUs to the AHHS II monitored concentrations of formaldehyde in American residential indoor air
with certainty. However, article-specific and aggregate scenario based indoor air concentrations of
formaldehyde were within the same order of magnitude as reported in the AHHS II study. Therefore, it
is reasonable to conclude these results support the hypothesis that the identified TSCA COUs are key
contributors to real-world concentrations of formaldehyde in residential indoor air.

3.1.8 Comparing Indoor to Outdoor Air

As has been presented, indoor air concentrations of formaldehyde can vary due to the materials used in
construction, the timing of when new articles are introduced, and the amount of ventilation present.
Indoor air concentrations tend to be much higher than those found outside. To understand the
quantitative difference in these numbers, EPA developed Figure 3-2 to show the distribution of
monitoring data in both settings.

Page 41 of 81


-------

-------
products designed to clean/purify indoor air of certain pollutants can be a source of formaldehyde
exposure (or other chemical exposures) in homes or residences (Salthammer. 2019).

3.2 Tier 1: CEM Modeling	

This section summarizes the CEM indoor air exposure modeling results for formaldehyde. A full
description of the general screening methodology can be found in Section 2.2. Supplemental CEM
results not presented in this section can be found in Appendix D. EPA did not model indoor air
concentrations of commercial buildings as residential homes were considered a more protective indoor
air scenario.

3.2.1 CEM Modeling Results

Modeling results for inhalation exposures estimated with CEM are summarized and presented,
respectively, in Table 3-4 and Figure 3-3Figure 3-3. , according to COU. Table 3-4 and Figure 3-3Figure
3-3. present the various TSCA sources contributing to formaldehyde concentrations in residential and
automobile indoor air environments. The largest contributor to the 1-year average daily concentration of
formaldehyde in a typical home is the building wood articles (423.47 (J,g/m3) scenario. The lowest
contributors to formaldehyde indoor air concentrations were furniture seat covers (4.01 (J,g/m3) and
clothing (5.19 (j,g/m3).

Table 3-4. Estimated Chronic Average Daily Formaldehyde Indoor Air Concentrations

According to CEM)

COU Subcategory

Scenario

Environment

CEM Calculated 1-Year
Average Daily
Concentration
(fig/m3)

Construction and building materials covering
large surface areas, including wood articles;
Construction and building materials covering
large surface areas, including paper articles;
metal articles; stone, plaster, cement, glass
and ceramic articles

Building /
Construction
Materials - Wood
Articles:

Hardwood Floors
(Residential)"

Living Room

423.47

Fabric, textile, and leather products not
covered elsewhere

Seat Covers
(Automobile)

Automobile

7.10

Fabric, textile, and leather products not
covered elsewhere

Furniture Seat
Covers (Residential)

Living Room

4.01

Fabric, textile, and leather products not
covered elsewhere

Fabrics: Clothing
(Residential)

Bedroom

5.19

Floor coverings; Foam seating and bedding
products; Cleaning and furniture care
products; Furniture & furnishings including
stone, plaster, cement, glass and ceramic
articles; metal articles; or rubber articles

Furniture &
Furnishings -Wood
Articles: Furniture
(Residential)

Living Room

108.62

Paper products; Plastic and rubber products;
Toys, playground, and sporting equipment

Paper-Based
Wallpaper

Living Room

18.05

" The bolded text are representative scenarios (as described in Section 2.2.2).

Page 43 of 81


-------
Construction and building materials covering
large surface areas, including wood articles;
Construction and building materials covering
large surface areas, including paper articles;
metal articles; stone, plaster, cement, glass and

ceramic articles

Floor coverings; Foam seating and bedding
products; Cleaning and furniture care products;
Furniture & furnishings including stone, plaster,
cement, glass and ceramic articles; metal
articles; or rubber articles

D

o
o

Paper products; Plastic and rubber products; Toys,
playground, and sporting equipment

Fabric, textile, and leather products not covered

elsewhere

Figure 3-3. CEM Estimated 1-Year Average Daily Formaldehyde Indoor Air Concentrations

3.2.1.1 Aggregate Exposure

As previously noted, per the indoor monitoring data, there is robust evidence that people are exposed to
multiple sources of formaldehyde simultaneously. As suggested by the SACC during the peer review of
the Draft Formaldehyde Indoor Air Exposure Assessment, EPA considered realistic scenarios where an
individual may be exposed to multiple TSCA COUs to better characterize indoor air exposures to
formaldehyde. For example, if aggregating all the representative scenarios (bolded text in Table 3-4)
including Building/Construction Materials - Wood Articles: Hardwood Floors (Residential), Fabrics:
Clothing (Residential), Furniture & Furnishings -Wood Articles: Furniture (Residential), and Paper-
Based Wallpaper, the total CEM-modeled indoor formaldehyde concentration may be approximately
555 |ig/m3.

3.2.2 CEM Modeling Discussion

Estimated air concentrations were driven by the emission rate per surface area in the expected room of
use, which is dependent on the emission rates taken from literature and the anticipated surface area of
the product in the assumed room of use. CEM accounts for some dissipation over time via air exchanges
between the room of use and the rest of the home, and between the home and outdoor air. Since the
CEM E5 emission model does not consider a decline in emission rates over time or chemical half-life
(EPA 2019a). COU-specific estimates likely represent formaldehyde air concentrations if new articles
were repeatedly introduced to a home or automobile. It should also be noted that while all age groups
were considered, there were no differences observed in formaldehyde 1-year average concentrations per
room of use across age groups.





A











A



Environment

± Residential

1

1 I . . .

. . 1 . . . .

, ,

101	102

Chronic Average Daily Concentration (|jg/m3)

Page 44 of 81


-------
3.2.2.1 Data Integration

From the COUs identified as significant contributors to the indoor air environment, the CEM modeling
results highlight potential COU-specific contributions of formaldehyde to indoor air were driven
primarily by the reported emission rates in literature along with the expected surface area of the
article(s) in the home. Higher emission rates and surface areas corresponded with higher air
concentrations. Central tendency product-specific emission rates (within a COU category) were used for
the CEM modeling to represent emission rates in the typical American home or automobile. Therefore,
it is conceivable that the estimated formaldehyde air concentrations would be lower if the lowest
emission rates were used or higher if the highest reported emission rates were used.

CEM does not allow the user to adjust the model according to a chemical-specific half-life. This was a
key source of uncertainty in the indoor air analysis of formaldehyde. In general, it is unclear whether the
CEM modeling results are reflective of most indoor air home environments in American residences as
there is uncertainty in the results from these models, especially given the CEM E5 emission condition
only allows for a constant rate of emissions instead of an exponential or power law decay rate (as is
done in IECCU modeling).

Considering the breadth of potential sources of exposure to formaldehyde in indoor settings, evaluating
exposures and associated risks resulting specifically from TSCA COUs in this setting is complex. As
such, the uncertainties in EPA's analysis of formaldehyde exposure in indoor air from TSCA COUs via
CEM should be recognized. Furthermore, assumptions made in CEM modeling created a conservative
exposure scenario, likely resulting in higher modeled concentrations than may typically be found in
indoor air. EPA characterized this uncertainty by comparing modeled concentrations to monitored
concentrations and found modeled concentrations for an individual TSCA COU typically fall within the
range of monitored values, and therefore are not unreasonable. However, most homes have more than
one source contributing to formaldehyde concentrations, and monitored concentrations may represent all
sources of formaldehyde. When multiple individual TSCA COU contributions are added together, the
total exposure to modeled concentrations is generally greater than monitored values. This supports
EPA's recognition that most CEM-modeled concentrations are conservative in nature and may not be
representative of actual exposures over extended periods of time. This recognition is particularly
important when considering off-gassing from various articles (e.g., wooden cabinets, flooring, carpet) as
typical dissipation curves associated with off-gassing show high concentrations immediately following
installation, which typically peak within the first week but fall off significantly to much lower
concentrations to which individuals are exposed over an extended period of time.

Through the CEM modeling, EPA assumes and uses the initial formaldehyde concentration from a
finished article, as reported by relevant SDSs and literature, as the initial concentration to which an
individual is exposed when in their residence. This results in a conservative emission rate and exposure
estimate via modeling since it is unlikely a newly manufactured product will be instantly installed in a
home and individuals will immediately be exposed to those higher concentrations early in their exposure
period. It is more likely that the actual emissions from any given product will be the rate following some
period of storage time where initial off-gassing at those high-rates would have occurred prior to being
installed in a residence. The IECCU modeling considered this lead time by foregoing the use of initial
weight fractions reported in article SDSs and relying on the emission factors reported from chamber
studies of finished articles.

The actual emission rate to which an individual is exposed over an extended period indoors would likely
be significantly lower than the initial measured emission rate following manufacturing of the product.
Nonetheless, the current assessment using CEM can be considered as a screening approach to ensure

Page 45 of 81


-------
potential exposures to formaldehyde via the indoor air pathway are not missed and to provide a
conservative exposure estimate that can be considered for characterizing exposures and associated risks
while recognizing the uncertainty around such estimates.

Although EPA modeled a scenario using measured emission rates from manufactured articles for this
indoor air exposure assessment within CEM, the Agency investigated other modeling approaches {i.e.,
IECCU) and emission rate values available to evaluate prior to finalizing its indoor air formaldehyde
exposure assessment. Ideally, these models and emission rates will consider multiple factors like the
rapid dissipation of off-gassing pollutants, and a more representative actual emission rate (per article-
specific surface areas) from off-gassing to model exposures.

A general residential dissipation curve of formaldehyde over time (in years) is presented in FigureApx
C-l.

3.2.2.2 Automobile Exposures

At the time of the CEM exposure assessment, EPA considered all potential TSCA related sources of
formaldehyde in typical automobile environments. The estimated formaldehyde concentration in an
automobile interior due to automobile seat covers was 7.1 [j,g/m3. By comparison, a study of automobile
formaldehyde concentration in the New York City Metropolitan Area reported that the measured
average concentration of formaldehyde inside automobiles during commutes was approximately 300
[j,g/m3 (Lawrvk and Weisel. 1996; Lawryketai. 1995). The two automobiles used in this study were a
1988 Chevrolet Celebrity and a 1987 Plymouth Horizon (Lawrvk and Weisel. 1996). While this is a
well-executed study, it is relatively dated. It is unclear how this monitoring data represent current
vehicular formaldehyde indoor air concentrations during a typical commute. Regarding the difference
between the modeled seat cover estimates and monitoring values, it is also unclear which sources other
than seat covers may be contributing to typical automobile indoor air exposures.

3.3 Tier 2: IECCU Modeling

This section summarizes the IECCU indoor air exposure assessment results for formaldehyde. A full
description of the modeling methodology can be found in Section 2.3. EPA did not model indoor air
concentrations of commercial buildings as residential homes were considered a more protective indoor
air scenario. The Agency also did not model indoor air concentrations of automobiles using IECCU
since this model is unable to estimate chemical concentrations in such indoor air settings.

3.3.1 IECCU Modeling Results

Modeled 15-minute peak, 3-month, and 1-year average indoor air concentrations for both single item
and aggregate scenarios are shown in Table 3-5 and Figure 3-4. For the TSCA COU covering building
materials with large surface areas, models were generated for both pressed wood cabinets and laminate
flooring. Though, as described in Section 2.2.2, only the highest exposure scenario was selected to
represent a COU. 15-minute peak air concentrations for the high exposure scenario {i.e., high "level"),
were 51 |ig/m3 for pressed wood cabinets and 142 |ig/m3 for laminate flooring. As such, model results
for laminate flooring are reported as the representative article for the building materials COU. Model
results for pressed wood cabinets are not reported in Table 3-5 or shown in Figure 3-4 because it did not
meet the criteria to represent the building materials COU. However, as previously noted pressed wood
cabinets are included in the modeled results for the new construction aggregate scenario. Also, note that
for items modeled in a whole home, a single zone model was used, and Zone 2 concentrations are
therefore recorded as N/A. In all, IECCU-modeled concentrations ranged from 0.00009 to 142 (J,g/m3,
across all individual COUs and exposure scenarios {i.e., "levels").

