United States
Environmental Protection
Agency
EPA/600/R-12/051I September 30, 2012 | www.epa.gov/ord
Polychlorinated Biphenyls
(PCBs) in School Buildings
Sources, Environmental
Levels, and Exposures
Kent Thomas, JianpingXue, Ronald Williams,
Paul Jones. Donald Whitaker
i
..LI (Mm |Nn PO T|
i i i
'•• !• at
IB.
Office of Research and Development
National Exposure Research Laboratory
-------
Notice
The information in this document has been funded in part by the United States
Environmental Protection Agency under Contract EP-D-10-070 to Alion, Inc.
It has been subjected to the Agency's peer and administrative review and has
been approved for publication as an EPA document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
-------
EPA/600/R-12/0511 September 30, 2012
United States
Environmental Protection
Agency
Polychlorinated Biphenyls
(PCBs) in School Buildings
Sources, Environmental
Levels, and Exposures
Kent Thomas, Jianping Xue, Ronald Williams,
Paul Jones, Donald Whitaker
Office of Research and Development
National Exposure Research Laboratory
-------
Acknowledgements
We would like to acknowledge and thank Dennis Santella
and James Haklar, U.S. EPA Region 2, and Kimberly Tisa,
U.S. EPA Region 1, for their assistance and their insight into
schools with PCBs. We would also like to acknowledge
the New York City School Construction Authority and TRC
Engineers, Inc. for designing and conducting the remedial
pilot school investigation program and for providing
measurement results and school information. We thank
Tamira Cousett, Carlton Witherspoon, Keith Kronmiller,
Paulette Yongue, and Hunter Daughtry of Alion Inc. for their
project management, field sampling, and data coordination
efforts. We thank the staff of NBA Pace Analytical for
the PCB Aroclor and congener sample analysis. We
appreciate the assistance and consultation regarding PCB
emissions, transport, and indoor modeling by Zhishi Guo
in the U.S. EPA National Risk Management Research
Laboratory. Finally, we would like to thank Peter Egeghy,
U.S. EPA National Exposure Research Laboratory, for his
efforts in collecting and organizing extant data from other
measurement efforts that were used in the SHEDS model
comparison.
-------
Executive Summary
E.I Background
Caulk containing polychlorinated biphenyls (PCBs) was
used in some buildings, including schools, in the 1950s
through the 1970s. PCBs were used as a plasticizer in
caulk, added either during manufacture or mixed on site
prior to installation. Other potential sources of PCBs, such
as fluorescent light ballast capacitors, were also used in
school buildings during that era. PCBs were used in other
types of capacitors, transformers, plasticizers, coatings, inks,
adhesives, and carbonless copy paper but the extent of these
uses in school buildings is not well known. Materials and
components containing PCBs are still present today in many
of these older buildings. PCBs are semi-volatile organic
chemicals and can be transported in and around buildings
through vaporization into the air and through absorption into
dust and materials. PCBs may be present in the air, dust, soil
and on surfaces in and around school buildings leading to the
potential for occupant exposure through multiple routes. In
September 2009 the U.S. EPA announced new guidance for
school administrators and building managers for managing
PCBs in caulk and to help minimize possible exposure.
However, there was limited information on PCBs in school
buildings in the United States. Neither the PCB sources,
nor the routes of exposure, have been well-characterized
in school buildings. As such, there remained considerable
uncertainty regarding the extent to which children and staff
members may be exposed to PCBs in school environments.
The EPA also announced in 2009 that additional research
would be performed by the Office of Research and
Development's (ORD) National Exposure Research
Laboratory (NERL) and the National Risk Management
Research Laboratory (NRMRL) to further study this issue.
The research was intended to help fill information gaps and
improve our understanding of PCBs in school buildings and
approaches for mitigating exposures.
E.2 Objectives
Information on sources of PCBs and levels in school
environments is needed to improve risk management
decision-making. ORD's NERL planned research to better
understand and characterize PCB sources, emissions,
environmental concentrations, and exposures in school
environments. Research was also planned by ORD's
NRMRL to perform laboratory studies of PCB sources and
transport and to evaluate selected mitigation approaches.
In order to better understand the significance of PCB-
containing caulk and other building materials and
components as a source of PCB exposures to children,
teachers, and staff in school buildings, the ORD's NERL
planned research to utilize a limited set of real-world
measurements to:
1. characterize PCB-contaminated caulk and other
potential primary and secondary sources of PCBs in
school buildings,
2. characterize levels of PCBs in school air, dust, soil, and
on surfaces and to investigate relationships between
potential PCB sources and environmental levels,
3. apply an exposure model for estimating children's
exposure to PCBs in schools with PCB sources,
4. evaluate which routes of exposure (e.g., inhalation,
contact with surfaces or dust) are likely to be most
important, and
5. provide information to assist in developing risk
management practices for reducing exposure to PCBs
in schools.
E.3 Approach
NERL exposure scientists and their contractor collected
multiple air, dust, soil, and surface wipe samples from a
school building scheduled for demolition. Scientists also
collected samples from building materials like caulk, tiles,
paints, mastics, and others. These environmental and
material samples were analyzed for PCB Aroclor mixtures
to determine total PCB concentrations. A subset of these
samples was then analyzed for PCB congeners (individual
PCB compounds) to provide information regarding the
presence and relative amounts of individual PCBs and
to evaluate relationships between different sources and
environmental media.
NERL scientists also used environmental and building
material PCB Aroclor measurement data from five
schools that were generated by the New York City School
Construction Authority under a pilot remedial investigation
plan developed under an agreement with U.S. EPA Region 2.
Measurement results and information from these six school
buildings provided a limited data set for characterizing
real-world PCB sources and environmental levels in schools
that were built from the late 1950s to the early 1970s.
Environmental measurement results were used to generate
distributions of estimated PCB exposures for several different
child age groups using the Stochastic Human Exposure and
Dose Simulation (SHEDS) model. The SHEDS model was
also used to estimate the relative contribution of inhalation,
dermal contact, and non-dietary ingestion of dust and soil
components of the total exposure.
-------
E.4 Results
Primary PCB Source Assessment
Primary sources of PCBs are denned here as those that
were manufactured containing PCBs or had PCBs added
during construction. Primary sources of PCBs that might be
currently found in buildings include caulk or other sealants,
window glazing, fluorescent light ballast capacitors, ceiling
tile coatings, and possibly other materials such as paints or
floor finishes. Numerous samples of caulk, window glazing,
and joint sealants were collected at the six schools. When
considering measurement results for these samples it is
important to remember that multiple samples of the same
type of material may have been collected from several places
in a school building, and that these schools were selected
because they had suspected or known PCB sources. Key
results for caulk and light ballast characterization at the six
schools are reported below.
Caulk
• All six schools contained some interior or exterior
caulk with PCB levels greater than 10,000 ppm.
8.6 % of the 427 samples of interior caulk, window
glazing, and building joint material from six
schools had total PCB levels > 10,000 ppm, while
82% had levels < 50 ppm. The maximum concentration
was 440,000 ppm (which is 44% PCBs by weight).
• 41% of the 73 samples of exterior caulk, window
glazing, and building joint material from three schools
had total PCB concentrations > 10,000 ppm, while
37% had levels < 50 ppm. The maximum concentration
was 328,000 ppm. No exterior caulk measurement results
were available for the other three schools.
• Aroclor 1254 was the Aroclor mixture reported for most
caulk samples, although the analytical laboratory often
reported an altered PCB pattern with Aroclor 1254 as
the best match. Caulk with an Aroclor 1260 pattern was
found at one school.
• Caulk with high PCB concentrations was generally found
to be intact and still somewhat flexible. Other flexible
caulks and sealants without high PCB levels were also
found, while dry and brittle sealants did not have high
levels of PCBs.
• Total PCB emission rates were estimated for caulks
with > 50,000 ppm PCBs collected in several rooms,
around the exterior of a window, and for an exterior joint.
Estimated emission rates ranged from 53 to 3100 ug/hr
and depended on the PCB concentration in the caulk and
the total surface area of the caulk in the specific room or
location. The effect of different temperatures was not
evaluated.
• Screening-level estimates of the range of indoor air
concentrations that might result from PCB emissions
from caulk inside or outside several rooms gave results
ranging from 10 to 1900 ng/m3, and depended on the
estimated emission rate, air exchange rate, and for the
exterior window caulk the percentage of emitted PCBs
estimated to enter the classroom. Measured indoor air
concentrations were higher than the screening-level
estimates in two of four rooms.
Light Ballasts
• Light ballast survey results were available for five
schools. The percentage of ballasts found likely to be
PCB-containing ranged from 24% to 95% across the
schools.
• Estimated total PCB emission rates for intact ballasts
in three rooms ranged from 1.2 to 290 ug/hr (based on
laboratory emissions tests of four ballasts at 45°C, near
the operating temperature of ballasts when the lights
are on) and from 0.08 to 18 ug/hr (based on laboratory
emissions tests of four ballasts at 23 °C, near the
temperature of ballasts when light are off). The estimates
depended on the number of PCB-containing ballasts in the
room and the lowest and highest emitting ballasts in the
chamber testing, which saw a nearly 80-fold difference in
emission rates.
• Screening-level estimates of the range of air
concentrations that might result from PCB emissions from
intact PCB-containing ballasts in three rooms gave results
ranging from 1.6 to 2400 ng/m3 (at a ballast temperature
of 45°C). At the median estimated emission rate the
screening-level air concentration estimates ranged from
2.3 to 44 ng/m3 and were lower than the measured air
concentrations that ranged from 690 to 1460 ng/m3.
• There was evidence of previous ballast failure in and
around some light fixtures in these schools and some
ballasts had been replaced with non-PCB containing
ballasts. Emissions of PCBs from light fixtures that
have been previously contaminated by leaking or failed
ballast capacitors have not been measured, but could be
substantial if PCB oil residues are present. Emission rates
have not been estimated for ballasts with capacitors that
are leaking or have burst, but based on the bursting of a
ballast capacitor during laboratory testing it is anticipated
that emissions would be substantially higher than those
measured for intact ballast capacitors and would have a
large impact on indoor air PCB concentrations.
Secondary PCB Source Assessment
Secondary sources of PCBs are defined here as those
materials that become contaminated due to absorption from
direct contact with primary PCB sources such as caulk, or
through absorption of PCBs in the indoor air that have been
emitted by primary sources such as caulk and light ballasts.
Materials such as paints, dust, masonry, floor and ceiling
tiles, and mastics may become secondary sources after years
of exposure to PCBs emitted from primary sources. These
materials should be considered sinks as well as sources, due
to their ability to absorb PCBs from direct contact or from
the air. When considering measurement results for these
samples it is important to remember that multiple samples
of the same type of material may have been collected from
-------
several places in a school building. Key results for secondary
source characterization at the three schools with material
measurements are reported below.
• 93% of the 411 building material samples collected at
three schools had measureable levels of PCBs. The
overall median total PCB concentration was 16 ppm
(interquartile range 6.1 - 39.6 ppm).
• Paint had the highest total PCB concentrations with a
median of 39 ppm (interquartile range 26 - 72 ppm).
Fiberboard had a median level of 31 ppm (interquartile
range 13 - 39 ppm), while lower levels were found in
some other materials that often have high surface areas
in buildings such as ceiling tile with a median of 7.6 ppm
(interquartile range 2.7 - 12 ppm) and floor tile with a
median of 4.4 ppm (interquartile range 1.4 - 8.7 ppm).
• It is impossible to be certain that these materials did not
contain some PCBs when originally installed. Measured
concentrations and relative levels among different
materials are consistent with these materials being "sinks"
that have absorbed PCBs emitted from primary sources.
• Screening-level estimates of emission rates for
multiple materials in several rooms were calculated
to evaluate their relative potential as secondary
sources of PCB emissions following removal of
primary sources. Estimated emission rates for
different materials in classrooms ranged from
< 1 to 100 ug/hr and cumulative totals for 20 materials
in a room ranged up to 270 ug/hr. Estimated emission
rates for different materials in gymnasiums ranged
from < 1 to 1100 ug/hr and cumulative totals for
16 materials ranged up to 2700 ug/hr. Estimated
emission rates depended on the surface area of the
material and concentration of PCBs in the material.
Paints and varnishes generally had the highest relative
potential emissions due to the combination of higher
PCB concentrations and high surface areas. (There are
considerable uncertainties in these estimates, which are
based on emission parameters derived from laboratory
emissions testing of caulks. Emission parameters for
the many different types of other materials could be
substantially different than those for caulk).
• It is difficult to estimate indoor air concentrations of
PCBs that might result from secondary sources following
removal of primary sources because of the large number
of different types of PCB-containing material in a room,
and because the source - sink dynamics for multiple
different materials are difficult to characterize. However,
the cumulative PCB emissions from secondary sources
could potentially result in indoor air PCB levels above
background in school rooms following mitigation of
primary sources, depending on relative emission rates,
sink rates, and ventilation rates by outdoor air.
• Surface dust and dust in ventilation systems may be
another secondary source for potential exposures to PCBs.
PCB Concentrations in the School Environment
PCBs may be transported from primary and secondary
sources to environmental media in and around buildings
including indoor air, dusts, and soils. PCBs may also be
found on surfaces inside buildings. Key results for PCB
concentration measurements in environmental media at six
schools are reported below.
• The median indoor air total PCB concentration
based on 64 measurements across six schools was
318 ng/m3 (interquartile range 59 - 732 ng/m3). There
was considerable variability between schools with
median air levels at individual schools ranging from
<50 to 807 ng/m3. There was considerable variability
within schools; for example, indoor air levels ranged from
236 to 2920 ng/m3 in different rooms at one school.
• Surface wipe samples were collected from high-
contact (desks, tables) and low-contact (walls, floors,
window sills) surfaces at six schools. Median total PCB
concentrations were 0.147 ug/100cm2 (interquartile
range O.100 - 0.330 ug/100cm2) for 72 high-contact
surfaces and 0.201 ug/100cm2 (interquartile range
0.128 - 0.419 ug/100cm2) for 78 low-contact surfaces.
Concentrations ranged from <0.100 to 2.84 ug/100cm2 for
high contact surfaces and O.100 to 2.30 ug/100cm2 for
low-contact surfaces. Median high-contact wipe levels
ranged from <0.100 to 0.380 ug/100cm2 at individual
schools. There was no consistent difference between
median high-contact and low-contact concentrations
across the six schools.
• There was a modest but significant correlation
between indoor air total PCB concentrations and total
PCB levels measured in high-contact surface wipes
(Spearman r = 0.531, p-value O.001) with a lower
but still significant correlation between PCBs in air
and low-contact surface wipes (Spearman r = 0.247,
p-value = 0.050).
• Soil samples were collected at six schools at distances
of 0.15, 0.91, and 2.44 m (0.5, 3, and 8 feet) from
the building. Only 33% of the samples had PCB
concentrations above the quantifiable limit. The median
total PCB concentration across all 309 soil samples
was less than the quantifiable limit. The 75th percentile
concentration was 0.98 ppm and the maximum value was
211 ppm. There was considerable variability between
schools; for example one school had only 10% of the
PCB levels above the quantifiable limit while another
school had 100%. The 75th percentiles of total PCB
concentrations across six schools at 0.15, 0.91, and
2.44 m from the building were 2.13, 0.55, and <0.5 ppm
respectively, supporting the idea that higher soil PCB
levels are likely to be found in closer proximity to
building sources including exterior window caulk and
building joints with PCBs.
-------
• Indoor dust samples were collected at only one school.
The median total PCB concentration was 22 ppm
(interquartile range 17 - 53 ppm) across samples collected
in seven rooms.
PCB Congener Analysis Results
Individual PCB congeners were measured in all of the air
samples and in a subset of surface wipe, indoor dust, soil,
caulk, and other building material samples at one school.
Examining patterns of individual congeners can provide
insight regarding relationships between PCBs sources and
environmental media, and can also provide information
useful in exposure and risk assessment. Key results for PCB
congener measurements at one school are reported below.
• Analysis of sample extracts for congeners resulted in total
PCB concentrations that were approximately 20% lower
than the analysis of the same sample extracts using an
Aroclor method.
• Air samples collected both vapor and particle-bound
PCBs, but these fractions were not analyzed separately.
The pattern of congeners in indoor air was more heavily
weighted towards more volatile congeners as compared
to non-aged Aroclor 1254 and as compared to the
PCB-containing caulk at the building. The pattern of
congeners in air was not as heavily weighted towards
more volatile congeners as would be predicted if they
were from vapor emissions from caulk alone, suggesting
that a portion of the PCBs in air may be associated with
airborne particles.
• The congener pattern in the PCB-containing caulk was
somewhat more heavily weighted towards less volatile
congeners as compared to non-aged Aroclor 1254. It
is possible that the more volatile congeners have been
depleted from sources such as the exterior caulk over a
period of 43 years.
• Many PCB-containing fluorescent light ballasts have been
shown to contain Aroclor 1242. However, Aroclor 1254
was found to be the capacitor oil in one NYC ballast and
was reportedly used in light ballast capacitors prior to
1952. The congener pattern in indoor air at the school
with congener measurements did not resemble a pattern
that would be expected if the predominant source of PCBs
was Aroclor 1242 from light ballasts, but the capacitor oil
was not analyzed to determine if it was, in fact, Aroclor
1242.
• Congener patterns in surface wipe, indoor dust, and other
building materials were generally similar to Aroclor 1254
and to the PCB-containing caulk. Because only four dust
samples were analyzed for congeners we did not assess
air - dust congener correlations, but we anticipate that
such a relationship is likely to exist.
• Soil samples had a congener pattern weighted towards
less volatile congeners as compared to either Aroclor
1254 and even compared to the PCB-containing exterior
caulk. The higher proportion of less volatile congeners
as compared to the likely source is possibly a result
of weathering and favored partitioning of less volatile
congeners into soil rather than air.
• Congener-specific analysis provides information that
may be useful for risk assessment. For example, between
5% (indoor air) and 14% (soil) of the total amount of
PCBs were comprised of the sum of the 12 dioxin-like
congeners. As additional information on PCB toxicity
accumulates, having congener-specific information
becomes more important.
Modeled Exposure and Dose Estimates
The PCB concentrations in the air, surface wipe, soil, and
estimates for dust were used in the SHEDS model to generate
distributions of estimated exposures to PCBs for children
in four age groups (4-5, 6-10, 11-13, and 14-18 years old).
The model incorporated environmental concentration
distributions based on the range of air, surface wipe, and soil
measurements across the six schools. Dust measurements
were not made at most schools and dust concentrations were
estimated based on air concentrations and estimated dust/air
partition coefficients. The model incorporated distributions
of activity levels appropriate for each age group in estimating
exposures. The model then generated distributions of
estimated absorbed doses of PCBs resulting from the
estimated exposures in the school environment for the four
age groups. The model provided information on the potential
relative contribution to total absorbed dose from inhalation,
dermal contact, and dust and soil ingestion. Various remedial
approaches were examined by the NYC School Construction
Authority at the five New York City schools, and pre-and
post-remediation absorbed dose estimates were prepared
using environmental measurements from these schools.
Absorbed dose estimation included only those exposures to
PCBs that would occur to children while at school and do not
include background exposure through the diet or inhalation
from indoor and outdoor air while away from the school
and at home. Sensitivity testing was performed to examine
the impact of uncertainty in the pulmonary PCB absorption
fraction and for different PCB levels in dust and soil. Key
results for estimated exposures and exposure pathways are
reported for the 6 - 10 year-old age group below.
• Estimated absorbed doses for 6-10 year-olds at the
pre-remediation time point were 0.022 ug/kg/day at the
50th percentile and 0.041 ug/kg/day at the 95th percentile
of the distribution. Estimated absorbed doses for 4-5,
11-13, and 14-18 year old age groups were somewhat
lower.
• After incorporating indoor air and surface wipe
measurements obtained following different remedial
activities in five schools, estimated absorbed doses
for 6-10 year-olds were 0.007 ug/kg/day at the 50th
percentile and 0.012 ug/kg/day at the 95th percentile.
These levels were approximately 64 - 69% lower than
those at the pre-remediation time point.
-------
• Up to the 90th percentile of the estimated absorbed
dose distribution, the predominant route of exposure
for all age groups at all time points in schools with the
measured environmental levels would be via inhalation.
It was estimated that, on average, 74% of the absorbed
dose would result from inhalation at the pre- and post-
remediation time points (assuming a 70% pulmonary
absorption fraction). However, above the 90th percentile
of the estimated absorbed dose distribution, intake
from dust was predicted to be greater than intake from
inhalation.
• Sensitivity analysis showed that the estimated absorbed
doses were sensitive to assumptions regarding the
pulmonary absorption fraction. When the pulmonary
absorption fraction was changed from the baseline
of 70% to 30% in the model, the median estimated
exposure for the 6 - 10 year old group decreased
49% to 0.011 ug/kg/day. When the pulmonary absorption
fraction was increased to 100%, the median estimated
exposure for the 6 - 10 year old group increased
35% to 0.028 ug/kg/day.
• As described in more detail in subsequent sections, there
are uncertainties and limitations in modeled estimates
of exposure distributions and contributions of relative
exposure pathways. There are currently insufficient data
to perform a full uncertainty analysis for the SHEDS PCB
exposure estimation.
• Exposure estimation was not performed for adults,
including teachers and staff, as part of this effort due to
the lack of personal activity data at school suitable for
SHEDS modeling, such as those available for children
in the Consolidated Human Activity Database. It is
anticipated that adults would spend more time in school
buildings, which would tend to lead to higher doses, but
less ingestion of dust and soil and higher relative body
masses would lead to lower doses.
E.5 Conclusions
Sources ofPCBs in School Buildings
• PCBs-containing caulk is a primary source of PCBs in
and around school buildings. PCBs from exterior caulks
around windows and mechanical ventilation system air
intakes can lead to elevated concentrations in indoor
spaces. PCBs in exterior caulk are likely to enter the
soil near school buildings with the highest soil PCB
levels found in closest proximity to the building. Caulk
containing PCBs was found to be mostly intact and still
somewhat flexible, but visual examination alone may
not be adequate for determining if PCBs are present and
testing is needed to determine if caulk or other sealants in
a building contain PCBs.
• PCB-containing fluorescent light ballasts remain in use
in some older school buildings and are a primary source
of PCBs. Emissions from intact ballast capacitors can
lead to the presence of PCBs in school environments.
PCB residues from previously failed ballast capacitors
may remain in fixtures even if the ballast is replaced.
Leaking or bursting capacitors are likely to substantially
elevate PCB levels in indoor environments when they
fail. Because these ballasts have exceeded their expected
operational lifetimes, failure and possible leakage will
continue and is likely to increase for ballasts remaining in
place.
• Several paint samples had total PCB concentrations above
100 ppm, up to 718 ppm. PCBs were used as plasticizers
or flame retardants in some paints, so it is possible that
these paints may have incorporated PCB when they were
originally applied. Thus, it is possible that paints could
be primary sources of PCBs in buildings based on our
definition. Although they were not encountered in this
study, window glazing and ceiling tile surface coatings
containing PCBs have been reported in school buildings
and would be considered primary sources.
• Other primary sources of PCBs may have been used in
school buildings but are no longer present today. For
example, carbonless copy paper and PCB-containing
capacitors in early computer video display terminals may
have been used in school buildings. The potential impact
of previously removed sources on current PCB levels in
building environments cannot be easily determined.
• Many of the building and furnishing materials in
schools were found to contain PCB levels in the
4 to 100 ppm range. It appears likely that these materials
have absorbed PCBs that have been emitted from
primary sources. While primary sources remain in
buildings these other materials are likely to be in quasi-
dynamic equilibrium, with PCB emission and absorption
roughly balancing. However, when primary sources
are removed, these materials may serve as secondary
sources for emissions of PCBs into the air in the building.
Paints may be the most significant secondary sources
given their large surface areas and relatively high PCB
concentrations, but other materials may be important as
well. Following mitigation of primary sources it may, in
some cases, be necessary to consider mitigation actions
for secondary sources.
School Environment PCB Levels and Exposures
• PCBs are present in indoor air, dust, and on surfaces in
school buildings with PCB-containing source materials,
and are likely to be present in the soil near buildings
with exterior PCB-containing caulk. Building occupants
would be exposed to PCBs through expected normal
contacts with these environmental media.
• Estimated average total absorbed doses that could
occur from the PCBs in school buildings with
environmental levels that were found in these six
schools were near the reference oral dose levels for
Aroclor 1254 (0.020 ug/kg/day), and the reference dose
adjusted for absorption (0.017 ug/kg/day) which is a more
direct comparison with modeled dose estimates from
SHEDS, which estimated absorbed dose. The reference
dose level was exceeded by 75% or more of the estimated
-------
distributions of absorbed doses for the two younger age
groups. These estimates do not include the additional
background exposures to PCBs that occur outside of the
school environment, including exposures from dietary
intake and inhalation of PCBs in outdoor and indoor air at
non-school locations.
• PCB concentrations in indoor air were found to exceed
EPA's 2009 public health guidance levels (ranging from
70 to 600 ng/m3 depending on age) in many school
locations. Inhalation was estimated to be responsible for
over 70% of the exposure that could occur in buildings
with the environmental levels of PCBs that were found
in these six schools. Mitigation efforts that focus on
reducing indoor air PCB concentrations are likely to
have the greatest impact on reducing exposures, although
cleaning to reduce dust will contribute to reducing
exposures, particularly for children on the upper end of
the dust exposure distribution. Based on information
from the five New York City schools, it appears that
mitigation efforts can be successful in substantially
reducing indoor air concentrations and exposures to
PCBs.
Complexity of PCBs in School Buildings
PCBs in school buildings present a complex problem from
exposure assessment, risk assessment, and mitigation
decision-making perspectives. Different aspects of this
complexity are summarized below.
• There may be multiple primary sources of PCBs in
school buildings. Numerous different kinds of caulks
and sealants may be present throughout many building
locations and they must be sampled to determine
whether they contain PCBs. Fluorescent light ballasts
containing PCBs may be present and light fixtures may
be contaminated with residues from ballasts that have
previously failed.
• PCBs are semi-volatile organic chemicals with a wide
range of vapor pressures that will vaporize from primary
sources and will be transported through indoor and into
outdoor environments. They are absorbed by dust and
soil which can result in additional transport and exposure.
• PCBs absorb into numerous materials in a building
resulting in a reservoir that remains even after primary
sources are removed. These secondary sources may result
in continuing exposures following removal or remediation
of primary sources.
• Over 120 different PCB congeners were measured
in indoor air. These different congeners have a wide
range of physical properties. In addition to adding
exposure and toxicity complexity, the large numbers
of congeners raises a challenge of selecting the most
appropriate measurement methods for assessing PCBs in
buildings. Aroclor measurements are the simplest and
least expensive, but they may suffer in accuracy since
the congener patterns in school environmental media do
not exactly match those in the Aroclors. Homolog and
congener-specific analyses allow for improved accuracy
and better understanding and are recommended for use
where possible.
• Ventilation of building spaces with outdoor air is an
important factor in the indoor air PCB concentrations that
will result from source emissions. However, ventilation
in some older school buildings has been found to be poor,
and ventilation in older buildings is often difficult to
assess and to improve.
E.6 Limitations
There are important limitations and uncertainties in the
information included in this report. Key limitations and
uncertainties are summarized below.
• PCB measurement results were available from only six
schools. These schools were selected because they had
known or suspected PCB sources. It is not known if these
results are representative of older schools nationwide,
both in terms of the presence of PCB-containing materials
and components and the environmental concentrations
measured in and around the school buildings.
• Materials and components containing PCBs were likely
to have been used in buildings other than schools. This
report does not address whether and to what extent PCBs
may be a potential problem in other types of buildings,
and if so, whether environmental concentrations and
exposures are likely to be similar.
• PCB emissions from materials and light ballasts were not
directly measured at the six schools. Modeled emission
estimates and the resulting predictions of indoor air and
dust concentrations have considerable uncertainties.
Emission parameters are likely to vary across different
materials and for different temperature and ventilation
conditions. Two different types of chambers were used to
test caulk and light ballasts emissions, possibly impacting
comparability. Emissions from light ballasts are likely
to vary depending on the lighting fixture design and the
condition of the ballast and capacitor. Emissions from
light fixtures potentially contaminated from previously
leaking or failed ballasts were not evaluated.
• Attributing the relative impact of PCB emissions from
caulk and light ballasts on PCB levels in the schools
was difficult because both sources were present in most
buildings, and the Aroclor mixture used in all light
ballasts was not identified. Several paint samples were
found to have several hundred ppm of PCBs, and it is not
clear whether these contained PCBs when installed and
might be considered primary sources.
• There is uncertainty in modeled estimates of PCB
exposures due to uncertainties in key exposure model
parameters. In particular, there is limited information for
pulmonary absorption fraction from the range of PCB
congeners in vapor and particle-bound forms. There is
also uncertainty in total PCB exposures because of the
lack of robust data for background exposures from dietary
and other non-school sources.
-------
Table of Contents
1.0 Introduction 1
1.1 Background 1
1.2 Need for Research 4
1.3 Research Objectives 5
1.4 About this Report 5
2.0 Methods 7
2.1 Generation of School Measurements 7
2.2 Sample Collection 7
2.3 Sample Extraction and Analysis 11
2.4 Caulk Emissions Modeling 11
2.5 Light Ballast Emissions Modeling 14
2.6 Building Material Emissions Modeling 16
2.7 SHEDS Exposure Modeling 16
3.0 Quality Assurance and Quality Control 23
3.1 Quality Control Results for Aroclor Analysis 23
3.2 Quality Control Results for Congener Analysis 26
3.3 Quality Assurance Assessments 28
4.0 Results 29
4.1 School Information 29
4.2 PCB Source Characterization 33
4.3 PCBs in Environmental Media 53
4.4 Congener and Homolog Measurements 65
4.5 SHEDS Exposure Modeling 82
5.0 Conclusions 95
5.1 Sources of PCBs in School Buildings 95
5.2 School Environment PCB Levels and Exposures 95
5.3 Complexity of PCBs in School Buildings 96
5.4 Study Limitations 96
6.0 References 97
Appendix A A-l
Appendix B B-l
Appendix C C-l
Appendix D D-l
Appendix E E-l
-------
List of Tables
Table 1-1. Several commonly used Aroclor mixtures, their chlorine weight percent values, and their highest five
congeners by weight percent 1
Table 1-2. PCB vapor pressure ranges: examples from three reports 3
Table 2-1. Summary of sample collection and analysis methods 8
Table 2-2. Sample collection information (numbers of samples by type and location) 9
Table 2-3. Vapor pressures for selected target congeners in Aroclor 1254 12
Table 2-4. Predictive error for the P-N correlations 13
Table 2-5. Estimated total PCB emission rates for four intact light ballasts in chamber testing 15
Table 2-6. Ratio of PCB emissions at different temperatures from four light ballasts tested in a laboratory chamber. . 16
Table 2-7. Input variables for the SHEDS-multimedia model 20
Table 2-8. Key input variables for the SHEDS-multimedia model 21
Table 2-9. Key input variables for the SHEDS-multimedia model 22
Table 2-10. Average inhalation rate (m3/day) 22
Table 3-1. Aroclor analysis: recovery of Aroclor 1254 from field and laboratory controls 23
Table 3-2. Aroclor analysis: recovery of Aroclor 1254 from laboratory matrix spikes 24
Table 3-3. Aroclor analysis: recovery of surrogate analytes from samples and duplicate samples 24
Table 3-4. Aroclor analysis: total PCBs measured on field and laboratory blanks 25
Table 3-5. Aroclor analysis: total PCBs measured in laboratory method blanks 25
Table 3-6. Aroclor analysis: precision results for duplicate sample measurements 26
Table 3-7. Congener analysis: total recovery of congeners in Aroclor 1254 from field controls 26
Table 3-8. Congener analysis: total recovery of congeners in Aroclor 1254 from laboratory matrix spikes 27
Table 3-9. Congener analysis: recovery of surrogate analytes from samples and duplicate samples 27
Table 3-10. Congener analysis: total PCBs measured on field blanks 27
Table 3-11. Congener analysis: total PCBs measured in laboratory method blanks 28
Table 3-12. Congener analysis: precision results for duplicate sample total PCB measurements 28
Table 4-1. School building information 29
Table 4-2. School building remediation activity and environmental sampling summary 31
Table 4-3. School building information at sampling time points used in exposure modeling 32
Table 4-4. Caulk total PCB measurement results for schools with available data 33
Table 4-5. Caulk total PCB measurement results by school 34
Table 4-6. Interior and exterior caulk and window glaze total PCB measurement results by concentration category . .35
Table 4-7. Estimates of total PCB emission rates for several examples of PCB-containing caulk 36
Table 4-8. Screening-level comparison of predicted air concentrations resulting from PCB emissions from caulk to
measured concentrations 38
Table 4-9. PCB-containing fluorescent light ballast survey results from five schools 40
Table 4-10. PCB-containing fluorescent light ballast survey results at School 6 40
Table 4-11. Screening-level comparison of predicted air concentrations resulting from PCB emissions from fluorescent
light ballasts to measured concentrations 42
Table 4-12. Total PCB measurement results for materials at three schools with available data 43
-------
Table 4-13. Total PCB measurement results for subsets of paint uses 44
Table 4-14. Screening-level estimates of total PCB emission rates for selected interior materials in three locations at
School 2 for relative comparisons 45
Table 4-15. Screening-level estimates of total PCB emission rates for selected interior materials in three rooms
at School 3 for relative comparisons 47
Table 4-16. Screening-level estimates of total PCB emission rates for selected interior materials in three rooms
at School 6 for relative comparisons 48
Table 4-17. Summary of environmental media total PCB measurement results for six schools 53
Table 4-18. Indoor air total PCB measurement results at six schools 54
Table 4-19. Outdoor air total PCB measurement results at six schools 55
Table 4-20. Surface wipe total PCB measurement results at six schools 56
Table 4-21. Indoor dust total PCB measurement results at School 6 57
Table 4-22. Soil total PCB measurement results at six schools 58
Table 4-23. Soil total PCB measurement results for two soil depths at School 1 59
Table 4-24. Pre- and post-remediation indoor air total PCB measurement results at five schools 60
Table 4-25. Pre- and post-remediation surface wipe total PCB measurement results at five schools 63
Table 4-26. Correlations between total PCB concentrations in selected school environment samples 64
Table 4-27. Comparison of Aroclor and congener measurement results for total PCBs at School 6 65
Table 4-28. Summary of average PCB congener concentrations at School 6 66
Table 4-29. Summary of average PCB congener weight percents at School 6 68
Table 4-30. Summary of average PCB homolog weight percents at School 6 75
Table 4-31. Differences in PCB congener emission estimates from exterior caulk assuming an un-weathered Aroclor
1254 pattern vs. measured congener concentrations 78
Table 4-32. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from six
schools (units: ug/kg/day) 83
Table 4-33. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from five
schoolsa at pre-remediation and post-remediation time points (units: ug/kg/day) 83
Table 4-34. Post-remediation/pre-remediation ratios of total absorbed PCB dose estimates based on measurement
data from five schools 84
Table 4-35. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from three
schoolsa at three time points (units: ug/kg/day) 86
Table 4-36. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from
six schools in this report and from extant measurement dataa (units: ug/kg/day) 87
Table 4-37. Proportion of mean estimated total absorbed PCB dose for 6-10 year olds for
inhalation, non-dietary ingestion, and dermal absorption routes of exposure based on measurements
from five schools 88
Table 4-38. Sensitivity test results of different pulmonary absorption rates on the distributions of estimated
total absorbed PCB dose for 6-10 year olds based on pre- and post-remediation measurements
from five schools 90
Table 4-39. Sensitivity test results for post-remediation decreases in dust and soil PCB concentrations on the
distributions of estimated total absorbed PCB dose for 6-10 year olds based on measurements from
five schools 91
Table A-l. Information for the 209 PCB congeners A-l
Table B-l. Interior caulk and window glaze total PCB measurement results by concentration category B-l
Table B-2. Exterior caulk and window glaze total PCB measurement results by concentration category B-2
Table B-3. Total PCB measurement results for selected materials by school B-3
-------
Table C-l. Air sample PCB congener concentrations at School 6 C-l
Table C-2. Surface wipe PCB congener concentrations at School 6 C-3
Table C-3. Dust sample PCB congener concentrations at School 6 C-5
Table C-4. Soil sample PCB congener concentrations at School 6 C-6
Table C-5. Caulk sample PCB congener concentrations at School 6 C-8
Table C-6. Other building material sample PCB congener concentrations at School 6 C-10
Table C-7. Air sample PCB homolog weight percents at School 6 C-12
Table C-8. Surface wipe sample PCB homolog weight percents at School 6 C-12
Table C-9. Dust sample PCB homolog weight percents at School 6 C-13
Table C-10. Soil sample PCB homolog weight percents at School 6 C-13
Table C-ll. Caulk sample PCB homolog weight percents at School 6 C-14
Table C-12. Building material PCB homolog weight percents at School 6 C-14
Table E-l. Distributions of total PCB concentrations on dust estimated from
solid/air partition coefficient estimation E-2
Table E-2. Distributions of SHEDS estimates of total PCB absorbed doses using two approaches for estimating
indoor dust levels at the six schools E-3
-------
List of Figures
Figure 1-1. Three of the 209 different PCB congeners showing chlorine substitution patterns and schematic
examples of non-planar (congeners 1 and 209) and co-planar (congener 77) orientations.
Source: PubChem Compound - http://www.ncbi.nlm.nih.gov/pccompound 1
Figure 2-1. Hypothetical exposure profile for an individual 17
Figure 2-2. Overview of SHEDS residential methodology 17
Figure 2-3. General overview of SHEDS multimedia exposure model 19
Figure 4-1. Estimated total PCB emission rates from caulk in several building locations 37
Figure 4-2. Example of the estimated range of total PCB emission rates from fluorescent light ballasts in a school
classroom using the lowest, median, and highest rates from chamber tests of four ballasts 41
Figure 4-3. Screening-level estimates of total PCB emission rates from materials in the gymnasium in School 2 49
Figure 4-4. Screening-level estimates of total PCB emission rates from materials in a classroom in School 3 50
Figure 4-5. Screening-level estimates of total PCB emission rates from materials in a classroom in School 6 50
Figure 4-6. Distributions of indoor air total PCB concentrations across all six schools and at
each individual school. The box plots show the median, 25th, and 75th percentiles.
The whiskers show the 10th and 90th percentiles 54
Figure 4-7. Public health levels of PCBs in school indoor air developed in 2009 by the U.S. EPA
(http://www.epa.gov/pcbsincaulk/maxconcentrations.htm) 55
Figure 4-8. Distributions of surface wipe total PCB concentrations collected from high-contact and low-contact
surfaces across all six schools and at each individual school. The box plots show the median, 25th,
and 75th percentiles. The whiskers show the 10th and 90th percentiles 55
Figure 4-9. Distributions of outdoor soil total PCB concentrations across all six schools and at each individual school.
The box plots show the median, 25th, and 75th percentiles. The whiskers show the 10th and 90th
percentiles 59
Figure 4-10. Pre- and post-remediation indoor air total PCB concentrations across five schools and at each individual
school. The box plots show the median, 25th, and 75th percentiles. The whiskers show the 10th
and 90th percentiles 61
Figure 4-11. Pre- and post-remediation indoor air total PCB concentrations across three schools with several
different remedial activities 61
Figure 4-12. Patterns of congeners in Aroclor 1254 (top) and exterior caulk collected at School 6 70
Figure 4-13. Patterns of congeners in indoor air (top) and indoor dust collected at School 6 72
Figure 4-14. Patterns of relative weight percent for selected congeners in Aroclor 1254 (top) and Aroclor 1242 (bottom)
compared to averages for exterior caulk and indoor air at School 6 73
Figure 4-15. Patterns of relative weight percent for selected congeners in Aroclors 1254 compared to average
values for indoor air, paint, and exterior caulk (top), and indoor air, surface wipes, dust, and
soil (bottom) at School 6 74
Figure 4-16. Relative weight percents of PCB chlorine-number homologs for Aroclors 1242 and 1254 compared to the
averages of the exterior PCB-containing caulk, the indoor building materials, and indoor air 76
Figure 4-17. Relative weight percents of PCB chlorine-number homologs for Aroclors 1242 and 1254 compared to the
averages of the environmental media 77
Figure 4-18. Patterns of estimated emission rates from exterior caulk collected at School 6 assuming the
caulk contains congeners in an Aroclor 1254 proportions (top) and using congener values measured
in the caulk (bottom) 79
Figure 4-19. Patterns of estimated emission rates from exterior caulk collected at School 6 (top) versus the congeners
measured in indoor air (bottom) 80
-------
Figure 4-20. Distributions of estimated absorbed total PCB doses from exposures at school for four age groups at
five schools for pre- and post-remediation time points. The box plots show the median, 25th, and 75th
percentiles. The whiskers show the 10th and 90th percentiles 84
Figure 4-21. Distributions of estimated absorbed total PCB doses from exposures at school for the 6-10 year old
age group at five schools for pre- and post-remediation time points 85
Figure 4-22. Distributions of estimated absorbed total PCB doses from exposures at three schools for the
6-10 year old age group at three pre- and post-remediation time points across two years 86
Figure 4-23. Contributions of different exposure routes towards total estimated absorbed PCB doses for the 6-10 year
old age group at different percentiles of the total dose estimate based on measurements at six schools
(assuming 70% pulmonary absorption) 88
Figure 4-24. Contributions of different exposure routes towards total estimated absorbed PCB doses for the 6-10 year
old age group at different percentiles of the total dose estimate based on pre-remediation (top) and
post-remediation (bottom) measurements at five schools (assuming 70% pulmonary absorption) 89
Figure 5-1. Illustration of the complexity of PCBs in school buildings 96
Figure A-l. Relative congener weight percents for Aroclors 1242 and 1254 (based on Frame et al. 1996) A-7
Figure D-l. Distributions of total PCB concentrations in several environmental media collected at school and
college buildings (GM is the geometric mean) D-2
-------
Acronyms and Abbreviations
AC Air conditioning
AER Air exchange rate
ANSI American National Standard Institute
ASHRAE American Society of Heating, Refrigerating,
and Air Conditioning Engineers
ASTM American Society for Testing and Materials
(nowASTM International)
BMR Basal metabolic rate
BVA Basal alveolar ventilation
CAFO Consent agreement and final order
CHAD Comprehensive Human Activity Database
CQCS Comprehensive quantitative congener specific
analysis
DCBP Decachlorobiphenyl
GC/ECD Gas chromatography/electron capture
detector
GM Geometric mean
EPA United States Environmental Protection
Agency
FDA Food and Drug Administration
HID High intensity discharge lamp ballast
HA/ Heating and ventilation
HVAC Heating, ventilation, and air conditioning
ND Not detected
NELAP National Environmental Laboratory
Accreditation Program
NERL National Exposure Research Laboratory
NIOSH National Institute for Occupational Safety and
Health
NRMRL National Risk Management Research
Laboratory
NYC SCA New York City School Construction Authority
ORD Office of Research and Development
PCB Polychlorinated biphenyl
PPM Parts per million (equivalent to mg/kg)
PBPK Physiologically-based pharmacokinetic
PK Pharmacokinetic
PUF Polyurethane foam
QA Quality assurance
QAPP Quality Assurance Project Plan
QC Quality control
QL Quantitation limit
QSAR Quantitative structure-activity relationship
RfD Reference dose
RPD Relative percent difference
RSD Relative standard deviation
SD Standard deviation
SHEDS Stochastic Human Exposure and Dose
Simulation model
SVOC Semi-volatile organic compound
TCDD 2,3,7,8-Tetrachlorodibenzodioxin
TCMX Tetrachloro-m-xylene
TEQ Toxic equivalence (as compared to TCDD)
TSCA Toxic Substances Control Act
-------
-------
1.0
Introduction
1.1 Background
1.1.1 About PCBs
Polychlorinated biphenyls (PCBs) are synthetic chemicals
that were manufactured in the United States between about
1930 and 1977 for use in various industrial and commercial
applications because of their non-flammability, chemical
stability, plasticizer, and electrical insulation properties
(ATSDR, 2000). PCBs were used in numerous products and
processes, including electrical, heat transfer, and hydraulic
equipment; as plasticizers in various products; in paints and
finishes; in pigments, dyes, and carbonless copy paper; and
in other industrial and commercial applications (Erickson and
Kaley, 2011; U.S. EPA, 1976). There are no known natural
sources of PCBs. PCBs that were produced commercially
were mixtures of individual chlorinated biphenyl compounds
known as congeners (see Figure 1-1). There are a total
of 209 different possible PCB congeners (Appendix A),
although many congeners did not appear in commercial
mixtures.
Most of the PCB mixtures manufactured for commercial use
in the United States are known by the trade name Aroclor.
Each commercially produced Aroclor contained mixtures
of some of the 209 congeners, with chlorine contents of
the different Aroclors ranging from 21% to 68% (Table
1-1). Between 1957 and 1971, 12 types of Aroclors were
produced (ATSDR, 2000). During this time PCBs were
used in completely closed systems (such as transformers and
209 different PCB congeners
PCB1
2-chlorobiphenyl
PCB77
3,3; 4, 4'-tetrachlorobiphenyl
PCB209
decachorobiphenyl
Figure 1-1. Three of the 209 different PCB congeners showing
chlorine substitution patterns and schematic examples of non-planar
(congeners 1 and 209) and co-planar (congener 77) orientations.
Source: PubChem Compound - http://www.ncbi.nlm.nih.gov/
pccompound.
capacitors), nominally closed systems (such as hydraulic
systems and vacuum pumps), and open systems (such as
plasticizers and paints) (Erickson and Kaley, 2011). In 1970,
the primary U.S. manufacturer decided to discontinue use of
Aroclors in open products and uses that could lead to direct
transfer into the environment (Erickson, 1997). Information
about commercial PCB mixtures, PCB production, and PCB
nomenclature is provided by Erickson and Kaley (2011).
Table 1-1. Several commonly used Aroclor mixtures, their chlorine weight percent values, and their highest five
congeners by weight percent
Common
Aroclors
Aroclor 1221
Aroclor 1232
Aroclor 101 6
Aroclor 1242
Aroclor 1248
Aroclor 1254
Aroclor 1260
Aroclor 1262
Chlorine
Weight %
21%
32%
41%
42%
48%
54%
60%
62%
Number of Congeners
Measured3
63
93
71
95b
95+
95+b
93
95
Congeners
Five Highest by Weight %
1,3,8,4,15
1,8,3,4,18
18,31,28,8,33
18,31,28,8,33
70, 66, 52, 44, 31
118, 110, 105, 138, 70
180, 153, 149, 138, 187
180, 187, 153, 149, 174
a Based on Frame et al., 1996.
b See Appendix A for more information about Aroclor 1242 and Aroclor 1254 congeners.
-------
1.1.2 PCB Exposure and Effects
Manufacture of PCBs was banned in the United States
with final rules published by the U.S. EPA in May, 1979.
Their use was phased out, except for certain limited uses,
because of evidence they are persistent in the environment
and can cause harmful health effects. PCBs can accumulate
in people over time, and PCB exposure over a long period
of time may be harmful to health. Short term exposure to
large amounts of PCBs can lead to skin conditions such as
acne and rashes and may be associated with decreased liver
function, neurological effects, and gastrointestinal effects
(ATSDR 2000). These types of acute toxic effects due to
high levels of exposure are generally rare in the general
population. Chronic exposure to lower levels of PCBs may
also cause health effects (ATSDR 2000; ATSDR 2011). In
animal studies, PCBs have been shown to cause effects on
the immune, reproductive, nervous, hepatic, and endocrine
systems. PCBs have also been shown to cause cancer
in animals. Some studies in humans provide supportive
evidence for some of these health effects. Studies also show
that PCBs in pregnant women can have an impact on their
children's birth weight, short-term memory, and learning.
Because of potential neurotoxic and endocrine effects, there
is concern regarding children's exposures to PCBs (ATSDR
2000; ATSDR 2011).
PCBs are highly persistent in the environment. As such, they
are still present in soils and sediments in some locations
today and may be found at low levels in ambient air and
water. PCBs can be released into the environment from
hazardous waste sites, illegal or improper disposal of
industrial wastes and consumer products, leaks from old
electrical transformers and capacitors containing PCBs,
and burning of some wastes in incinerators. PCBs undergo
bioaccumulation and may eventually enter foods that people
consume. Foods with the highest PCB levels are typically
fish, meat, and dairy products. Dietary consumption of
contaminated foods is believed to be an important route of
background exposure to PCBs for people in the United States
(ATSDR, 2000).
1.1.3 PCB Uses in Buildings
Additional exposure to PCBs may occur for people who
spend time in buildings where PCB-containing materials and
equipment are present. A number of products manufactured
before the late-1970s contained PCBs (Erickson and Kaley,
2011; U.S. EPA, 1976). Products and equipment that may
contain PCBs include:
• dielectric fluid in transformers and capacitors,
• other electrical equipment, including voltage regulators,
switches, circuit breakers, reclosers, bushings, and
electromagnets,
• oil used in motors and hydraulic systems,
• microscope oil,
• old electrical devices or appliances containing capacitors
with PCBs,
• fluorescent light ballast capacitors,
• cable insulation,
• thermal insulation material, including fiberglass, felt,
foam, and cork,
• adhesives and tapes,
• oil-based paints,
• caulk, window glazing, and other sealants,
• plastics,
• carbonless copy paper,
• ceiling tile coatings, and
• floor finish.
Some of these materials can still be found in buildings,
particularly those constructed or renovated between about
1950 and 1978.
Production of PCBs used as plasticizers in the United States
ranged from approximately 1.4 million kg in 1957 to 8.6
million kg in 1969, decreasing to zero by 1971 (NIOSH,
1975). Caulk and other sealant materials that incorporated
PCBs as plasticizers have been examined as a potential
source of exposure to building occupants. Kohler et al.
(2005) reported on concentrations of PCBs in more than
1300 samples of joint sealants collected from buildings in
Switzerland built between 1950 and 1980. Nearly half (48%)
contained PCBs, and levels exceeding 50 ppm were found in
42% of the samples. Concentrations exceeding 10,000 ppm
were found in 21% of the samples, whereas concentrations
exceeding 100,000 ppm were found in 9.6% of the samples.
PCB content of 100,000 ppm in caulk means that the caulk
contains 10% PCBs by weight. Chlorine content was
examined in a subset of Swiss samples, and more than 90%
had results consistent with mixtures of Aroclors 1248, 1254,
1260, and 1262.
PCB mixtures were used as the dielectric fluid in fluorescent
light ballast capacitors used in standard fluorescent light
fixtures and in high-intensity discharge (HID) ballasts
manufactured in North America from the 1960s through the
late 1970s (Environment Canada, 1991). Staiff etal. (1974)
reported uses of PCBs in ballasts before 1952. The earliest
date of use is not clear, but PCB use in ballast capacitors
continued to 1978, a year of transition in which alternative
capacitor dielectric fluids began to be used. By 1979, no
ballasts capacitors were manufactured with PCBs. The
most commonly used Aroclor mixtures were Aroclor 1242
and Aroclor 1016 (which has a similar chlorine content and
congener mixture as Aroclor 1242) (Erickson and Kaley,
2011) but the use of Aroclors 1221 and 1254 in capacitors
has also been reported (U.S. EPA, 1976). According to a
citation in Staiff et al. (1974) Aroclor 1254 was primarily
used in ballasts prior to 1952 with Aroclor 1242 thereafter.
Capacitors used in ballasts for fluorescent light fixtures
containing two 4-foot lamps contained approximately 10 -
24 g of PCBs, while HID ballast capacitors could contain
between 90 and 390 g (Environment Canada, 1991; General
Electric, 2004).
-------
Table 1-2. PCB vapor pressure ranges: examples from three reports
Del le Site, 1997a
Pa at 25° C
Homolog Series (Number of Congeners)
Monochlorobiphenyls
Dichlorobiphenyls
Trichlorobiphenyls
Tetrachlorobiphenyls
Pentachlorobiphenyls
Hexachlorobiphenyls
Heptachlorobiphenyls
Octachlorobiphenyls
Nonachlorobiphenyls
Decachlorobiphenyl
7.9E-2to2.1E-1 (3)
7.4E-4 to 3.2E-1 (5)
4.8E-3 to 7.6E-2 (5)
1.8E-5to2.2E-2(5)
4.0E-4to2.2E-3(1)
2.9E-6to1.6E-3(3)
Not reported
2.4E-6to3.0E-5(1)
Not reported
2.9E-9to1.4E-5(1)
Falconer and
Bidleman, 1994b
Pa at 20° C or 25° C
Not reported
1.0E-1to2.5E-1
6.3E-3 to 4.0E-2
3.2E-3to1.6E-2
5.0E-4 to 2.5E-3
2.5E-4 to 7.9E-4
5.0E-5 to 2.5E-4
Not reported
Not reported
Not reported
Holmes et al., 1993C
Pa at 25° C
(Number of Congeners)
3.2E-1 to 9.3E-1 (3)
5.1E-2to4.2E-1 (12)
8.4E-3to1.7E-1 (24)
1.4E-3to6.6E-2(42)
2.7E-4to1.7E-2(46)
5.4E-5 to 6.4E-3 (35)
1.4E-4to1.6E-3(24)
3.8E-5to6.2E-4(12)
1.0E-4to1.3E-4(3)
2.8E-5(1)
aDelle Site, 1997: Compiled vapor pressures for selected congeners from multiple references using direct, indirect, and
prediction methods.
b Falconer and Bidleman, 1994: Liquid saturation vapor pressures; range across average vapor pressures of different levels of
othro-substituted chlorines within homolog (estimated from Figure 2).
cHolmes et al., 1993: Originally from Buckhard et al., 1985,
rather than from solid phase.
Most fluorescent light ballasts are designed for about 50,000
hours of operation under standard conditions (Natural
Resources Canada, 2009) although an earlier report suggested
approximate lifetimes ranging from 10 to 15 years of normal
service at operating temperatures of 85 - 90°C (U.S. EPA,
1976). Ballast lifetime calculations from one manufacturer
indicate that at an operating temperature of 65°C ballasts
can be expected to operate for 50,000 hours with only
10% failure, but that the failure rate is likely to increase
substantially after this period due to drying up of liquid
electrolytic fluid in the capacitor and degradation of soldered
contacts (Philips Lighting, undated). Staiff et al. (1974) gave
a normal operating ballast surface temperature of 67 - 72°C
and cited a 1966 lighting handbook stating that ballast life
was normally estimated at around 12 years. If lights are
operated in a building for 10 hours per day, 250 days per
year, then 50,000 hours would cover approximately 20 years
of use. PCB-containing light ballasts have been present in
school buildings in many cases for over 40 years.
As previously noted, PCBs were used in many types of
materials and components. It is not clear to what extent
other types of PCB-containing materials remain in buildings
today. Ceiling tiles with surface coatings containing PCBs
have been reported (personal communication, U.S. EPA
Region 1). It is possible that PCB-containing microscope oil
remains in some buildings, and other electrical or hydraulic
system components with PCBs could be present. PCBs
were sometimes used as a plasticizer in paints and other
types of coatings. While these coatings were used in marine,
military, and grain silo applications it not clear whether
ES&T, 22:503-509. Based on liquid or sub-cooled liquid values
PCB-containing paints and coatings were used extensively
in buildings. There is one report of PCBs in floor finish in a
residential application (Rudel et al., 2008).
1.1.4 PCBs in Building Environments
PCBs from caulk and other materials containing PCBs may
move into the air and dust indoors and into the soil around
older buildings, leading to the potential for exposures to
building occupants (Currado and Harrad, 1998; Kohler
et al., 2002, 2005; Priha et al., 2005; Hazrati and Harrad,
2006; Heinzow et al., 2007; Harrad et al., 2010; Herrick et
al., 2004, 2007; Zhang et al., 2011; Macintosh et al., 2012).
The different vapor pressures of the 209 PCB congeners and
the effects of weathering over 40 or more years may affect
which congeners are present and available for exposure from
different environmental media (Heinzow et al., 2007; Harrad
et al., 2009). The extent of exposure to PCBs in indoor
environments may depend on their vaporization into indoor
air in combination with degradation of materials resulting
in contaminated particles available for contact. Many
researchers have measured or estimated PCB congeners
vapor pressures, and results vary considerably depending on
the method (Erickson, 1997). In general, congener vapor
pressures decrease with increasing levels of chlorination
(Table 1-2, Appendix A). Within chlorine-number homolog
groups, vapor pressures increase with increasing levels of
chlorination in ortho positions (Falconer and Bidleman,
1994). Multimedia fugacity modeling has been applied to
PCBs in residential and office environments to examine
emissions and fate in indoor environments
(Zhang etal., 2011).
-------
Based on vapor pressure characteristics, it might be
anticipated that inhalation exposures to PCBs will be
weighted to congeners with higher vapor pressures and lower
chlorine numbers. However, PCBs in indoor environments
may also be bound to particles that can become suspended in
the indoor air (Zhang et al., 2011). PCB congeners measured
in indoor air may not match congener patterns found in the
Aroclor mixtures used in materials in a building because
of the differences in congener vapor pressures and because
congeners are absorbed onto particles differentially based on
their vapor pressures and solid/air partition coefficients.
Harrad et al. (2009) noted that the combination of residential
indoor air inhalation and dust ingestion could exceed dietary
intake of PCBs in some scenarios. Exposures to PCBs in
air and dust from contaminated nonresidential buildings
could substantially increase exposures above those seen in
residential environments. For example, levels of PCBs in the
indoor air of some office buildings with PCB-contaminated
sealants (up to 6000 ng/m3 [Kohler et al., 2005]) exceed
the levels reported in residential indoor air (maximum of
14 ng/m3 total PCBs [Harrad et al., 2009] and 35 ng/m3 for
congeners 52, 105, and 153 in a contaminated home [Rudel
etal.,2008]).
Of particular concern is the potential for school children's
exposures to PCBs in older school buildings. Schools
constructed or renovated between the 1950s and 1970s may
contain both lighting fixtures with PCB capacitors and caulk
or other sealants that incorporated PCBs as a plasticizer.
Caulk containing PCBs may have been used around exterior
windows and doors, exterior building joints, and in interior
locations. PCBs may vaporize from the caulk and become
airborne. PCBs may be absorbed onto (or into) other
building surfaces, materials, or dust (Herrick et al., 2004).
Caulk may degrade or suffer abrasion wear that can create
PCB-containing dust that is then available for transport in
indoor areas. PCBs from exterior caulk may be transported
into soils near the building (Herrick et al., 2007).
7.1.5 PCBs in School Buildings
There have been few systematic efforts to characterize
PCB sources and environmental levels at schools across
the United States. Measurements of PCBs in caulk have
been made for several college buildings and a number of
primary and secondary schools. Environmental samples,
including air, dust, wipe, and soil samples, also have been
collected at a number of school buildings. Measurement
results for PCBs in schools in the United States have not
been widely published in the scientific literature, although
Herrick et al. (2007) reported on PCBs in soil collected near
buildings with PCB-contaminated caulk or joint material
and Macintosh et al. (2012) published on environmental
levels and remediation results for a primary school building.
Measurement results have been reported in presentation
and report formats (e.g., Coghlan et al., 2002; Sullivan,
2008; TRC, 2006, 2008, 2009). From available extant data
examined through 2009 (Appendix D), total PCBs in caulk in
U.S. college and school buildings ranged from not detected
to over 200,000 ppm. Concentrations in school indoor air
ranged up to approximately 1000 ng/m3. Total PCB levels
in dust ranged up to approximately 80 ppm. Wipe samples
ranged up to approximately 1 ug/100cm2 total PCBs. Soil
concentrations in samples collected next to buildings ranged
up to approximately 80 ppm of total PCBs.
Although caulk is believed to be a primary source of PCBs
in some older schools, there is still considerable uncertainty
regarding the extent to which PCBs in other materials used in
schools might contribute to exposures (Coghlan et al., 2002).
Other primary sources (materials manufactured with PCBs,
or materials to which PCBs were added during construction)
or secondary sources (sources that have absorbed PCBs
emitted from primary sources) may be present in some
schools. For example, window glazing has been found in
several locations to contain levels of PCBs greater than
50 ppm. Fluorescent light ballasts with PCB-containing
capacitors remain in some older buildings. Secondary
sources might include surfaces, materials, and dust that
have been contaminated through transport and absorption
of PCBs from caulk or other primary sources. An in-depth
investigation in a high school found PCBs in numerous
materials, including but not limited to laminate adhesive,
mastics, paint, gasket, carpet, foam padding, and bulk dust
(TRC, 2008; Sullivan, 2008; TRC, 2009). Some of these
materials had concentrations exceeding 10 ppm, ranging up
to more than 250 ppm. The potential for secondary PCB
sources to contribute to exposures to children and other
building occupants is not well understood. To make sound
decisions regarding mitigation of exposures to PCBs in
schools, it is important to understand the range of potential
sources of PCBs in schools; their contributions to PCBs in
air, dust, soil and on surfaces; and the magnitude of potential
exposures to children and staff in school environments.
1.2 Need for Research
In September 2009 the U.S. EPA announced new guidance
for school administrators and building managers for
managing PCBs in caulk and to help minimize possible
exposure. However, there was limited information on
PCBs in school buildings in the United States. Neither the
sources, including PCB-contaminated caulk, light ballasts,
and secondary sources; nor the potential magnitude and
routes of exposure have been systematically characterized in
schools. There remains considerable uncertainty regarding
the extent to which children and staff members may
exposed to PCBs in school environments. Research was
needed to help fill these information gaps to improve our
understanding of exposure to PCBs in schools. Therefore,
-------
the EPA also announced in 2009 that research would be
performed by the Office of Research and Development's
(ORD) National Exposure Research Laboratory (NERL)
and the National Risk Management Research Laboratory
(NRMRL) to further study this issue. The research was
intended to help fill information gaps and improve our
understanding of PCBs in school buildings and approaches
for mitigating exposures. Information on sources of PCBs
and levels in school environments is needed to improve risk
management decision-making. NERL developed a research
plan to better understand and characterize PCB sources,
emissions, environmental concentrations, and exposures in
school environments (U.S. EPA NERL, 2010). A research
plan was also developed by NRMRL to perform laboratory
studies of PCB sources and transport and to evaluate selected
mitigation approaches (U.S. EPANRMRL, 2010).
1.3 Research Objectives
In order to better understand the significance of PCB-
containing caulk and other building materials and
components as a source of PCB exposures to children,
teachers, and staff in school buildings, NERL planned
research to utilize a limited set of real-world measurements
to:
1) characterize PCB-contaminated caulk and other
potential primary and secondary sources of PCBs in
school buildings,
2) characterize levels of PCBs in school air, dust, soil,
and on surfaces and to investigate relationships
between potential PCB sources and environmental
levels,
3) apply an exposure model for estimating children's
exposure to PCBs in schools with PCB sources,
4) evaluate which routes of exposure (e.g., inhalation,
contact with surfaces or dust) are likely to be most
important, and,
5) provide information to assist in developing risk
management practices for reducing exposure to PCBs
in schools.
1.4 About this Report
A limited set of measurement data and information
collected using a systematic approach at six schools was
used to characterize primary sources of PCBs, to evaluate
whether secondary sources were present and their relative
importance, to describe PCB concentrations in school
environmental media, and to prepare modeled estimates
of exposure and characterize the relative importance of
different routes of exposure. Section 2 describes the
methods for sample collection, sample analysis, emissions
modeling, and exposure modeling. Section 3 provides
information regarding quality control and quality assurance.
Section 4 describes the results for source characterization,
environmental media PCB concentrations, and exposure
modeling for several child age-groups. Conclusions are
presented in Section 5.
-------
-------
2.0
Methods
2.1 Generation of School Measurements
NERL's objective was to recruit up to nine schools to
participate in the research by allowing NERL and contractor
scientists to collect air, surface wipe, dust, soil, caulk,
and other building material samples in multiple indoor
and outdoor locations (U.S. EPANERL, 2010). With the
assistance of EPA's Office of Science Policy and several EPA
Regional Offices, contacts were made with numerous state
and local educational agencies and school districts. Although
there was often interest in understanding the problem of
PCBs in school buildings at state and local levels, most
organizations decided not to participate in the research study
due to a variety of potential concerns and issues. NERL
reached preliminary agreement with two school districts to
participate in the study. But, due to factors outside of their
control, the NERL researchers ultimately were not able to
enroll schools from these districts into the study. NERL was
able to implement the research plan and collect samples at
one school building scheduled for demolition.