Page 46 of 81


-------
Table 3-5.15-Minute Peak, Mean 3-Month (Intermediate) and Mean 1-Year (Chronic)
Formaldehyde Concentrations (jig/m3) in Indoor Air for Single Representative Article and
Aggregate Model Scenarios 					

COU(s)

Representative
Scenario"

Level

15-Minute Peak
Air

Concentration
Zone 1 (jig/m3)

3-Month
Average Air
Concentration
Zone 1 (jig/m3)

1-Year Average
Air

Concentration
Zone 1 (jig/m3)

15-Minute Peak
Air

Concentration
Zone 2 (jig/m3)

Construction and
building materials
covering large
surface areas,
including wood
articles; Construction
and building
materials covering
large surface areas,
including paper
articles; metal
articles; stone,
plaster, cement, glass
and ceramic articles

Laminate
Flooring

High

142

27.5

6.1

N/A

Med

22

4.3

1

N/A

Low

1

0.2

0.05

N/A

New

Construction
(Aggregate)b

High

160

31.1

6.87

N/A

Med

64

12.3

2.72

N/A

Low

21

4.1

0.9

N/A

Fabric, textile, and
leather products not
covered elsewhere

Textile

Furniture

Covers

High

18

3.5

0.8

0.3

Med

2

0.4

0.1

0.03

Low

0.007

0.001

0.0003

0.00009

Floor coverings;
Foam seating and
bedding products;
Cleaning and
furniture care
products; Furniture &
furnishings including
stone, plaster,
cement, glass and
ceramic articles;
metal articles; or
rubber articles

Pressed Wood
Furniture

High

26

5

1.1

0.4

Med

7

1.4

0.3

0.1

Low

2

0.3

0.1

0.03

Living Room
Decor Change
(Aggregate)c

High

68

12.9

2.9

0.9

Med

16

3.1

0.7

0.2

Low

3

0.5

0.1

0.04

Paper products;
Plastic and rubber
products; Toys,
playground, and
sporting equipment

Wallpaper



12

2.3

0.5

N/A

" Representative scenarios are defined in Section 2.2.2.

b New construction aggregate scenario includes laminate flooring, pressed wood cabinets, and carpet.
c Decor change aggregate scenario includes pressed wood furniture, indoor furniture (foam components), and indoor
furniture (textile components).

Page 47 of 81


-------
Construction and building materials covering large surface areas,
including wood articles; Construction and building materials
covering large surface areas, including paper articles; metal
articles; stone, plaster, cement, glass and ceramic articles

Fabric, textile, and leather products not covered elsewhere

| Single Item
Aggregate

V

0

A

V

0

Low Exposure Scenario
Medium Exposure Scenario
High Exposure Scenario

A

Floor coverings; Foam seating and bedding products; Cleaning
and furniture care products; Furniture & furnishings including
stone, plaster, cement, glass and ceramic articles; metal articles;
or rubber articles

Paper products; Plastic and rubber products; Toys, playground,
and sporting equipment

v 0

A

25

50

75

100

125

150

175

Figure 3-4. 15-Minute Peak Concentrations (jig/m3) of Formaldehyde in Indoor Air for TSCA COU Representative Article and
Aggregate Models

Page 48 of 81


-------
Except for wallpaper, modeled air concentrations for each representative scenario, as described in
Section 2.2.2, showed a significant range in values. This was driven largely by variability in emissions
factors and estimated surface areas from actual articles in the literature (Table 2-4). However, these
parameters likely exhibit significant variability due to differences in materials, manufacturing practices,
and purchasing preferences. As such, EPA considers it reasonable that the range of estimated
concentrations reflects real-world conditions for each COU assessed.

Figure 3-5 shows air concentrations over the full duration of modeling (10,000 hours) for the high-end
models (see Table 2-4 for high, med, low-end scenarios) and for each representative scenario as
described in Section 2.2.2. The modeled concentrations of formaldehyde in air peaked on the first day
the article was installed in the home. Then, the concentrations in indoor air declined rapidly,
approaching zero |ig/m3 after a period of approximately 3 months.

—	Aggregate-Decor Change — Fabric Furniture Covers — Pressed Wood Furniture

—	Aggregate-New Build — Laminate Flooring — Wallpaper

150

CO

E

D>

3

<
c

O 100
-*—•

03

i—

C



33

CO

E

I—

o

LL

0

1 2 3 4 5 6 7 8 9 10 11 12
Time after Installation in Home (Months)

Figure 3-5. Formaldehyde Concentrations in Indoor Air (jig/m3) for TSCA COU Representative
Article and Aggregate Scenarios over the Course of ~1 Year

(10,000-hour simulation duration)

3.3.2 IECCU Modeling Discussion

3.3.2.1 Residential

3.3.2.1.1 Data Integration	

As previously noted, EPA considered monitoring data in its risk evaluation to characterize real-world
exposures to formaldehyde in indoor air. However, to assess the TSCA COU specific contributions to
indoor air, EPA used modeling tools to estimate formaldehyde exposures.

Page 49 of 81


-------
Mobile homes had generally higher concentrations of formaldehyde in indoor air across all monitoring
studies (Murphy et at.. l , < <1 x < 2008) identified and also had notably higher concentrations of
formaldehyde compared to IECCU modeling results. In addition, according to a 2006 ATSDR study of
96 unoccupied FEMA trailers (ATSDR. 20071 measured formaldehyde concentrations were as high as
4,500 |ig/m3 for some trailers, with an average of 1,280 |ig/m3; and were significantly higher than the
concentration of 124 [j,g/m3 reported by AHHS II for American homes.

However, drawing conclusions about the relationship between the modeled values and measured values
in homes is difficult partly due to the relatively rapid changes in emission rate after placing an article in
the home. Unless indoor air monitoring measurements were intentionally conducted after new building
materials or articles were placed in a home, it is unlikely that monitoring data will capture the highest
period of emissions from these items, and associated peak air concentrations, as displayed in Table 3-5,
Figure 3-4, and Figure 3-5.

The reported AHHS II U.S. indoor air measured concentrations of formaldehyde are expected to reflect
typical long-term formaldehyde concentrations, and generally does not reflect initial peak concentrations
when an article is first introduced into the home. In general, the range of modeled formaldehyde
residential indoor air concentrations were within an order of magnitude of the range of monitoring
values (0.007 to 142 (J,g/m3, across all individual COUs and scenarios, compared to 0.3 to 124 (J,g/m3,
respectively) from the AHHS II residential indoor monitoring study. From a nationally representative
sample of 689 homes, the measured concentration of formaldehyde in American homes ranged from 0.3
to 124 (J,g/m3, with 19.77 [j,g/m3 as the 50th percentile. Per Figure 3-1, most homes had a formaldehyde
air concentration that was 40 [j,g/m3 or less (Table 3-3). Potential factors that might have impacted the
AHHS II and other monitoring results include temperature and humidity (seasonality), and ventilation.
For some articles, (e.g., particle boards), an increase in temperature and humidity may increase
formaldehyde off-gassing rate (Pickrell etai. 1984); and indoor environments with poor ventilation can
lead to higher concentrations of measured formaldehyde in indoor air (EPA. 2016). It should be noted
that AHHS II monitoring results cannot be apportioned according to TSCA COUs. Furthermore, the
AHHS II air sampling was not performed throughout the entire home, and across multiple seasons.

It should also be reiterated that formaldehyde emission rates decrease over time. Generally, it is
expected that after the installation of formaldehyde-bearing materials in a home, there is an initial rise of
formaldehyde concentration, followed by a leveling-off period that may be as brief as 30 days or less,
depending on the article. This is followed by a significant decline of formaldehyde concentration over
time (EPA. 2016; Park and Ikeda. 2006). EPA does not expect, nor assumes (via IECCU modeling), that
newly manufactured articles are added immediately to a home. It is possible that air concentrations
resulting from solid article emissions may remain elevated longer than the IECCU models indicate. For
example, new single-family homes in California are built relatively air-tight, and because the windows
and doors were kept shut during the duration of one residential indoor air study, the indoor-outdoor air
exchange rates were generally low (i.e., 0.2 air exchanges per hour). This resulted in significantly
elevated indoor air concentrations of formaldehyde (Offermann et at.. 2008). Furthermore, if home
ventilation rates are lower than those used in the IECCU models, peak air concentrations would be
somewhat higher than model results and air concentrations would remain elevated above background
levels for a longer duration of time.

Although IECCU incorporates formaldehyde's first-order exponential decay from finished articles,
according to the literature and as described in Section 2.3.4, it does not incorporate indoor sinks that
may capture and re-emit formaldehyde. The model also does not incorporate various forms of barriers
(e.g., lamination, coatings, article thickness, etc.) that may accelerate, delay, and/or prolong

Page 50 of 81


-------
formaldehyde emissions over time, or other factors including moisture content within the article (for
wood), room temperature, and humidity fluctuations, which may differ across housing units, seasons and
regions. Also, in some instances formaldehyde emitted from articles with large surface areas may
contribute significantly to concentrations measured in homes. Though the extent of this contribution is
variable, depending on the article, consumer preferences, room of use, home size and configuration,
ventilation rates, and relevant meteorological parameters {i.e., temperature and humidity). The modeled
concentrations are also likely dependent on potential replacement rates of articles whose data are
currently lacking and could not be integrated into the model nor used to further characterize previously
discussed monitoring studies. Thus, while EPA is confident that the range of indoor air concentrations
for each modeled scenario provides a reasonable estimate of formaldehyde content that may result from
the installation of new solid articles in a home, the modeling results are not intended nor expected to be
predictive of measured concentrations of formaldehyde in indoor air.

Due to the highlighted uncertainties and limitations within and across data types, as shown in Figure 3-5,
the decay of formaldehyde concentrations in indoor air likely differs depending on the articles added to
the home, and modeling tools used. Although IECCU indoor air concentrations were estimated up to a
year, modeled results began approaching 0 [j,g/m3 starting around the third month. Yet measured
concentrations from AHHS II show typical concentrations of nearly 23 [j,g/m3 for homes of various ages.
This highlights a gap between IEECU modeling and AHHS II results. It is possible that the IECCU
model slightly underestimates long term indoor air exposures from TSCA COUs.

After formaldehyde volatilizes rapidly from the top layer of a laminated or relatively thick article, there
is a significant decrease in the off-gassing of formaldehyde. Some of the remaining formaldehyde may
be released slowly over time through a delayed release (or second phase). This delayed release may be
triggered by the decomposition of unstable urea-formaldehyde or lignins present in the wood article as a
result of hydrolysis over an extended period (Docket ID: EPA-HQ-OPPT-2023-0613-0270-A2).
According to He et al. (2019). of the two causes of delayed long-term release, hydrolysis is the primary
cause based on chamber studies of finished articles from approximately 3 to 18 months. This hydrolysis
typically occurs through crevasses located at the core of wood articles (He et al.. ). Therefore, the
decay curves of formaldehyde concentrations for wood articles are driven mainly by volatility upon
installation. Then, as the product ages, hydrolysis becomes the primary cause of further formaldehyde
release (Docket ID: EPA-HQ-QPPT-202:	70-A2Y

Figure 3-6, from He et A > depicts this initial rapid release followed by a gradual decrease of
formaldehyde release measured over time. Though only four samples were measured from this chamber
study, it is interesting to note the similarity of the high-end peak exposures between the IECCU
modeling (Figure 3-5) and the He et al. C experimental study (Figure 3-6). In the long term, though
the IECCU modeled decay curves (Figure 3-5) share aspects of the measured formaldehyde
concentration decay curves, per the above discussion, potentially due to an inability to account for
hydrolysis or a breakdown of urea-formaldehyde, the IECCU models assume a much more rapid
approach to 0 (J,g/m3, which may lead to an underestimation of formaldehyde exposures from wood
articles in the long-term, following peak off-gassing.