At the same time that NERL was planning research on PCBs
in schools, the New York City School Construction Authority
(NYC SCA) reached an agreement regarding the assessment
and remediation of caulk containing PCBs in public school
buildings with the United States Environmental Protection
Agency Region 2, under a Consent Agreement and Final
Order (CAFO, Docket Number TSCA-02-2010-9201). The
goal of the CAFO was to develop a city-wide approach to
assessing and managing caulk containing PCBs in schools
built between 1950 and 1978. As a result of the agreement,
New York City initiated a comprehensive pilot study during
the summer of 2010, when students were absent, to evaluate
the possible presence of PCB-containing caulk and preferred
remedial actions in three schools, with evaluations for two
additional schools conducted in 2011.
A remedial investigation plan was developed by the SCA
and TRC Engineers Inc. describing the selection of the pilot
schools, the approach for measuring PCBs in and around
school buildings, and the caulk remediation approaches to be
investigated (NYC SCA, 2010). The sampling and analysis
approach was similar in many respects to the NERL research
plan and sampling strategy but added remedial investigation
steps. Pre-remediation samples of caulk, indoor and
outdoor air, indoor surface wipes, and soils were collected
in and around three NYC elementary schools during 2010.
Different remedial approaches were then instituted at each
school including caulk patch and repair, caulk removal and
replacement, and caulk encapsulation. Post-remediation
indoor and outdoor air and indoor surface wipe samples
were collected to evaluate the remedial effect on PCB levels
in the school environment. Analysis of both the pre- and
post-remediation air samples showed levels of PCBs at
some indoor locations were greater than Public Health
Levels recommended by the U.S. EPA (http://www.epa.gov/
pcbsincaulk/maxconcentrations .pdf). Several actions were
subsequently taken at the three schools to investigate and
reduce the elevated PCB concentrations in air. Examination
of materials that could potentially serve as additional
sources of PCBs was conducted. As part of this effort it was
determined that PCB-containing fluorescent light ballasts
were present throughout the schools. An additional set of air
samples was collected at each school following removal of
the PCB-containing light fixtures and a period of ventilation
with outdoor air. The NYC SCA remedial investigation
continued with two additional pilot schools in 2011, and
additional work was also performed in the three previous
schools. Investigation reports and measurement results are
available at the NYC PCB Program web site maintained by
the NYC SCA (http://www.nycsca.org/Community/Programs/
EPA-NYC-PCB/Pages/default.aspxI
U.S. EPA Region 2 requested the assistance of EPA's
NERL in characterizing potential exposures associated with
environmental levels of PCBs measured at the New York
City Schools. Region 2 specifically requested application of
NERL's Stochastic Human Exposure and Dose Simulation
(SHEDS) model for estimating multi-pathway exposure
distributions. Region 2 also requested assistance in
understanding and characterization of primary and potential
secondary sources of PCBs. Because the NYC SCA data
were collected in a systematic fashion using strategies
and methods similar to the NERL research plan, it was
determined that the measurement results could be used to
address the Region 2 interests and to meet many of the NERL
objectives for source and environmental characterization and
exposure modeling. Measurement results from the five NYC
schools and one school sampled by NERL are used in this
report.
2.2 Sample Collection
The sample collection strategy and methods used by the NYC
SCA in the pilot remedial investigation efforts are described
elsewhere (NYC SCA, 2010). Both the approach and the
sampling methods used by the SCA were generally similar to
those used by NERL.
Methods for sample collection and analysis used by NERL
are summarized in Table 2-1; these methods are described in
more detail in the following sections. The samples collected
by NERL at one school are summarized in Table 2-2.
NERL's contractor, Alion Inc., was responsible for preparing
-------
for and conducting field sampling operations. They were
joined at the field site during sample collection activities by
NERL scientists.
2.2.1 Indoor and Outdoor Air Sample Collection
U.S. EPA Method TO10A was used to collect total PCBs in
air, using a low-volume sampling approach to minimize the
size and noise of the pumps to be used in school buildings.
Sample filters were pre-cleaned open-cell polyurethane foam
(PUF) in 30-mm x 70-mm tubes. Total suspended particle
quartz filters were used as part of the sample filter assembly;
however, the filter and the PUF were extracted and analyzed
together as a single sample. Separate particle- and gas-phase
air concentrations were not obtained. Collection of inline
backup filters to assess potential break-through were not
used based on results from the NYC SCA sampling, which
showed no breakthrough in almost all cases.
Sample collection tubes were situated with the inlet in a
downward facing position at a height of 1 m from ground
or floor level. A sampling stand was used to secure the
sample filter. An active air-flow pump, capable of unattended
24-hr operation was used to provide a flow of 3.5 to 5.0 liters
per minute (1pm) through the PUF. Flow measurements
were performed and recorded at the initiation of sampling
and then again at the completion of the nominal 24-hr
sampling period. Start and stop times were recorded, so
that a cumulative amount of time at an average flow rate
could be calculated to yield the volume of air sampled.
Total air volumes sampled through the filters ranged from
5.3 to 6.5 m3. Indoor and outdoor air temperatures were
measured periodically through the sample collection period.
2.2.2 Surface Wipe Sample Collection
Two types of indoor surface wipe samples were collected.
The first type was individual samples collected from two
different surfaces that might be contacted routinely by a
student. These "high-contact" surface sampling locations
included the tops of desks or tables. The second type was
individual samples collected from building surfaces. These
"low-contact" surface sampling locations included floors,
walls, and window sills (where present). Several wipe
samples were also collected from light fixtures including
from the outside cover, diffuser, exterior of the ballast cover,
and interior of the ballast cover.
Wipe samples were collected based on the wipe sample
collection procedure described in ASTM Method
D6661-01 [2006], "Standard Practice for Field Collection of
Organic Compounds from Surfaces Using Wipe Sampling."
PCB-free gauze wipes were wetted with 5 mL of hexane
to collect surface wipe samples. Each wipe sample was
collected in a 100-cm2 area as defined by a template secured
to the surface. Transfer of PCBs from the surface occurred
through physical wiping of the defined surface area while
applying moderate pressure in a serpentine pattern. Wipes
were handled using appropriate chemical-resistant gloves,
followed by storage in a clean amber glass jar with a Teflon-
lined cap.
2.2.3 Dust Sample Collection
There is the potential for loose (accumulated and visible)
dust to be present on surfaces in school rooms. A sample
size of 2 g for each sample was preferred, whereas 1 g was
considered to be the minimum sample size. To obtain
Table 2-1. Summary of sample collection and analysis methods
Sample
Type
Air
Sample Collection
Method
EPA Method TO-10A
Sample
Extraction
Method
EPA Method 3540C
Aroclor Sample
Analysis Method
EPA Method 10A/EPA 8082
Congener Sample
Analysis Method
NEACQCS Method
(Sampling duration
approx. 24-hr)
(Modified EPA Method 8082)
Surface wipe
ASTM D6661-01
EPA Method 3540C
EPA SW-846 8082
NEACQCS Method
(Modified EPA Method 8082)
Dust (indoor)
Soil (outdoor)
Research Operating
Protocol3: Procedure for
Collecting Loose Dust
for PCB Analysis
Research Operating
Protocol: Procedure for
Collecting Soils for PCB
Analysis
EPA Method 3540C
EPA Method 3540C
EPA SW-846 8082 NEA CQCS Method
(Modified EPA Method 8082)
EPA SW-846 8082 NEA CQCS Method
(Modified EPA Method 8082)
-------
Table 2-1. Summary of sample collection and analysis methods (continued)
Sample
Type
Caulk and
window glaze
Sample Collection
Method
Research Operating
Protocol: Procedure for
Screening Collection
of Caulk and Window
Glazing From Buildings
for PCB Analysis
Sample
Extraction
Method
EPA Method 3540C
Aroclor Sample
Analysis Method
EPA SW-846 8082
Congener Sample
Analysis Method
NEACQCS Method
(Modified EPA Method 8082)
Bulk
materials'"
Research Operating
Protocol: Procedure
for Collecting Select
Materials from Buildings,
Fixtures, and Associated
Items for PCB Analysis
EPA Method 3540C
EPA SW-846 8082
NEACQCS Method
(Modified EPA Method 8082)
a Research operating protocols were developed for the U.S. EPA/ORD research study on PCBs in schools.
b Bulk materials include paints, floor tiles, ceiling tiles, foam, cove molding, etc.
Table 2-2. Sample collection information (numbers of samples by type and location)3
Collection
Location
Air
Surface Interior Exteri-
Wipe Dust or Soil
Interior Caulk
and Glazing
Exterior
Caulk
Bulk
Materials0
Indoor Locations
Classroom 1
Lab Classroom 2
Classroom 3
Classroom 4
Shop Classroom 5
Cafeteria
Gymnasium
1
1
1
1
2b
2b
6
7b
7
6b
6
6
5b
1
1
2b
2b
1
1
1
3b
3
6b
4b
2
5b
2
20
12
15
Outdoor Locations
Rear side of building
Front side of building
50' from front side
5b
7b
Total Number of Samples
11
42
25
12
47
a Quality control field blank and field control samples are not included in this table.
b Includes duplicate samples.
0 Bulk materials include samples of building materials and furnishings.
-------
suitable sample sizes in most rooms, multiple surfaces had
to be sampled including surfaces not ordinarily contacted,
including cabinet tops and the tops of door molding. Where
sufficient quantities of dust were available, the dust was
collected using a cartridge filter assembly (pre-cleaned dust
collection cassettes, Forensic Source part # 4-3109) and
specialized vacuum system (3M Trace Evidence Vacuum
System, Forensic Source part # 4-3005). Following
collection, the dust was transferred from the filter into a
clean amber vial with Teflon-lined cap. Dust samples were
collected following the completion of air sample collection.
2.2.4 Outdoor Soil Sample Collection
Soil samples were collected at two locations adjacent to
the building, below one joint containing caulk and below a
window assembly with exterior caulk. At each of the two
locations, soil samples were collected at three distances from
the building: 0.15, 0.91, and 2.44 m (0.5 ft, 3 ft, and 8 ft).
Each soil sample was collected to a 5-cm (2-inch) depth after
carefully removing any vegetation. A total sample size of
approximately 100 g was collected at each sampling location.
A pre-cleaned stainless scoop was used to collect the soil.
The soil was placed into a clean amber glass container with a
Teflon-lined cap.
Another soil sample was collected from an area further from
the school at a location that could be contacted by students.
In this case, the sample came from an island in the parking
area in front of the school. This sample was not intended
to fully characterize a school yard but is intended to assess
whether there is a potential for soil exposures in areas that
would need to be considered in exposure modeling efforts.
An area 3.05 m x 3.05 m (10 ft x 10 ft) was identified. Soil
samples were collected from the 0 to 5-cm depth at five
locations (corners and center) within the designated area.
Soil samples were combined into a single container and
mixed. A total sample size of approximately 100 g was
collected.
2.2.5 Caulk Sample Collection
Caulk samples (including caulk, window glaze, and joint
sealant) were collected from interior locations in which
environmental samples were collected and from selected
exterior locations including from around window frames,
building joints, and an entranceway joint. Sample collection
generally followed the standard operating procedure,
"Sampling Porous Surfaces for Polychlorinated Biphenyls
(PCBs)" (U.S. EPA, 2008). However, only one sample of
each selected type of caulk was collected, rather than three
to form a composite. Duplicate samples were collected at
designated locations to examine variability. Also, a sample
size of 2 to 4 g was collected, rather than 10 g. Caulk
samples were collected by physically removing sections of
caulk using clean knives, scalpels, and tweezers (or other
clean implements, as needed). Pieces of caulk were removed
from the site of interest and placed in a single pre-cleaned
amber glass jar with a Teflon-lined cap. Caulk samples were
collected as they existed in the selected rooms; no emphasis
was made on collecting samples of deteriorating caulk
because intact caulk potentially may contain higher levels of
PCBs. Attempts were made to collect a sample of each type
of caulk, window glazing, and joint sealant in each room.
Additional descriptive information was recorded including
its location, use, current condition, and the presence of
any paints or coatings. The length, width, and depth of the
sampled caulk was also measured and recorded. Caulk
samples were collected after air sampling was completed.
2.2.6 Other Material Sample Collection
Selected materials other than caulk were collected from
three rooms in the school building (two classrooms and the
gymnasium), when they were available and accessible, as
potential primary or secondary sources of PCBs in school
buildings. Where possible, 10 g of material was the preferred
material sample size, with a minimum sample size of 2 g.
Samples were collected using pre-cleaned instruments such
as a scalpel, razor, spatula, utility knife, paint scraper, putty
knife, or other hand tool, as needed. Wall concrete block
was collected from three distances from an interior joint
caulk, with the material collected up to 0.5-inch depth using
an impact drill. Materials collected for PCB analysis were
stored in clean amber glass containers with Teflon-lined lids.
Material samples were collected following the completion of
air sample collection.
2.2.7 Light Ballast Survey
Fluorescent light ballasts were surveyed in the rooms in
which samples were collected to determine if PCB-containing
ballasts might be present. A visual survey of a subset of
fixtures was conducted by removing the light fixture lamps,
removing the ballast cover, and examining the ballast label.
Information from the ballast label was recorded. Many of
the ballast labels stated "No PCBs" and were considered
to be PCB-free. A smaller number of ballast labels did not
state "No PCBs" and these ballasts were considered likely to
contain PCBs. Ballasts were not surveyed in the gymnasium.
2.2.$ Sample Transport and Storage
All samples were transported and stored for analysis under
conditions appropriate for minimizing contamination by
PCBs or losses of PCBs. At the field collection site all
samples were stored in coolers with ice packs sufficient to
maintain a temperature of approximately 4°C. Samples
were transported to the laboratory with ice packs sufficient
to maintain a temperature of approximately 4° C. Storage
of samples and sample extracts at the analytical laboratory
was at general freezer temperatures of approximately -20°C.
Samples with potentially high levels of PCBs (such as caulks
or other primary PCB source materials) were stored and
transported in separate coolers to minimize any potential
cross-contamination of samples with low levels of PCBs
(such as air and wipe samples).
-------
2.3 Sample Extraction and Analysis
All samples were analyzed by NEA, a Division of Pace
Analytical Services, Inc. NEA is a National Environmental
Laboratory Accreditation Program (NELAP) accredited
laboratory and maintains certifications for New York
(EPA: NY00906, ELAP: 11078), New Jersey (NY026),
Connecticut (PH-0337), Massachusetts (M-NY906), and
North Carolina (668).
2.3.1 Sample Extraction
All samples were extracted following EPA Method 3540C.
Method 3540C is a Soxhlet procedure for extracting
non-volatile and semi-volatile organic compounds from
solid materials. The surrogate recovery compounds
tetrachloro-meta-xlyene (TCMX) and decachlorobiphenyl
(DBCP - a PCB congener) were added to the sample media
prior to extraction. Sulfuric acid, Florisil, and mercury shake
clean-up steps were used to prepare extracts for analysis.
Extracts were concentrated to different final volumes
depending on the sample type and detection limit goals.
2.3.2 Sample Analysis for Aroclors
All sample extracts were analyzed for Aroclor mixtures
following EPA Method 8082 using dual column gas
chromatography with electron capture detection. Multi-
level calibrations were performed for Aroclors 1016, 1221,
1232, 1242, 1248, 1254, 1260, 1262, and 1268. Calibration
levels of 20, 100, 250, 500, and 1000 parts per billion
were used for each Aroclor. Continuing calibration check
standards were analyzed periodically to verify calibration
stability. Sample extracts were injected on two different
gas chromatography columns, Phenomenex, ZebronZB-1,
30 m, 0.25 mm ID, 0.25 um and Phenomenex, Zebron
ZB-5, 30 m, 0.25 mm ID, 0.25 um. Some sample extracts
required dilution to bring the extract concentration into
the calibration range. Aroclor quantitation was based on
relative response factors for five chromatographic peaks
following electron capture detection. A 40% relative percent
difference acceptability limit was used for comparing the
results obtained for the two chromatography columns. The
highest value from the two columns was used for reporting
when the acceptability criterion was met. In some cases
sample results were reported as exhibiting an altered Aroclor
pattern. In some cases, particularly for indoor air samples,
more than one Aroclor (typically Aroclors 1248 and 1254
for air samples) was used to report an altered PCB pattern
exhibited by the sample. Actual Aroclor 1248 was not present
in the samples, but was reported to more accurately quantify
PCB present in sample that had undergone environmental
alteration. When multiple Aroclor values were reported, they
were summed to generate total PCB concentrations in the
samples. The laboratory used procedures to estimate relative
contributions from different Aroclors for shared congeners
so as not to overestimate the concentration attributed to each
Aroclor mixture. In a few cases for samples with very high
concentrations (exterior caulks for example) the dilution was
so large that it was not possible to determine the surrogate
recoveries. A final concentration was calculated using the
extract concentration and applying air volume, surface area,
or solid material mass and dilution factors as appropriate.
2.3.3 Sample Analysis for PCB Congeners
All of the air samples and a subset of other samples types
were analyzed for PCB congeners using an NEA adaptation
of EPA Method 8082 titled "Comprehensive Quantitative
Congener Specific" (CQCS) analysis. The same sample
extracts that had been previously analyzed for Aroclors were
also analyzed using the CQCS method. The method utilizes
high resolution gas chromatography with electron capture
detection. Multi-level calibrations were performed using
calibration mixtures containing all 209 PCB congeners.
Calibration levels of 0.1, 0.5, 5.0, 25, and 50 ng/mL were
used for each congener. 1,2,3,4,5,6,7,8-octachloronaphtalene
was added to all calibration solutions and sample extracts
as the internal standard. Continuing calibration check
standards were analyzed periodically to verify calibration
stability. Calibration and sample extracts were injected
onto a Chrompack CP-Sil5/C18, 50 m length, 0.25 mm ID,
and 0.10 micron phase thickness gas chromatography
column. Of the 209 congeners, approximately 62 do not
occur in Aroclor mixtures or are present at trace level
(<0.05% by weight percentage). The chromatography
column resolves 146 chromatographic peaks, providing
complete separation for 99 congeners and co-elution of
congeners for the remaining 47 chromatographic peaks.
Of those 47 peaks with co-eluting congeners, 20 include a
congener that is not found in Aroclor mixtures. Congener
quantitation was based on relative response factors for each
chromatographic peaks following electron capture detection
Peaks which are known to represent co-eluting congeners
were mathematically quantitated as individual congeners
using established ratios from Frame et al. (1996). A final
concentration of each congener was calculated using the
extract concentration and applying air volume, surface area,
or solid material mass and dilution factors as appropriate.
Chlorine-number homolog concentrations were calculated by
summing the individual congener concentrations belonging
to each homolog group.
2.4 Caulk Emissions Modeling
Caulk containing high concentrations of PCBs has been
found in older buildings, including school buildings. It
is of interest to understand the potential impact that PCB-
containing caulk may have on indoor air levels of PCBs in
school buildings. Total emission rates of PCBs from PCB-
containing caulk collected at six locations at two schools
were estimated using emission parameters generated from
laboratory chamber emission rate test results for caulk
described by Guo et al. (2011). The laboratory testing was
performed using microchambers to minimize loss of PCBs to
chamber surfaces. There were a limited number of locations
that had both PCB-containing caulk and the information
available to calculate parameters used in the estimation,
including the surface area of the caulk that was present and
the room volume. These six locations were used as examples
to highlight a range of emissions that might be encountered.
Emission rates of chemicals from solid materials are
primarily controlled by the chemical's solid/air partition
coefficient and the diffusion coefficient for the chemical
-------
in the solid material. Guo et al. (2011) described the
relationship between a normalized emission factor and a
chemical's vapor pressure:
lnNK=b1+b2lnPi (Eq. 2.1.1)
where
NE. = emission factor for congener i (ug/m2/hr)
normalized to a constant (1000 ug/g) concentration of a
PCB congener in a material.
R = vapor pressure for congener i (torr)
bj, b2 = constants
Guo et al. (2011) measured the emission rates for ten PCB
congeners (17, 52, 66, 77, 101, 105, 110, 118, 154, 187) from
12 samples of caulk, each tested from 4 to 7 times at 23°C, to
generate information for calculating the emission constants
(bj and b2 in Equation 2.1.1). Both constants (bl and b2) were
observed to be consistent among different caulk samples,
which indicated that a single correlation could be applied
to all caulk samples. The average values for bl and b2 were
14.02 and 0.976, respectively (Guo et al., 2011), resulting in
the following relationship:
In ME. = 14.02 + 0.976 In P. (Eq. 2.1.2)
Equation 2.1.2 can be used to estimate the emission rate for
a PCB congener from caulk when its concentration in the
caulk is known and the vapor pressure is known or can be
estimated.
PCB Aroclor concentrations were measured in the caulks
from six locations for which the surface area of the caulk
at that location could also be calculated. Aroclor 1254 was
the reported Aroclor in each caulk. The concentration of
each congener in the caulk was estimated by multiplying the
weight fraction of each congener as reported by Frame et al.
(1996) for the G4 Aroclor 1254 standard (see Appendix A) by
the total measured Aroclor concentration in the material.
Several sets of vapor pressure data for PCBs congeners
are available in the literature for many of the 209 PCB
congeners. Two publications were used to obtain vapor
pressures for as many congeners as possible and for all
of the congeners reported by Frame et al. in Aroclor 1254
(Fischer et al. 1992, Methods A and B; Foreman and
Bidleman 1985, Methods A and B). For some congeners,
vapor pressure data have not been reported in the literature
but have been estimated using physical/chemical properties
approaches [U.S. EPA Estimation Programs Interface Suite™
(MPBWIN™ model as accessed via ChemSpider)]. Vapor
pressures were selected for emission rate estimation in the
following descending priority scheme, where all of the
congener vapor pressures available from one source were
used prior to moving to the next source:
1. Fischer etal. Method B,
2. Fischer etal. Method A,
3. Foreman and Bidleman Method A
4. Forman and Bidleman Method B, and
5. EPI Suite™.
Examples of PCB congener vapor pressures are described in
Table 2-3 (for a more complete list see Appendix A). The
congeners in Table 2-3 are those that were measured during
laboratory chamber emissions measurement testing (Guo et
al, 2011). There is a large range in vapor pressures among
the congeners, with congener 17 having a vapor pressure
approximately 200-fold higher than congener 187.
The calculation to estimate the total emission rate of PCBs
from a material in a room includes multiple steps. The steps
are described below along with an example calculation:
Table 2-3. Vapor pressures for selected target congeners in Aroclor 1254
PCB Congener
17
52
66
77
101
105
110
118
154
187
Number of Chlorines
3
4
4
4
5
5
5
5
6
7
P (torr)
5.82x1 0-4
1.50x10-4
4.42x1 0-5
1.43x10-5
2.99x1 0-5
5.82x1 0-6
1.68x10-5
8.42x1 0-6
1.36x10-5
2.79x1 0-6
-------
Step 1: Calculate the concentration of each congener in a
material [X. in (ug/g)]:
Xj = Weight percent of congener in
Aroclor 1254 x Aroclor 1254 concentration
Congener 52 example in a gymnasium door
frame caulk:
X = 5.38% x 117,000 ug/g = 6290 ug/g
Step 2: Obtain the vapor pressure for the congener
Congener 52 example: 1.5 xlO'4torr
(Fisher et al. 1992, Method B)
Step 3: Calculate the normalized emission factor (ME.) from
Equation 2.1.2:
Congener 52 example:
InNE. = 14.02 + 0.976 In 1.5 x 10'4 = 5.42
NE. = 226 (ug/m2/hr)
Step 4: Convert the normalized emission factor to the
emission factor (E):
E. = (NE. - 1000 ug/g) x x.
Congener 52 example:
E = (226 ug/mVhr-^- 1000 ug/g) x
6290 ug/g = 1420 ug/m2/hr
Step 5: Multiply the emission factor by the surface area
of the material in the room to obtain the total estimated
congener emission rate from the material in that room:
R (ug/hr) = E (ug/m2/hr) x S (m2)
Congener 52 example:
R = 1420 ug/m2/hr x 0.143 m2 = 203 ug/hr
Step 6: Sum the estimated emission rates across all of the
congeners expected to be present in Aroclor 1254 (based on
Frame etal., 1996):
(Eq.2.1.3)
R.
n
4=1*,
/ = !
Step 7: Adjust the total estimated Aroclor 1254 emission
rate. Equation 2.1.3 estimates the total Aroclor 1254
emission rate assuming that the congener proportions in the
caulk are equivalent to those in unaged Aroclor 1254. As
shown in Section 4.5, caulk that has been in a building for
over 40 years has an altered congener pattern. Based on
congener measurements for caulk in School 6, Table 4-31
shows that the total emission rate was overestimated by a
factor of 1.8 assuming unaged Aroclor 1254 as compared
to the actual congener proportions measured in the caulk.
Another factor to consider is that the caulk chamber testing
might have overestimated emission rates due the presence
of freshly cut surfaces. Three tests used to examine this
(Guo et al., 2011) showed an average of 19% higher
emissions for freshly cut surfaces in chamber testing. The
two factors, when combined (0.55 x 0.81 = 0.45), yield
an adjustment factor of 0.45 that can be applied to total
emissions estimates:
This approach assumes that the emission factors generated
for the emissions of PCBs from caulk in the laboratory test
chamber experiments are applicable to the caulks sampled
in the building. It also assumes that the temperature and air
flow conditions used in chamber testing are applicable to the
conditions in the rooms.
Guo et al. (2011) reported the predictive errors for the P-N
calculation method for the ten congeners measured in their
chamber emission rates studies. The error was calculated
by using Equation 2.1.4, and the results are presented in
Table 2-4. It is anticipated that the average error would be
applicable across estimated total PCB emission rates using
the P-N calculation method.
£ =
-F
^
E,.
xlOO%
(Eq.2.1.4)
where
e = predictive error (%)
Ep = predicted emission factor (ug/m2/hr)
Em = measured emission factor (ug/m2/hr)
Table 2-4. Predictive error for the P-N correlations3
Congener ID
P-N Correlation Error
#52
#66
#101
#105
#110
#118
#154
Average
30.0%
40.8%
32.3%
21.6%
31.2%
29.7%
59.4%
35.0%
a (Guo etal. 2011 Eq. 4.6)
Screening-level estimates of the concentration of PCBs in
indoor air that might result from emission from caulk were
generated for four school building locations. The following
equation, including the adjustment factor described in Step
7 above, was used to estimate steady-state indoor air PCB
concentrations resulting from emissions from caulk, if there
was only one type of caulk with PCBs in the location:
c =
Q
X0.45 (Eq.2.1.5)
R
where:
C = total PCB concentration in room air (ug/m3)
R. = emission rate from caulk from congener i (ug/hr)
n = number of congeners with emission rate estimates
Q = ventilation rate for the room with outdoor air (mVhr)
(based on room volume x ACH)
"totadj
-------
If there were multiple kinds of caulk with PCBs in a location,
the following equation was used:
c =
Q
xO.45
(Eq.2.1.6)
where
R = emission rate for congener i from source j
m = number of congeners with emission rate estimates
n = number of caulk types in the room
The estimates are based on numerous assumptions, and
therefore should be considered only screening level.
Assumptions include:
• well mixed air in room
• constant temperature
• temperature equivalent to chamber conditions that
generated caulk emission parameters (23 °C)
• constant ventilation rate
• steady-state emission
• steady state and approximately equal absoption/
desorption of PCBs in other materials in the room
• no chemical reactions of PCBs
• PCBs from other school spaces are not impacting the
levels in air for the room of interest
• emission parameters for caulk in the room are the same as
for the caulk tested in lab chambers
2.5 Light Ballast Emissions Modeling
PCB-containing fluorescent light ballasts remain in operation
in some older buildings, including school buildings. The
capacitors in these ballasts are not perfectly sealed and small
amounts of PCBs can be emitted from intact capacitors (Guo
et al., 2011). Measured emission rates from apparently intact
light ballasts are highly dependent on the ballast temperature,
with very low emissions at room temperature and much
greater emissions, relatively, at temperatures of 45 - 50°C,
approaching normal ballast operating temperatures
(Guo et al., 2011; Hosomi, 2005). There is interest in
understanding the impact that light ballasts may have on PCB
concentrations in indoor spaces.
Laboratory chamber emissions tests were performed for
four intact PCB-containing ballasts at temperatures from
23 to 45°C (Guo et al., 2011). These emission tests were
performed using 55-L stainless steel chambers, as compared
to the microchambers used for caulk emission testing.
Emission measurement results were used here to perform
screening-level estimates of a range of total emission rates
that might be encountered in school classrooms and the
resulting air concentrations.
Light ballasts capacitors tested in the chambers were found
to contain Aroclor 1242 (Guo et al., 2011). Guo et al.
measured congeners 13, 15, 17, 18, 22, 44, 49, 52 in chamber
testing as important components of the Aroclor 1242 mixture.
Measured emission rates for each congener are presented
in Table 2-5 for the four ballasts tested at 45°C. In order to
use the congener emission results for estimating total PCB
emissions (assuming Aroclor 1242) several steps were taken.
First, the congeners were grouped by chlorine-number
homolog and the emission rates were summed. Estimates
for the total emission for all congeners in the homolog
group were generated by multiplying summed emission
rate for measured congeners at 45°C by the ratio of total
weight percent in the homolog group to the weight percent
represented by the measured congeners (Table 2-5). Next,
the sum of the total emissions for the 2, 3, and 4-homolog
groups was calculated to estimate a value approaching the
total PCB emission rate for Aroclor 1242. The combined
weight percent of congeners in those three homolog groups
represents over 95% of the total weight percent for Aroclor
1242 (see Appendix A) based on Frame et al. (1996). Finally,
the estimated total PCB emission rate for each chamber-
tested ballast was multiplied by the number of PCB-
containing ballasts in three building rooms as an estimate
of total PCB emission rates that might be encountered in
rooms where intact PCB-containing ballasts are present
and are in operation (lights are on). Because there was an
almost 80-fold difference in the highest and lowest estimated
emission rate across the four chamber-tested ballasts, the
lowest, median, and highest total emission rates were used to
generate a range of results.
It is also of interest to estimate ballast emissions for the
condition when the lights are off and the ballasts are at room
temperature. The four ballasts were tested in the laboratory
chamber at several temperatures, including 23 °C which is
similar to room temperature. Because the emission rates
were much lower at 23°C as compared to 45°C during
chamber tests, many of the congener analyses had results
below the detection limit (Table 2-6). Congeners 17 and
18 had 50% and 100% measurable results at 23°C. Ratios
of emission rates at 45°C vs. 23°C were calculated. The
average of the ratios for congeners 17 and 18 were used to
estimate the overall total PCB emission rates at 23 °C by
dividing the emission rates previously estimated at 45°C by
the average ratio of 16.2 (Table 2-6). The estimated total
PCB emission rate for each chamber-tested ballast at 23 °C
was multiplied by the number of PCB-containing ballasts in
three building rooms as an estimate of total PCB emission
rates that might be encountered in rooms where intact PCB-
containing ballasts are present and are not in operation (lights
are off).
Screening-level estimates of the ranges of concentrations
of PCBs in indoor air that might result from emission from
intact light ballasts were generated for three school building
locations. These three school rooms were selected only as
examples to highlight the ranges of total emissions that might
be found based simply on the number of ballasts present
in the room. The following equation was used to estimate
-------
steady-state indoor air PCB concentrations resulting from
emissions from ballasts using the lowest, median, and highest
rates across the four chamber-tested ballast emission rates:
(Eq.2.1.7)
where:
C = total PCB concentration in room air (ug/m3)
P^ = emission rate from ballast b (ug/hr)
n = number of PCB-containing ballasts in the room
Q = ventilation rate for the room with outdoor air (m3/hr)
(based on room volume x ACH)
The estimates are based on numerous assumptions, and
therefore should be considered only screening level.
Assumptions include:
• well mixed air in room
• constant temperature
• temperature equivalent to chamber conditions that
generated emission rates
• constant ventilation rate
• steady-state emission
• steady state and approximately equal absoption/
desorption of PCBs in other materials in the room
• no chemical reactions of PCBs
Table 2-5. Estimated total PCB emission rates for four intact light ballasts in chamber testing
Ballasts Tested in Emission Chamber at 45°C
Ballast 1 Ballast 2 Ballast 3 Ballast 4
Congener 13 (pg/hr)
Congener 15 (pg/hr)
Z Congeners (pg/hr)
Weight Percent of Measured Congeners in Aroclor 1242
Total 2-Chlorine Homolog Weight Percent in Aroclor 1242
Estimated Emission Rate for 2-Chlorine Homologs (pg/hr)
NDa
0.0072
0.0072
2.12
13.4
0.0459
0.0029
0.0120
0.0149
2.12
13.4
0.0941
0.0049
0.0210
0.0259
2.12
13.4
0.164
0.224
0.953
1.18
2.12
13.4
7.45
Congener 17 (pg/hr) 0.0256 0.038 0.0313 1.72
Congener 18 (pg/hr) 0.0832 0.114 0.0856 5.42
Congener 22 (pg/hr) 0.0050 0.0113 0.0058 0.364
Z Congeners (pg/hr) 0.114 0.162 0.123 7.51
Weight Percent of Measured Congeners in Aroclor 1242 15.5 15.5 15.5 15.5
Total 3-Chlorine Homolog Weight Percent in Aroclor 1242 48.0 48.0 48.0 48.0
Estimated Emission Rate for 3-Chlorine Homologs (pg/hr) 0.352 0.504 0.380 23.3
Congener 44 (pg/hr) ND 0.0059 0.0024 0.161
Congener 49 (pg/hr) ND 0.0045 ND 0.154
Congener 52 (pg/hr 0.0055 0.0091 0.0036 0.254
Z Congeners 0.0055 0.0195 0.0060 0.569
Weight Percent of Measured Congeners in Aroclor 1242 9.7 9.7 9.7 9.7
Total 4-Chlorine Homolog Weight Percent in Aroclor 1242 32.7 32.7 32.7 32.7
Estimated Emission Rate for 4-Chlorine Homologs (pg/hr) 0.0184 0.0658 0.0203 1.92
Z 2, 3, and 4-Chlorine Homologs (pg/hr)
0.416
0.664
0.564
32.7
Median of Total (pg/hr)
0.614
Mean ± Standard Deviation of Total (pg/hr)
8.57 ± 16.0
"ND = Not detected.
-------
• PCBs from other school spaces are not impacting the
levels in air for the room of interest
• ballasts contains congeners equivalent to Aroclor 1242
• emission rates for ballasts in the rooms are the same as
for the ballasts tested in lab chambers
Because emission testing for caulk and light ballasts were
performed in two different types of chambers, it is possible
that surface velocity conditions were different, possibly
affecting the emission rates due to different boundary layer
conditions. Losses to the chamber walls were also possible
for the larger chambers used for light ballast emissions,
potentially resulting in underestimation of total emission
rates. Thus, some caution is warranted in making direct
comparisons between estimated caulk and light ballast
emission rates.
2.6 Building Material Emissions Modeling
Most of the building materials collected in three school
buildings had measurable levels of PCBs. There is interest
in understanding whether these materials might be important
as secondary sources of PCB emissions, particularly once
primary sources have been removed or otherwise mitigated.
Screening-level emission rates were calculated to provide a
relative sense of potential emissions from different building
materials and to better understand whether these materials
are potentially important sources for exposures to PCBs in
school buildings.
Screening-level emission estimates for materials in nine
building rooms were generated using the material PCB
concentration and the measured surface area of the material
in the room. These rooms were selected as examples
based on the availability of information regarding the air
concentration, the surface areas of multiple materials in
those rooms, and PCB concentration measurements for the
materials. The approach described in Section 2.4 for caulk
was applied to the materials. The estimated emission rates
should only be considered screening-level estimates because
emission parameters generated for caulk in laboratory
chamber tests were applied to all of the other materials. No
emission parameter data are available for PCB congeners for
the many different materials that were sampled in the school
rooms. It is not clear whether, and how well, the caulk
emission parameters apply to the other materials. It is likely
that are considerable differences in emission parameters
for materials that have different physical and chemical
properties, different thicknesses, and different surface areas.
The screening-level estimates generated for the building
materials were not used to generate estimated room air PCB
concentrations for two reasons. First, as noted above, the
estimated emission rates have considerable uncertainties.
Second, the relative dynamics of absorption and desorption
(the materials acting both as sources and sinks) have not been
well characterized for school room environments, particularly
when there are multiple materials serving as sources and
sinks at the same time. This makes it difficult to accurately
predict the air concentration that will result from PCBs in
these materials.
2.7 SHEDS Exposure Modeling
2.7.7 SHEDS Background Information
SHEDS-Residential is one of modules of the SHEDS-
Multimedia human exposure/dose model (http://
www.epa.gov/heasd/products/sheds multimedia/files/
Table 2-6. Ratio of PCB emissions at different temperatures from four light ballasts tested in a laboratory chamber
PCB
Congener
13
15
17
18
22
44
49
52
% Results >
23°C
0
0
50
100
0
0
0
0
Detection
35°C
0
75
100
100
50
25
25
25
Limit3
45°C
75
100
100
100
100
75
50
100
Average Ratio of
Emission Rates
45/23°C
NCb
NC
15.9
16.4
NC
NC
NC
NC
Average Ratio of
Emission Rates
45/35°C
NC
4.0
3.0
2.8
2.2
NC
NC
NC
Average of Ratios
16.2
3.0
a For four intact light ballasts with no visible leaks that were tested for emissions at several different temperatures in a
laboratory test chamber.
b NC = Not Calculated. Ratio calculations were used only when at least 50% of the measurements were greater than the
detection limit.
-------
SHEDSResidential_TechManual_2012.pdf: Zartarian et
al., 2012; Glen et al., 2012). The primary function of the
SHEDS-residential model is to estimate the exposure of
a population to one or more specified chemicals from
inhalation, ingestion (by mouthing of hands or objects), or
dermal contact in a residential setting. SHEDS uses the
Monte Carlo statistical method to simulate a population of
individuals based on time-location-activity diaries in EPA's
Consolidated Human Activity Database (CHAD; www.epa.
gov/chadnetl) and weights from the U.S. Census. These
individuals are not specific persons, but are stochastically
created synthetic persons whose collective properties
reflect the simulated population and input distributions for
exposure-related variables. For each individual, SHEDS
constructs a sequence of activities, media concentrations, and
the resulting exposures over the selected simulation period,
which may range from one day to a year or more (although
simulation time steps can range from 1 minute to 1 hour
within a day). These individual exposure time series may be
stored or exported, or aggregated over time to give time-
integrated or time-averaged exposures (Figure 2-1). They
may also be input to a dose model, either internal or external
to SHEDS, to follow the fate of the chemical after it enters
the human body. Exposure is defined in this model as the
contact between a chemical agent and a simulated human
target at the external body surface, either the skin surface
or the oral/nasal boundary. Dose is defined in this model as
Time-Averaged
Instantaneous
Time-Integrated
Figure 2-1. Hypothetical exposure profile for an individual
the amount of chemical that enters the target after crossing
the exposure surfaces. Details regarding the pathways,
distributional functions, and exposure/dose equations are
provided in the SHEDS Technical Manual (Glen et al., 2012).
SHEDS can be used for various purposes, including
estimating population distributions of exposure and dose;
understanding intensity, duration, frequency, and timing of
exposures; identifying critical media, exposure routes, and
factors; considering how to identify and address greatest
uncertainties; and comparing modeled estimates against
real-world data. Figures 2-2 and 2-3 illustrate the SHEDS
methodology. The model estimates the exposure and/or
dose of individuals in a user-specified population cohort
EPA's Consolidated
Human Activity Database:
Time-Location-Activity Diaries
GENERATE
POPULATION * *
FOR
SIMULATION
lift*
SIMULATE LONGITUDINAL ACTIVITY DIARIES
Winter
Weekday
Winter
Weekend
Spring
Weekday
Spring
Weekend
Summer
Weekday
Summer
Weekend
Fall
Weekday
Fall
Weekend
UNCERTAINTY:
Sample N sets of
parameter
distributions
180
Day of Year
360
VARIABILITY:
perform M iterations
from each input
distribution
SAMPLE FROM
INPUT DIST'NS
N
CALCULATE INDIVIDUAL
EXPOSURE AND DOSE
CALCULATE EXPOSURE
OR DOSE
FOR SIMULATED
POPULATION
Figure 2-2. Overview of SHEDS residential methodology
-------
to a particular chemical via three primary exposure routes:
inhalation, non-dietary ingestion (i.e., via soil/dust ingestion,
hand mouthing, or object mouthing pathways), and dermal
contact in a residential setting. To do this, it simulates the
daily activities and locations of individuals using sequential
time/location/activity diaries from EPA's Consolidated
Human Activity Database (CHAD) (McCurdy et al.
2000). SHEDS utilizes the Xue et al. (2004) approach for
longitudinal diary assembly.
For each individual in a SHEDS-Residential run, the
following general steps are applied (see the SHEDS Technical
Manual for more detail: Glen et al., 2012):
1. Randomly select the age, gender, and other
demographic properties of interest, given the
distribution of the target population.
2. Generate a longitudinal activity diary, which indicates
the sequence and duration of activities and locations
for that person. For the residential module, these
are based on sequential time-location-activity diaries
from EPA's CHAD database.
3. Generate concentration time series for each potential
contact medium (e.g., indoor air, indoor smooth
surfaces, indoor textured surfaces, outdoor lawn).
4. Simulate the contacts between the individual and the
affected media. These depend on the diary activity/
location information and contact probabilities derived
from user-specified inputs.
5. Calculate pathway-specific exposure time series for
the individual, using the results of the prior two steps
and user-specified distributions for exposure factors.
6. Generate an approximation for the components of the
intake or absorbed dose time series and export these
for use in a simple pharmacokinetic (PK) or more
complex physiologically-based pharmacokinetic
(PBPK) model.
7. Time-aggregate to daily totals of absorbed dose.
The SHEDS-Residential model was applied to the dose
modeling estimation for school PCB environmental
measurement data because the school environment is,
in many ways, similar to the residential environment,
particularly with regard to multiple exposure pathways.
Rather than using residential activity data, school activity
data from CHAD were used in this PCBs assessment. Dose
estimation was not performed for adults, including teachers
and staff, as part of this effort due to the lack of school
activity data such as those available for children in the
Consolidated Human Exposure Database.
2.7.2 Input Data for SHEDS School PCB Modeling
Key inputs for PCBs exposure simulation are concentrations
of PCBs in various media. PCB environmental measurement
data from the schools were pooled and fitted to lognormal
distributions for indoor air, soil and wipe sample
concentrations for total PCBs (Table 2-7). The hypothesis for
a normal distribution of log-transformed measurement results
for air was not rejected based on the Shapio-Wilk test (0.05
level). Significance levels were slightly exceeded for surface
wipe and dust measurements but this was likely due to the
higher number of non-detect values; the log distribution
appeared to best represent the data for SHEDS modeling
purposes.
Only those soil results for samples collected from the
0 - 5 cm (0 - 2") depth were used as inputs to the model.
Dust measurements were not made at most schools
and dust concentrations were estimated based on air
concentrations measured in each room and estimated dust/
air partition coefficients (see Appendix E). Also, outdoor air
concentrations from other studies (Appendix D) were used
for PCB exposure simulation because there were insufficient
measurements to fit distributions and the quantitation limit
for the school measurements was high relative to typical
outdoor air levels. Outdoor air measurement results used in
this analysis had a median of 0.4 ng/m3 (mean of 18 ± 25 ng/
m3 total PCBs). Outdoor air concentrations were applied to
the fraction of estimated time spent outdoors while at school.
Absorption parameters of PCBs by humans are another set
of important inputs (Table 2-8, information from ATSDR,
2000). Absorption information and their application in
SHEDS is described below.
Inhalation Absorption
PCBs, when administered orally, are well absorbed by
experimental animals and at generally high fractions by
humans (ATSDR, 2000). There is very limited information
available for PCB pulmonary absorption. A recent study of
inhalation of vapor phase PCB congeners by rats estimated
that 33 ug of the 40 ug inhalation exposure was present
across multiple tissues, suggesting at least 82% of the PCBs
were absorbed (Hu et al., 2010). Another study examined
pulmonary absorption of a chemical similar to PCBs, 2,3,7,8-
TCDD, following instillation in rats and estimated 95%
absorption of the administered dose (Dilberto et al., 1996).
A portion of the PCBs in air are likely absorbed to dust and/
or soil particles. A study examined the relative absorption
of 2,3,7,8-TCDD on soil as compared to a different substrate
following pulmonary instillation and found 100% relative
absorption (Nessel et al, 1992). These studies suggest a
relatively high pulmonary absorption rate, but this is still
uncertain for humans across a range of congeners and
different vapor/particle phase fractions. Volckens and Leith
examined the deposition of inhaled SVOCs and found for
up to 7-chlorine PCB congeners that deposition to the lung
as vapor dominated (2002). A value of 70% pulmonary
absorption was used for this PCBs exposure simulation, and
sensitivity analyses were conducted ranging from 35% to
100% to examine the impact on modeled dose estimates.
-------
SHEDS-Multimedia v4: Overview
User
Specified
Inputs
CHEMICAL CONCENTRATIONS
DecayKHsperston - fugacity
• applications)
n»*-*« i«s BeWstuth tat raferanec (oappBcatjo™
SHEDS-Multimedia v4 Program Modules
Figure 2-3. General overview of SHEDS multimedia exposure model
Gastrointestinal Absorption
As with fats and other fat-soluble chemicals, PCBs are
most likely absorbed from the gut via lymphatic circulation
and consequently avoid first-pass metabolism in the liver
(Hansen 1999). Price et al. (1972) found that 88% of
the ingested PCBs were not excreted, and were therefore
assumed to be retained in the body (7-9 year old girls). This
estimate of PCB absorption in young girls is supported by
the more comprehensive, congener specific mass balance
study of Schlummer et al. (1998). Retention was estimated
to be >90% and 85.4% of the administered dose in monkeys
(Allen et al. 1974b) and ferrets (Bleavins et al. 1984),
respectively. An absorption value of 85% was used for the
simulation.
Dermal Absorption
Experimental data on the percutaneous absorption of PCBs in
humans is limited to in-vitro studies that used human cadaver
skin (Wester et al. 1990, 1993) with 14C-labeled Aroclor 1242
and 1254. Over a 24-hour period, 2.6, 10, and 43% of the
dose was retained in human skin when the Aroclor 1242 was
formulated in soil, mineral oil, or water, respectively. Similar
results were observed with Aroclor 1254, with 1.6, 6.4, and
44.3% of the dose retained in human skin, following PCB
exposure in a soil, mineral oil, or water vehicle, respectively.
The in-vitro data indicate that PCBs readily enter human skin
and are available for systemic absorption, and that the dosing
vehicle has a major role in regulating the relative retention of
PCBs in human skin.
In a related study, Wester et al. (1990, 1993) assessed the
in-vivo percutaneous absorption of PCBs in adult female
Rhesus monkeys. Topical administration of Aroclor
1242 resulted in 14, 20, 18, and 21% absorption of the
administered dose when formulated in soil, mineral oil,
trichlorobenzene, or acetone, respectively. In contrast to the
above in-vitro results with human skin, the vehicle had little
effect on the systemic absorption of the PCBs applied to the
skin of monkeys. This may be due to the uncertain viability
of the human skin used in the in-vitro studies and the fact
that the in-vitro study primarily assessed retention of PCBs
in human skin and could not estimate systemic absorption.
Absorption efficiency ranged from 0.15 to 34% of the
applied radioactivity in the monkeys and averaged 33% (42%
chlorine) and 56% (54% chlorine) of the applied radioactivity
in the guinea pigs.
For this simulation, 2% was used for dermal absorption rate
per day for dust or soil using a uniform distribution, with
10% and 40% for the daily dermal absorption rate for the
residues.
Other default inputs are listed in Table 2-9 (from Appendix
G - default values for non-chemical specific variables from
the SHEDS-Multimedia version 4 Technical Manual; Glen,
2012). The U.S. EPA Child-specific Exposure Factors
Handbook was consulted in selecting input values, but
relevant data for fitting distributions for soil and dust contact
and ingestion were available from Kissel et al. (1996),
Holmes et al. (1999), and Ozkaynak et al. (2011) and were
used in this analysis. The object mouthing rates shown in
Table 2-9 were used in conjunction with the residue data
from the dermal wipe samples.
Inhalation rate
Short-term inhalation rates are generated for SHEDS based
on several factors (Glen et al., 2012). The basal metabolic
rate (bmr) is calculated from a regression equation using
body weight as the independent variable. The units for bmr
are megajoules per day. The slope, intercept, and standard
-------
Table 2-7. Input variables for the SHEDS-multimedia model
PCBs Environmental Concentration Inputs
Distributions Used for Modelina All 6 Schools (pre-remediation)
Indoor air PCBs
Outdoor air PCBs
Wipe PCBs in high contact area
Soil PCBs
Estimated dust PCBs
Units
|jg/m3
|jg/m3
pg/cm2
M9/kg
M9/kg
Distribution
lognormal
lognormal
lognormal
lognormal
lognormal
Geo. Mean
2.29E-01
2.08E-03
1.51E-03
524.6
7032
Geo. SD
4.26
12.94
2.85
3.71
4.26
Distributions Used for Modelina 5 Schools tore- and post-remediation)
Indoor air PCBs for 5 schools (pre-remediation)
Outdoor air PCBs (pre- and post-remediation)
Wipe PCBs in high contact area (pre-remediation)
Soil PCBs for 5 schools (pre- and post-remediation)
Estimated dust PCBs (pre-remediation)
Indoor air PCBs (post-remediation)
Wipe PCBs in high contact area (post-remediation)
Estimated dust PCBs (post-remediation)
|jg/m3
pg/m3
pg/cm2
|jg/m3
pg/cm2
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
2.03E-01
2.08E-03
1.18E-03
535.6
6236
7.45E-02
1.08E-03
2294
4.44
12.94
2.45
3.65
4.44
2.87
2.47
2.87
Distributions Used for Modelina 3 Schools (Year 1 pre- and post-remediation and Year 2 pre-remediaiton)
Year 1 1ndoor air PCBs (pre-remediation)
Year 1 Wipe PCBs in high contact area (pre-remediation)
Soil PCBs
Year 1 Estimated dust PCBs (pre-remediation)
Year 1 Air PCBs (post-remediation)
Year 1 Wipe PCBs in high contact area (post-remediation)
Soil PCBs
Year 1 Estimated dust PCBs (post-remediation)
Year 2 Indoor air PCBs (pre-remediation)
Year 2 Wipe PCBs in high contact area (pre-remediation)
Soil PCBs
Year 2 dust PCBs (pre-remediation)
|jg/m3
pg/cm2
M9/kg
M9/kg
|jg/m3
pg/cm2
M9/kg
M9/kg
|jg/m3
pg/cm2
M9/kg
Mg/kg
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
lognormal
3.19E-01
1.29E-03
524.6
9820
7.63E-02
1.03E-03
524.6
2347
9.96E-02
8.46E-04
524.6
3065
3.86
2.31
3.71
3.86
2.87
2.38
3.71
2.87
3.07
2.43
3.71
3.07
deviation of the residual are taken from the body weight and
surface area files by age and gender. A minimum value of
0.1 megajoules per day is permitted. The basal inhalation
rate is the rate in effect for activities with a METS of one and
has units of cubic meters of air per hour. The basal alveolar
ventilation rate, bva, is related to the basal metabolic rate:
bva = bmr x 0.166 x 0.01963 x (0.20 +0.01 x u) x 60
The factor 0.166 converts from megajoules per day to
kilocalories per minute. The factor 0.01963 converts from
liters of oxygen consumed to cubic meters of air inhaled.
The variable "u" is uniformly distributed between zero and
one, and then term (0.20 + 0.01 x u) represents the metabolic
efficiency (liters of oxygen consumed per kilocalorie
expended). The final factor of 60 converts the per minute
rate to the per hour rate. Multiplication of metabolic ratio of
energy expenditure for an activity to the resting rate (METS)
and bva leads to the inhalation rate for SHEDS. In this way,
we link age, body weight and activity levels with inhalation
rate. SHEDS is using macro activity, therefore, we only use
short-term inhalation rates. Table 2-10 displays summary
statistics of average inhalation rates by age groups.
-------
Table 2-8. Key input variables for the SHEDS-multimedia model
Concentration or Process Inputs
Absorption fraction for lungs
Dermal absorption rate per day for dust or soil
Dermal absorption rate per day for surface residues
Gl tract absorption rate per day for dust or soil
Gl tract absorption rate per day for surface residues
Bioavailability fraction for dust/soil
Bioavailability fraction for surface residues
Residue-skin transfer efficiency
Soil-skin adherence factor
Body-surface fractional contact rate
Hand-surface fractional contact rate
Fraction of body unclothed
Surface-skin transfer coefficient for body (unclothed)
Surface-skin transfer coefficient for hand
Dust ingestion rate (indoor, direct only, 4=age =5)
Dust ingestion rate (indoor, direct only, 6<=age <=10)
Dust ingestion rate (indoor, direct only, age >=11)
Soil ingestion rate (outdoor, direct only, 4=age=5)
Soil ingestion rate (outdoor, direct only, 6<=age <=10)
Soil ingestion rate (outdoor, direct only, age >=11)
Units
H
1/day
1/day
1/day
1/day
H
H
H
mg/cm2
1/20min
1/20min
H
cm2/hr
cm2/hr
mg/hr
mg/hr
mg/hr
mg/hr
mg/hr
mg/hr
Distribution
Form
point
point
uniform
point
point
point
point
normal
lognormal
beta
Weibull
beta
lognormal
lognormal
lognormal
lognormal
point
lognormal
lognormal
point
Distribution
v1
0.7
0.02
0.1
0.85
0.85
1
1
0.051
0.11
42
10
3
3070
3070
0.706
0.446
0
0.722
0.276
0
Parameters3
v2
0.43
0.022
2
166
2.5
6.7
1.68
1.68
4.009
8.011
6.293
9.774
a Distributional parameters (vl, v2): lognormal (geometric mean, geometric standard deviation); normal (mean, standard
deviation); uniform (minimum, maximum); Weibull (shape, scale); beta (a, (3).
Time and activity in school
The simulated population of 6-18 year-old children was
generated using -35,000 person-days from the new CHAD
database; time-location-activity diaries were selected
according to age and school attendance information.
Longitudinal activity diaries of the simulated schoolchildren
were generated using a published method to optimize inter-
and intra- individual variability (that uses 8 CHAD person-
days by season and weekday/weekend for each age/gender
cohort; Xue et al., 2004). Applying this method generated
an average 6.34 hours indoor and 0.2 hours outdoor during
school time. Higher ventilation rates were applied for the
outdoor time due to an expected higher level of physical
activity. The longitudinal activity patterns for each individual
were then combined with available PCB concentration data
and exposure factors and inserted into exposure pathway
equations as described in the SHED S-Multimedia technical
manual.
Only PCB exposures incurred during school hours (in/
around the school) were modeled; neither dietary intake nor
intake away from school was considered. Routes considered
were inhalation, dermal contact, and soil ingestion. For
dermal contact, wipe data were used; these likely include
-------
Table 2-9. Key input variables for the SHEDS-multimedia model
Concentration or Process Inputs
Hand mouthing events per hr (indoor, 4=age=5)
Hand mouthing events per hr (outdoor, 4=age=5)
Hand mouthing events per hr (indoor, 6<=age <=10)
Hand mouthing events per hr (outdoor, 6<=age <=10)
Hand mouthing events per hr (age >=10)
Fraction of surface of one hand that enters mouth
Object mouthing events per hr (indoor, 4=age=5)
Object mouthing events per hr (outdoor, 4=age =5)
Object mouthing events per hr (indoor, 6<=age <=10)
Object mouthing events per hr (outdoor, 6<=age <=10)
Object mouthing events per hr (age >=11)
Object-surface concentration ratio
Object-mouth contact area
Object-mouth transfer efficiency
Transfer coefficient for object mouthing (age >=4)
Removal efficiency during bath/shower
Removal efficiency during events without water
Removal efficiency during mouthing
Removal efficiency during hand washing
Mean # hand washes/day per person
Maximum dermal loading for body
Maximum dermal loading for hands
Units
events/hr
events/hr
events/hr
events/hr
events/hr
H
events/hr
events/hr
events/hr
events/hr
events/hr
H
cm2
H
cm2/hr
H
1/hr
H
H
1/day
pg/cm2
pg/cm2
Distribution
Form
Weibull
Weibull
Weibull
Weibull
point
beta
Weibull
Weibull
Weibull
Weibull
point
uniform
exponential
beta
point
uniform
point
beta
uniform
lognormal
triangle
triangle
Distribution
v1
0.75
0.55
1.36
0.49
0
3.7
0.58
0.55
0.84
0.55
0
0
1
2
0
0.9
0
2
0.3
3.74
0.1
0.1
Parameters3
v,
12.59
5.53
7.34
1.47
25
6.9
5.38
1.2
1.1
0.2
10
8
1
8
0.9
2.63
0.6
1
a Distributional parameters (vv v2): lognormal (geometric mean, geometric standard deviation); normal (mean, standard
deviation); uniform (minimum, maximum); triangle (minimum, mode, maximum); Weibull (shape, scale); beta (a, (3).
Table 2-10. Average inhalation rate (nrYday)
Age group
(yr)
06-10
11-13
14-18
Mean
8.20
10.98
12.86
SD
1.85
2.50
3.01
Percent! les
p5
5.82
7.56
8.60
p25
6.96
9.16
10.78
p50
7.91
10.68
12.50
p75
9.01
12.34
14.60
p95
11.95
15.72
18.24
p99
14.46
18.21
21.06
both PCB residues and PCBs bound to dust. We assumed
children 11 years and older had no soil/dust ingestion due to
lack of data, however, this is likely to result in only a very
small underestimation in the total exposure for children 11-
18 years old. Direct dermal contact with and ingestion of
caulk was also not included due to an absence of information
on relevant contact rates and how much PCBs would be
available for dermal transfer from caulk-bound PCBs.
Handling of Values Below the Quantifiable Limit
Quantifiable limits (QLs) for air were usually about 50 ng/
m3. The QL for soil was 0.5 mg/kg for most samples and
0.1 ug/100 cm2 for all wipe samples. A value of one-half of
the QL was substituted for samples with values
-------
3.0
Quality Assurance and Quality Control
Quality assurance (QA) and quality control (QC)
procedures were implemented for the NERL PCBs in
schools measurement study by following the guidelines
and procedures detailed in the project's Quality Assurance
Project Plan (QAPP), "PCBs in Schools - Field and
Laboratory Data Collections" Quality control samples
were used where appropriate and available for assessing
potential contamination of field sampling materials, and
spiked samples were used to assess recovery of PCB from
sampling materials. Duplicate samples were collected where
appropriate to assess precision. Laboratory quality control
and quality assurance procedures and analyses were used by
NBA-Pace Analytical to ensure data of known quality were
produced. Data quality reviews were conducted by EPA's
research contractor (Alion Inc.) and by EPA staff. The QA
and QC results described below apply to the measurements
made at one school by NERL and NERL contractors.
Quality assurance and quality control procedures and results
for the New York City School Construction Authority
remedial investigation pilot study for the five New York City
schools are documented in the remedial investigation plan
that was developed by the SCA and TRC Engineers Inc.
(NYC SCA, 2010) and in the remedial investigation reports
(NYC SCA, 2011; NYC SCA 2012). Overall, the NYC
SCA reported that a very high percentage of measurements
(>99%) were found to be acceptable for use, and in almost
all cases < 2% of the measurement data were reported to be
qualified across multiple QA/QC criteria. The NYC SCA
quality assurance and quality control results are not included
in this report, the reader is referred to the original primary
documentation (NYC SCA, 2011; NYC SCA 2012).
3.1 Quality Control Results for Aroclor Analysis
Field and laboratory quality control samples were prepared
and analyzed to assess recovery of target chemicals and
the recovery of surrogate compounds added to all samples
prior to analysis, to evaluate the potential for contamination
of sampling and analysis materials by target analytes,
and to examine precision in sampling and analysis. The
completeness of scheduled sample collection and analysis
was also examined. This section describes results for
analysis of PCB Aroclors.
All of the air, surface wipe, dust, soil, and caulk samples
scheduled for collection were successfully collected and
analyzed for Aroclors (see Table 2-2). Three of the other
building material samples scheduled for collection were
not collected. These included samples of blackboard,
whiteboard, and spray insulation. Over 98% of the scheduled
samples were successfully collected and analyzed.
Field controls were prepared for air and surface wipe samples
by adding known amounts of Aroclor 1254 to sampling filters
and wipes. Field controls were transported to the sampling
site and then were transported and stored with the samples
until extraction and analysis. Recovery of Aroclor 1254 from
field controls is reported in Table 3-1. Average recoveries
were 85.3 ± 11.0% for air filter media and 86.0 ± 8.9% for
wipe media. Laboratory controls were prepared for air and
surface wipes by adding known amounts of Aroclor 1254
to sampling filters and wipes. Laboratory controls were
prepared at the same time as the field controls, but then
were stored at the analytical laboratory until extraction and
analysis along with the samples and field controls. Recovery
of Aroclor 1254 from laboratory controls is reported in Table
Table 3-1. Aroclor analysis: recovery of Aroclor 1254 from field and laboratory controls
Spiking
Level
Field Controls3
% Recovery
Laboratory Controls"
% Recovery
Media
Air
Surface Wipe
N
3
3
ng/sample
1000
1000
Mean
85.3
86.0
SD
11.0
8.9
Mean
92.7
112
SD
3.9
0.6
a Sample media were spiked, transported to and from the field site, and stored with samples until analysis.
b Sample media were spiked at the same time as field controls and stored in the laboratory until analysis.
-------
Table 3-2. Aroclor analysis: recovery of Aroclor 1254 from laboratory matrix spikes3
Media
Air
Surface Wipe
Dust
Soil
Caulk
Other Materials
N
4
5
2
2
4
4
Spiking
Level
1000
12.5
0.245-1.25
1.21-1.24
1.20-1.25
1.24-1.25
Spiking
Units
ng/sample
pg/sample
M9/9
M9/9
M9/9
M9/9
% Recovery
Mean
98.5
106
97.2
96.4
108
99.8
SD
3.5
13.0
1.8
6.3
5.6
5.6
1 Sample media were spiked in the laboratory immediately prior to extraction.
Table 3-3. Aroclor analysis: recovery of surrogate analytes from samples and duplicate samples3
Spiking Level
jjg/sample
Media
Air
Surface Wipe
Dustd
Soil
Caulkd
Other Materials
N
11
44
8
9
26
47
TCMXb
0.025
0.25
0.05
0.25
0.25
0.25
DCBP
0.25
2.5
0.5
2.5
2.5
2.5
TCMX
% Recovery
Mean
82.1
94.2
79.4
96.2
100
98.3
SD
6.8
6.5
22.8
6.0
6.3
12.5
DCBP
% Recovery
Mean
87.6
99.3
89.4
106
102
101
SD
5.8
13.0
15.1
6.2
16.2
15.5
a Two surrogate analytes were spiked on every sample and duplicate sample prior to extraction.
b TCMX = Tetrachloro-meta-xylene.
c DCBP = Decachlorobiphenyl (PCB Congener #209).
dOne dust sample and 12 caulk samples were diluted significantly for analysis; surrogate recoveries not available for these
samples and are not included in the recovery statistics.
3-1. Average recoveries were 92.7 ± 3.9% for air filter media
and 112 ± 0.6% for wipe media. Laboratory matrix spikes
were used to assess recovery of Aroclor 1254 from all media
at the laboratory. Unused sampling media or surrogates of
each type of sample matrix (PUF filter, wipe material, dust,
soil, and caulk) was fortified with known amounts of Aroclor
1254 and the matrix spike samples were extracted and
analyzed along with the samples. Recovery of Aroclor 1254
are reported in Table 3-2 and ranged from 98.5 ± 3.5% for
air filter media to 108 ± 5.6% for caulk media. All recovery
results met the 80 - 120% QAPP data quality objective.