Page 51 of 81


-------
160
140

a 120

1	ioo

o
a>

-a

43
CJ

2

IS

o

Uh

k

I.

80
60
40
20
0

HE1
—a—HE2
HE3
—e>--HE4









€>- ©O -®" ©» O. ^ -G-©-©- O -Q.

r

3-mon

T

\ - -Q B- Q- €> -O -Q -B - B- Q- &- -G ¦
	1	'	1	

0.75-yr 1-yr

Time

-SS3B

0.5-yr

1.25-yr 1.5-yr

Figure 3-6. Figure from He et al. (2019 Displaying Formaldehyde Concentration Decay Curves
from Finished Wood Article Specimens Over Time, Using Test Chamber

HE1 and HE2 represent test specimens without exposed cut edges and seams while HE3 and HE4
represent test specimens with a selected proportion of cut edges and seams exposed (Te et al.. 2019).
According to He et al. (2019). for the HE4 specimen, the exposed perimeter cut edges were sealed at day
57 of the chamber testing. It is possible that the IECCU results for wood articles are best represented by
the first portion of the figure from 0 to ~2 months (reading from left to right). IECCU was unable to
consider a biphasic distribution. As such, the model would underestimate concentrations resulting from
the second and longer emission phase.

Although lECCU-estimated air concentrations from residential articles and AHHS II are individually
informative—especially if the modeled home is new or if new articles are introduced to the modeled
home—caution should be exercised when comparing the results from these two data types as there are
key and impactful aspects to each, presented above. Generally, IECCU results provide a snapshot of the
potential relative formaldehyde indoor air contributions from four COUs when they are initially added to
a home, individually or all at once (i.e., the aggregate scenario). On the other hand, AHHS II results
provide the real-world concentrations of formaldehyde in residential indoor air at the time of sampling,
among American households. There are uncertainties and limitations associated with both data types, as
illustrated above, but together they provide a fuller depiction of formaldehyde residential indoor air
exposures, as opposed to individually.

3.3.2.2 Aggregate Exposure

EPA defines aggregate exposure as "the combined exposures from a chemical substance across multiple
routes and across multiple pathways" (40 CFR 702.33). The reported formaldehyde concentrations from
the monitoring data may represent aggregate formaldehyde indoor air concentrations, as presented in the
AHHS II study across U.S. households (JuanTech. 2021). assuming either at least a 3-hour TWA, or the

Page 52 of 81


-------
typical indoor air concentration of formaldehyde in residential environments. An aggregate exposure to
formaldehyde via the COUs assessed may occur in the home in which an individual resides.

IECCU modeling estimated aggregate scenario (TSCA COUs only) indoor air concentrations as high as
160 [j,g/m3 for a new construction scenario and 68 [j,g/m3 for a living room decor change scenario, as
described in Section 2.3.5. Using the AHHS II measured average concentration of formaldehyde (-23
(j,g/m3) as a baseline residential concentration and considering the addition of new laminate flooring
yielding concentrations as high as 142 (j,g/m3, a resident's aggregate exposure may be as high as 165
[j,g/m3. The estimated aggregate exposures suggest potential yet reasonable estimates of total exposures
an individual may receive in indoor air. These aggregate exposures may differ according to conditions of
use considered, an individual's activity patterns, purchase habits, etc.

Page 53 of 81


-------
4 INDOOR AIR EXPOSURE: WEIGHT OF SCIENTIFIC EVIDENCE

EPA used a combination of modeling and monitoring lines of evidence to characterize formaldehyde
indoor air exposures while considering the relative contributions of TSCA COUs to the indoor air
environment. EPA assessed indoor air exposures for four COUs expected to be significant sources of
formaldehyde in indoor air. Monitoring data were further used to characterize indoor air exposures and
provide context for estimated concentrations.

EPA first used CEM as a tier 1 modeling tool to model long-term indoor air exposures to TSCA COUs.
In the Draft Indoor Air Exposure Assessment, peak exposures were not estimated. However, due to
uncertainties associated with CEM's potential overestimation of long-term exposures given CEM's E5
emission condition which only allows for a constant rate of emissions, reliance on initial formaldehyde
weight fraction in article formulation, along with an inability to consider first-order exponential decay
for articles, this assessment also utilized the IECCU as a tier 2 modeling tool to characterize 15-minute
peak, 3-month average, and 1-year average formaldehyde residential indoor air concentrations.

The 1-year results for IECCU are significantly lower than CEM and are likely an underestimate of actual
indoor air concentrations. Thus, there is uncertainty in the precise estimates of long-term concentrations.
A study by He et al. (2019) suggests a biphasic emission profile (rapid emission of formaldehyde when
the product is new followed by a much slower emission of formaldehyde) for laminated wood products
that is not captured in either modeling results. This biphasic emission profile may also occur for other
urea-formaldehyde based products; however, data are not available to confirm. Therefore, CEM was
used along with IECCU to characterize 1-year average indoor air concentrations and provide the
potential range of longer-term formaldehyde exposures.

In addition, monitoring data were incorporated into this assessment to characterize formaldehyde
concentrations most people are exposed to via indoor air. As noted in Section 2.1.1, EPA used publicly
available data acquired through systematic review to characterize indoor air concentrations of
formaldehyde. Unfortunately, these monitoring data do not differentiate between TSCA COUs and other
sources of formaldehyde like cigarette smoke or gas stoves. Monitoring data may also not capture peak
concentrations or fluctuations in indoor air concentrations of formaldehyde but provide insight into
long-term and aggregate exposures to formaldehyde. Monitoring data also cannot fully reflect how and
when (following manufacture) formaldehyde-emitting materials (including imported articles from places
with varying wood standards) are installed. Similarly, monitoring data cannot explain how frequently
these materials are replaced. Lastly, monitoring data may not reflect changes in energy efficiency home
improvements that reduce ventilation (e.g., leaks). Considering limitations in the monitoring data, it is
reasonable to rely on modeled concentrations according to TSCA COUs.

Furthermore, currently available indoor air monitoring data may not represent future potential exposures
in homes. Congress established formaldehyde emission standards for composite wood products which
began to go into effect on June 1, 2018, pursuant to the Formaldehyde Standards for Composite Wood
Products regulations (40 CFR Part 770). which implement TSCA Title VI. Some of the relatively high
monitoring concentrations seen in the currently available data are anticipated to be addressed by
enactment of the composite wood standards for formaldehyde emissions as they are being implemented,
but comprehensive monitoring data are not yet available for confirmation.

Although EPA has uncertainty in either modeling tool's long-term estimates, when used together, CEM
and IECCU modeled estimates provide confidence in the potential range of indoor formaldehyde
concentrations from TSCA COUs. This assumption is supported by the available monitoring data. While
the exposure durations and scenarios do not perfectly align, consideration of all of the available data

Page 54 of 81


-------
suggests model estimated and measured formaldehyde indoor concentrations are in reasonable
agreement. Specifically, model estimated formaldehyde concentrations were within the same order of
magnitude as measured concentration data (e.g., HUD's AHHS II). In addition, the estimated aggregate
exposures fall within the range of available monitoring data. This suggests that TSCA COUs are
contributors to real-world concentrations of formaldehyde in indoor air. Formaldehyde concentrations
are expected to be highest for newly constructed residences with formaldehyde-based materials
(including laminate flooring); and new formaldehyde-based articles added to a residence (including
furniture covers).

Based on consideration of the weight of scientific evidence summarized above, including the strengths
and limitations of the available lines of evidence, EPA has high confidence in the indoor air exposure
assessment of formaldehyde based on the available monitoring data and the estimated short- and long-
term exposures from TSCA COUs. However, the precise concentrations on a long-term basis are
uncertain and are expected to be highly variable.

See Appendix E's Table Apx E-l for a tabular summary of the weight of scientific evidence for the
indoor air exposure assessments.

Page 55 of 81


-------
REFERENCES

Allaire. S; Bourbeai isson. C. (1997). Prevalence of the sick building syndrome symptoms in

office workers before and six months and three years after being exposed to a building with an
improved ventilation system. Occup Environ Med 54: 49-53.
http://dx.doi.on	?em.54.1.49

AT SDR. (1999). Toxicological profile for formaldehyde [ATSDR Tox Profile], Atlanta, GA: U.S.
Department of Health and Human Services, Public Health Service.
http://www.atsdr.cdc.eov/toxprofiles/

ATSDR. (2007). An update and revision of ATSDR's February 2007 health consultation: Formaldehyde
sampling of FEMA temporary-housing trailers (Baton Rouge, Louisiana, September-October,
2006). Atlanta, GA: U.S. Department of Health and Human Services.

https://www.atsdr.cdc.eov/siibstances/formaldehyde/pdfs/revised formaldehyde report. 1007.pd
f

Azuroa. K; Ikeda. K; Kaei. N: Yanagi sawa. H. (2017). Physicochemical risk factors for building-
related symptoms in air-conditioned office buildings: Ambient particles and combined exposure
to indoor air pollutants. Sci Total Environ 616-617: 1649-1655.

http: //dx. doi. or e/10.1016/i. scitotenv. ^

Azuiii i I 'i hiyariu i ;|;eda. K. (2005). The risk management for indoor air pollution caused by

formaldehyde in housing: The historical perspectives on early warnings and actions. Facilities
24: 420-429. http://dx.doi.ore/ 8/02632770610684909
Baez. AP; Padilla. 1\ a i. KM >1 ^'Unoni KM << >t res Mdet. C. (2004). Measurements of carbonyls
in a 13-story building. Environ Sci Pollut Res Int 11: 400-404.
https://heronet.epa.eov/heronet/index.cfm/reference/download/reference id/106705
Beckett. EM; Millei rice. K; Russman. E; Pierce. IS. (2022). Evaluation of volatile organic
compound (VOC) emissions from memory foam mattresses and potential implications for
consumer health risk. Chemosphere 303: 134945.
http://dx.doi.ore/10.1016/i .chemosphere.2022.134945
Blondel. A; Plaisance. H. (2011). Screening of formaldehyde indoor sources and quantification of their
emission using a passive sampler. Build Environ 46: 1284-1291.
https://heronet.epa.eov/heronet/index.cfm/reference/download/reference id/814332
Board. CAR. (2004). Report to the California Legislature: Environmental health conditions in
California's portable classrooms. Sacramento, CA: CalEPA.
https://ww2.arb.ca.gov/sites/default/files/classic/research/apr/reports/13QQ6.pdf
Canada. H. (2005). Proposed residential indoor air quality guidelines for formaldehyde. Ottawa, Canada:

Minister of Health. https://publications.gc.ca/collections/Collection/H128-l-05-432E.pdf
Ceballos. DM; Burr. GA. (2012). Evaluating a persistent nuisance odor in an office building. J Occup

Environ Hyg 9: D1-D6. http://dx.doij 10 IQSQ/15459624.201. > ^ I > I
Chang. T; Wang. JH; Lu. JO; Shen. ZX; Hua»y \ v ^i. J; Xu. HM; Wang. X; R-m i ao. II.