The two surrogate standard compounds tetrachloro-meta-
xylene (TCMX) and decachlorobiphenyl (PCB Congener
#209; DCBP) were added to every sample prior to analysis
to assess recovery through the laboratory extraction and
analysis process. Surrogate analyte recovery results are
reported in Table 3-3. Average recoveries ranged from 82.1
± 6.8% for TCMX in air samples to 106 ± 6.2% for DCBP
in soil samples. Surrogate samples in one dust sample
and in 12 caulk samples were diluted significantly and the
surrogate standard recoveries could not be measured. One
wipe sample, one caulk sample, and two materials samples
had recoveries below the 60% acceptance level for DBCP,
and one dust sample had a TCMX recovery lower than the
acceptance level. In those samples, the analytical result for
the other surrogate analyte met the acceptance level and the
sample analysis results were accepted.
-------
Field blanks were used to assess potential contamination by
PCBs. Air filters and wipe sampling media were used along
with PCB-free caulk and soil materials. The media were
placed in the same type of containers used for samples. Field
blanks were transported to the sampling site and then were
transported and stored with samples until extraction and
analysis. Field blank results are shown in Table 3-4. Mean
total PCB concentrations on all types of field blanks were
lower than the quantifiable limits (QLs). Laboratory blanks
were prepared for air filters and surface wipes by storing
unused media at the laboratory until extraction and analysis
along with samples. Laboratory blank results are shown in
Table 3-4. Mean total PCB concentrations were lower than
the QL. Laboratory method blanks were prepared to assess
potential PCB contamination in the extraction materials
and methods through the instrumental analysis procedure.
Unfortified extraction solvent was carried through the
extraction and analysis procedures along with the samples.
Laboratory method blank results are shown for all media in
Table 3-5. Except for caulk, all laboratory blank results were
lower than the QLs. For caulk the average measured total
PCB values on the method blanks was 0.063 ± 0.092 ug/g
which was slightly higher than the 0.05 ug/g QL. Overall,
the background contamination levels as measured in field
blanks, laboratory blanks, and laboratory method blanks
were judged to have no impact on the use and interpretation
of measurement results. Although the laboratory method
blank result for caulk was slightly greater than the QL value,
almost all caulk samples had concentrations well above the
QL, and values of the samples of most interest had levels
hundreds of times higher than that measured in the method
blanks. Sample measurement results were not adjusted for
background concentrations.
Table 3-4. Aroclor analysis: total PCBs measured on field and laboratory blanks
Field Blanks3
Laboratory Blanks"
Media
Aird
Surface Wipe
Dust
Soil
N
3
3
3
2
QLC
50
0.1
0.047
0.050
Units
ng/m3
|jg/100cm2
ppm
ppm
Mean
3.1
NDe
0.024
0.012
SD
0.58
ND
0.022
0.002
Mean
10.5
0.001
NPf
NP
SD
12.1
0.002
NP
NP
a Unfortified sample media, transported to and from the field site, and stored with samples until analysis.
b Unfortified sample media, prepared at the same time as field blanks and stored in the laboratory until analysis.
0 QL = quantitation limit.
d Based on an assumed nominal air volume of 5.5 m3.
eND = not detected.
f NP = none prepared.
Table 3-5. Aroclor analysis: total PCBs measured in laboratory method blanks
a Unfortified solvent taken through extraction and analysis.
b QL = quantitation limit.
0 Based on an assumed nominal air volume of 5.5 m3.
Lab Method Blanks3
Media
Air
Surface Wipe
Dust
Soil
Caulk
Other Materials
N
2
5
2
2
4
4
QLb
50
0.1
0.047
0.050
0.050
0.050
Units
ng/m3
|jg/100cm2
M9/9
M9/9
M9/9
M9/9
Mean
10.3
0.027
0.010
0.003
0.063
0.008
SD
10.0
0.028
0.011
0.004
0.092
0.008
-------
Precision was examined using duplicate sample collection
and analysis as well as the precision in measurements for
various quality control samples and analyses. Precision
results for duplicate sample collection and analysis are
reported in Table 3-6. Results are reported as the relative
percent difference (RPD). Duplicate air samples collected
in close proximity provide a good measure of sampling and
analysis precision because pollutants in indoor air collected
side-by-side within a room are typically homogeneous.
The mean RPD for duplicate air samples was 13.6 ±
6.0% which met the precision QAPP objective of ± 20%.
Duplicate sample collection for surface wipe, dust, soil,
and caulk materials includes a sample non-homogeneity
component in addition to sampling and analysis method
precision. One of the purposes of this work was to provide
information regarding variability of PCBs in these media
in school buildings. The mean RPD ranged from 26.0 ±
12.0% for surface wipes to 70.8 ± 31.1% for dust. The
dust collected for duplicate samples was collected in the
same room, but obtaining sufficient sample sizes required
vacuuming multiple surfaces in the room. The results
suggest the potential for considerable variability in dust PCB
concentrations for different locations within the same school
room.
Table 3-6. Aroclor analysis: precision results for
duplicate sample measurements3
Relative % Difference"
Media
Air
Surface
Wipe
Dust
Soil
Caulk
Other
Materials0
N
3
3
2
2
9
0
Mean
13.6
26.0
70.8
43.0
29.8
—
SD
6.0
12.0
31.1
38.5
18.8
—
a Duplicate sample results serve as an indicator of
measurement precision for relatively uniform media
such as air; for other media the duplicate sample results
include elements of both measurement precision and non-
homogeneity of analytes in the environment.
b Relative % difference calculated as
2x(|X1-X2|)/(X1+X2)x 100.
0 No duplicate samples collected.
3.2 Quality Control Results for
Congener Analysis
All of the indoor air samples, the outdoor air sample, and a
subset of the remaining sample extracts previously analyzed
for Aroclors were scheduled for congener-specific analysis
and were successfully analyzed.
Recovery of the sum of the congeners in Aroclor 1254 in
the air filter media field controls is reported in Table 3-7.
Average recoveries were 85.9 ± 12.3%. Laboratory matrix
spike recovery results are reported in Table 3-8. Recovery
of the sum of the congener concentrations in Aroclor 1254
congeners ranged from 91.1% for caulk to 112% for dust.
Recoveries of the surrogate compounds TCMX and DCBP
are shown in Table 3-9 and ranged from 72.6 ± 28.6% for
caulk TCMX to 153 ± 13.2% for DCBP in dust samples. The
remaining average recovery results ranged from 81 to 117%.
Air media field blank results are reported in Table 3-10. No
congeners were measured in the air field blanks at detectable
levels (the congener QL was 0.5 ng/sample). Laboratory
method blank results are shown in Table 3-11. Results were
below detectable levels for all congeners for the air, surface
wipe, dust, soil, and other materials blanks. The method
blank value for the sum of measured congeners in the single
caulk method blank was 0.03 ug/g which was far below the
total PCB concentrations measured in most caulk samples.
Precision was examined using duplicate samples collected
for air and caulk. Precision results for duplicate sample
collection and analysis are reported in Table 3-12. The
average RPD for three air sample duplicates was 17.9 ± 4.5%
while the RPD for the single caulk duplicate sample analyzed
for congeners was 41.7%.
Table 3-7. Congener analysis: total recovery of
congeners in Aroclor 1254 from field controls
Spiking Field Controls3
Level % Recovery
Media N
Air 3
ng/
sample Mean
1000 85.9
SD
12.3
" Sample media were spiked, transported to and from the
field site, and stored with samples until analysis.
-------
Table 3-8. Congener analysis: total recovery of congeners in Aroclor 1254 from laboratory matrix spikes3
Media
Air
Surface Wipe
Dust
Soil
Caulk
Other Materials
N
4
2
1
2
1
1
Spiking
Level
1000
12.5
0.245
1.21
1.23
1.25
Spiking
Units
ng/sample
pg/sample
M9/9
M9/9
M9/9
M9/9
% Recovery
Mean
96.1
94.2
112
95.0
91.1
95.2
SD
2.6
24.1
—
7.0
—
—
a Sample media were spiked in the laboratory immediately prior to extraction.
Table 3-9. Congener analysis: recovery of surrogate analytes from samples and duplicate samples3
Spiking Level
jjg/sample
Media
Air
Surface Wipe
Dust
Soil
Caulkd
Other Materials
N
11
10
4
3
5
18
TCMXb
0.025
0.25
0.05
0.25
0.25
0.25
DCBPC
0.25
2.5
0.5
2.5
2.5
2.5
TCMX
% Recovery
Mean
93.1
81.0
103
94.9
72.6
93.3
SD
6.9
5.1
3.2
1.0
28.6
8.7
DCBP
% Recovery
Mean
104
85.0
153
109
88.2
117
SD
10.3
4.2
13.2
12.7
27.5
20.1
a Two surrogate analytes were spiked on every sample and duplicate sample prior to extraction.
bTCMX = Tetrachloro-meta-xylene.
c DCBP = Decachlorobiphenyl (PCB Congener #209).
dFour caulk samples were diluted significantly for analysis; surrogate recoveries not available for these samples and are not
included in the recovery statistics.
Table 3-10. Congener analysis: total PCBs measured on field blanks
Field Blanks3
Media N QLb Units Mean SD
Air 3 0.5 ng/m3 NDC
a Unfortified sample media, transported to and from the field site, and stored with samples until analysis.
b QL = quantitation limit.
c ND = not detected.
-------
Table 3-11. Congener analysis: total PCBs measured in laboratory method blanks
Lab Method Blanks3
Media
Air
Surface Wipe
Dust
Soil
Caulk
Other Materials
N
2
2
1
2
1
4
QLb
0.5
0.0025
0.00025
0.00125
0.00125
0.00125
Units
ng/m3
|jg/100cm2
ppm
ppm
ppm
ppm
Mean
NDC
ND
ND
ND
0.030
ND
SD
—
—
—
—
—
—
a Unfortified solvent taken through extraction and analysis.
b QL = quantitation limit.
c ND = not detected.
Table 3-12. Congener analysis: precision results for duplicate sample total PCB measurements3
Media
Air
Caulk
N
3
1
Mean
17.9
41.7
Relative % Difference"
SD
4.5
"Duplicate sample results serve as an indicator of measurement precision for relatively uniform media such as air; for other
media the duplicate sample results include elements of both measurement precision and non-homogeneity of analytes in the
environment.
Relative % difference calculated as 2 x (|x, -
, + X2) x 100
3.3 Quality Assurance Assessments
Quality assurance assessments of field and laboratory data
collection and analyses were performed at multiple levels. A
summary of reviews and outcomes is provided below.
• On-site QA assessment of field sampling procedures and
adherence to protocols was performed. Air sampling flow
rate measurements were verified using audit flow devices.
Corrective action was taken when it was determined that
air sampling pumps were not being started at targeted
flow rates.
• All field sampling data were QA reviewed to ensure
accuracy and completeness. Corrective action was taken
for three air samples that did not have the correct total
sampled air volume calculated correctly. The correct
air volume data were applied to Aroclor and congener
analysis results.
• The analytical laboratory performed ongoing review
of calibration, continuing calibration checks, QC
recovery and background assessments, and instrumental
performance parameters. Two groups of sample extracts
were reanalyzed based on data quality review.
Analytical laboratory results were first QA reviewed by
NERL's contractor, and then again by NERL scientists
to ensure that the results were complete and accurate.
Quality control results were summarized and examined to
ensure overall data quality objectives were met.
NRMRL experts reviewed the caulk/materials emission
model and model calculations.
A NERL QA review was performed to ensure that the
measurement data were accurately transcribed into
data analysis files and that calculations were correctly
performed. Transcription accuracy into the report tables
was also assessed.
-------
4.0
Results
4.1 School Information
Basic information regarding the six schools with PCB
measurement results used in this report is shown in Table 4-1.
Five of the schools were primary schools and two of those
contained pre-kindergarten classes. The sixth school was a
secondary school. The school buildings were constructed
between 1959 and 1972 and all but one had 3 floors. The
secondary school had multiple wings with 2 to 3 floors per
wing. Only one school (School 1) had heating, ventilation,
and air conditioning (HVAC) forced air units serving the
Table 4-1. School building information
entire building. For the other five schools, different types of
heating and ventilation systems were present in the schools
including exhaust ventilation ducts only, single room unit
ventilators, and zoned heating/ventilation systems for specific
spaces such as gymnasiums. These schools did not have
building-wide air conditioning systems and in many cases
relied on natural ventilation, including opening windows,
in warmer months. Several schools had window air
conditioners in some rooms.
Grade Year
School Levels Constructed
Floors Ventilation System(s)
1 K-5/PK-123
1972
- 6 HVAC units serving different building zones
- Window glazing previously replaced
K-5
1962
3 PK-5/K-1
PK-5
PK-8
1963
1959
Additions 1
1968,2005
1961 3
9-12
1968
2 to 3
• Classroom/bathroom exhaust systems with 19 roof exhaust fans
• Separate HA/ systems for gymnasium and auditorium
• Window-mounted AC units in most classrooms
• Windows/frames previously replaced
• Classroom/bathroom exhaust systems with 15 roof exhaust fans
• Basement fans provide HA/ for gymnasium and auditorium
• Window-mounted AC units in most classrooms
• Windows/frames previously replaced
• Exhaust systems vented to roof in older building areas
• Office window-mounted AC units
• 3 HVAC units in new construction
• Four HVAC systems - not operable
• 11 roof exhaust fans
• 10 window and 10 portable AC units
• Unit ventilators in classrooms and cafeteria
• HA/ systems in auditorium and gymnasium
• Several window AC units
1K = kindergarten; PK = pre-kindergarten
-------
Five schools were part of the NYC remedial pilot
investigation while the sixth school was scheduled for
demolition and had no measurements associated with
remedial activities. Information regarding the sampling
time points and different remedial and sampling activities is
summarized in Table 4-2. Much more detailed information
regarding remedial activities and outcomes at the five NYC
schools is provided in the NYC SCA remedial investigation
reports (NYC SCA 2011; NYC SCA 2012). The remedial
investigation is ongoing and further information will be
available in the future (http://www. nycsea. org/Community/
Programs/EPA-NYC-PCB/Pages/default.aspxX The
measurement time points that were used in SHEDS exposure/
dose modeling are shown in Table 4-2. It is important to
recognize that different types of remedial actions were taken
at the different school buildings, and that caulk remediation
completed during the first year at Schools 1, 2, and 3 only
occurred in the rooms or transitory areas to be sampled.
The conditions at each of the schools at the measurement
time points used in the SHEDS exposure/dose modeling
are shown in Table 4-3. All of the measurements at these
time points occurred between late spring and early fall, with
outdoor temperatures ranging from 69 to 94°F and indoor
temperatures ranging from 70 to 84°F. The status of the
operation of ventilation systems and whether doors and
windows were open or closed in the measurement rooms
is also reported in Table 4-3. At School 1, it was learned
following the post-remedial sampling time points that the
HVAC system controllers were not operating correctly on
all units and that exterior air had not been incorporated
at the designed ventilation rates. This may have been a
factor in indoor air PCB concentrations at that school. The
gymnasium heating/ventilation system at School 6 was not
operating during the indoor air sample collection period
and the lack of forced ventilation with outdoor air may have
impacted indoor air concentrations in the gym.
-------
Table 4-2. School building remediation activity and environmental sampling summary
School Activity
1 Pre-remediation
Caulk patch and repair
Ventilation with outdoor air
Soil cover and access restriction
Cleaning & light fixture removal
Supplemental cleaning
HVAC evaluation and repair
Pre-remediation
Encapsulate exterior caulk
Soil removal/replacement
Caulk patch and repair
2 Pre-remediation
Caulk removal
Ventilation with outdoor air
Soil cover and access restriction
Heating/ventilation cleaning
Cleaning & light fixture removal
Caulk encapsulation
Exterior caulk removal
Pre-remediation
Soil removal/replacement
Caulk removal
3 Pre-remediation
Caulk encapsulation
Ventilation with outdoor air
Soil cover and access restriction
Heating/ventilation cleaning
Cleaning & light fixture removal
Pre-remediation
Soil removal/replacement
Caulk encapsulation
Re-Cleaning
4 Pre-remediation
Soil cover and access restriction
Light fixture removal
Caulk encapsulation
5 Pre-remediation
Soil cover and access restriction
Remove/replace windows
Re-cleaning
6 Pre-demolition
Additional Information
Sampled spaces only
24-hr high vol., filtered
>1 ppm PCB
Pre-K, K only
>1 ppm PCB
Whole building
Sampled spaces only
24-hr high vol., filtered
>1 ppm PCB
Whole building
Whole building
Whole building
>1 ppm PCB
Whole building
Sampled spaces only
24-hr high vol., filtered
>1 ppm PCB
Whole building
Whole building
>1 ppm PCB
Whole building
Selected rooms
>1 ppm PCB
Whole building
One stairwell
>1 ppm PCB
Selected areas w old windows
Sampled spaces
Date
July 2010
July 2010
Aug2010
Aug2010
Aug2010
Aug2010
Sept 2010
June 2011
July 2011
Aug2011
Aug2011
July 2010
Aug2010
Aug2010
Aug2010
Aug2010
Aug2010
Sept 2010
Apr 2011
June 2011
Aug2011
Aug2011
July 2010
Aug2010
Aug2010
Aug2010
Aug2010
Aug2010
June 2011
July 2011
Aug2011
Aug2011
May 20 11
2011
Aug2011
Sept 2011
June 2011
2011
Aug2011
Sept 2011
July 2011
Samples3
A,W, S
A,W
A
A
A
A
A.W.S
A,W
A.W.S
A,W
A
A
A
A,W
A,W
A,W,S
A,W
A
A
A,W
A,W
A
A,W,S
A,W
A
A,W,S
A,W
A,W
A,W,S,D
SHEDS"
Modeling
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
1A = air samples, W = surface wipe samples, S = soil samples, D = dust samples.
b SHEDS exposure modeling using data at this time point.
-------
Table 4-3. School building information at sampling time points used in exposure modeling
Average Outdoor
Sample Time Sample Indoor High
School Point Date Temp. °F Temp. °Fa Conditions
1
Pre-Remediation July 2010 74
87
- HVAC systems operational
- Windows and doors closed
Post-Remediation
(Caulk patch/repair & Aug2010 70
light removal)
- HVAC systems operational
69 - Outdoor air dampers found closed in some units
- Windows and doors closed
Pre-Remediation June 2011 74
72
- HVAC systems operational
- Doors closed
Pre-Remediation July 2010 83
94
- Exhaust systems operational; window ACs off
- Windows and doors closed
Post-Remediation
(Caulk removal Slight Aug2010 76
removal)
Pre-Remediation June 2011 77
79
83
- Exhaust systems operational; window ACs off
- Windows opened; doors closed
- Exhaust systems operational
- Doors closed, windows opened slightly
Pre-Remediation July 20 10
77
83
- Exhaust systems operational; window ACs off
- Windows and doors closed
Post-Remediation
(Caulk encapsulation & Aug2010 74
light removal)
Pre-Remediation June 2011
78
Pre-Remediation
4 May 2011 83
(Light removal)
82
80
84
- Exhaust systems operational; window ACs off
- Windows opened; doors closed
- Exhaust systems operational
- Doors closed, windows opened slightly
- Exhaust systems operational; window AC units on
- Windows opened slightly where no AC units
- Doors closed
Post-Remediation 72
Pre-Remediation May 2011 84
Post-Remediation Aug/Sept
(Window replacement) 2011
80/78
79
84/83
- Exhaust systems operational; window AC units on
- Windows opened slightly where no AC units
- Doors closed
- Exhaust systems operational
- Doors closed, windows opened slightly
- Exhaust systems operational; window AC units on
- Windows opened slightly where no AC units
- Doors closed
Pre-Demolition July 2011
81
- Room unit ventilators operating
86/91 - Gymnasium HA/not operating
- Doors and windows closed; lights on
1 Outdoor air high temperature on sampling day(s) from a nearby National Weather Service reporting station.
-------
4.2 PCB Source Characterization
4.2.1 Caulk
Between the 1950s and early 1970s PCBs were sometimes
added to caulk and other sealants as a plasticizer. In some
cases PCBs were added to caulk at the construction site
to improve its application properties. PCB-containing
sealants have been shown to be present in buildings, and
buildings with PCB-containing sealants have higher indoor
air PCB concentrations than other buildings. A total of
427 samples of interior caulk and other sealants such as
window glazing and building joint material were collected
at five of the six schools. A total of 73 samples of exterior
caulks and sealants were collected at three of the schools.
A summary of total PCB measurement results is reported
in Table 4-4. Measurement results reported in Table 4-4
were divided into samples with values greater than or less
than 100 ppm because only a few samples had measurement
results between 100 and 1000 ppm, and very few had values
between 100 and 200 ppm. The analysis laboratory reported
that Aroclor 1254 was the PCB mixture best matched in most
of the caulk samples, with Aroclor 1260 reported for some
caulk samples at School 1. The median concentration for
interior caulks with values <100 ppm was 6.9 ppm, while
the median level for caulks with >100 ppm was 102,000
ppm (or about 10% PCBs by weight). The highest measured
concentration in indoor caulk was 440,000 ppm (44% PCBs
by weight) from caulk around a school display case. In
exterior caulks the median for <100 ppm caulks was 5.9 ppm
and for >100 ppm caulks the median was 130,000 ppm.
The highest concentration measured in exterior caulk was
328,000 ppm.
Caulk total PCB measurement results are shown separately
by school for interior and exterior caulks in Table 4-5.
When considering caulk measurement results from these
schools it is important to note that multiple samples of the
same kind of caulk may have been collected from different
locations in and around the building. Schools 1 and 6 had
lower levels of PCBs in interior caulks than the other three
schools. The median interior caulk value for School 1 was
309 ppm for samples with >100 ppm, while School 6 did
not have any interior caulks with >100 ppm. At the other
three schools the median interior caulk values for caulks
with >100 ppm ranged from 127,000 to 232,000 ppm. The
maximum value measured at School 1 was 90,700, which
was at least 60% lower than the maximum levels at Schools
2, 3, and 4. For three schools with exterior caulks, median
total PCB concentrations for caulks with <100 ppm ranged
from 3.5 to 19 ppm. Median values ranged from 77,500 to
193,000 for caulks with >100 ppm.
Another way to show the range of caulk PCB concentrations
is by concentration category (Table 4-6). Over 82%
of the 427 interior caulk samples had concentrations
<50 ppm while 6% had concentrations greater than
100,000 ppm. Only 37% of the exterior caulk samples had
concentrations <50 ppm while 41% had concentrations
Table 4-4. Caulk total PCB measurement results for schools with available data
Total PCB Levels in Caulka'b
Interior Caulks
< 100 ppm
> 100 ppm
N
Schools
5d
5
N
Samples
375
52
%
>QLC
86
100
Median
ppm
6.91
102,000
Inter-Quartile
Range
ppm
2.90 - 17.4
2,110 - 233,000
Overall
Range
ppm
100 ppm
3e
3
27
46
96
100
5.88
130,000
2.16 - 9.94
3,870 - 248,000
-------
Table 4-5. Caulk total PCB measurement results by school
Total PCB
N
% > QLC
Median
ppm
Levels in Caulka'b
Inter-Quartile
Range ppm
Overall Range
ppm
Interior Caulks <100 ppm
School
School
School
School
School
1
2
3
5
6
97
101
106
51
20
62
95
93
92
100
1.78
10.9
9.33
5.84
11.5
100 ppm
School
School
School
School
School
1
2
3
5
6
14
12
14
12
0
100
100
100
100
--
309
127,000
217,000
232,000
--
204 -
83,000
82,600
138,000
19,300
- 151,000
- 284,000
- 264,000
--
114 -
103 -
1,430 -
1,600 -
90,700
243,000
440,000
306,000
--
Exterior Caulks <100 ppm
School
School
School
4
5
6
8
18
1
100
94
100
3.50
6.62
19.1
2.07
2.34
- 5.51
- 10.7
-
1.68
1 00 ppm
School
School
School
4
5
6
7
31
8
100
100
100
77,500
193,000
138,000
5,690
2,560 -
118,000
- 91,100
288,000
- 144,000
126 -
319 -
84,400
226,000
328,000
- 152,000
a Reported as total PCBs from Aroclor measurements.
b When duplicate samples were collected, the average of the duplicates was used.
°QL = quantitation limit; sample size dependent, typically < 1 ppm (range 0.3 - 79 ppm).
>100,000 ppm. Caulk concentrations by category are
shown for individual schools in Appendix B. Insufficient
information was available for most of the schools to
determine how many discrete types of caulk were present at
each school. Approximately six different kinds of interior
caulk, window glazing, and joint material were found at
School 6 across the seven rooms that were sampled, with
most of those appearing in multiple rooms. Two types of
exterior caulk were found at School 6 where 8 samples
collected from around windows and in building joints
contained >80,000 ppm of PCBs and were likely the same
material while a ninth exterior sample from an entranceway
brick/masonry seam contained 19 ppm.
None of the interior caulks at School 6 (which included
caulk, window glazing, and joint material) contained high
levels of PCBs. However, no sealants were found in the
hallways and stairwells. In the seven rooms that were
sampled, at least one sample of every type of sealant present
in the room was collected, and it appeared that the materials
that were collected were common in rooms throughout the
building. Given the number of sealants present inside large
school buildings it is possible that a interior sealant with
high PCB levels was not collected, but it is unlikely that
the sampling failed to identify one that was widely used in
accessible areas throughout the building.
All of the caulk samples with high PCB levels collected
in NERL studies were still at least somewhat flexible and
largely intact. All of the brittle or dry caulk and other sealant
materials had levels of PCBs <50 ppm. However, intact and
flexible caulks and other sealants were found that also had
<50 ppm PCBs. It is difficult to determine whether a caulk
or other sealant is likely to contain high levels of PCBs,
-------
Table 4-6. Interior and exterior caulk and window glaze total PCB measurement results by concentration category
Concentration
Category
Interior Caulk and Window
Glaze
Five Schools3
Exterior Caulk and Window Glaze
Three Schools"
All samples
< 50 ppm
50 - 999 ppm
1,000 -9,999 ppm
10, 000 -99,000 ppm
100,000 -199,999 ppm
200,000 -299,999 ppm
300,000 -399,999 ppm
400,000 -499,999 ppm
< 50 ppm
50 - 999 ppm
1,000 -9,999 ppm
10, 000 -99,000 ppm
100,000 -199,999 ppm
200,000 -299,999 ppm
300,000 -399,999 ppm
400,000 -499,999 ppm
427
351
33
6
11
10
14
1
1
82.2
7.7
1.4
2.6
2.3
3.3
0.2
0.2
Number of Samples
73
27
5
11
5
9
11
5
0
Percentage of Samples
37.0
6.8
15.1
6.8
12.3
15.1
6.8
0
a Schools 1,2,3,5,6.
b Schools 4,5,6.
particularly when that material is still somewhat flexible.
Sampling and analysis in a laboratory is the only sure way to
know at this time. Development of field-portable screening
methods would allow more rapid characterization of sealants
in buildings. For example, hand-held x-ray fluorescence
(XRF) devices could be evaluated for their ability to detect
PCBs in materials such as caulk in-situ, but issues regarding
caulk dimensions and interferences or materials that would
give false positives would need to be examined.
A total PCB emission modeling approach was used to
estimate PCB emissions from caulk collected at several
locations at two schools. The total concentration of PCBs
in caulk was measured as Aroclors and information was
available for the caulks at these locations for calculation of
their total surface area. The method used to estimate the
total PCB emission rate was described in Section 2, and
assumes that the PCB content in the caulk is equivalent to
the congener mixture in Aroclor 1254, which was the Aroclor
reported by the analysis laboratory.
Emission rate estimates shown in Table 4-7 are based on the
emission parameters derived from chamber tests of different
caulk samples at 23 °C. Estimates of total PCB emission
rates from several caulks collected in the gymnasium and
cafeteria ranged from 140 to 600 ug/hr. Estimates from three
types of caulk collected in a third floor corridor at School
2 ranged from 53 to 3100 ug/hr. Estimates of emissions
from caulk at exterior window locations at two classrooms
at School 6 ranged from 830 to 940 ug/hr. These window
locations included caulk around the window frame, around
the concrete/brick seams below the window, and from around
the unit ventilator air intake grill. Finally, a total PCB
emission rate of 320 ug/hr was estimated for caulk collected
from a single 2-story building joint at School 6. Graphical
representations of the relative emissions for several of the
sets of caulk are shown in Figure 4-1. There are uncertainties
in these estimates because it is not known if the emission
parameters for the measured caulks match those tested in the
chamber (although emission parameters for the 12 caulks
that were tested were consistent). Also, the actual emission
-------
Table 4-7. Estimates of total PCB emission rates for several examples of PCB-containing caulk
School Room
2 Gymnasium
2 Gymnasium
2 Cafeteria
2 Cafeteria
2 Cafeteria
2 Corridor
2 Corridor
2 Corridor
6 Classroom 3
6 Classroom 4
Material
Interior Door Frame Caulk
Interior Bay Door Caulk
Door Frame Caulk
Metal Door Frame Caulk
Bay Door Frame Caulk
Interior Door Caulk
Interior Wall Panel Caulk
Interior Metal Panel Caulk
Exterior Window Caulkb
Exterior Window Caulkb
Linear
Length of
Caulk
m (ft)
33.5 (110)
5.2 (17)
25.6 (84)
23.2 (76)
4.3 (14)
110 (360)
3.7 (12)
2.7 (9)
28.3 (93)
29.9 (98)
Surface
Area
m2
0.1431
0.0442
0.219
0.198
0.036
0.468
0.0468
0.0117
0.271
0.286
Estimated
Total PCB Emission Rate3
ppm ug/hr
117,000
137,000
57,100
112,000
146,000
243,000
217,000
165,000
112,000
120,000
460
160
340
600
140
3,100
280
53
830
940
N/A
Exterior Joint Caulk0
6.4 (21)
0.0813
142,000
320
1 Based on PCB emission parameters derived from chamber measurements of other caulks at 23°C.
b Calculated using entire length of caulk around both window frame units, the single unit ventilator inlet, and concrete-brick seams below
both windows.
c Calculated for one joint that was 2 stories high. There were multiple joints around the building exterior.
rate is likely to depend on the temperature. Guo et al. (2011)
found in chamber testing that emission rates increased by
approximately 6-fold with a 10°C increase in temperature.
The temperature of caulk on the exterior of a building may
also become higher than ambient temperature due to radiant
heating from the sun and/or, in the winter because of heat
from the building.
The emission rates estimated for caulks at four of the building
locations (Table 4-7) were used to generate screening level
estimates of the indoor air PCB concentrations that could
potentially result under different assumed conditions of
ventilation with outdoor air. The approach for calculating
screening-level PCB indoor air concentration estimates was
described in Section 2, using Equation 2.1.6. Numerous
assumptions apply to this estimation approach, and results
reported in Table 4-8 are considered to be only screening-
level estimates. One important factor is the room air
exchange rate.
While recognized as important for indoor pollutant
modeling, accurate measurement of ventilation conditions
including both indoor/outdoor air exchange and inter-
zonal flows in large old buildings - and for specific rooms
within those buildings - is difficult. Information on this
approach and the conditions that must be met to do this
have been described (Persily, 1997) and an ASTM method
has been developed, Standard Guide for Using Indoor
Carbon Dioxide Concentrations to Evaluate Indoor Air
Quality and Ventilation, ASTM D6245-07. A central tenet
of the procedure for producing accurate ventilation rate
information, whether using decay or equilibrium procedures,
is that the space to be evaluated is a single zone wherein the
tracer concentration is uniform and that only exchanges air
with the outdoors. By definition, the tracer concentration
must not differ by more than 10% from the average across all
locations in the building or building zone. For spaces that do
not meet the 10% criterion, it must be demonstrated that there
is no significant airflow from other building spaces, such as
hallways or other rooms, into the test space. It is important
that the indoor pollutant of interest be measured at the same
time that AER is being measured since ventilation rates in a
room can change rapidly under different ventilation system
and door/window use conditions.
Other researchers have measured air exchange rates
(AERs) in school building rooms and a wide range of AERs
have been reported, ranging from less than 0.5 to 12 hr1
(Macintosh et al., 2012; Nazaroff et al., 2010; Godwin and
Batterman, 2007; Bartlett et al., 2004; Scheff et al., 2000).
AER measurements have been based on measurements
of ventilator airflow rates, equilibrium or decaying CO2,
or SF6 decay. These methods allow an assessment of the
-------
q>
LU
CD
O
LU
600
D)
^
J3
TO
ir
o
tn
tn
1
LU
CQ
O
Q_
1 200
400-
£
ro
1/5
LU
School 2 Gymnasium
«? c- .w-
^x ,.<$•
fr x
^
v
X X
^ .^
^
•IF
' J
9f
A
Sf
4000
I
TO
tn
tn
E
LU
CO
o
0-
8
o
3000
2000-
TO
E
•i=
CO
LU
1000
1000 -
Figure 4-1. Estimated total PCB emission rates from caulk in several building locations
-------
Table 4-8. Screening-level comparison of predicted air concentrations resulting from PCB emissions from caulk to
measured concentrations
Estimated
Room Emission
Volume Rate"
Room3 m3 ug/hr
Gymnasium 2666 620
Cafeteria 1134 1080
Classroom 3 201 830
Classroom 4 182 940
%
Emissions
Entering
Roomc
100
100
1
5
10
1
5
10
Predicted Total PCB Air
Concentration in Rooms at Different Measured
Air Exchange Ratesd Total
PCBs
"9/m3 In Room
AER =
0.5
460
1900
83
410
830
100
520
1000
AER =
1.0
230
950
41
210
410
52
260
520
AER =
2.0
120
480
21
100
210
26
130
260
AER= Air8
4.0 ng/m3
58 1360
240 600
10 950
52
100
13 690
65
130
a Gymnasium and cafeteria at School 2; Classroom 3 and Classrom 4 at School 6.
b Estimated total PCB emission rates from caulk based on PCB emission parameters derived from chamber measurements of
other caulks at 23 °C.
0 Percent of the total PCB emissions entering the room air. For the two classrooms, the PCB-containing caulk was around the
two exterior windows, one exterior unit ventilator intake grill, and exterior concrete/brick joints below the windows. Values
of 1%, 5%, and 10% of the PCB emissions entering the classrooms from the exterior caulk were used. For the gymnasium
and cafeteria, all caulk was located inside the rooms and a value of 100% was used.
d Predicted air concentrations at several possible air exchange rates for the room. Calculations based on the estimated
PCB emission rate and ventilation rate (seeEq. 2.1.6). Assumes 23°C, steady state emission and ventilation conditions,
steady-state absorption/desorption from other room materials, complete room air mixing, the concentration of PCBs in the
ventilation air is zero, caulk is the only source of PCBs, and that the PCB mixture in the caulk is equivalent to Aroclor 1254.
e Air samples were collected under different temperature conditions; gym and cafeteria @ 28°C; classrooms 3 and 4 @
27°C. The exterior temperature for classrooms 1 and 2 ranged from an overnight low of 19°C and a daytime high of 33°C.
The estimated emission rates and predicted room air concentrations may be underestimates relative to actual conditions
because they are based on emission parameters generated at 23 °C.
-------
overall ventilation rate for a room, but do not always provide
accurate information regarding how much of the total
ventilation air flow is occurring between the room being
assessed and other building spaces. This is important for
indoor pollutants such as PCBs, which can be generated
in other building spaces and transported between building
spaces - it can't be assumed that the ventilation air coming
into a room from other interior building spaces is PCB-free.
It is also important to recognize that AERs for a room
can change rapidly depending on changes in operation of
mechanical ventilation systems and opening and closing of
doors and windows.
Actual ventilation rates were not measured in the schools
included in this report at the time of sample collection, and
ventilation rates may change substantially under different
conditions of mechanical ventilation system operation,
window and door opening, human activity, and temperature.
Screening-level estimates were prepared using air exchange
rates (AER) of 0.5, 1.0, 2.0, and 4.0 air changes per hour
(assuming that all of the ventilation air was from outdoors
with low/no PCBs). For comparison, current ANSI/
ASHRAE standards call for 7.4 L/s (15 cubic feet per minute)
per person for outdoor air ventilation rates in classrooms with
5-8 year old children; giving an overall AER with outdoor
air of 3.3 ach for 25 people in a 200 m3 classroom. Forty- to
fifty-year-old school buildings may not meet current ANSI/
ASHRAE standards. The gymnasium described in Table
4-8 had a ventilation system design AER of 4.33 and had
measured AERs of 1.75 and 1.52 using ventilator airflow
measurement and tracer decay measurement, respectively
(NYC SCA, 2012). Measurement-based values are not
available for the other three rooms used in this screening-
level example, but it is anticipated that the range of 0.5 - 4.0
AER will cover the range of likely ventilation rates under
most use conditions. Thus, screening level estimates were
generated for a range of possible AER.
Exterior caulk with high PCB levels was present at two
windows for each of two of the classrooms in Table 4-8.
These classrooms had no interior caulk with high PCB
levels. It is not known what fraction of the PCBs emitted
from outdoor installations will enter indoor spaces. The
amount is likely quite variable depending on the location and
extent of the PCB-containing caulk, the temperature, wind
speed, whether windows are open or closed, whether caulk
is present at ventilation system intake openings, and the
operation of the ventilation system. In this screening-level
estimation exercise, indoor air concentrations were estimated
using a range of penetration rates including 1, 5, and 10% of
total PCB emissions.
Screening-level estimates of indoor air total PCB
concentrations under different AER conditions are shown in
Table 4-8 for four rooms. A wide range of possible indoor
air PCB concentrations resulting from emissions from PCB-
containing caulk were found under the range of assumed
AER and penetration conditions. Indoor air concentrations
measured at one time point are shown for comparison. In
general, measured concentrations were consistent with
indoor air levels that might be present from caulk emissions
at the Cafeteria and Classroom 4. Measured concentrations
were higher than the largest estimated concentrations for the
Gymnasium and Classroom 3. There are many assumptions
in these estimates and the results can be considered only
screening-level. Reasons for the differences could include
higher room temperatures as compared to chamber testing
temperatures, higher penetration rates from exterior caulk
emissions, the presence of other sources (such as emissions
from light ballasts or fixtures), or ventilation from more
highly contaminated spaces within the building.
4.2.2 Light Ballasts
Fluorescent light fixtures installed in buildings prior to
the late 1970s often used ballasts with PCB-containing
capacitors. Some of those fixtures and ballasts may remain in
school buildings today. PCB emissions from intact ballasts
have been demonstrated in laboratory chamber studies and
have been shown to be highly temperature dependent (Guo et
al., 2011; Hosomi et al., 2005). PCB-containing ballasts are
likely to have exceeded their expected operational lifetime,
and capacitors in ballasts can fail with PCBs rapidly released
into the building environment. One ballast capacitor burst
during laboratory chamber emissions testing, resulting in
very high levels of PCBs in the air and on chamber surfaces
(Guo et al., 2011). Fluorescent light ballasts may be
important sources of PCBs in building environments.
Light ballast capacitors for schools of this age are most
likely to contain Aroclor 1242 or Aroclor 1016, which means
that emissions of more volatile, lower chlorine-number
congeners would be expected to be released into the school
environment. However, use of Aroclors 1221 and 1254 in
capacitors was also reported (U.S. EPA, 1976). Capacitor
oil was analyzed in three ballasts tested by Guo et al. (2011)
and all were identified as Aroclor 1242. NYC SCA analysis
of school light ballast capacitor oil was analyzed showed
that Aroclor 1242 was present in the single ballasts tested
from Schools 1, 3 and 4, while Aroclor 1254 was present in
a ballast from School 2. Ballast capacitor oil measurements
were not available for Schools 5 and 6.
Survey results are available for fluorescent light fixtures
at five schools (Table 4-9). Between 24 and 95% of the
surveyed ballasts were likely to be PCB-containing. The
total number of ballasts that were likely to be PCB-
containing ranged as high as 879 at one school. A breakdown
of the ballast survey result by room at School 6 is shown in
Table 4-10. Temperatures at the surfaces of several operating
ballasts (with lights on) were measured at School 6 and
ranged from 48°C to 54°C.
No emission estimates have been made for two conditions
that might be present in buildings. First, some ballast
capacitors may have previously failed and leaked contents
into light fixtures. Residues may serve as a source of PCB
emissions. Second, some ballasts may be leaking and have
emission rates higher than those measured in chamber
testing. At this time there are no data suitable for estimating
emission rates for these two conditions. It is not clear
-------
Table 4-9. PCB-containing fluorescent light ballast survey results from five schools
School
1
2
3
4
6a
Ballasts
Not Containing
PCBs
310
114
344
48
25
Ballasts
Likely Containing
PCBs
417
373
275
879
8
Total
Ballasts in
School
727
487
619
927
33
% Ballasts
Likely Containing
PCBs
57%
77%
44%
95%
24%
1 Only a subset of ballasts in School 6 were surveyed.
Table 4-10. PCB-containing fluorescent light ballast survey results at School 6
Location
Classroom 1
Lab Classroom 2
Classroom 3
Classroom 4
Shop Classroom 5
Cafeteria
Gymnasium
Total across locations
Number of Ballasts
Visually Examined3
1
5
5
16
2
4
0
33
Number of Ballasts
Not Containing PCBsb
1
4
5
13
0
2
--
25
Number of Ballasts
Likely Containing
PCBsc
0
1
0
3
2
2
--
8
1 In-place visual examination of ballasts to examine ballast label.
b Ballast labels explicitly state "No PCBs". Ballasts included Sylvania Quicktronic QT 2X32/120 IS; Advance REL-2P32-SC; Advance
R-2S40-1-TP; Advance RQM-2S40-3-TP; Phillips Advance ICN-2P32-N; Universal 446-L-SLH-TC-P
c Ballast labels did not state "No PCBs". All ballasts were General Electric 7G1020B.
-------
whether the one ballast tested in laboratory chambers by
Guo et al. (2011) that had a much higher emission rate than
the other three ballasts that were tested represents a leaking
ballast condition.
Total PCB emission rates were estimated from four intact
ballasts that were tested at 23 °C and 45°C in a laboratory
chamber (Section 2). Those emission rates were used
to generate a range of estimated total emission rates for
likely PCB-containing ballasts in three classrooms, with
an assumption that the emission rates from the ballasts in
the classroom would fall within the range of tested ballasts.
Total estimated emission rates at 45°C ranged from 1.2 ug/hr
from three ballasts at the lowest emission rate, up to 290 ug/
hr from nine ballasts at the highest emission rate. The range
of potential emission rates is shown in Figure 4-2. Total
estimated emission rates at 23°C ranged from 0.08 ug/hr
from three ballasts at the lowest emission rate, and up to 18
ug/hr from nine ballasts at the highest emission rate.
The estimated emission rates for intact light ballasts in the
three classrooms were used to generate screening level
estimates of the resulting indoor air PCB concentrations
that might occur under different assumed conditions of
ventilation with outdoor air. The approach for calculating
screening-level PCB indoor air concentration estimates was
described in Section 2, using Equation 2.1.7. Numerous
assumptions apply to this estimation approach, and results
reported in Table 4-11 are considered to be only screening-
level because information is not available to understand
whether all of the assumptions are correct for the light
ballasts and conditions in the school rooms. Screening-level
estimates of total PCB indoor air concentrations ranged from
1.6 to 2400 ng/m3 for ballasts at 45°C, with median values
ranging from 2.3 to 44 ng/m3. Screening-level estimates
of total PCB indoor air concentrations ranged from 0.1 to
150 ng/m3 for ballasts at 23°C, with median values ranging
from 0.14 to 2.8 ng/m3. The lowest and median screening-
level estimates of total PCB indoor air concentration are
substantially lower than concentrations measured at one
point in time, which ranged from 690 to 1460 ng/m3. If all
of the ballasts in a room had emission rates similar to the
highest-emitting ballast in the chamber tests, it is possible
that emissions could result in indoor air concentrations
approaching those that were measured.
4.2.3 Secondary Source Characterization
Secondary sources of PCBs are defined here as materials
that have become contaminated due to absorption of PCBs,
either from direct contact with primary sources such as
caulk, or through absorption of PCBs in the indoor air that
have been emitted by caulk, light ballasts, or other primary
sources. There are numerous materials and furnishings in
buildings that have the potential to absorb semi-volatile
organic chemical like PCBs. These materials can include
paints, dust, foam, masonry, floor and ceiling tiles, mastics,
wood, cork and pin board, and many others. After years
of exposure to PCBs from primary sources, enough PCBs
300
School 2 Classroom
With 9 PCB Light Ballasts
250 -
200 -
LU
m
O
D.
"ro
'o
Figure 4-2. Example of the estimated range of total PCB emission
rates from fluorescent light ballasts in a school classroom using the
lowest, median, and highest rates from chamber tests of four ballasts
may have become absorbed to be become emission sources
once the primary sources have been removed or otherwise
remediated. These materials are also likely to continue to
serve sinks as well as sources, due to their continued ability
to absorb PCBs from the air and their often large surface
areas. The materials' net effect as secondary sources depends
on source-sink dynamics, which is affected by a number of
factors including PCB concentration in materials and indoor
air, diffusion and partition coefficient parameters, ventilation
and room air flow rates, and temperature.
A wide range of building material samples was collected
at Schools 2, 3, and 6. Materials were analyzed for total
PCBs as Aroclors; Aroclor 1254 was reported for most
materials, sometimes with an altered Aroclor pattern. Total
PCB measurement results for 411 materials are shown in
Table 4-12. When considering measurement results for these
samples it is important to remember that multiple samples
of the same type of material might have been collected from
several places in a school building. Across all 411 materials,
93% had PCB levels higher than the quantitation limit, with
a median concentration of 16.1 ppm (interquartile range 6.1
- 39.6 ppm) and a maximum value of 718 ppm. Paint had
the highest total PCB concentrations with a median of 39
ppm (interquartile range 25.9 - 71.7 ppm). Fiberboard had
a median level of 30.9 ppm (interquartile range 13.0 - 38.7
ppm), while lower levels were found in some other materials
that often have high surface areas in buildings such as ceiling
tile with a median 7.6 ppm (interquartile range 2.7-11.8
ppm) and floor tile with a median of 4.4 ppm (interquartile
range of 1.4 - 8.7 ppm). Results for several material types
-------
Table 4-11. Screening-level comparison of predicted air concentrations resulting from PCB emissions from fluorescent
light ballasts to measured concentrations
Room and Room
Balllast Volume
Condition3 m3
Usina Emission Estimates at 45°C
Classroom 1 231
Lowest
Median
Highest
Classroom 2 249
Lowest
Median
Highest
Classroom 4 182
Lowest
Median
Highest
Usina Emission Estimates at 23°C
Number Estimated
of PCB Emission
Predicted Total PCB Air
Concentration in Rooms at Measured
Different Air Exchange Rates0 Total PCBs
ng/m3 In Room
Containing Rate" AER= AER =
Ballasts ug/hr 0.5 1.0
(near the temperature of liaht ballasts when
5
0.416
0.614
32.7
9
0.416
0.614
32.7
3
0.416
0.614
32.7
(near the temperature of liaht ballasts when
liahts are on)
18
27
1400
30
44
2400
12
18
980
liahts are off)
9.0
13
710
15
22
1200
6.2
9.2
490
AER =
2.0
4.5
6.6
350
7.5
11
590
3.1
4.6
240
AER= Air"
4.0 ng/m3
1460
2.2
3.3
180
859
3.8
5.6
300
690
1.6
2.3
120
Classroom 1 231 5
Lowest 0.026 1.1 0.56 0.28 0.14
Median 0.038 1.7 0.83 0.42 0.21
Highest 2.04 88 44 22 11
1460
Classroom 2 249 9
Lowest
Median
Highest
Classroom 4 182 3
Lowest
Median
Highest
0.026
0.038
2.04
0.026
0.038
2.04
1.9
2.8
150
0.78
1.1
61
0.94
1.4
74
0.39
0.57
31
0.47
0.69
37
0.19
0.29
15
859
0.23
0.35
18
690
0.10
0.14
7.6
a Classrooms 1 and 2 at School 2; Classroom 4 at School 6. Several estimates are made using the lowest, highest, and median
emission rates from four ballasts tested in a chamber. Many light fixtures in schools show evidence of leaking or ballasts
that had previously failed. There is insufficient information to estimate emissions when these conditions are present.
bEstimated total PCB emission rates from light ballasts based on PCB emission rates measured for several congeners from
chamber measurements of four intact ballasts at several temperatures.
0 Predicted air concentrations at several possible air exchange rates for the room. Calculations based on the estimated PCB
emission rate and ventilation rate (see Eq 2.1.7). Assumes steady state emission and ventilation conditions, steady-state
absorption/desorption from other room materials, complete room air mixing, the concentration of PCBs in the ventilation
air is zero, the only source of PCBs is the ballasts, and that the PCB mixture in the caulk is equivalent to Aroclor 1242. No
caulk with high PCB levels was found inside these classrooms. Caulk with PCBs was in hallways outside Classrooms 1 and
2 and on the exterior window and unit ventilator intake grill for Classroom 4.
d Air samples were collected under different temperature conditions; Classrooms 1 and 2 @ 28°C; classroom 4 @ 27°C.
-------
Table 4-12. Total PCB measurement results for materials at three schools with available data
Total PCB Levels in Other Materials3
Material Category
All Material Samples
Paint
Fiberboard
Chair Rail/Radiator Cover
Baseboard Cove Molding
Foams/Pinboard/Corkboard
Particle Board
Varnish
Mastics (Tile and Molding)
Ceiling Tile
Wood
Laminate
Floor Tile
Ventilator Insulation
Oils and Cleaner (liquids)
Wall Concrete Block
N
411
143
28
6
5
8
19
30
65
8
6
29
56
2
3
3
QLb
93
100
100
100
100
100
100
97
92
100
100
76
82
100
0
0
Median
ppm
16.1
39.1
30.9
30.5
28.0
15.5
13.5
11.4
7.83
7.59
7.46
5.35
4.43
1.75
-------
Table 4-13. Total PCB measurement results for subsets of paint uses
Total PCB Levels
Material Category
All Paints
Wall Paint
Ceiling Paint
Metal Paint (radiator, locker, etc)
Floor Paint
Door and Door Trim Paint
Basketball Backboard Paint
Handrail Paint
N
143
36
28
31
4
33
2
9
% > QLb
100
100
100
100
100
100
100
100
Median
ppm
39.1
29.1
30.5
36.3
54.9
55.0
63.3
121
Inter-Quartile
Range ppm
25.9 -
21.8 -
14.5 -
27.6 -
34.6 -
37.5 -
-
57.2 -
71.7
41.8
46.6
78.0
83.3
87.0
132
in Paints3
Overall Range
ppm
3.31
3.31
3.43
7.00
32.6
5.54
50.8
49.1
- 718
- 129
- 227
- 382
- 110
- 718
- 75.8
- 172
a Reported as total PCBs from Aroclor measurements for materials at Schools 2,3, and 6.
b QL = quantitation limit; sample size dependent, typically < 1 ppm.
amount of PUF in most older school buildings is relatively
small, particularly when compared to residential settings.
Interestingly, the the concentration of PCBs in PUF for the
few samples that were collected were generally lower than
the concentrations in paints in the same buildings.
Understanding the relative potential of different secondary
sources to emit PCBs is of interest for informing remedial
action, should it be needed. Information about materials
in nine interior spaces across three school buildings was
organized to better understand both the relative potential
source impact as well as the cumulative potential impact from
PCB emissions from secondary sources. Screening-level
estimates of emission rates for multiple materials in these
rooms were calculated following the approach described for
caulk in Section 2. Material descriptions, surface areas, total
measured PCB concentrations, and screening-level estimates
of emission rates are shown in Tables 4-14, 4-15, and 4-16.
Figures 4-3, 4-4, and 4-5 show the relative screening-
level emission rates for materials in three of the rooms. It
is important to note that the emission rates estimated for
secondary sources are applicable only after primary sources
have been removed or otherwise mitigated. There are
considerable uncertainties in these estimates, in part because
they assume that the emission parameters derived for caulk
in chamber testing apply to the wide range of different
materials. In practice, the emission parameters for PCBs
may be considerably different for many materials. Therefore,
these are only screening level estimates for the purposes of
relative comparisons.
Screening-level estimated emission rates for different
materials in classrooms ranged from < 1 up to 100 ug/
hr and cumulative totals for 20 materials in a room ranged
up to 270 ug/hr. Estimated emission rates for different
materials in gymnasiums ranged from < 1 up to 1100 ug/
hr and cumulative totals for 16 materials ranged up to 2700
ug/hr. Estimated emission rates depended on the surface
area of the material and concentration of PCBs in the
material. Paints and varnishes generally had the highest
relative potential emissions due to the combination of higher
PCB concentrations and high surface areas. There are
considerable uncertainties in these estimates, which are based
on emission parameters derived from laboratory emissions
testing of caulk. Emission parameters for the many different
types of other materials could be substantially different
than those for caulk. It is difficult to estimate indoor air
concentrations of PCBs that might result from secondary
sources following removal of primary sources because
of the large number of different types of PCB-containing
material in a room, and because the cumulative source -
sink dynamics for multiple different materials is difficult
to characterize. However, the cumulative emission rates
from secondary sources could potentially result in indoor air
PCB levels above ambient air background levels in school
rooms following mitigation of primary sources, depending
on relative emission rates, sink rates, and rates of ventilation
from indoor and outdoor air.
4.2.4 Source Assessment Uncertainties and Limitations
Characterizing the source(s) of PCBs in and around school
buildings is important because it will inform remediation
approaches for cases where exposure reduction decisions
need to be made. PCB source assessment for buildings can
be difficult because there may be multiple primary sources,
and transport of the semi-volatile congeners through air
can contaminate dust and soil and create secondary sources
of other materials in a building. Information from six
school buildings examined in this work was used to try
to characterize and to understand the relative potential of
various PCB sources. There remain important uncertainties
and limitations in this information and the emission rate
estimates as discussed below.
Attribution of primary sources - The school buildings
examined in this work had both caulk with high PCB
concentrations and PCB-containing light ballasts. It would
be helpful to understand whether one of the primary sources
-------
Table 4-14. Screening-level estimates of total PCB emission rates for selected interior materials in three locations at
School 2 for relative comparisons"bc
School
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Room
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Classroom
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Corridor
Corridor
Corridor
Material
Wall Paint - Blue
Door Paint - Blue
Wall Paint - Cream
Floor Tile - Grey
Radiator Paint - Blue
Door Varnish
Door Frame Paint - Blue
Closet Door Varnish
Fiberboard
Ceiling Tile
Ceiling Paint -White
Cove Molding
Ceiling Paint -Blue
Sink Door Varnish
Door Frame Varnish
Particle Board
Wormhole Ceiling Tile
Transom Glaze
Electrical Penetration Caulk
Toilet Caulk
Wall Paint - Cream
Wood Floor Varnish
Fiber Ceiling Tile
Ceiling Beam Paint - Cream
Backboard Paint - White
Radiator Paint - Blue
Baseboard Paint - Black
Duct Paint - Cream
Bench - Varnish
Bay Door Paint - Gray
Vent Paint - Blue
Door Paint - Blue
Door Frame Paint - Blue
Door Varnish
Bay Door Soffit Paint - Gray
Door Window Glaze
Electrical Penetration Caulk
Sink Caulk
Door Paint - Blue
Radiator Paint - Blue
Floor Tile -White
Surface Area
m2
102
3.90
19.5
59.5
9.66
7.80
3.34
8.36
12.4
52.0
4.65
3.21
2.79
1.86
0.743
0.929
1.21
0.0064
0.0052
0.0006
358
401
402
191
25.6
20.0
5.88
46.5
10.0
4.37
3.90
5.20
1.89
9.29
0.929
0.0520
0.0020
0.0117
23.8
20.1
121
Total PCB
ppm
36.7
444
51.5
16.5
52.0
48.2
102
39.5
26.5
4.48
32.0
42.3
29.6
13.4
28.9
10.3
2.11
51.0
5.30
9.50
51.5
32.2
10.7
20.4
50.8
52.0
115
12.1
34.8
59.1
52.0
37.5
102
12.9
59.1
15.0
68.0
7.40
116
92.0
12.1
Estimated
Emission Rate
ug/hr
100
47
27
27
14
10
9.3
9.0
9.0
6.4
4.1
3.7
2.3
0.68
0.59
0.26
0.070
0.009
0.001
0.0002
500
350
120
110
36
28
18
15
9.5
7.0
5.5
5.3
5.3
3.3
1.5
0.021
0.004
0.002
75
50
40
-------
Table 4-14. Screening-level estimates of total PCB emission rates for selected interior materials in three locations at
School 2 for relative comparisons"bc (continued)
School
2
2
2
2
2
2
2
2
2
Room
Corridor
Corridor
Corridor
Corridor
Corridor
Corridor
Corridor
Corridor
Corridor
Material
Ceiling Tile
Door Frame Paint - Blue
Door Varnish
Door Louver Paint - Blue
Fiberboard -Gray
Fiberboard - Brown
Floor Tile - Cream 1
Interior Door Window Glaze
Floor Tile - Cream 2
Surface Area
m2
133
5.79
18.6
0.465
7.43
5.57
9.29
0.0988
0.929
Total PCB
ppm
4.48
102
16.5
444
20.7
26.5
4.60
88.0
6.79
Estimated
Emission Rate
ug/hr
16
16
8.4
5.6
4.2
4.0
1.2
0.24
0.17
aNot all materials in each room were sampled. Materials such as mastics are not included here since they are not directly
exposed to room air.
b Based on chamber-derived emission parameters for caulk - these may not apply well to all materials.
0 These materials also act as sinks, absorbing PCBs from the room air. The estimated emission rates shown here cannot be
simply used to estimate total indoor air concentrations.
is much more important than the other with regard to
increasing environmental levels and potential exposures to
PCBs. There appears to be evidence that both caulk with
high PCB levels and PCB-containing light ballasts are likely
to be important sources of PCBs in older school buildings.
However, given the limitations of this study, it is difficult to
determine whether one source is likely to be more important
than another in school buildings.
One approach might be to examine congener patterns to
try to determine if a source signature can be elucidated. A
limitation was that most of the samples were analyzed for
Aroclors and not specific congeners. Some of the caulk
samples collected at the six buildings contained high
concentrations of PCBs (in the range of 1% to 44% by
weight). Aroclor analysis of these high-PCB caulks showed
that most contained a pattern consistent with Aroclor 1254, or
in the case of one school, a pattern resembling Aroclor 1260.
The capacitor fluids in several light ballasts were analyzed;
three of these contained Aroclor 1242 and one contained
Aroclor 1254. Because these schools were constructed
after 1952, when capacitor fluids reportedly transitioned
from Aroclor 1254 to Aroclor 1242, one might expect that
most ballasts contained Aroclor 1242. But with hundreds
of unmeasured light ballasts in each school and the finding
of a ballast with Aroclor 1254, it is not clear to what extent
different Aroclor capacitor fluids may have been present
in each building. Most of the air samples were reported
to contain an altered Aroclor pattern with characteristics
of Aroclors 1248 and 1254. Given that Aroclor 1242 is
primarily composed of 2-, 3-, and 4-chlorine homologs, and
Aroclor 1254 is primarily composed of 4-, 5-, and 6- chlorine
homologs, it would be surprising - based on the air PCB
composition - if sources with Aroclor 1242 (i.e. light ballasts)
were contributing a much higher amount of PCBs into the
indoor air than sources with Aroclor 1254 or 1260 (caulk
and possibly some light ballasts). This is particularly true
because the 2-, and 3- chlorine congeners have much greater
vapor pressures that the 5-chlorine congeners, and Aroclor
1242 sources would contribute an even higher fraction of
their PCBs to air than Aroclor 1254 sources. Congener-
specific measurements were obtained for one school (see
Section 4.4) and provide information for more directly
examining source and environmental relationships.
Estimates of total PCB emission rates for several examples
of school building caulk were made. As discussed later,
there are limitations and uncertainties in the estimates. But,
if those estimates are reasonable, then emission rates of
hundreds to thousands of micrograms of PCBs per hour at
specific school locations would appear to be possible. For
interior caulks, these levels of emissions would contribute
to increased indoor air PCB concentrations, and could serve
as a source for partitioning into dust and other materials.
Even assuming low penetration rates for emissions from
exterior caulks around windows and ventilation intakes,
exterior caulk could also contribute to increased PCB levels
in indoor air, but this may be highly dependent on ambient
temperature and wind conditions, as well as specific locations
and uses of windows for ventilation. Given the estimates of
emission rates from caulk, a question arises as to whether
all of the available PCBs might have been depleted over
the course of 40 - 50 years in a building, and whether these
rates have been overestimated. Guo et al. (2011) used a
modeling approach to show that over 50% of congener 52
would remain after 50 years in a building, and even higher
proportions of less volatile congeners would remain. It is
possible that some of the more volatile congeners, such as 8
and 18 might have been largely depleted in that time frame,
but these two comprise < 1% of all congeners, by weight, in
-------
Table 4-15. Screening-level estimates of total PCB emission rates for selected interior materials in three rooms at School
3 for relative comparisonsa'b'c
School Room
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
Classroom 1
Classroom 1
Classroom 1
Classroom 1
Classroom 1
Classroom 1
Classroom 1
Classroom 1
Classroom 2
Classroom 2
Classroom 2
Classroom 2
Classroom 2
Classroom 2
Classroom 2
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Surface Area
Material m2
Wall Paint - Cream
Ceiling Paint - Gray
Door Paint - Orange
Cove Base - Black
Floor Tile- Green
Fiber Board
Radiator Paint - Light Orange
Fiber Board - Green
Floor Tile - Gray
Wall Paint - Yellow
Ceiling Paint -White
Cove Base Paint - Black
Radiator Paint - Yellow
Fiber Board - Brown
Door Paint - Yellow
Wall Paint - Yellow/Green
Ceiling Paint -White
Cove Base Paint - Light Gray
Fiber Board - Brown
Door Paint - Light Gray
Particle Board - Brown
Floor Tile -Beige
53.9
85.5
5.57
6.04
89.2
11.1
4.83
3.34
82.9
59.5
62.3
6.69
14.4
9.29
1.86
51.1
60.4
3.53
10.2
1.86
0.93
7.43
Estimated Emission
Total PCB Rate
ppm ug/hr
29.4
10.7
108
40.9
1.73
13.0
27.5
21
9.69
13.5
10.5
40.9
9.71
13.0
29.6
27.1
9.47
35.9
10.5
29.7
12.9
-------
Table 4-16. Screening-level estimates of total PCB emission rates for selected interior materials in three rooms at School
6 for relative comparisons"bc
School
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Room
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 3
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Classroom 4
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Gymnasium
Material
Locker Paint - Tan
Wall Paint - Beige
Wall Paint -White
Ventilator Paint - Beige
Pin Board
Door Paint - Brown
Ceiling Tile
Floor Tile - Grey
Floor Tile - Beige
Chair PUF Foam
Cove Molding
Interior Window Glaze 1
Ventilator Gasket Foam
Interior Window Glaze 2
Interior Joint Caulk
Ventilator Insulation
Door Window Glaze
Wall Concrete Block
Wall Paint -White
Locker Paint - Tan
Door Paint - Brown
Ventilator Paint - Brown
Ceiling Tile
Floor Tile - Grey
Pin Board (2 boards)
Cove Molding
Cork Board
Interior Window Glaze 2
Interior Window Glaze 1
Interior Joint Caulk
Ventilator Insulation
Wood Floor Varnish
Wall Paint 1
Wall Paint 2
Floor Underlayment
Gym Floor Wood
Backboard Paint
Door Paint - Black
Bleacher Seat Wood
Cove Molding
Exercise Mat PUF Foam
Surface Area
m2
17.9
29.2
11.8
5.71
7.32
4.08
68.5
58.1
19.4
0.534
2.36
0.109
0.0277
0.0176
0.197
0.0987
0.0064
41.0
40.1
18.9
6.19
5.68
62.2
70.1
4.94
2.38
1.11
0.0697
0.0998
0.0955
0.0987
668
524
189
668
668
18.9
25.6
194
10.3
32.4
Total PCB
ppm
79.7
38.3
64.9
71.9
48.6
71.4
2.06
2.07
5.73
42.9
2.44
9.22
35.6
40.4
2.40
2.23
11.7
-------
Table 4-16. Screening-level estimates of total PCB emission rates for selected interior materials in three rooms at
School 6 for relative comparisons"bc (continued)
School Room
6 Gymnasium
6 Gymnasium
6 Gymnasium
6 Gymnasium
6 Gymnasium
6 Gymnasium
Material
Backboard Wood
Bleacher End Wood
Plywood Subfloor
Interior Joint Caulk
Pin Board
Interior Caulk - Fountain
Surface Area
m2
18.9
8.43
668
0.627
1.11
0.0108
Total PCB
ppm
9.53
14.8
0.125
21.7
8.79
5.21
Estimated
Emission Rate
ug/hr
4.9
3.4
2.3
0.37
0.27
0.002
a Not all materials in each room were sampled. Materials such as mastics are not included here since they are not directly
exposed to room air. The gym sub-floor and underlayment are included because it is not clear whether and to what extent
their emissions might impact room air.
b Based on chamber-derived emission parameters for caulk - these may not apply well to all materials.