(2019). Evaluation of Indoor Air Pollution during the Decorating Process and Inhalation Health
Risks in Xi'an, China: A Case Study. Aerosol Air Qual Res 19: 854-864.
https://heronet.epa.gov/heronet/index.cfrn/reference/download/reference id/5973739
Chen. W; Mendell. M; Li. Na; Kumaeai. K. (2018). Formaldehyde emissions from seams and cut edges
of laminate flooring: Implications for emission testing protocols and exposure estimation. Build
Environ 143: 652-660.

https://heronet.epa.gov/heronet/index.cfrn/reference/download/reference id/5952378
Dai. W; Zhong. H; Li. L; Cao. J; Huam \ Shen. M; Warn 1 1 >ong. i \ SSH; Ho. KF.
(2018). Characterization and health risk assessment of airborne pollutants in commercial
restaurants in northwestern China: Under a low ventilation condition in wintertime. Sci Total

Page 56 of 81


-------
Environ 633: 308-316.

https://heronet.epa.eov/heronet/index.cfm/reference/download/referer	7

Derbez. M; Wv;	ier. E; Ramalho. O; Riberon. J; Mandin. C. (2017). Indoor air quality in

energy-efficient dwellings: Levels and sources of pollutants. Indoor Air 28: 318-338.
https://heronet.epa.eov/heronet/index.cfm/reference/download/referer
Dingle. P; Tapsell. P; Hu. S. (2000). Reducing formaldehyde exposure in office environments using
plants. Bull Environ Contam Toxicol 64: 302-308. http://dx.doi.ore/10.1007/s001289910044
Dodson. ^ t houseman < \ < ^\;\ v peneler. JD; Shine. IP; Benn-'n 1 U I (2007). Measured and

modeled personal exposures to and risks from volatile organic compounds. Environ Sci Technol
41: 8498-8505. http://dx.doi.ore/10.1021/es(

Dodson. " 1 iesk\ in Col ton. MP; McCaulev. M; Camann. DE; Yau \\ \tltmkiewicz. G; Rudel.
RA. (2017). Chemical exposures in recently renovated low-income housing: Influence of
building materials and occupant activities. Environ Int 109: 114-127.
https://heronet.epa.eov/heronet/index.cfm/reference/download/reference id/5755270
ECHA. (2019). Annex XV restriction report, proposal for a restriction: Formaldehyde and formaldehyde
releasers. Helsinki, Finland: European Union, European Chemicals Agency.
https://echa.eiiropa.eii/documents,	: formaldehyde axvreport en.pdf/2c798a08-

! 19-8180-a3c5a0 >

« « \ \ v (2011). Exposure factors handbook: 2011 edition [EPA Report], (EPA/600/R-090/052F).

Washington, DC: U.S. Environmental Protection Agency, Office of Research and Development,
National Center for Environmental Assessment.

https://nepis.epa. eov/Exe/ZyPIJRL.cei?Dockev=P 100F2QS.txt

(2012). Formaldehyde indoor air model - Pressed wood products, version 2.0 (FIAM-pwp
v2.0): Model user manual and documentation. Washington, DC.

http://www.epa.eov/sites/production/files/2015-09/dociiments/fiam-p-	iser euide.pdf

(2016). Formaldehyde from composite wood products: Exposure assessment for TSCA Title
VI Final Rule. Washington, DC: Risk Assessment Division, Office of Pollution Prevention and
Toxics, Office of Chemical Safety and Pollution Prevention.
https://heronet.epa.eov/heronet/index.cfm/reference/download/referer
EPA. US. (2019a). Consumer Exposure Model (CEM) 2.1 User Guide. (EPA Contract # EP-W-12-010).
Washington, DC.

(2019b). IECCU 1.1 User's Guide. In Simulation Program for Estimating Chemical Emissions
from Sources and Related Changes to Indoor Environmental Concentrations in Buildings with
Conditioned and Unconditioned Zones (IECCU). (EPA Contract # EP-W-12-010). Washington,
DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics.

(2019c). Synthetic turf field recycled tire crumb rubber research under the Federal Research
Action Plan, Final report part 1: Tire crumb rubber characterization volume 1. (EPA/600/R-
19/051.1). Washington, DC: U.S. Environmental Protection Agency, ATSDR, CDC.
https://www.epa.gov/sites/default/files/2Q19-

08/documents/synthetic turf field recycled tire crumb rubber research under the federal res
earch action plan final report p /plume1 .pdf
i t \ I 1 (2020). Risk evaluation for 1-bromopropane (n-Propyl bromide), CASRN: 106-94-5 [EPA

Report], (#740-Rl-8013). Washington, DC: Office of Chemical Safety and Pollution Prevention.

https://www.reeiilations.eov/dociiment/EPA-HQ-OPPT-2019-0235-0085
i t \ I 1 (2021a). About the Exposure Factors Handbook [Website],

https://www.epa.eov/expobox/aboiit-exposiire-factors-handbook

(2021b). Draft systematic review protocol supporting TSCA risk evaluations for chemical
substances, Version 1.0: A generic TSCA systematic review protocol with chemical-specific
methodologies. (EPA Document #EPA-D-20-031). Washington, DC: Office of Chemical Safety

Page 57 of 81


-------
and Pollution Prevention. https://www.regulations.gov/document/< " \ ^ IQ-OPPT-2Q. I Ojl {¦

0005

(2023a). Draft Risk Evaluation for Formaldehyde - Systematic Review Protocol. Washington,
DC: Office of Pollution Prevention and Toxics, Office of Chemical Safety and Pollution
Prevention.

(2023b). Draft Risk Evaluation for Formaldehyde - Systematic Review Supplemental File:
Data Quality Evaluation Information for General Population, Consumer, and Environmental
Exposure. Washington, DC: Office of Pollution Prevention and Toxics, Office of Chemical
Safety and Pollution Prevention.

(2023c). Summarized data of the Building Assessment Survey and Evaluation (BASE) Study
[Website], https://www.epa.gov/indoor-air-qualitv-iaq/summarized-data-building-assessment-
survev-and-evaluation-studv

(2024a). Draft Chemistry, Fate, and Transport Assessment for Formaldehyde. Washington,
DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics.

(2024b). Draft Consumer Exposure Assessment for Formaldehyde. Washington, DC: U.S.
Environmental Protection Agency, Office of Pollution Prevention and Toxics.

(2019). Technical Issues - Formaldehyde Emission Standards for Composite Wood Products.
Washington: Federal Information & News Dispatch Inc.
https://search.proquest.com/docview/2276861950?accountid=171501
Gam mage. RB; Hawthorne. AR. (1985). Current status of measurement techniques and concentrations
of formaldehyde in residences. In V Turoski (Ed.), Advances in Chemistry (pp. 117-130).
Washington, DC: ACS Publications, http://dx.doi.org/10.1021 /ba-198	309

Gilbert. NL; Gauvii nay. M; Heroux. ME; Dupuis. G; Legris. M; Chan. CC: Dietz. RN: Levesque.

(2006). Housing characteristics and indoor concentrations of nitrogen dioxide and
formaldehyde in Quebec City, Canada. Environ Res 102: 1-8.
https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/95801
GUben 's 1 uay. U 1 Kt /id Miller. J; Judel v i han. CC: Dales. RE. (2005). Levels and determinants
of formaldehyde, acetaldehyde, and acrolein in residential indoor air in Prince Edward Island,
Canada. Environ Res 99: 11-17. http://dx.doi.or	i .envres.2004.09.009

Girman. JR: Apte. I	or. GW: Allen. 1R: Hollowell. CD. (1982). Pollutant emission rates from

indoor combustion appliances and sidestream cigarette smoke. Environ Int 8: 213-221.
https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/34608
Glas. H, ] cvin. JO: Stenbei^ r% Stenlund. H: Sunesson \1 (2004). Variability of personal chemical
exposure in eight office buildings in Sweden. J Expo Anal Environ Epidemiol 14: S49-S57.
http://dx.doi.oo	ii.iea.7500358

Glas. B: Stenben <1' Stenlund. H: Sunesson \1 (2014). Exposure to formaldehyde, nitrogen dioxide,
ozone, and terpenes among office workers and associations with reported symptoms. Int Arch
Occup Environ Health 88: 613-622. http://dx.doi.org/10.1007/sQ0420 )85~y
Haghighat. F: Donnini. G. (1999). Impact of psycho-social factors on perception of the indoor air
environment studies in 12 office buildings. Build Environ 34: 479-503.
http://dx.doi.on	50360-1323(98100034-1

Hanazato. M: Suzuki. N: Koga. C: Nakaoka. H: Mo (2018). A study of design for reduction and

monitoring of volatile organic compounds in indoor air: The case of a commercial bank in Japan.
Journal of Asian Architecture and Building Engineering 17: 573-579.
http: //dx. doi. or g/10.313 0/i aab 2
He. Z: Xiong. J: Kumagai. K: Chen. W. (2019). An improved mechanism-based model for predicting
the long-term formaldehyde emissions from composite wood products with exposed edges and
seams. Environ Int 132: 105086. http://dx.doi.org/10.1016/i.envint.2019.105086

Page 58 of 81


-------
Hedg	:son. WA; Rubin. G. (1995). Individual and occupational correlates of the sick building

syndrome. Indoor Air 5: 10-21. http://dx.doi.ore	>0-0668.1995.tO 1 -3-00003,x

Ho. S; Cheng. Ya; Bai. Yi: Ho. K: Dai. W; (	iang. Yu: Ip. H; Deng. W; Guo. We.

(2016). Risk Assessment of Indoor Formaldehyde and Other Carbonyls in Campus Environments
in Northwestern China. Aerosol Air Qual Res 16: 1967-1980.
https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/3476957
Ho. SSH; Ip. HI' iioi'l V I Pi 1 Ut W i i i i ban. CS; Ho 1 Cs (2014). Evaluation of
hazardous airborne carbonyls on a university campus in southern China. J Air Waste Manag
Assoc 64: 903-916.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/2535460
Hodgson. AT; Rude! U d 1 * • handra. S. (2000). Volatile organic compound concentrations and
emission rates in new manufactured and site-built houses. Indoor Air 10: 178-192.
https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/2994962
Hodgson. AT; Shendell. DG: Fisk	MG. (2004). Comparison of predicted and derived

measures of volatile organic compounds inside four new relocatable classrooms. Indoor Air 14:
135-144. http://dx.doi.org/10J 1 i 1/11600-0668.2004.00 ; I ¦> \

Hong. W; Meng. MW; Xie. JL; Gao. DP: Zem \ Cs \i Ui i ii^n. CO; Huam ' \ lion. ZM. (2017).
Investigation of the Pollution Level and Affecting Factors of Formaldehyde in Typical Public
Places in Guangxi, China. Aerosol Air Qual Res 17: 2816-2828.
https://heronet.epa.gov/heronet/index.cfm/reference/download/referer
Huariv 1 < >ian. Hu; Deng. S; Gu<» < It \ iu<< \\ \\t^ N (2018). Urban residential indoor volatile
organic compounds in summer, Beijing: Profile, concentration and source characterization.