0 These materials also act as sinks, absorbing PCBs from the room air. The estimated emission rates shown here cannot be
simply used to estimate total indoor air concentrations.
d QL = quantifiable limit.
50
I
03
cr
40 -
.<2 30
LJJ
CD
O
?: 20
I
ro 10 -
|
"GO
LU
School 6 Classroom 3
D
**$$$$$&
2i >t-
V
Figure 4-3. Screening-level estimates of total PCB emission rates from materials in the gymnasium in School 2
-------
Estimated Total PCB Emission Rate (i^g/h)
-j- w co .&. a
5 O O O O C
School 3 Classroom 1
U i i 1 i . . I 1
.- j/V
o* o*
^ y
cF
Figure 4-5. Screening-level estimates of total PCB emission rates from materials in a classroom in School 6
-------
Aroclor 1254. We have performed additional calculations
(not shown) based on our estimated emission rates, and
for reasonable initial total PCB concentrations that are
within the range that have been measured in caulks now,
substantial fractions of PCBs would still be present in the
caulk after more than 40 years. Other research has shown
that sealants with PCBs are an important contributor to
PCB levels in indoor spaces. For example, the average
total PCB concentration in indoor air in apartments
with PCB-containing sealants was 1030 ng/m3 (range
168 - 3843 ng/m3) (Frederiksen et al., 2012). Macintosh
et al. (2012) reported an average indoor air total PCB
concentration of 533 ng/m3 (range 299 - 1800 ng/m3) in
an elementary school with PCB-containing sealants but
no PCB-containing light ballasts. These results show that
emissions of PCBs from caulk, in the absence of light
ballast sources, are sufficient to create indoor air total PCB
concentrations two orders of magnitude or more higher
than outdoor ambient concentrations.
Estimates of PCB emissions from light ballasts were also
made. These estimates are limited and some caution is
recommended regarding their interpretation. Estimates
were based on the testing of four intact light ballasts at
multiple temperatures by Guo et al. (2011). One of the
four ballasts showed substantially higher emissions than
the other three, approximately 80-fold greater than the
ballast with the lowest emissions. Thus, it is not certain
what ballast-related emissions are likely to occur across the
large number of ballasts that may be present in a building.
If a larger percentage of ballasts emit at the highest rate,
then emissions from intact ballasts alone could have an
important impact on indoor air PCB concentrations. If
most of the intact ballasts emit at the lower levels, then the
impact on levels in air would be more modest. Perhaps
most importantly, this work was not able to provide
information on emissions resulting from failing ballasts
and for light fixtures and other building components
that may have been contaminated from previously failed
ballasts. Ballast capacitors that burst or suddenly fail and
leak their PCB contents will clearly have a substantial
impact on PCB levels in indoor spaces. The impact of
contaminated fixtures and other components is less certain
but could be important. The New York City remedial
investigation showed that remedial measures could
substantially reduce indoor air PCB concentrations for
schools with initially elevated levels, and a considerable
part of that decrease occurred after a school cleaning
step and removal of PCB-containing light fixtures (see
Figure 4-11). The Agency has prepared a guide for school
administrators and maintenance personnel titled "Proper
Maintenance, Removal, and Disposal of PCB-Containing
Fluorescent Light Ballasts" (http://www.epa. gov/epawaste/
hazard/tsd/pcbs/pubs/ballasts.htm) that recommends
removal.
Another uncertainty regarding sources of PCBs in school
buildings is that other primary sources of PCBs may
have been used in school buildings that are no longer
present today. For example, carbonless copy paper and
PCB-containing capacitors in early computer video display
terminals may have been present in school buildings during
a period of the school's history. The potential impact of
previously removed sources on current PCB levels in
building environments cannot be easily determined.
Importance of secondary sources for exposure - PCBs were
found to be widespread throughout many materials collected
in three buildings, mostly in the range of 4 to 100 ppm.
Given the indoor air PCB concentrations and the chamber
sink test results from Guo et al. (2012) these levels would
appear to be consistent with the materials being sinks that
have absorbed PCBs from the air over many years. A few
of the paint samples had higher levels of PCBs, up to 718
ppm, and it can't be ruled out that some of the materials
contained PCBs when they were installed in the building.
Regardless of their provenance, these materials represent a
reservoir of PCBs in a building that might, in some cases,
need to be considered as part of building mitigation efforts
to reduce environmental levels of PCBs. While screening-
level emission estimation approaches were used to assess
the relative potential importance of various materials,
the report is limited about what it can say quantitatively
with regard to the impact these secondary sources will
have on environmental levels of PCBs following removal
or mitigation of primary sources. This is because the
cumulative source-sink dynamics are difficult to predict
for multiple widespread materials with various PCB
concentrations, particularly given the lack of important
diffusion and partition parameters for most of the materials
and the range of PCB congeners.
Emission estimation uncertainties - Emission estimates in
this report are largely based on the work of Guo et al. (2011)
and their laboratory chamber testing of PCB-containing
caulk and light ballasts, combined with measurements and
characterization of PCBs in materials and components in the
schools. Caulk emission rate parameters were determined by
Guo et al. in the laboratory using micro-chambers because
it was determined that large losses of PCB congeners to
chamber walls were occurring for caulk emissions testing
in larger chambers. This problem did not occur with the
micro-chambers. Numerous samples of caulk obtained from
older buildings were tested for emissions using the micro-
chambers, and consistent congener emission parameters were
obtained from the different caulks. Several light ballasts
were obtained from older buildings and were tested in larger
55-L chambers. Four of these ballasts were tested at multiple
temperatures. Congener emission rates were found to be
variable across the different ballasts. Because these ballasts
contained Aroclor 1242, with proportionally more volatile
congeners, loss to chamber walls was less significant than for
the less-volatile congeners from Aroclor 1254 in the caulks,
but the emissions estimates from light ballasts in chamber
tests may still be somewhat underestimated due to losses.
Because the caulk and light ballasts were tested in two
different chamber systems, there could be some limitations
in making direct comparisons using emission factors derived
from the testing. However, the micro-chambers have been
-------
shown to produce emissions estimates comparable to those
from larger chambers for VOCs and SVOCs (Scheff et al.,
2000). Material/air boundary conditions are affected by
surface air flow velocities and these could be different across
the two types of chambers and between the chambers and the
surfaces in school rooms (which are typically in the range
of 2 to 25 cm/s). The micro-chamber manufacturer reports
that air velocity is in a range of 0.5 cm/s at 50 mL/min and
5 cm/s at 350 mL/min. At caulk test flow rate of 449 mL/
min, the estimated velocity would be about 6.4 mL/min. A
published paper (Scheff et al., 2000) showed an empty micro
chamber likely had velocities less than 10 cm/s. However,
it is possible that the caulk testing configuration could have
resulted in somewhat higher velocities.
Even if the boundary air flow conditions during caulk
emissions testing in micro-chambers was substantially
different than the surface air velocities across caulk in school
spaces, the impact on estimated emissions likely would
not be large. The rate determining step for evaporation of
a chemical from a pure liquid into air is governed largely
by mass transfer in the boundary layer immediately above
the liquid. The depth of the boundary layer depends in part
on the air velocity across the surface, and this can be an
important factor in measuring and estimating emissions.
For solid materials, the emission rate is controlled by
two factors, the boundary layer conditions and the rate of
chemical diffusion from inside the material to the surface of
the material. The boundary layer affects the emission rate at
early times of the emissions and the effect diminishes over
time. For long-term emissions, particularly for SVOCs,
the internal diffusion becomes more important as the rate-
determining step and the boundary layer has more limited
effect on the emission rate (Qian et al., 2007).
Efforts were made to better understand whether the predicted
caulk emission rates were realistic as compared to estimates
of PCB emissions from evaporation from pure PCB liquids.
Emission estimates from caulks in school buildings in this
report would appear to higher than would be expected
based on an early QSAR-based estimate of the evaporation
rate of pure Aroclor 1254 (U.S. EPA, 1976) but would not
overestimate evaporation rates derived from more recent
modeling approaches (Hummel et al., 1996; Guo, 2000) (data
not shown).
There are still considerable limitations and uncertainties
regarding the predicted emission rates and air concentration
estimates presented in this report. The actual emission rates
were not directly measured and the estimations were based
on a number of assumptions. For caulk, it was assumed that
the PCB emission parameters for the school caulks were
the same as those in the chamber testing and estimates were
based on an assumed temperature equivalent to that used
in the chamber testing. The chamber testing showed that
PCB emissions from caulk were sensitive to temperature,
with an approximate six-fold increase in emissions for
a 10°C increase in temperature. In estimating room air
concentrations resulting from caulk, many other assumptions
were made that might not be true or might change from time
to time in a school building. Assumptions include:
• well mixed air in room
• constant temperature
• temperature equivalent to chamber conditions that
generated caulk emission parameters (23°C)
• constant ventilation rate
• steady-state emission
• steady state and approximately equal absoption/
desorption of PCBs in other materials in the room
• no chemical reactions of PCBs
• PCBs from other school spaces are not impacting the
levels in air for the room of interest
• emission parameters for caulk in the room are the same as
for the caulk tested in lab chambers
Because congener-specific measurements were not performed
for most of the caulk samples, we estimated total PCB
emission rates by first assuming that the caulk contained an
un-aged Aroclor 1254 mixture. It is likely that over time
the congener composition has changed and current emission
rates may not match those made assuming an un-aged
Aroclor 1254. We examined this where we had congener
specific data and found that the emission rate estimated
using actual congener composition was 45% lower than
using an assumption of un-aged Aroclor 1254. Also, in the
emissions testing by Guo et al., three caulks were tested to
assess the effects of having freshly cut surfaces. Emissions
from freshly cut surfaces were, on average, about 19% higher
than those from the original surface. To limit potential
overestimation of PCB emissions from caulk, we elected to
adjust (reduce) the total PCB emission rates by 55% based on
the available information. However, this adjustment factor
was based on limited data.
As noted earlier, the information available for estimating
PCB emissions from light ballasts was limited, and it is not
clear how well those emission rates apply to the hundreds
of ballasts in the schools. Likewise, the emissions of PCBs
from contaminated light fixtures and other components have
not been characterized, so it is possible that screening-level
estimates of PCB concentrations school room air could
be underestimated. Most of the assumptions listed for air
concentrations estimated from caulk emissions also apply to
estimates from light ballasts.
-------
4.3 PCBs in Environmental Media
PCBs are semi-volatile organic chemicals and may
experience transport from sources into and throughout the
environment in and around school buildings where they
can become available for human contact and exposure. In
order to understand the potential for exposure and ranges
of possible exposures, it is important to characterize
PCB concentrations in environmental media. Indoor air,
surface wipe, and outdoor soil samples were collected from
multiple locations at the six schools. Indoor dust samples
were collected from multiple locations in one school
building. Outdoor air samples were collected at all school
buildings. At the five NYC schools, indoor air and surface
wipe samples were collected before and following different
remedial actions. Measurement results are summarized
here to characterize the magnitude, range, and within- and
between-school variability in PCB concentrations in school
environments. A summary of total PCB concentrations for
indoor and outdoor air, surface wipes, soil, and dust is shown
in Table 4-17.
4.3.1 Indoor Air
Indoor air samples were collected in several locations in each
school. Collection locations included classrooms, cafeterias,
gymnasiums, and transitory spaces. The median indoor air
total PCB concentration based on 64 measurements across
six schools was 318 ng/m3 (interquartile range of 59.4 - 732
ng/m3) with a maximum concentration of 2920 ng/m3 (Table
4-18). These samples were collected prior to any of the
reported remedial activities as shown in Table 4-2. There
was considerable variability between schools with median
air levels at individual schools ranging from <50 ng/m3 at
School 4 to 807 ng/m3 at School 2. The distribution of indoor
air PCB concentrations is shown in Figure 4-6. There was
considerable variability within schools; for example, indoor
air levels ranged from 236 to 2920 ng/m3 in different rooms
at School 3.
Indoor air concentrations at these schools can be compared
to measurements at other buildings. Macintosh et al. (2012)
reported a median pre-remediation value of 429 ng/m3 at
a U.S. primary school for samples collected in nine indoor
locations. Concentrations ranging from 2 to 310 ng/m3 were
measured in a U.S. secondary school (TRC, 2006). Coughlan
et al., (2002) reported indoor concentrations ranging from
111 to 393 ng/m3 inaU.S. university building. Indoor PCB
concentrations measured at three of the six schools in this
report had higher concentrations than those reported by
Macintosh et al., TRC, and Herrick et al., while the indoor
levels for the other three schools in this report were similar to
or below those previously reported.
The U.S. EPA developed information in 2009 on public
health levels of PCBs in school indoor air (Figure 4-7). If
school indoor air levels are kept below these concentrations,
Table 4-17. Summary of environmental media total PCB measurement results for six schools3
Total PCB Levels"'0
Environmental Medium (units) Nd
%
>QLe
Mean
QL
Median
Inter-Quartile
Range
Overall
Range
Indoor Air (ng/m3
64
77
47
318
59.4 - 732
-------
Table 4-18. Indoor air total PCB measurement results at six schools
School/
Condition
All Six Schools
School 1
School 2
School 3e
School 4
School 5
School 6
Nc
64
11
12
14
9
11
7
QLd
ng/m3
47
49
50
51
50
50
24
Total PCB Levels in Air3'"
%
>QL
77
54
100
100
44
54
100
Median
ng/m3
318
58
807
504
-------
Public Health Levels of PCBs in School Indoor Air (ngiii3)
Assuming a background scenario of 110 significant PCB contamination in building materials and
average exposure from other sources, these concentrations should keep total exposure below the
reuce dose of 20 112 PCB/kg-day.
Age
l-<2 yr
70
Age
2-<3 yr
70
Age
3-<6 yi
100
Age
6-<12yr
Elementary
School
300
Age
12-<15yr
Middle
School
450
Age
15-<19yr
High
School
600
Age
19+yr
Adult
450
Figure 4-7. Public health levels of PCBs in school indoor air developed in 2009 by the U.S. EPA
(http://www.epa.gov/pcbsincaulk/maxconcentrations.htm)
Table 4-19. Outdoor air total PCB measurement results
at six schools3
School/
Condition N
School 1 1
School 2 1
School 3 1
School 4 1
School 5 1
School 6 1
Mean
QLb
ng/m3
49
50
51
50
49
17
%
>QL
0
0
0
0
0
0
Total PCB
Levels in
Air
ng/m3
t 1-2'
-------
Table 4-20. Surface wipe total PCB measurement results at six schools
Total PCB Levels in Surface Wipes3'"
School/Condition
All Six Schools
High-contact surfaces
Low-contact surfaces
School 1
High-contact surfaces
Low-contact surfaces
School 2
High-contact surfaces
Low-contact surfaces
School 3
High-contact surfaces
Low-contact surfaces
School 4
High-contact surfaces
Low-contact surfaces
School 5
High-contact surfaces
Low-contact surfaces
School 6
High-contact surfaces
Low-contact surfaces
Nc
72
78
11
11
12
12
14
14
9
9
12
12
14
20
QLd
62
80
54
46
83
83
57
86
44
100
33
83
93
80
Median
ug/100cm2
0.147
0.201
0.106
-------
Table 4-21. Indoor dust total PCB measurement results at School 6
Total PCB Levels in Indoor Dusta'b
School
School 6e
Nc QLd
7 100
Median
ppm
22.0
Inter-Quartile
Range
ppm
16.6 - 53.4
Overall Range
ppm
11.6 - 86.8
a Reported as total PCBs from Aroclor measurements in dust.
b When duplicate samples were collected, the average of the duplicates was used.
0 Number of samples and number of rooms sampled.
dQL = Quantitation limit; ranged from 0.54 to 14.2 ppm depending on sample size
eThis school building had not been cleaned in the approximately 5 weeks prior to sample collection; the residence time for
PCB absorption may have been longer than for most school situations.
-------
Table 4-22. Soil total PCB measurement results at six schools
Total PCB Levels in Soila'b
School/Condition
All Six Schools
All soil samples
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
School 1
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
School 2
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
School 3
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
School 4
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
School 5
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
3.7/4.9m from building; 0-5cm depth
School 6
0.15m from building; 0-5cm depth
0.91m from building; 0-5cm depth
2.44m from building; 0-5cm depth
15.2m from building; 0-5cm depth
Nc
309
99
102
105
12
12
26
30
29
27
8
9
14
35
38
31
12
12
5
2
2
2
2
1
%
>
QLd
33
48
28
21
50
8
8
13
7
11
100
100
100
49
18
0
92
67
40
100
100
100
50
0
Median
ppm
-------
total PCB concentrations across six schools at 0.15, 0.91,
and 2.44 m from the building were 2.13, 0.55, and <0.5 ppm
respectively. Soil PCB concentrations decreased with
increasing distance from the school building. There was
considerable variability between schools; for example one
school had only 10% of the PCB levels above the quantifiable
limit while another school had 100%. Distributions of the
soil concentrations are shown by school in Figure 4-9 where
there were sufficient measurable results.
Additional soil samples were collected from the 5 to 10
cm (2 to 4 inch) depth at School 1. A comparison of the
results from the 0-5 and 5-10 cm samples are shown in
Table 4-23. The results are difficult to interpret. At both the
0.15 and 0.91 m distances, the median concentration at the
5 - 10 cm depth was greater than the median concentration
at the 0 - 5 cm depth. On the other hand, the maximum
concentrations at the 0 - 5 cm depth were greater than the
maximum values at the 5 - 10 cm depth for samples collected
0.15 and 2.44 m from the building. Concentrations at the
0 - 5 cm depth are likely to be more relevant for assessing
exposure potential.
The geometric mean total PCB concentration for school soils
from extant data (Appendix D) was 0.98 ppm with a range
up to about 80 ppm. The median value for the six schools
in this report was less than the QL (a value of 0.5 ppm for
most samples), lower than the extant data geometric mean.
On the other hand, the maximum value from these schools
(211 ppm) was greater than the extant data maximum. There
are few reports of PCB levels in urban soils in the United
States. One recent report from Cedar Rapids, IA showed a
HNote:
/
1 0
2 -
The 75th, 50th, 25th,
and 10th percentile values
for Schools 1, 2, and 4 were
all below the quantifiable limi
Only the 90th percentile value
are shown.
\
All 6
234
School
Figure 4-9. Distributions of outdoor soil total PCB concentrations
across all six schools and at each individual school. The box plots
show the median, 25th, and 75th percentiles. The whiskers show the
10th and 90th percentiles.
mean total PCB level of 0.056 ±0.16 ppm with a maximum
value of 1.2 ppm (Martinez et al., 2012). The authors report
that the mean is about 10 times higher than the world-wide
background levels based on the sum of 29 congeners. The
75th percentile level of PCBs in soils adjacent to school
buildings in this study is considerably higher than the 0.045
ppm 75th percentile value for urban soils in Cedar Rapids, and
much greater than background non-urban levels.
4.3.6 Pre- and Post-Remediation Air and Wipes
As part of the NYC SCA remedial pilot investigation,
various remedial actions for reducing potential exposures
to PCBs were examined in five NYC schools. These
Table 4-23. Soil total PCB measurement results for two soil depths at School 1
Total PCB Levels
School/Condition
School 1
0.15m from building; 0-5cm depth
0.15m from building; 5-1 Ocm depth
0.91m from building; 0-5cm depth
0.91m from building; 5-1 Ocm depth
2.44m from building; 0-5cm depth
2.44m from building; 5-1 Ocm depth
Nc
12
12
12
12
26
26
>QLd
50
67
8
58
8
12
Median
ppm
0.910
1.43
-------
remedial activities are briefly summarized in Table 4-2.
It is not within the scope of this report to examine and
explain the remedial activities and results in detail. That
information has been well described in other reports
(NYC SCA2011; NYC SCA2012). However, it is of
interest to understand how the changes in school environment
PCB concentrations resulting from remedial actions may
impact the potential exposures to building occupants. Pre-
remediation and post-remediation air and surface wipe
measurement results at selected remedial time points were
assembled for use in the SHEDS modeling effort.
Summaries of the indoor air PCB concentrations at pre- and
post-remediation time points for five schools are shown in
Table 4-24. Distributions of indoor air PCB levels at pre- and
post-remediation time points are shown in Figure 4-10. (The
post-remediation air values in Table 4-24 and in Figure 4-10
are based on the time point following light fixture removal
at Schools 1-3 and the final remediation time point for
Schools 4 and 5; see Table 4-2 for details). Across all five
schools, the median indoor air concentrations decreased 72%
from the pre- to post-remediation time point and average
levels decreased 74%. The magnitude of the decrease was
different across the five schools. For example, at School 3
the median levels decreased from 504 ng/m3 to < 50 ng/m3
while at School 5 the median value decreased from 154 to
76 ng/m3. The pre-remediation indoor air levels at School
1 were relatively low at 58 ng/m3 and showed only a small
decrease. However, problems were found with the HVAC
outdoor air intake controls at School land it is not clear
whether ventilation conditions were the same at both the
pre- and post-remediation time points.
Table 4-24. Pre- and post-remediation indoor air total PCB measurement results at five schools
Total PCB Levels in Air3'"
School/Condition Nc
Mean
QLd
ng/m3
%
>QL
Median
ng/m3
Inter-Quartile
Range
ng/m3
Overall Range
ng/m3
II Five Schools
Pre-Remediation 57 50 74
Post-Remediation 163 48 58
257
73
-------
2500 T
2000 -\
1500^
1000
T
75)
750
g 500 H
CL
250
Pre-Remediation
Post-Remediation
i ,
All 5
1
School
Figure 4-10. Pre- and post-remediation indoor air total PCB concentrations across five schools and at
each individual school. The box plots show the median, 25th, and 75th percentiles. The whiskers show
the 10th and 90th percentiles.
o
o
CO
CD
O
0.
"nj
-*—'
o
^tuu •
2200 •
/
1400 -
1200 -
1000 -
800 -
600 -
400 -
200 •
n
i i Pre-Remediation
i i Post-Caulk Remediation
i i Post-High Ventilation
i i Post-Cleaning & Light Fixture Removal
J
\
!i
I
_
T J
90th Percentile ~
75th Percentile^
Median -
25th Percentile L
10th Percentile -L
. TT
!!„.
-i- /
-,
J
~:h
All 3 Schools
School 1
School 2
School 3
Figure 4-11. Pre- and post-remediation indoor air total PCB concentrations across three schools with several
different remedial activities.
-------
Air measurements were conducted at several stages of
remedial activities at at the three 2010 NYC pilot schools.
Summaries of the indoor air total PCB levels in air are shown
in Figure 4-11. For School 1, pre-remediation concentrations
were relatively low as compared to Schools 2 and 3, but
appeared to increase somewhat at two remedial time points.
This may have been a result with problems subsequently
found in the ventilation system outside air inlet. For Schools
2 and 3, there appeared to be substantial reductions in indoor
air levels at each stage of remedial action. Caulk remediation
included removal at School 2 and encapsulation at School 3,
but only in those spaces that had indoor air measurements.
It appeared that cleaning/ventilation steps and light fixture
removal also contributed to decreased indoor air levels at
subsequent time points in these buildings. The reductions
in air concentrations occurred based on measurements
taken within days or weeks following the specific remedial
activities.
Pre- and post-remediation results for surface wipe samples
are shown in Table 4-25. (The post-remediation wipe
samples were collected following the caulk remediation step,
while the air results shown in Table 4-24 and Figure 4-11
were collected following the additional ventilation, cleaning,
and light removal steps). There was no clear pattern of pre-
to post-remediation change in surface wipe measurement
results. Median measurement values across the five schools
showed virtually no change from pre- to post-remediation
for the high-contact surfaces, while the median and 75th
percentile values at the post-remediation samples were
approximately 40% lower than pre-remediation levels for
low-contact surfaces. PCBs collected by the surface wipes
may reflect PCB concentrations in the underlying materials
which could serve as reservoirs for longer time periods
that the few weeks between the pre- and post-remediation
sample collection. PCB concentrations in many materials
were reported in Section 4.3. Based on modeling estimates
by Guo et al. (2012) it is likely that PCB concentrations in
these secondary sources will persist for some time following
removal of primary sources. This may be a factor in the lack
of large differences in pre- and post-remediation surface wipe
concentrations measured over this time frame. On the other
hand, the surface wipe measurements were generally low
even at the pre-remediation time point.
4.3.7 Relationships Between School Environment
PCB Concentrations
Relationships between total PCB concentrations in different
school environmental media can be examined for those
situations with sufficient numbers of samples collected from
the same school building rooms at the same time. A total
of 64 sets of collocated indoor air, high-contact surface
wipe, and low-contact surface wipe samples were collected
across the six schools. Modest but significant correlations
were found between indoor air PCB concentrations and
high-contact surface wipe levels for samples collected at the
six schools (Table 4-26). Higher correlation values were
found for the Spearman correlation approach, which uses
rank orders and may be more appropriate for data exhibiting
logarithmic distributions. While some association was
shown to exist, it is difficult to interpret in a larger exposure
context because the air levels are relatively large and the
surface wipe levels relatively low. Most of the exposure
comes from the indoor air, not contact with surfaces. It
is also not certain whether higher indoor air levels lead
to associated higher surface concentrations or vice-versa,
and whether the relationship might suggest some level of
dynamic equilibrium between PCBs in the air and at surface
boundaries.
A total of seven sets of collocated indoor air, surface wipe,
and dust samples were collected at School 6. Pearson and
Spearman correlation coefficients among these media are
reported in Table 4-26. The correlations between indoor
air and high-contact surface wipe PCB levels was similar
to those for all six schools, but with only seven samples the
correlation was not significant. High and significant Pearson
correlations were found between indoor air and dust levels,
and between high-contact and low-contact surface wipe
values. On the other hand, Spearman correlations were lower
and not significant for these media combinations. Pearson
correlations for small sample sizes must be interpreted
cautiously because they can be highly influenced by one or
two measurements.
-------
Table 4-25. Pre- and post-remediation surface wipe total PCB measurement results at five schools
School/Condition
All Five Schools
Pre-Remediation (H)e
Pre-Remediation (L)
Post-Remediation (H)
Post-Remediation (L)
School 1
Pre-Remediation (H)
Pre-Remediation (L)
Post-Caulk Remediation (H)
Post-Caulk Remediation (L)
School 2
Pre-Remediation (H)
Pre-Remediation (L)
Post-Caulk Remediation (H)
Post-Caulk Remediation (L)
School 3
Pre-Remediation (H)
Pre-Remediation (L)
Post-Caulk Remediation (H)
Post-Caulk Remediation (L)
School 4
Pre-Remediation (H)
Pre-Remediation (L)
Post-Caulk & Light Remediation (H)
Post-Caulk & Light Remediation (L)
School 5
Pre-Remediation (H)
Pre-Remediation (L)
Post-Caulk & Light Remediation (H)
Post-Caulk & Light Remediation (L)
Nc
58
58
58
58
11
11
11
11
12
12
12
12
14
14
14
14
9
9
9
9
12
12
12
12
%
>
QLd
55
79
50
55
54
46
9
54
83
83
75
92
57
86
57
43
44
100
33
33
33
83
67
50
Total
Median
|jg/100cm2
0.121
0.201
-------
Table 4-26. Correlations between total PCB concentrations in selected school environment samples"b
Schools/Sample Media
Schools 1 - 6
Indoor Air
High-Contact Surface Wipe
Indoor Air
Low-Contact Surface Wipe
High-Contact Surface Wipe
Low-Contact Surface Wipe
School 6
Indoor Air
High-Contact Surface Wipe
Indoor Air
Low-Contact Surface Wipe
Indoor Air
Dust
High-Contact Surface Wipe
Low-Contact Surface Wipe
High-Contact Surface Wipe
Dust
Low-Contact Surface Wipe
Dust
Pearson Correlation
N r p-value
64 0.256 0.041
64 0.104 0.415
64 0.270 0.031
7 0.258 0.577
7 0.270 0.558
7 0.805 0.029
7 0.840 0.018
7 0.170 0.716
7 0.010 0.983
Spearman
r
0.531
0.247
0.220
0.500
0.500
0.536
0.357
0.393
-0.179
Correlation
p-value
O.001
0.050
0.081
0.253
0.253
0.215
0.432
0.383
0.702
aBased on total PCBs from Aroclor measurements.
b When duplicate samples were collected, the average of the duplicates was used.
-------
4.4 Congener and Homolog Measurements
Examining patterns of individual congeners can provide
insight regarding relationships between PCBs sources and
environmental media, and can also provide information
useful in exposure and risk assessment. Individual PCB
congeners were measured in all of the air samples and in a
subset of surface wipe, indoor dust, soil, caulk, and other
building material samples collected at School 6.
Many of the building measurements that have been
performed for PCB assessment have used an Aroclor analysis
approach as a way to keep costs and complexity relatively
low as compared to performing congener-specific analysis.
There is concern, however, that building environmental
media and materials may not contain PCB congeners in
the mixture proportions found in Aroclors. Congener
mixtures in indoor air may reflect changes associated with
vapor pressure and the resulting emission differences from
sources such as caulk. PCBs in both materials and soil may
show characteristics of "weathering" where the congener
mixture has changed over time due to losses or perhaps even
depletion of more volatile components over the 40 - 60
year residence time. It is possible that analyzing building
environmental media and materials as Aroclors could provide
an inaccurate result for total PCB concentrations because of
the way analytical quantitation is done for Aroclor analyses.
It is possible that Aroclor results could be biased high or low,
depending on the specifics of which congeners are used for
quantitation and how the quantitation is handled. Because of
the improved information obtained from congener analysis,
and because of the potential biases in applying Aroclor
analysis to environmentally altered mixtures, congener
specific analysis would be preferred in most situations.
A comparison can be made for the total PCB concentration
determined by analysis using both Aroclor and congener
quantitation approaches for the samples collected at School
6. Results are shown in Table 4-27 for the several types of
samples that were collected. Analysis of sample extracts
for congeners resulted in total PCB concentrations that were
approximately 20% lower (range 14 to 46% lower) than
the analysis of the same sample extracts using an Aroclor
method. The indoor air analysis results are of greatest
interest because it is likely, if emissions from PCBs in caulk
with an un-weathered Aroclor 1254 pattern are the primary
source, that the congener pattern in air would be substantially
different than the Aroclor 1254 pattern. This effect was seen
in chamber emissions testing of caulk (Guo et al., 2011)
where the pattern in chamber air was shifted substantially
towards higher weight percentages of more volatile
congeners.