Atmos Environ 188: 1-11.

https://heronet.epa.gov/heronet/index.cfm/reference/download/referer
Huanv \ Mi i W.iiig. L; Wjiiv	\ Ka W 1^ ' C; Cao. J; Ho. SSH. (2019). Evaluation

and characterization of volatile air toxics indoors in a heavy polluted city of northwestern China
in wintertime. Sci Total Environ 662: 470-480.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/.c S I i 3
IPCS. (2002). Concise International Chemical Assessment Document 40: Formaldehyde. Geneva,
Switzerland: World Health Organization.
https://inchem.org/documents/cicads/cicads/cicad40.htm
Jung. C; Mahmoud. NSA. (2022). Extracting the critical points of formaldehyde (HCHO) emission
model in hot desert climate. Air, Soil and Water Research 15: 1-15.
http://dx.doi.on	362212211Q5Q82

Kaiser. EA; Sylvair (1994). Health hazard evaluation report No. HETA-94-0023-2473, Veterans
Administration, White River Junction, Vermont (pp. 94-0023). (HETA-94-0023-2473). National
Institute for Occupational Safety and Health.

Kaunelie	sauskas. T; Krugly. E; Stasiulaitiene. I; Ciuzas. D; Seduikyte. L; Martuzevicius. D

(2016). Indoor air quality in low energy residential buildings in Lithuania. Build Environ 108:
63-72. https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/3458632
Kelly. TJ; Smith. PL; Satola. J. (1999). Emission rates of formaldehyde from materials and consumer
products found in California homes. Environ Sci Technol 33: 81-88.
https://heronet.epa.gov/heronet/index.cfm/reference/download/referer	58

Klepeis. NE; Nelson \\ ¦>s" U I Vinson. JP; Tsang. AM; Switzer. P; Behar J\ . Hern. SC;

Engelmann. WH. (2001). The National Human Activity Pattern Survey (NHAPS): A resource
for assessing exposure to environmental pollutants. J Expo Anal Environ Epidemiol 11: 231-252.
http://dx.doi.on 10 10 'A sije-i ->00 i ^

Lawrvk. NJ; Lie	sel. CP. (1995). Exposure to volatile organic compounds in the passenger

compartment of automobiles during periods of normal and malfunctioning operation. J Expo

Page 59 of 81


-------
Anal Environ Epidemiol 5: 511-531.

https://heronet.epa.eov/heronet/index.cfm/reference/download/reference id/S
Lawrvk. NJ: Weisel. CP. (1996). Concentrations of volatile organic compounds in the passenger
compartments of automobiles. Environ Sci Technol 30: 810-816.
https://heronet.epa.eov/heronet/index.cfm/reference/download/reference id/79703
LBNL. (2008). Aldehyde and other volatile organic chemical emissions in four FEMA temporary
housing units - final report. (LBNL-254E). Berkley, CA.
https://www.cdc.gov/air/trailerstudv/pdfs/lbnl-254e.pdf
Lee. K: Choi. JH; Lee. S: Park. HI; 01. U i na. GB; Lee. WS; Son. BS. (20181 Indoor levels of

volatile organic compounds and formaldehyde from emission sources at elderly care centers in
Korea. PLoS ONE 13: eO 197495.

https://heronet.epa.gov/heronet/index.cfiTi/reference/download/reference id/6200076
Liu. 0; Liu. Y; Zhang. M. (2014). Source Apportionment of Personal Exposure to Carbonyl Compounds
and BTEX at Homes in Beijing, China. Aerosol Air Qual Res 14: 330-337.
https://heronet.epa.gov/heronet/index.cfrn/reference/download/reference id/2394125
Liu. W: Zham < ihang. L; Turpi n r% < \\ ^'isel. CP: Morandi. MT; Stock. TH; Colome. S: Korn. LR.

(2006). Estimating contributions of indoor and outdoor sources to indoor carbonyl concentrations
in three urban areas of the United States. Atmos Environ 40: 2202-2214.
http://dx.doi.off	.atmosenv.2005.12.005

L /lason. MA: Guo. Z: Krebs. KA: Roache. NF. (2015). Source emission and model evaluation of
formaldehyde from composite and solid wood furniture in a full-scale chamber. Atmos Environ
122: 561-568. http://dx.doi.org/10 J 016/i.atmosenv._01 0° 052
Luke i ^ lotti. TL: Franklin. DE: Bun \ (2014). Indoor Environmental and Air Quality

Characteristics, Building-Related Health Symptoms, and Worker Productivity in a Federal
Government Building Complex. Arch Environ Occup Health 71:0.
http://dx.doi.on BO/19338244.2014.965246
Maddalena. R: Russell. M: Sullivan. DP: Apte. MG. (2009). Formaldehyde and other volatile organic

chemical emissions in four FEMA temporary housing units. Environ Sci Technol 43: 5626-5632.
https://heronet.epa.gov/heronet/index.cfrn/reference/download/reference \<3J1 2
Matthews. TG: Hawthorne. AR: Daffron. CR: C	Schrimsher. JM. (1984).

Formaldehyde surface emission monitor. Anal Chem 56: 448-454.
https://heronet.epa.gov/heronet/index.cfrn/reference/download/referer
Menzies. 1): Tamblyn. RM: Nunet i	< < .tmbh ii K i (1996). Exposure to varying levels of

contaminants and symptoms among workers in two office buildings. Am J Public Health 86:
1629-1633. http://dx.doi.ore/10.2105/aiph.8C 29
Murphy. MW: Lando. IF: Kieszak. SM: Sutter. ME: Noonan. GP: Brunkard. JM: McGeehin. MA.

(2013). Formaldehyde levels in FEMA-supplied travel trailers, park models, and mobile homes
in Louisiana and Mississippi. Indoor Air 23: 134-141. http://dx.doi.ore/10.1 1 1 1 /j. 1600-
0668.2012.00800.x

Page. E: Couch. J. (2014). Evaluation of employee health concern and suspected contamination at an
office complex. (Report No. 2010-0061-3206). Washington, DC: National Institute for
Occupational Safety and Health.

Park. JS: Ikeda. K. (2006). Variations of formaldehyde and VOC levels during 3 years in new and older

homes. Indoor Air 16: 129-135. http://dx.doi.off	^.2005.00408.x

Persil /lusser. A: Emmerich. SJ. (2010). Modeled infiltration rate distributions for U.S. housing.

Indoor Air 20: 473-485. http://dx.doi.	'0-0668.2010.00669.x

Pickn'U 1 \ > otts. LC: Mokln r%\ i anapilly. GM: Hobbs. CH. (1984). Formaldehyde release from
selected consumer products: influence of chamber loading, multiple products, relative humidity,

Page 60 of 81


-------
and temperature. Environ Sci Technol 18: 682-686.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/22466
Pickrell. J A; Mokle	abbs. CH. (1983). Formaldehyde release rate coefficients from

selected consumer products. Environ Sci Technol 17: 753-757.

http://dx.doi.on 10 10 J I esQQl 18.iO IJ
Poulhet. G; Dusanter. S: Crunaire. S: Locoge. N: Gaudion. V: Merlen. C: Kaluzny. P; Coddeville. P.
(2014). Investigation of formaldehyde sources in French schools using a passive flux sampler.
Build Environ 71: 111-120.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/2.k-N8
OuanTech. (2021). American Healthy Homes Survey, final report: Data documentation. (AHHS11).
Washington, DC: U.S. Department of Housing and Urban Development.
https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/9416854
Rovira. J; Roig. N: Nadal. M: Schuhmacher. M: Domingo. JL. (2016). Human health risks of

formaldehyde indoor levels: An issue of concern. J Environ Sci Health A Tox Hazard Subst
Environ Eng 51: 357-363.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/3362355
Salthammer. T. (2019). Formaldehyde sources, formaldehyde concentrations and air exchange rates in
European housings. Build Environ 150: 219-232.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/5960808
Salthammer. T; Mentese. S: Marutzkv. R (2010). Formaldehyde in the indoor environment. Chem Rev

110: 2536-2572. http://dx.doi.ni y 10 tO:t >T800399g
Sax. SN; Bennett 1>M. Chillrud. SN; Kinney. PL; Spengler JP (2004). Differences in source emission
rates of volatile organic compounds in inner-city residences of New York City and Los Angeles.
J Expo Anal Environ Epidemiol 14: S95-S109. http://dx.doi.org/10.1038/si.iea.7500364
Sheehan. P; Singhal. A: Bogen. KT; Macintosh. D; Kalmes. RM; McCarthy. J. (2017). Potential

exposure and cancer risk from formaldehyde emissions from installed Chinese manufactured
laminate flooring. Risk Anal 38: 1 128-1 142. http://dx.doi.org/	ia. 12926

Shendell. DG: Winer. AM; Stock. TH; Zhang. L; Zhang. II; Maberti. S: Colome. SD. (2004). Air

concentrations of VOCs in portable and traditional classrooms: results of a pilot study in Los
Angeles County. J Expo Anal Environ Epidemiol 14: 44-59.

https://heronet.epa.gov/heronet/index.cfm/reference/download/referer	5334

Shin. SH; Jo. WK. (2012). Volatile organic compound concentrations, emission rates, and source

apportionment in newly-built apartments at pre-occupancy stage. Chemosphere 89: 569-578.

https://heronet.epa.gov/heronet/index.cfm/reference/download/referer	7932

Shinohara. N; Tokumura. M; Kazama. M; Yonemoto. Y; Yoshioka. M; Kagi. N; Hasegawa. K; Yoshino.
H; Yanagi. U. (2014). Indoor air quality and thermal comfort in temporary houses occupied after
the Great East Japan Earthquake. Indoor Air 24: 425-437.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/2231518
Sundell. J; Anderss>	ersson. K; Lindvall. T. (1993). Volatile organic compounds in ventilating

air in buildings at different sampling points in the buildings and their relationship with the
prevalence of occupant symptoms. Indoor Air 3: 82-93. http://dx.doij	'h

0668	Ix

Weng. M; Z	g. K; Chen. S. (2009). Levels and health risks of carbonyl compounds in selected

public places in Hangzhou, China. J Hazard Mater 164: 700-706.

https://heronet.epa.gov/heronet/index.cfm/reference/download/reference id/2598599
Westat. (1987). Household solvent products: A national usage survey [EPA Report], (EPA-OTS 560/5-
87-005). Washington, DC: Office of Toxic Substances, Office of Pesticides and Toxic
Substances. https://nepis.epa.gov/Exe/ZyPIJRL.cgi?Dockey=Pl 00754Q.txt

Page 61 of 81


-------
WHO. (2010). WHO guidelines for indoor air quality: Selected pollutants. Copenhagen, Denmark:
WHO Regional Office for Europe. https://www.who.int/publications/i/item/9789289002134
Yu. C: Cam (1998). A review of the emission of VOCs from polymeric materials used in buildings
[Review], Build Environ 33: 357-374. http://dx.doi.org/10 J 016/S0360 I >(97100055-3

Page 62 of 81


-------
APPENDICES

Appendix A SYSTEMATIC REVIEW PRIORITIZATION FOR
FORMALDEHYDE DATA

As noted in Section 2.1.1, EPA applied a fit-for-purpose approach to its systematic review of indoor air
data for the formaldehyde exposure assessment. This appendix serves as a supplement to the discussion
provided in Section 2.1.1.

Impacts of SR Approach

The extracted data provides a high-level of confidence in the supporting data that is available for
formaldehyde's exposure analysis, while improving the efficiency of the systematic review of
formaldehyde exposure studies and data. This required the reassignment of EPA and contractors to the
formaldehyde systematic review project, as necessary. This approach facilitated the ability to meet the
necessary deadlines to complete the formaldehyde exposure assessments.

Administrative Actions Involved

EPA and contractors assigned appropriate staff to support the proposed approach to review and extract
formaldehyde data of interest. As directed, for the review of formaldehyde data, EPA and the contractor
prioritized the evaluation and extraction of COU-specific air concentration and emission rate (and other
supporting exposure modeling parameters) data.

A.l Formaldehyde Data Needs

Within the Exposure study pool are six key study types: monitoring, experimental, modeling, completed
assessment, database, and survey.

•	Monitoring: The Formaldehyde Assessment Team determined that measured indoor and ambient
air data associated with formaldehyde COUs from the monitoring study type are most relevant to
the formaldehyde exposure assessment. This is because the primary media of exposure for
formaldehyde is air. Some monitoring studies contain air concentration data that may be used to
compare with formaldehyde exposure modeling results. In addition, modeling parameters such as
room ventilation rates, may also be useful for the refinement of models such as the CEM or the
execution of higher tier models like the IECCU Model. This monitoring data has been identified
as the top priority for formaldehyde. This data has been identified as important to extract.

•	Completed Assessments: Completed assessments may contain completed risk evaluations of
formaldehyde, this study type can be informational and may be referred to for contextual
information (e.g., methodologies, conclusions, and other information). Some completed
assessment studies contain modeling parameters which may be used for the formaldehyde
exposure analysis—namely, product-specific formaldehyde emission rates (and room ventilation
rates, if available) useful in CEM modeling refinements or higher tier models like the IECCU
model. Under the current systematic review protocol for Exposure, completed assessments are
extracted as monitoring or modeling studies. Completed assessments typically make use of
secondary data that are not extracted for any study type. However, if completed assessments
have been deemed to use primary monitoring data that are COU-specific, extract this data.
However, do not extract any other data for this study type as it is not a critical need for the
formaldehyde exposure assessment.

•	Databases: Databases may provide quantitative or supplementary information often useful for
exposure analyses. These may include datasets that contain air or water concentration data (e.g.,

Page 63 of 81


-------
monitoring data) such as the Water Quality Portal (WQP). Data from such source streams may
be referenced or potentially used for comparison to EPA modeled concentrations in its
evaluation of formaldehyde exposures. Key datasets of need including the Toxics Release
Inventory, Discharge Monitoring Report (which contain data from the WQP), and National
Emissions Inventory and other datasets which provide direct inputs to EPA modeling efforts for
formaldehyde have already been extracted and provided by ECRAD engineers per the Draft
Systematic Review Protocol (EPA. 2 ). Thus, there is currently no need for any other
datasets for the formaldehyde exposure assessment. Relevant data evaluation, QC, and extraction
for databases which may contain monitoring data relevant to the ambient air, indoor air, and
water pathways relevant to formaldehyde COUs has been completed.

•	Experimental: Modeling parameters typically found in experimental studies such as permeability
coefficients, absorption fractions, were identified through systematic review for formaldehyde.
However, COU-specific emission rates, room ventilation rates and others, via chamber studies,
for instance, are typically found in experimental study types. Such modeling parameters are
useful in CEM modeling refinements or higher tier models like the IECCU Model. This
experimental data has been identified as the top priority for formaldehyde and such data has been
extracted as needed, to support the formaldehyde exposure assessment.

•	Modeling: Similar to experimental studies, modeling studies are needed for the formaldehyde
risk evaluation. Because such COU-specific-modeling parameters (e.g., emission rates) typically
found in these study types have been identified as essential to the refinement of CEM modeling
of consumer products and articles or the execution of the IECCU Model for the formaldehyde
exposure assessment. This modeling data has been identified as a top priority for formaldehyde
and such data has been extracted as appropriate, to support the formaldehyde exposure
assessment.

•	Survey: No survey data specific to formaldehyde were identified.

A.2 Boolean Search Terms

The following is a list of search terms derived from the formaldehyde TSCA COUs presented in the
Final Scope o f the Risk Evaluation for Formaldehyde; CASRN 50-00-0 (	2020):

Fertilizer, paint, vinyl wallpaper, fiber glass wallpaper, fiberglass, latex paint, glue, building, wood,
hardwood floor, furniture, pressed wood products, particle board, plywood, bare urea-formaldehyde
wood product, coated urea-formaldehyde wood product, bare phenol-formaldehyde wood product,
adhesive, caulk, sealant, vinyl covering, concrete, cement, plaster, PVC foam wallpaper, PVC wall
covering, vapor barriers (bituminous tar), drain cleaner, toilet cleaner, multi-purpose cleaner, cleaner,
stain remover, waterproofing agent, leather tanning, electronic, electronic appliance, furniture cover, car
seat cover, tablecloth, textile wall, acoustic partitions, office chair, chair, textile, clothing, new clothing,
fabric, permanent press fabric, varnish, floor finishes, floor coverings, decorative laminates,
commercially applied urea-formaldehyde floor finish, foam insulation, insulation products, insulation,
mineral wool insulation batt, glass wool fibrous insulation, insulant, PVC, liquid fuel, motor oil, oil,
hardwood floor, furniture, chair, sofa, ink, toner, laundry detergent, dishwashing soap, soap, hand soap,
liquid soap, liquid hand soap, lubricant, grease, paper, diaper, wipe, newspaper, magazine, paper towel,
paper plates, paper cups, paper grocery bag, glues/adhesives (already noted above), fingernail hardener,
photographic supplies, liquid photographic processing solutions, photographic processing solutions,
photographic solutions, plastic, rubber, flooring, carpet, rubber mats, vinyl tiles, soft plastic flooring,
cork floor tiles, plastic laminated board, black rubber trim, jointing, baby bottle nipple, pacifier, toy, car
wax, polish, foam block, foam, tent, fish tank, water treatment product, drinking water treatment

Page 64 of 81


-------
product, embalming, taxidermy [and] air, indoor air, ambient air, air pollution, air release, emission,
emission rate, emission flux, flux, inhalation, atmosphere, fume, fugitive, gas, release, release rate.

The following section provides a summary of select extracted data for additional context to the
formaldehyde indoor air exposure assessment.

A.3 Indoor Air Data Extracted from Systematic Review for Consideration
into the Indoor Air Exposure Assessment	

A.3.1 Indoor Air (ng/m3) - Vapor/Gas Fraction

Measured concentrations of FDH in Indoor Air with unit of ng/m3, extracted from 21 sources, are
summarized in FigureApx A-l and supplemental information is provided in Table Apx A-l. Overall,
concentrations ranged from not detected to 928,000 ng/m3 from 9,007 samples collected between 1998
and 2017 in 9 countries, CA, CN, ES, FR, JP, KR, LT, MX and US. Location types were categorized as
General Population (Background) and Near Facility (Highly Exposed). Reported detection frequency
ranged from 0.65 to 1.

US Vapor/Gas

NonUS Vapor/Gas

5755270 - Dodson et al., 2017 - US
2591662 - Maddalena et al., 2009 - US
1986334 - Shendell et al., 2004 - US

5431563 - Huang et al., 2019 - CN
5973739 - Chang et al., 2019 - CN
4460157 - Hong et al., 2017 - CN
3362355 - Rovira et al., 2016 - ES
3458632 - Kauneliene et al., 2016 - LT
4166147 - Derbez et al., 2017 - FR
4914425 - Huang et al., 2018 - CN
1787932 - Shin and Jo, 2012 - KR
2231518 - Shinohara et al., 2014 - JP
4641587 - Dai et al., 2018 - CN
2331688 - Poulhet et al., 2014 - FR
3476957 - Ho et al., 2016 - CN
814332 - Blondel and Plaisance, 2011 - FR
2394125 - Liu et al., 2014 - CN
2598599 - Weng et al., 2009 - CN
6200076 - Lee et al., 2018 - KR
95801 - Gilbert et al., 2006 - CA
106705 - Baez et al., 2004 - MX

0.1

| I General Population (Background)

Near Facility (Highly Exposed)

V Lognormal Distribution (CT and 90th percentile)
A Normal Distribution (CT and 90th percentile)

W

OA



~ A

V V

Dv

A A

100	1000	10A4

Concentration (ng/m3)

10A5

10A6

10*7

Figure Apx A-l. Concentrations of FDH (ng/m3) in the Vapor/Gas Fraction of Indoor Air from
1998 to 2017

Page 65 of 81


-------
TableApx A-l. Summary of Peer-Reviewed Literature that Measured FDH (ng/m3) Levels in the
Vapor/Gas Fraction of Indoor Air					

Citation

Country

Location Type

Sampling
Years

Sample Size
(Frequency of
Detection)

Detection
Limit
(ng/m3)

Overall Quality
Level

Dodson et

US

General

2013-2014

34(1)

0.056

High

al. (2017)



Population
(Background)









Maddalena

US

General

2007

8(1)

600

Medium

et al.
(2009)



Population
(Background)









Shendell et

us

General

2000-2001

100 (1)

10

High

al. (2004)



Population
(Background)









Huans et
al. (2019)

CN

General

Population

(Background)

2016-2017

88 (N/R)

N/R

High

Chang et

CN

General

2014-2017

176 (0.65)

100000

High

al. (2019)



Population
(Background)









Hons et al.

CN

General

2014-2015

2284(1)

N/R

High

(2017)



Population
(Background)









Rovira et

ES

General

2014

30(1)

200

High

al. (2016)



Population
(Background)









Kauneliene

LT

General

2014

11(1)

100

High

et al.
(2016)



Population
(Background)









Derbez et

FR

General

2013-2014

130 (0.89)

N/R

High

al. (2017)



Population
(Background)









Huans et
al. (2018)

CN

General

Population

(Background)

2013

27(1)

420

High

Shin and Jo

KR

General

2012

107 (0.96)

10

High

(2012)



Population
(Background)









Shinohara

JP

Near Facility

2012

128 (1)

N/R

High

et al.
(2014)



(Highly Exposed)









Dai et al.

CN

General

2011-2012

15(1)

N/R

High

(2018)



Population
(Background)









Poulhet et

FR

General

2011

8(1)

30

High

al. (2014)



Population
(Background)









Page 66 of 81


-------
Citation

Country

Location Type

Sampling
Years

Sample Size
(Frequency of
Detection)

Detection
Limit
(ng/m3)

Overall Quality
Level

Ho et al.
(2016)

CN

General

Population

(Background)

2010-2011

152 (N/R)

N/R

High

Blondel
and

Plaisance
(2011)

FR

General

Population

(Background)

2009-2010

48(1)

N/R

High

Liu et al.
(2014)

CN

General

Population

(Background)

2009

510 (N/R)

10

High

Wens et al.
(2009)

CN

General

Population

(Background)

2006-2007

263 (N/R)

150

High

Lee et al.
(2018)

KR

General

Population

(Background)

2007

4702 (1)

1750

High

Gilbert et
al. (2006)

CA

General

Population

(Background)

2005

96(1)

1000

High

Baez et al.
(2004)

MX

General

Population

(Background)

1998

90(1)

660

High

Abbreviations: N/R, Not reported

A.3.2 Indoor Air (ppb) - Vapor/Gas Fraction	

Measured concentrations of FDH in Indoor Air with unit of ppb, extracted from 1 source, are
summarized in FigureApx A-2 and supplemental information is provided in Table Apx A-2. Overall,
concentrations ranged from 4.67 to 107 ppb from 840 samples collected between 2012 and 2013 in 1
country, CN. Location types were categorized as General Population (Background). Reported detection
frequency was not reported.

NonUS Vapor/Gas

2535460 - Ho et al., 2014 - CN





General Population (Background)
A Normal Distribution (CT and 90th percentile)





0.1

10

1000

10A5 10A7 10*9
Concentration (ppb)

10A11

Figure Apx A-2. Concentrations of FDH (ppb) in the Vapor/Gas Fraction of Indoor Air in
General Population (Background) Locations from 2012 to 2013

Page 67 of 81


-------
TableApx A-2. Summary of Peer-Reviewed Literature that Measured FDH (ppb) Levels in the
Vapor/Gas Fraction of Indoor Air					

Citation

Country

Location Type

Sampling
Years

Sample Size
(Frequency of
Detection)

Detection
Limit (ppb)

Overall Quality
Level

Ho et al.
(2014)

CN

General
Population
(Background)

2012-2013

840 (N/R)

N/R

High

Abbreviations: N/R, Not reported

Page 68 of 81


-------
Appendix B AHHS II SUMMARY OF DATA COLLECTION
METHODOLOGY

To collect the data at each dwelling unit, a two-person team consisting of an interviewer and a
technician was used. AHHS II data were captured using three form sets and a tablet survey: a
Recruitment Questionnaire Form Set; a Resident Questionnaire Form Set and tablet Resident
Questionnaire; and a Technician Form Set [Information Collection Review (ICR) Reference No:
201912-2539-0011. The Recruitment Questionnaire was used by the interviewer to determine whether
the dwelling unit could be recruited into the survey. Once recruited, the interviewer used the tablet
survey, supplemented by the Resident Questionnaire Form Set, to collect data on the unit. The
Technician Form Set was used to collect data such as lead and formaldehyde. All data collected on paper
forms were double keyed, reviewed, and entered into the datasets (OuanTech. 2021).

Residential indoor air samples of formaldehyde were collected in absorption tubes within SGS Galson
air sampling pumps, in a frequently used location (commonly the living room). Samples ranged from 1
to 15 L of air at 0.03 to 1.5 L/minute. The air pump was run throughout the data collection home visit.
Sampling time was not provided, and it likely varied between residences. However, the environmental
sampling in AHHS II, while different in some respects from that of AHHS I, was expected to require a
similar amount of time based on the AHHS II ICR |"ICR Reference No: 201912-2539-0011. Per sampling
times reported in AHHS I, the targeted sampling time for AHHS II was approximately 3.5 hours
(OuanTech. 2021). As such, it may be reasonable to expect that the air sampling pump was typically on
for 3.5 hours. This means that formaldehyde air monitoring air concentrations from the AHHS II were at
least 3-hour TWAs. Formaldehyde air samples were then frozen and sent directly to SGS Galson, the
provider of the sampling pumps, for analysis. Air samples were analyzed using modified NIOSH 2016
(HPLC - UV detection). The detection limit for formaldehyde air concentrations was 0.15 |ig/m3 for 3-
hour sample at 1.5 L/min (which was at or near the maximum capability of the air sampling pump)
(Table Apx B-l). Detailed study methodology and results from the AHHS II are published in a series of
reports available from the HUD Office of Healthy Homes and Lead Hazard Control (OuanTech. 2021).

Table Apx B-l. Summary of Environmental Sampling and Analytical Method

Data Element

Description

ID17

T1

Information Captured or Target Analyte

Formaldehyde in air

Data Collection Method or Sampling Media

Absorption tube

Tests or Samples per Dwelling Unit

1 plus 1 blank/primary sampling unit (PSU)

Special Handling Requirements

Frozen after collection

Maximum Media Count

956

Sample Preparation

None

Analytical Method

Modified NIOSH 2016 (HPLC - UV detection)

Detection Limits

0.15 (ig/m3 for 3-hour sample at 1.5 LPM

Notes

Count includes 1 spiked QC/PSU

"Identifies the protocol containing detailed instructions for the tests or sample collection (OuanTech. 2021).

Page 69 of 81


-------
Appendix C FORMALDEHYDE RESIDENTIAL INDOOR AIR
HALF-LIFE

Residential indoor air formaldehyde concentrations are generally expected to decrease over time
following a first-order exponential process as the reservoir of formaldehyde from articles is depleted
over time (EPA. 2016). This means that a new home with new formaldehyde-based articles (e.g.,
hardwood floors, furniture, furniture covers, etc.) is expected to have a high initial contribution of
formaldehyde off-gassing to the indoor air environment. This is followed by a gradual decrease in
formaldehyde off-gassing as formaldehyde sources are gradually depleted over time; then, a tapering off
effect over an extended period, if no new formaldehyde-based articles are added to the home (EPA.
2016V

In a study of newer mobile homes (Gammaee and Hawthoi !5), there were significantly higher
measured mean concentrations of formaldehyde compared to older mobile homes—1,032 |ig |ig/m3 and
308 |ig/m3 respectively (Gammaee and Hawthoi !5). These highest reported concentrations are
slightly higher than the aggregated high-end modeled concentrations from TSCA COUs representing
new articles added to a home. It should also be noted that in addition to new materials added to a home,
other activities that may affect indoor concentrations of formaldehyde include ripping out drywall,
fixtures, and using various sources of combustion indoors (e.g., wood burning fireplace).

According to the 2016 Formaldehyde Exposure Assessment Report TSCA Title VI Final Rule (J
2016). the half-life of formaldehyde in indoor air (i.e., the amount of time for formaldehyde
concentrations to decrease by half) is expected to be approximately between 1.5 and 3 years. From a
cross-sectional study of homes with varying ages, authors noted that, if new formaldehyde-emitting
articles were being added over time within the homes assessed, an estimated half-life would be close to
2.92 years. The latter is assumed to be an upper bound estimate for formaldehyde's half-life in
residential indoor air. Instead, according to an analysis of various chamber studies of pressed wood
articles as they aged, the authors expect that the residential indoor air half-life of formaldehyde should
be approximately 1.5 years in most cases (	).

In newer homes built after 1990, due to improved insulation and relatively less air circulation in certain
homes, formaldehyde indoor air concentrations may persist longer (Persily et at.. 2010). However,
formaldehyde concentrations in remodeled or newly built homes, especially in wooden-framed homes,
were found to decrease to mean levels comparable to older homes levels within 2 years (Park and Ikeda.
2006). This is likely because formaldehyde found in newer articles is mostly released within that time
frame (Park and Ikeda. 2006).

C,1 General Formaldehyde Dissipation Curve

Figure Apx C-l displays the general formaldehyde dissipation in residential indoor air. The figure
shows an initial spike in concentration from off-gassing following initial installation of new articles.

This is followed by a rapid decrease in concentrations over the first few months. In each building
configuration, the living area has less fluctuations in concentrations after the initial concentration spike
following installation compared to other areas. Similarly, the basement in the attic/living space/basement
building configuration has less fluctuations in concentrations after the initial concentration spike
following installation. The higher variability in concentrations seen in the attic of both building
configurations and the crawlspace of the attic/living space/crawl space building configuration reflect the
sensitivity of off-gassing to temperature in unconditioned zones within the two building configurations.

Page 70 of 81


-------






























	Attic













	Basement























A















A.—A	

A—,





^	

0.0

1.0

2.0 3.0 4.0
Elapsed Time (in Years)

5.0

6.0

7.0

FigureApx C-l. General Formaldehyde Dissipation in a Residence

Page 71 of 81


-------
Appendix D CEM SUPPLEMENTAL ANALYSES

D.l CEM Screening Combustion Assessment

Because formaldehyde is a combustion byproduct/AT SDR. 1999). some home with various sources of
combustion (e.g., wood fireplace) may contribute to indoor air concentrations of formaldehyde. The
purpose of this supplementary indoor air assessment is to contextualize modeled formaldehyde
concentrations in the indoor environment and provide confidence in modeled concentrations while
accounting for combustion sources of exposure.

D.l.l Methods	

EPA used the AHHS II dataset to identify homes with and without sources of combustion. EPA
combined the AHHS2_Hazard and the ResidentQ datafiles from AHHS II, after sorting the data
according to dwelling unit id (duid), to collate data on formaldehyde concentration and reported sources
of combustion including the presence of smoking (including frequency of smoking events), combustion
furnace, gas stove, wood fireplace, gas hot water heater, gas dryer, gas cool stove/oven, portable fuel-
fired heater or other combustion sources. Presence or absence of combustion sources in each home was
identified based according to yes or no resident response to the relevant question. The data was
organized according to homes with at least one reported source of combustion and homes with no
reported sources of combustion. Measured formaldehyde indoor air concentrations were analyzed
according to this distinction.

CEM was used to model formaldehyde indoor air concentrations for new materials in a home; especially
for articles (for the relevant TSCA COUs) that are identified as the biggest emitters of formaldehyde
relative to others in the formaldehyde consumer exposure assessment. Such exposures were extrapolated
to a year of exposure and further as a lifetime average daily concentration for the identified TSCA
COUs

EPA generated box and whisker plots for a summary of all three sets of data including measured
formaldehyde indoor air concentrations for homes with and without reported combustion sources, and
estimated formaldehyde indoor air concentrations from TSCA COUs.

D.1.2 Comparison of Formaldehyde Indoor Air Estimates from CEM Modeling of TSCA
COUs Relative to Homes with and without Reported Combustion Sources,
According to AHHS II

Regarding the comparison of formaldehyde indoor air concentrations there are fundamental differences
between the modeled and the monitoring data. Caution should be applied when comparing modeled to
monitoring results, as this is not a 1:1 comparison, due to the following:

1. Assumed total exposure - The assessment of exposures in the indoor air environment is an
aggregate assessment. This means that the measured indoor air of formaldehyde from AHHS II
represent indoor air exposures from all formaldehyde sources across U.S. homes. Despite
controlling for combustion sources for exposure, there may be other sources of exposure that
could not be accounted for. Through the indoor air exposure assessment of formaldehyde, EPA
conducted a targeted assessment of the largest emitters of formaldehyde from TSCA-based
sources of exposure. Therefore, the aggregated modeled indoor air concentrations of
formaldehyde based on TSCA COUs may not be a directly comparable to AHHS II
concentrations of formaldehyde after removing homes without combustion sources.

Page 72 of 81


-------
2.	Dissipation over time - While measured formaldehyde concentrations from AHHS II represent
homes that have a combination of new and old materials that have off-gassed over time (and
potentially several decades), CEM does not incorporate chemical half-life (EPA. 2019a)

a. COU-specific estimates represent formaldehyde air concentrations from new articles only

i. Hence, total modeled estimates may represent formaldehyde air concentrations
from a newly built home (or automobile), based on the TSCA COUs assessed.

3.	Room of use - First, CEM models according to the most likely room of use for a given article
per TSCA COU. CEM also assumes a typical home has a building volume of 492 m3 and
specific default room sizes (e.g., 50 m3 for a living room). AHHS II measured formaldehyde in
the most frequently used room in each home, which differed from one to another. For instance,
formaldehyde may have been measured a living room for one home and in a kitchen for another.
Therefore, the measured monitoring concentration may be from a different room of use or an
entire home type than CEM considered. Similarly, CEM also assumes specific interzone
ventilation rates and air exchange rates in a residential area per hour, which may differ
depending on the home type and size of the home. AHHSII considered homes of varying types
(and, therefore, home of varying sizes) including detached single-family homes, mobile homes
and apartments in buildings with five or more units. Thus, it may also be assumed that the
interzonal ventilation rates and air exchange rates would differ between homes in the AHHS II
survey and CEM defaults (EPA. 2019a).

4.	Humidity and temperature - Increased indoor air temperatures and humidity levels have been
demonstrated to correlate with increased formaldehyde indoor air concentrations (Murphy et at..
2013). The degree to which humidity and temperature impacted the measured formaldehyde
indoor air concentrations in AHHS II is unknown. CEM cannot yet account for or vary
temperature and humidity but is an area of future improvement.

5.	Exposure duration - CEM assumes durations of exposure specific to TSCA COUs assessed,
from which a lifetime average daily concentration is estimated. However, the measurement of
formaldehyde indoor air concentrations in the AHHS II survey was according to a 3.5-hour
TWA.

Within the AHHS II survey, some homes were reported to have sources of combustion ranging from
tobacco smoke to wood fireplaces, which are known to produce formaldehyde as a byproduct. EPA
analyzed the formaldehyde concentrations in AHHS II from homes with and without at least one
combustion source of formaldehyde (Figure Apx D-l). Some agreement can be observed in the spread
of the two datasets mostly in the lower quartiles of the figure. However, there were more homes in the
upper quartile of formaldehyde indoor air concentrations where there was at least one source of
combustion compared to when there were none. In addition, some homes with at least one reported
combustion source had considerably higher measured formaldehyde indoor air concentrations compared
to homes with no reported combustion sources.

Page 73 of 81


-------
None

_Q

E
o
O

(0

(D
o

1_

o
w

At Least One Source

25

50

75

100

Concentration (ppb)

FigureApx D-l. Comparing the Relative Concentrations of Homes with and Without Sources of
Combustion for Formaldehyde in AHHS II

In general, the range of modeled formaldehyde residential indoor air concentrations were within an
order of magnitude of the range of monitoring values (4-423 [j,g/m3 compared to 0.3-124.2 (J,g/m3,
respectively) from a nationally representative studies of formaldehyde in residential indoor air, via the
AHHS II residential indoor monitoring study (Figure Apx D-2). Also, a few homes with at least one
source of combustion had at least double the concentration of formaldehyde compared to homes with no
reported sources of combustion.

Page 74 of 81


-------
Monitored No Combustion (AHHS)-

<1)
Q_

Modeled (CEM)

At Least One Source of Combustion (AHHS) -



100	200	300

Concentration (|jg/m3)

400

FigureApx D-2. Comparison of AHHS II Monitoring to Modeling Estimates of Indoor Air
Concentrations

Page 75 of 81


-------
D.2 CEM Screening Composite Wood Article Assessment

EPA conducted screening assessment of wood articles focusing on estimating potential exposure levels
based upon composite wood article emission limits set under TSCA Title VI.

D.2.1 Methods

This supplementary assessment was conducted using the following general steps:

1.	Identify emission standards set for hardwood plywood (HWPW), medium density fiberboard
(MDF), and particleboard (PD)

2.	Use the identified emission limits to estimate composite wood article-specific emission rates

a.	First, by converting the product-specific emission standards to air
concentration(Table_Apx D-l)

b.	Then, using that estimated air concentration to generate emission rates using
EquationApx D-1 (	) (Table Apx D-2)

EquationApx D-l.

Where:

[CH20]ss
[CH20]out

b

Area
PEX

VOL
D
m

Assuming that [CH20]out is zero, substituting for D in Equation Apx D-l, and denoting PEX * VOL as
Q (i.e., the airflow rate in/out of the chamber, in m3/hr), we can solve for b as follows:

Equation Apx D-2.

(	Area\ ( Q \

b = [CH20]ss,(l + m> — ).(—)

3. Model indoor air exposures using other key parameters highlighted in Section 2.2.3
a. Central tendency weight fractions for building wood articles were used

b * Area
PEX * VOL + LCrt2^Jout

[Ltl20\ss —	—

: steady-state formaldehyde concentration inside the compartment (mg/m3)
: steady-state formaldehyde concentration outside the compartment (mg/m3)
: the emission rate at zero CH2O concentration in the air (mg/m2-hr)

: Exposed surface area of the source (m2)

: the compartment's air exchange rate with outdoors (hr"1), assuming a mixing
factor equal to unity
: the volume of the compartment (m3)

^ m*Area

PEX*VOL

the mass transfer coefficient (m/hr)

Page 76 of 81


-------
Table Apx D-l. Estimated Concentrations in mg/m3 from Emission Standarc

s in ppm

Pressed Wood Articles

Emissions Standard

(PPm)

Molecular
Weight

Constant
(Volume of 1
mole at 1 atm)

Concentration
(mg/m3)

Hardwood plywood

0.05

30.03

24.45

0.061411043

Medium density fiberboard

0.11

30.03

24.45

0.135104294

Particleboard

0.09

30.03

24.45

0.110539877

Table Apx D-2. Estimating Emission Rates from Product Specific Concentrations

Pressed
Wood
Articles

Concentration
(mg/m3)

Assumed Slope
or Mass
Transfer
Coefficient
(m/hr)

Assumed
Chamber
Volume
(m3)

PEX "

Q

(m3/hr) b

Surface
Area (m2)

Emission

Rate
(mg/m2/hr)

Hardwood
Plywood

0.061411043

0.27

100

0.5

50

26

0.134679141

Medium

Density

Fiberboard

0.135104294

1.06

100

0.5

50

26

0.403026503

Particleboard

0.110539877

0.7

100

0.5

50

26

0.289954601

11 PEX = compartment's air exchange rate with outdoors (per hour), assuming a mixing factor equal to unity
'' Q = the airflow rate in/out of the chamber

The exposure scenario modeled was for an individual who spends two hours per day, every day, in a
living room with flooring made with engineered wood flooring. Formaldehyde inhalation exposures are
assumed to stem from emissions from HWPW, MDF, or PB.

D.2.2 Results

The estimated yearly average daily indoor air concentrations from pressed wood articles ranged from
5.53 to 16.58 pg/m3 (Table_Apx D-3).

Page 77 of 81


-------
Table Apx D-3. CEM

Estimated Average E

•aily Concentrat

ion Over 1 Year

Condition of Use
Subcategory

Scenario

Environment

CEM Calculated Average
Daily Concentration
(ppm)

CEM
Calculated
Average Daily
Concentration
(|ig/m3)

Construction and
building materials
covering large surface
areas, including wood
articles; Construction
and building materials
covering large surface
areas, including paper
articles; metal articles;
stone, plaster, cement,
glass and ceramic
articles

Building/C onstruction
Materials - Wood
Articles: Hardwood
Plywood (residential)

Residential
(Living Room)

4.50E-03

5.53

Construction and
building materials
covering large surface
areas, including wood
articles; Construction
and building materials
covering large surface
areas, including paper
articles; metal articles;
stone, plaster, cement,
glass and ceramic
articles

Building /
Construction
Materials - Wood
Articles: Medium
Density Fiberboard
(residential)

Residential
(Living Room)

1.35E-02

16.58

Construction and
building materials
covering large surface
areas, including wood
articles; Construction
and building materials
covering large surface
areas, including paper
articles; metal articles;
stone, plaster, cement,
glass and ceramic
articles

Building /

Construction

Materials - Wood

Articles:

Particleboard

(residential)

Residential
(Living Room)

9.69E-03

11.9

D.2.3 Conclusion

EPA qualitatively assessed potential exposures from HWPW, MDF, and PB, according to the emission
standards established under TSCA Title VI, using the best available information and tools. The degree to
which the estimated indoor air concentrations from the modeled composite wood articles are reflective
of real-world scenarios is unknown. Due to the following key uncertainties, EPA has a medium
confidence in this assessment:

Page 78 of 81


-------
The identified emission standards were assumed to be equivalent to a product specific indoor
concentration, but it is unknown to what degree this is reflective of composite wood articles
currently on the market.

Whether the assessed wood articles are made entirely of HWPW, MDF, or PB.

o Wood articles on the market may be composed of a combination of composite wood
layers.

Whether the identified articles are compliant with the relevant emission standards.

Whether the approach to estimating emission rates from the set emission limits sufficiently
represent articles on the consumer market.

Page 79 of 81


-------
Appendix E SUMMARY OF WEIGHT OF SCIENTIFIC EVIDENCE

Table Apx E-l summarizes the weight of scientific evidence for the indoor air exposure assessments, as
described in Section 4. The weight of scientific evidence for the indoor air exposure assessment of
formaldehyde was driven by studies that included article-specific emission factors (Table 2-2, Table 2-3)
which were key components of the modeling results. Only studies and datasets useful in generating a
quantitative assessment (e.g., via IECCU modeling) progress from data evaluation to data extraction. In
the case of the formaldehyde pool of studies, there were several COU-specific studies that did not report
any concentrations of formaldehyde but provided emission rates; those are labeled as "supplemental
studies" in Table 2-2 and Table 2-3. Because emission rate data for the COUs assessed were generally
scarce, from the exposure systematic review pool of studies, emission rates from supplemental studies
were used in addition to those with a study rating criteria (low to high-rated studies) (EPA. 2023b).

Nine studies were used to compile COU-specific emission rates used to apply user-defined, COU-
specific emission rates to model formaldehyde indoor air concentrations using IECCU—a peer-
reviewed, high-tier model that has been used in previous TSCA risk assessments, per the exposure
systematic review criteria (EPA. ^ ). Twenty indoor air monitoring studies were used to compare
measured formaldehyde concentrations against modeled concentrations from TSCA COUs. This
includes a robust nationally representative monitoring study of formaldehyde in indoor air via the AHHS
II, jointly sponsored by EPA and HUD.

Page 80 of 81


-------
Table Apx E-l. Weight of Scientific Evidence Conclusions for the Indoor Air Exposure
Assessments

Consumer
Route
(Assessment)

Confidence
in Model
Used"

Confidence
in Model
Default
Values6

Confidence in Key Modeling
Inputsc

Key Sources of
Indoor Air
Data®

Weight of
Scientific
Evidence
Conclusion'1

Emission
Factors'7

Article
Surface
Areas2

Room
of Use^

Inhalation
(Indoor Air)

High

High

Medium

High

Medium

9 studies

incorporated

into modeling

(using emission

factors)

20 indoor air

monitoring

High

11 "Confidence in Model Used" considers whether model has been peer reviewed, as well as whether it is being applied
in a manner appropriate to its design and objective. IECCU has been peer reviewed, is publicly available, and has
been applied in a manner intended; that is, to exposures associated with uses of household products and articles.
IECCU was the best available tool to assess indoor air exposure for formaldehyde.

h "Confidence in Model Default Values" considers default value data source(s) such as building and room volumes,
interzonal ventilation rates, and air exchange rates in IECCU (similar to CEM) (EPA. 2021a).
c "Confidence in User-Selected Varied Inputs" considers the quality of their data sources, as well as relevance of the
inputs for the selected consumer condition of use.

''"Emission Factors" is primarily sourced from high quality studies used to develop IECCU's COU/article-specific
modeling. Without better data, EPA assumed that composite woods currently on the market adhere to the TSCA Title
VI formaldehyde emission standards for composite woods. This contributed to the basis for the selection of
"Medium" confidence.

'' "Article Surface Areas" of formaldehyde in articles is sourced from various sources including literature, model
defaults (i.e., CEM and FIAM) and the EPA Exposure Factors Handbook (EPA. 2011).

'"Room of Use" (location of the exposure scenario) is informed by responses in the Westat (1987) survey, which
received a high-quality rating during data evaluation, although professional judgment is also applied for some
scenarios. The room of use, for example, was selected according to professional judgement. It is conceivable that
different rooms of article use may apply across homes. The reasonableness of these judgements contributed to the
basis for the selection of "Medium" confidence.

g In addition, while emission rates from nine studies were extracted from systematic review and incorporated into
IECCU modeling, over a dozen others were used to characterize the indoor air concentrations of formaldehyde.
h See the Draft Systematic Review Protocol for a detailed description of weight of scientific evidence ratings (EPA.
2023a).

Page 81 of 81


-------