Averages of the congener-specific measurement results
are shown for 45 selected congeners in Table 4-28 for all
of the different types of samples that were analyzed. The
45 congeners were selected to represent a wide range of
chlorination and vapor pressures; because they are congeners
that have been widely reported in other studies; and because
the sum of the congeners accounted for 82.9% or more of the
sum of all 209 congeners. Table 4-28 shows the individual
congener concentrations, the sum of all 209 congeners, the
sum and weight percent of the 45 congeners in the table,
the sum and weight percent of the 12 dioxin-like congeners
(which are often used as a comparative toxicity estimate), and
calculated toxic equivalence concentration for the 12 dioxin-
like congeners (U.S. EPA, 2010; WHO, 2006). Results for
the congener measurements calculated as weight percents
are shown in Table 4-29. The weight percent results allow
an easier comparison of patterns across the different sample
types. Congener measurement results are reported in more
detail by sample type in Appendix C.
A comparison of the measured congener concentration
pattern in the PCB-containing exterior caulk collected at
School 6 to the congener weight percent pattern in Aroclor
1254 as reported by Frame et al. (1996) is shown in Figure
4-12. The congener pattern in caulk appears to be shifted
towards higher percentages of congeners with lower vapor
pressure as compared to Aroclor 1254. It is not known with
any certainty whether this is a result of weathering over 43
years or possibly because the original congener mixture was
different than Aroclor 1254. However, the pattern did not
appear to be similar to 1260; for example the weight percent
of the 7-chlorine congeners in the caulk was 4.8% whereas it
is 38% in Aroclor 1260. The analysis laboratory reported the
result of the Aroclor analysis as Aroclor 1254 with an altered
Aroclor pattern, and that is consistent with the congener
specific results shown in Figure 4-12.
Table 4-27. Comparison of Aroclor and congener measurement results for total PCBs at School 6
Measurement
Indoor Air
Surface Wipe
Indoor Dust
Outdoor Soil
Interior Caulk
Exterior Caulk
Other Materials
N
7
10
4
3
5
3
18
Units
ng/m3
ug/100 cm2
ppm
ppm
ppm
ppm
ppm
Aroclor
Analysis
Mean ± SD
633 ± 189
0.507 ± 0.404
36.0 ± 24.6
1.50 ± 1.38
21.2 ± 12.3
143000 ±8180
47.2 ± 25.2
Congener
Analysis
Mean ± SD
500 ± 154
0.407 ± 0.379
30.9 ± 18.6
1.09 ± 0.95
11.5 ± 5.9
114000 ± 2230
36.9 ± 19.9
% Difference3
21
20
14
27
46
20
22
1 Calculated as (Aroclor Mean - Congener Mean)/Aroclor Mean) x 100.
-------
Table 4-28. Summary of average PCB congener concentrations at School 6
PCB
Congener #
4
8
17
18
28
31
44
47
49
52
56
64
66
70
74
82
84
85
87
91
92
95
97
99
101
105
110
118
128
130
132
138
141
146
149
151
153
156
158
163
170
174
Indoor
Air
N = 7
ng/m3
0.34
1.5
2.1
3.8
3.8
3.7
19
4.8
15
45
2.6
5.5
6.9
25
12
3.9
14
5.4
17
5.8
10
50
11
15
54
5.5
39
19
1.4
0.83
6.6
8.3
2.2
1.2
13
5.8
9.6
0.49
1.2
1.4
0.24
0.70
Outdoor
Air
N = 1
ng/m3
-------
Table 4-28. Summary of average PCB congener concentrations at School 6 (continued)
PCB
Congener #
180
187
206
Z 209 Cong.3
Z 45 Cong."
Z DLC Cong.0
DLC TEQd
% 45 Cong.8
% DLC Cong.'
Indoor
Air
N = 7
ng/m3
0.76
0.83
0.19
500
456
26
7.88E-04
91.0
5.0
Outdoor
Air
N = 1
ng/m3
-------
Table 4-29. Summary of average PCB congener weight percents at School 6a
PCB
Congener #
4
8
17
18
28
31
44
47
49
52
56
64
66
70
74
82
84
85
87
91
92
95
97
99
101
105
110
118
128
130
132
138
141
146
149
151
153
156
158
163
Indoor
Air
N = 7
0.07
0.3
0.4
0.8
0.8
0.7
3.8
1.0
3.0
8.9
0.5
1.1
1.4
5.1
2.4
0.8
2.8
1.1
3.5
1.2
2.1
10
2.2
3.0
11
1.1
7.8
3.7
0.3
0.2
1.3
1.6
0.4
0.2
2.5
1.2
1.9
0.1
0.2
0.3
Outdoor
Air
N = 1
-------
Table 4-29. Summary of average PCB congener weight percents at School 6a (continued)
PCB
Congener #
170
174
180
187
206
Z 45 Cong. %b
Z 209 Cong.0
Indoor
Air
N = 7
0.05
0.1
0.2
0.2
0.04
91
ng/m3
500
Outdoor
Air
N = 1
-------
10
I
I
CD 8
0.
0)
CD
CO
CD
CD
D)
§
o
LO
CM
_o
o
6 -
2 -
110
101
95
52
44
18
87
70
105
Aroclor 1254
118
\ \ \
138
132
128
ill II „ null
149
153
163
180
|nJj|J n
10 20 30 40 50 60 70 80 90 100110120130140150160170180190200
PCB Congener Number
E
Q.
g
2
-!—'
I
O
O
CD
C
CD
O)
c
o
O
G3
O
s
LJJ
uuuu -
8000 -
6000 -
4000 -
2000 -
n
101
95
87
52 70
44
8 18 j
. n II III II
H n J
84
ll
I
i
105
I
10
118
u
138 Exterior Caulk
1C
128
nil
2
111
153
149
ll
163 18°
I,
nil n L ^
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
PCB Congener Number
Figure 4-12. Patterns of congeners in Aroclor 1254 (top) and exterior caulk collected at School 6
-------
Figure 4-13 shows congener concentration patterns for
indoor air and dust samples collected in School 6. The
indoor air pattern appears to be shifted toward higher
percentages of congeners with higher vapor pressures as
compared to the dust, caulk, and Aroclor 1254 congener
patterns. This is consistent with vapor emission as a source
ofthePCBsinair.
Comparisons of congener patterns between indoor air, the
exterior caulk, and Aroclors 1242 and 1254 are shown in
Figure 4-14 for selected congeners ordered by decreasing
vapor pressure. The congeners are shown on a weight
percent basis so direct comparisons can be made. Once
again, the comparison shows congeners in indoor air shifted
towards more volatile congeners and those in caulk shifted
towards less volatile congeners as compared to the Aroclor
1254 pattern. The indoor air pattern does not match the
overall Aroclor 1242 pattern, but some contribution of
the more volatile Aroclor 1242 congeners in air can't be
ruled out. Congener weight percent patterns are shown for
a number of different sample media in Figure 4-15. The
congener pattern in paint more closely matches the Aroclor
1254 pattern than does the caulk. As Guo et al., (2012)
demonstrated in chamber studies, this might be expected
as a result of greater emission of more volatile PCBs from
caulk, but preferential absorption of less volatile PCBs
into the paint, resulting in a pattern that looks remarkably
like Aroclor 1254. Surface wipe and dust samples had
lower percentages of more volatile congeners as compared
to indoor air and Aroclor 1254 but as the congener vapor
pressures decreased their patterns became more similar to the
Aroclor 1254 pattern.
Another way to examine patterns and differences in patterns
of PCBs in different types of environmental media and
materials is by examining PCB homologs. Congener
measurement results can be used to calculate PCB homolog
concentrations by summing all of the results for congeners
with the same number of chlorines. A summary of the
average chlorine-number homolog measurement results
is shown as weight percents in Table 4-30. Results for
the individual sample types are shown in more detail in
Appendix C.
Homolog weight percents for the PCB-containing exterior
caulk and the indoor building materials are compared to
homolog patterns in Aroclors 1242 and 1254 and indoor air
in Figure 4-16. The caulk contains higher fractions of 6- and
7-chlorine homologs as compared to Aroclor 1254 and a
smaller fraction of 4-chlorine homologs compared both to
Aroclor 1254 and the other materials. Homolog patterns for
the air, wipe, dust, and soil samples are compared to patterns
for Aroclors 1242 and 1254 in Figure 4-17. The indoor
air has substantially higher percentages of the 2-, 3-, and
4-chlorine homologs and less of the 6-, 7-, and 8-chlorine
homologs compared to wipes, dust, soil, and Aroclor 1254.
Surface wipe and dust patterns look similar, which is not
surprising since the wipes likely have a dust component.
The higher fraction of 6-chlorine homolog in dust as
compared to air shows, as expected, that higher molecular
weight congeners are more likely to be particle-bound. The
outdoor soil has higher fractions of the 6-, 7-, and 8-chlorine
homologs than Aroclor 1254 and the other media, possibly
as a result of weathering loss of the more volatile congeners
over time.
Estimates of PCB emissions from caulk were reported in
Section 4.2.1 based on an assumption that the congener
pattern in the caulk was the same as the congener pattern in
Aroclor 1254. This was a requirement since only Aroclor
measurement results were available for most of the caulk
samples. However, the availability of congener specific
measurements for caulk at School 6 allows examination
of the assumption. Table 4-31 shows a comparison of the
estimated emission rates for selected congeners and total
PCBs using first using the assumption that an Aroclor 1254
pattern is present, and then again using the actual measured
congeners in the exterior caulk.
Because the relative concentrations of the more volatile
congeners such as congener 52 were lower in the caulk
than in Aroclor 1254, the assumed approach leads to higher
emission estimates for the more volatile congeners. The
opposite is true for the less volatile congeners such as
congener 153. Overall, the estimated total PCB emission
rate assuming an Aroclor 1254 mixture was 80% higher
than the rate using the actual measurements of congener
concentrations in the caulk. The impact is illustrated in
Figure 4-18, which shows the estimated congener emission
rates using the assumption of an Aroclor 1254 mixture
and the measured congener concentrations in the caulk.
Figure 4-19 shows the congener emission rates based on the
congeners measured in the exterior caulk at School 6 and
the concentration of congeners measured in indoor air. The
indoor air pattern more closely matches the pattern using
the measured congener concentrations than for the estimated
emission rates assuming an un-weathered Arcolor 1254
mixture as shown in Figure 4-18. A caution is warranted in
this assessment. It is not certain whether the most volatile
congeners were not detected in the School 6 caulk because
they were not present or because of the level of dilution that
was used for these sample extracts (due to the very high
PCB concentrations). If those more volatile congeners were
actually present but unmeasured, the total PCB emission rate
could be underestimated.
A summary of the key findings from the assessment of
congener and homolog data using samples collected at one
school is provided below.
• The pattern of congeners in indoor air was more heavily
weighted towards more volatile congeners as compared to
Aroclor 1254 and compared to the PCB-containing caulk
at the building. The pattern of congeners in air was not
as heavily weighted towards more volatile congeners as
would be predicted if they were from vapor emissions
from caulk alone, suggesting that a portion of the PCBs
in air may be associated with airborne particles, but
this could also reflect in part a higher proportion of less
volatile congeners in the caulk.
-------
2
"c
100
80 -
60-
c
o
O
CD
§ 40
O)
I
20 -
101
95
C
44
18
8 j ||
1,
2
70
1
JLj J.L
87
84
1
I
1
Indoor Air
110
105
n
n L
118
149
153
138
132 |
I
128 163 18Q
.n n . j, MO Mill II IL „ » . [In . _ . _ , „ n.
10 20 30 40 50 60 70 80 90 100110120130140150160170180190200
PCB Congener Number
Q.
£_
C
O
•-S 3H
c
o
O
i—
CD
C
CD
C
o
O
"GO
^
Q
O
o
T3
101
87
95
52
70
8 18
44
li
110
105
Indoor Dust
118
138
132
153
149
ill
163
180
M. llJIl |,
r^-H—
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
PCB Congener Number
Figure 4-13. Patterns of congeners in indoor air (top) and indoor dust collected at School 6
-------
I
ro
-
c:
=3
.Q
0
_
0)
o:
-i— •
0
e
0
Q.
12 -
10 -
8 -
6 -
4 -
2 -
n -
Congener Vapor Pressure (torr)
6.4 E-4 2.1 E-5 5.8 E-7
.1 .1
1
1
1
|
^H Aroclor 1254
^H Caulk
^^H Indoor Air
1
I i ,1
8
c
cc
T3
C
=3
.a
0
-»— '
_03
0
c
0
Q
0
Q.
18 28 52 44 95 92 101 87 110 149 118 105 153 138 128 180 170
Congener Number
(in order of decreasing vapor pressure)
19 -
10 -
8 -
6 -
4 -
2 -
n -
Congener Vapor Pressure (torr)
6.4 E-4 2.1 E-5 5.8 E-7
1
.1
1
1
I,
I
I
I
I
1.
^H Aroclor 1 242
^•l Caulk
^H Indoor Air
II
1. 1.
18 28 52 44 95 92 101 87 110 149 118 105 153 138 128 180 170
Congener Number
(in order of decreasing vapor pressure)
Figure 4-14. Patterns of relative weight percent for selected congeners in Aroclor 1254 (top) and Aroclor 1242 (bottom) compared to averages for
exterior caulk and indoor air at School 6
-------
,2-
10 -
CD
O
c
CO
T3 o
c o -
.D
CD
£ 6-
CO
Q)
1 4-
o
CD
Q.
2 -
n -
6.4 E-4
..1 .„!
Congener Vapor Pressure (torr)
2.1 E-5
I
||
I
1
1
5.8 E-7
I
^H Aroclor 1254
I^H Caulk
I I Wall Paint
I^H Indoor Air
1. ill. lln
18 28 52 44 95 92 101 87 110 149 118 105 153 138 128 180 170
Congener Number (in order of decreasing vapor pressure)
12 -
6.4 E-4
Congener Vapor Pressure (torr)
2.1 E-5
5.8 E-7
I
CD
0_
10 -
8
03
I
CD
> 6 ^
1
CD
OC.
4 -
2 -
J 1 II
Aroclor 1254
Indoor Air
Surface Wipe
Indoor Dust
Outdoor Soil
I
18 28 52 44 95 92 101 87 110 149 118 105 153 138 128 180 170
Congener Number (in order of decreasing vapor pressure)
Figure 4-15. Patterns of relative weight percent for selected congeners in Aroclors 1254 compared to average values for indoor air, paint, and
exterior caulk (top), and indoor air, surface wipes, dust, and soil (bottom) at School 6
-------
Table 4-30. Summary of average PCB homolog weight percents at School 6a
PCB
Homolog
1 -Chlorine
2-Chlorine
3-Chlorine
4-Chlorine
5-Chlorine
6-Chlorine
7-Chlorine
8-Chlorine
9-Chlorine
10-Chlorine
I 209 Cong.b
Indoor
Air
N = 7
0.02
0.6
4.3
31.1
51.1
11.9
0.8
0.1
0.04
-------
Aroclor1242
5-CI 1-CL
5.2% ,-0.5%
Aroclor1254
7-CI 3-CI
2.7%
Indoor Materials
7-CI 8'CI 3-CI
4.6% _ °-6% _2.6%
6-CI
24.5%
I
4-CI
19.9%
ExteriorCaulk
7-CI
4.8%
IndoorAir
7-C| 2-CI 3.C|
0.8% . 0-6% 4.3%
Figure 4-16. Relative weight percents of PCB chlorine-number homologs for Aroclors 1242 and 1254 compared to the averages of the exterior
PCB-containing caulk, the indoor building materials, and indoor air
-------
Aroclor1242
5-CI 1-CL
5.2%
Aroclor1254
7-CI 3-CI
2.7% 1.2%
IndoorAir
7-CI 2-CI 3_ci
0.8%
7-CI
7.1%
Surface Wipes
8-CI 9-CI 3_C|
1.7%^ —
7 r.
IndoorDust
8-CI 3-CI
OutdoorSoil
8-CI 3-CI
7-CI 1.0% _0.2%
8.2%
Figure 4-17. Relative weight percents of PCB chlorine-number homologs for Aroclors 1242 and 1254 compared to the averages of the
environmental media
-------
Table 4-31. Differences in PCB congener emission estimates from exterior caulk assuming an un-weathered Aroclor
1254 pattern vs. measured congener concentrations
Concentration Estimated
Congener Weight In Caulk Emission Rate
Congener Percent ppm ug/hr
Assuming Caulk Contains Aroclor 1254 Congener Mixture, Total PCBs 112,000 ppm, Surface Area 0.271 m2
8 0.13 146 67.6
18 0.25 280 70.9
28 0.19 213 21.0
44 2.31 2,590 121
52 5.38 6,030 370
70 3.49 3,910 78.1
110 9.29 10,400 75.6
153 3.77 4,220 10.2
180 0.67 750 0.31
1209 Cong. 100 112,000 1,841
Using Congeners Measured in Exterior Caulk with Total PCBs 112,000 ppm, Surface Area 0.271 m2
8
18
28
44
52
70
110
153
180
1209 Cong.
18
28
44
52
70
110
153
180
1209 Cong.
-------
tuu -
^~
.*_
"g 300 -
&
CO
or
c
0
'co 200 -
CO
£
HI
•o
"co
E 100 -
to
HI
n
I
44
8 18
i
1 ,Jy
nffll
52
Estimated Emission Rates
Assuming Caulk Contains
Aroclor 1254
95
70
I_J
84
87
II n
J ill
101
110
118
! n 1321QO 149
1 n^, i -jo
128 H 1R9
III „ n ..I o.jiH . , Ln . 1?3
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
PCB Congener Number
c
160
140
120
100 -
80 -
0)
-*— '
ro
E 40
.i— '
CO
HI
20 -
52
44
70
Estimated Emission Rates
Using Measured Congeners
101 in Caulk
95
87
110
105
128
118
149
132
138
163
.n n .
180
I I \
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
PCB Congener Number
Figure 4-18. Patterns of estimated emission rates from exterior caulk collected at School 6 assuming the caulk contains congeners in an Aroclor
1254 proportions (top) and using congener values measured in the caulk (bottom)
-------
160
140 -
120 -
§_ 100 -
g
'w 80 -
I «H
CD
-I—•
03
.1 40 -
'•4—'
c/)
HI
20 -
0
52
44
70
101
Estimated Emission Rates
Using Measured Congeners
in Caulk
95
87
110
105
128
118
149
138
132
jliU nil
163
180
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
PCB Congener Number
100
5
44
18
8 n I ||
..1 nln Jib . idL
J
(1
2
70
1
101
95
87
84
, I
Indoor Air
110
105
n
0 n
118
149
153
138
1f I
il
128 , 163 180
juLu yy ol II n IL II „ n On _ , „ . „ „ n
o
o
T3
80 -
g
"TO
8 60
o
O
CD
O 40
o
O
20 -
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
PCB Congener Number
Figure 4-19. Patterns of estimated emission rates from exterior caulk collected at School 6 (top) versus the congeners measured in indoor
air (bottom)
-------
• The congener pattern in the PCB-containing caulk was
somewhat more heavily weighted towards less volatile
congeners as compared to Aroclor 1254. It is possible
that the more volatile congeners have been depleted from
sources such as the exterior caulk over a period of 43
years.
• The congener pattern in indoor air in this school did not
resemble Aroclor 1242, the Aroclor most likely to be
present in light ballasts. However, the oil content in the
ballasts in this school was not measured and Aroclor 1254
can't be ruled out given that Aroclor 1254 was found in a
ballast in one of the NYC schools.
• Congener patterns in surface wipe, indoor dust, and other
building materials were generally similar to Aroclor 1254
and to the PCB-containing caulk.
• Soil samples had a congener pattern weighted towards less
volatile congeners as compared to the PCB-containing
caulk and Aroclor 1254, possibly as a result of weathering.
• The congener patterns in some environmental media,
particularly indoor air, did not match typical congener
patterns in the source materials. Total PCBs measured
as Aroclors were approximately 20% higher than total
PCBs based on congener analysis. A homolog analysis
approach could provide more accurate total PCB
measurements than Aroclor measurements, but at a
somewhat higher cost. Homolog measurements are likely
to be less costly than full congener-specific analyses.
• Congener-specific analysis provides information that
can be used in risk assessment. For example, between
5% (indoor air) and 14% (soil) of the total amount of
PCBs were comprised of the sum of the 12 dioxin-like
congeners. However, the congener analysis was limited
to samples from one school. This was not sufficient for
probabilistic analysis in the SHEDS model.
-------
4.5 SHEDS Exposure Modeling
4.5.1 Distributions of Exposure Estimates
The Stochastic Human Exposure and Dose Simulation
(SHEDS) model was used to generate estimated PCB
absorbed dose distributions resulting from exposures to
environmental levels measured at the six schools. The
model was run for four age groups (4-5, 6-10, 11-14, and
14-18 years old). Mean estimates of absorbed dose, and the
estimated absorbed doses across selected percentiles of the
modeled distribution are shown in Table 4-32. For the 6-10
year-old age group, the estimated absorbed dose was 0.022
ug/kg/day at the 50th percentile of the distribution and 0.041
ug/kg/day at the 95th percentile. Estimated absorbed dose
levels were lower for the other three age groups relative to
the 6-10 year old group, and were 55% and 56% lower for
the 14 - 18 year old age group at the 50th and 95th percentiles,
respectively.
Estimates of absorbed PCB doses were generated using
environmental measurements at five schools with available
pre- and post-remediation measurements. Mean estimates
of total absorbed dose, and the estimated absorbed doses
across selected percentiles of the modeled distribution are
shown for the four age groups at pre- and post-remediation
time points in Table 4-33. The ratios of post-remediation
to pre-remediation absorbed dose estimates are shown in
Table 4-34. For the 6-10 year-old age group the
post-remediation absorbed dose estimates were 64% lower
than pre-remediation at the 50th percentile and 69% lower
at the 95th percentile. Similar pre- to post-remediation
decreases in estimated absorbed dose were found for all
age groups. Box plots of the pre- and post-remediation
estimated absorbed doses by age group are shown in
Figure 4-20 and percentile distribution plots are shown for
the 6-10 year-old group in Figure 4-21.
These results indicate that remedial actions are likely to
result in decreased exposures. At the pre-remediation time
point, 25 to >75% of the distribution was lower than the
adjusted RfD value of 17 ng/kg/day, depending on age group.
At the post-remediation time point, >95 to >99% of the
distribution was less than the adjusted RfD, for school-related
exposures only. It is important to remember that different
types and scales of remedial action were taken at each of
the five schools (see Table 4-2) and that different relative
decreases in indoor air levels were seen across the schools
(see Table 4-24). It is possible that other factors, such as
differences in temperature and ventilation conditions at the
pre- and post-remediation time points could have affected the
environmental concentrations and estimated exposures.
Estimates of absorbed PCB doses were also generated using
environmental measurements at three schools with available
pre- and post-remediation measurements taken during the
same year and with subsequent pre-remediation indoor
air and surface wipe measurements taken the following
year. The purpose of this assessment is to evaluate whether
reductions in exposure may be expected to be sustained over
time following remedial actions. SHEDS model estimates
of absorbed doses using measurements from the second-year
post-remediation time point were not performed because the
indoor air and surface wipe pre- and post remediation results
were not significantly different (NYC SCA, 2012). Mean
estimates of absorbed dose, and the estimated absorbed doses
across selected percentiles of the modeled distribution are
shown for the four age groups at the three time points in
Table 4-35. Percentile distribution plots are shown for the
6-10 year-old group in Figure 4-22. Overall, the estimated
absorbed dose levels for the pre-remediation time point in the
second year were slightly higher than the post-remediation
time point from the first year, but were still substantially
lower than the first year pre-remediation time point. These
results suggest reductions in exposure resulting from
remedial actions may be retained over time. It is important to
remember that different types and scales of remedial action
were taken at these three schools and that some activities
were performed between the first year post-remediation time
point and the second year pre-remediation time point (see
Table 4-2). It is also possible that differences in temperature
and ventilation conditions at the different time points could
have affected the environmental concentrations and estimated
exposures, although temperatures were generally within 6°F
or less across the time points (see Table 4-3).
SHEDS model estimates of PCB total absorbed dose
generated using the environmental measurement data from
the six schools examined in this report were compared to
estimates generated using extant PCB data not associated
with the six schools measurements. These data were
gleaned from several reports and internet sources in 2009
and included measurements for indoor air, dust, surface
wipes, and soil (see Appendix D for the data sources and
environmental measurement distributions). The estimated
absorbed doses using the other extant data are shown in
Table 4-36 for comparison with results generated using
results from the six schools in this report. Comparisons
were made only for the 6-10, 11-13, and 14-18 year old age
groups. In general, the estimated absorbed doses using data
from the six schools are similar to estimates using other data.
Overall, the mean and median estimated absorbed doses
were similar for the two sets of data when comparisons are
for the pre-remediation time point for the six school sets of
measurements in this report. The fraction of absorbed dose
resulting from inhalation was higher for the results used in
this report than for the other school data at most percentile
levels largely because the indoor air PCB concentrations
in the six schools were about 3-fold higher at the median
than the other extant school data. On the other hand,
concentrations in wipe samples in the six schools were about
one-third of the levels obtained from the other extant school
data, with the dermal and non-dietary ingestion representing
a lower proportion of the intake at the six schools used in this
report.
4.5.2 Estimated Exposures by Exposure Route
Information on the relative importance and contribution of
different exposure pathways to the total exposure can help
inform mitigation decision-making. The SHEDS model
provides estimates of the school PCB exposure from each
relevant route. Table 4-37 shows the apportionment of
-------
Table 4-32. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from six schools
(units: jig/kg/day)
Child Age
Group
4-5 year olds
6 -10 year olds
11 - 13 year olds
14- 18 year olds
Percentiles of the Distribution of
Mean
0.023
0.027
0.015
0.011
SD
0.012
0.032
0.005
0.005
p5
0.012
0.015
0.009
0.004
p25 p50
0.017 0.021
0.019 0.022
0.012 0.015
0.007 0.010
Table 4-33. Distributions of total absorbed PCB dose estimated by SHEDS based on
schools3 at pre-remediation and post-remediation time points (units: jig/kg/day)
Child Age
Group
4-5 vear olds
Pre-remediation
Post-remediation
6 -10 vear olds
Pre-remediation
Post-remediation
11 - 13 vear olds
Pre-remediation
Post-remediation
14 -18 vear olds
Pre-remediation
Post-remediation
p75
0.026
0.027
0.018
0.014
Dose Estimates
p95
0.036
0.041
0.022
0.018
measurement data from
Percentiles of the Distribution of
Mean
0.019
0.007
0.023
0.008
0.014
0.005
0.010
0.003
SD
0.006
0.002
0.015
0.004
0.004
0.001
0.004
0.002
p5
0.010
0.004
0.014
0.005
0.008
0.003
0.003
0.001
p25 p50
0.016 0.019
0.006 0.007
0.018 0.021
0.006 0.007
0.011 0.013
0.004 0.005
0.007 0.010
0.002 0.003
p75
0.022
0.008
0.025
0.009
0.016
0.006
0.013
0.005
p99
0.061
0.125
0.028
0.023
five
Dose Estimates
p95
0.029
0.010
0.037
0.012
0.021
0.007
0.018
0.006
p99
0.032
0.014
0.071
0.018
0.023
0.009
0.022
0.008
1 Schools 1, 2, 3, 4, and 5.
-------
Table 4-34. Post-remediation/pre-remediation ratios of total absorbed PCB dose estimates based on measurement data
from five schools
Child Age
Group
4-5 year olds
6- 10 year olds
11 - 13 year olds
14- 18 year olds
Post/Pre-Remediation Ratio for the Distribution of Dose Estimates
Mean
0.37
0.34
0.37
0.35
p5
0.37
0.37
0.37
0.34
p25
0.37
0.36
0.36
0.34
p50
0.37
0.36
0.37
0.35
p75
0.38
0.35
0.37
0.36
p95
0.35
0.31
0.36
0.36
p99
0.43
0.26
0.38
0.35
0.10-
0.08-
_ 0.06 -
a
^
HI
^
0)
~ 0.04 -
0.02-
0.00-
8
0
0
0
8
o a o
i I ° • i
0 1
^ 8
1 B .JL.
Y t3
8 T T ^ Jf ^
pre 4-5 post 4-5 pie 6-10 post 6-10 pre11-13 post 11-1 3 pre14-18 post 14-1 8
Scenario and Age Group
Figure 4-20. Distributions of estimated absorbed total PCB doses from exposures at school for four age
groups at five schools for pre- and post-remediation time points. The box plots show the median, 25th, and
75th percentiles. The whiskers show the 10th and 90th percentiles.
-------
100 -i—
80
Aroclor 1254 RfD = 0.020 (jg/kg/day
RfD Adjusted for Absorption = 0.017 m/kg/day
0.01
0.02
0.03
0.04
0.05
ug/kg/day
Figure 4-21. Distributions of estimated absorbed total PCB doses from exposures at school
for the 6-10 year old age group at five schools for pre- and post-remediation time points
inhalation, dermal absorption, and non-dietary ingestion
routes for the estimated PCB absorbed doses for the 6-10
year old age group based on pre- and post-remediation
measurements at five schools. The results in Table 4-37
were estimated using several estimates of PCB pulmonary
absorption since the actual value is not well known.
Figures 4-23 and 4-24 show contributions from different
exposure routes at different percentiles of the distributions for
the 6-10 year-old age group.
Overall, the inhalation pathway would appear to be the
predominant route of exposure based on data from these
schools. Inhalation exposures would account for over 73%
of the total absorbed dose for 6-10 year-olds at the pre-
remediation time point (when PCB concentrations in air
were greatest) if pulmonary absorption fractions are 70% or
higher. If pulmonary absorption were as low as 30%, then
inhalation would be esimtated to account for 47% of the
total absorption, with dust/soil ingestion accounting for 41%
and dermal absorption 12%. At the post-remediation time
point, the fraction from each exposure route was similar, in
most cases, to the pre-remedition time point, although the
relative contribution for non-dietary ingestion decreased
when 30% pulomonary absorption is assumed. Similar
patterns were seen for the other age groups. The contribution
of non-dietary ingestion was not modeled for the two older
age groups due to the lack of hand-to-mouth data, but would
be expected to be lower than for 6-10 year olds because of
reduced hand-to-mouth activity.
Assuming a 70% pulmonary absorption, the fraction of the
overall absorbed dose from non-dietary ingestion of PCBs
in dust an soil remained relatively small (< 12%) for 6 - 10
year-olds for most of the percentiles of the distribution of
estimated absorbed doses (Figures 4-23 and 4-24). However,
in the highest 10% of the distribution, over 50% of the
pre-remediation total absorbed dose is predicted to result
from non-dietary ingestion, greater than the amount from
inhalation. This result occurs because of a combination
of relatively high dust ingestion at the upper end of the
activity distribution for the 6 - 10 year-old age group and
relatively high PCB concentrations in dust and soil in the
upper ends of the distributions. This could represent a
relatively highly exposed sub-group of children, as the total
estimated absorbed dose is also higher at the upper end of
the distribution. Some caution is needed in interpreting the
upper ends of modeled exposure distributions, and more
information would be needed to determine whether more
highly exposed sub-groups of children occur.
4.5.3 Sensitivity Testing
Model sensitivity analyses are used to assess the relative
impact and importance of uncertainties in model parameters
and input data. Limited sensitivity analysis was conducted
for two important but uncertain parameters used in
the SHEDS PCB model. These include the fraction of
pulmonary absorption of PCBs following inhalation, and the
concentration of PCBs in dust and soil in the schools.
To the best of our knowledge, the pulmonary absorption
fraction for PCBs following inhalation has not been
determined for humans. Extant rat data for PCBs and dioxin
in the vapor phase, and dioxin bound to soil, suggest that the
pulmonary absorbed fraction is likely to be high (Hu et al.,
2010; Dilberto et al., 1996 Nessel et al, 1992 - all resulted
in values >80%). On the other hand, three biomonitoring
studies suggest that people exposed to PCBs in contaminated
buildings may show increases in blood levels of the more
volatile congeners, but overall circulating blood levels of
total PCBs had only small differences compared to those in
unexposed groups (Herrick et al., 2011; Liebel et al., 2004;
Gabio et al., 2000). A value of 70% pulmonary absorbed
fraction was assumed for the SHEDS PCB analysis.
To examine the modeled absorbed dose estimate sensitivity
associated with uncertainty in pulmonary absorbed fraction,
sensitivity testing was performed on the impact of using
30%, 80%, or 100% values for pulmonary absorbed fraction
for 6-10 year-olds at the pre- and post-remediation time
points for five schools (Table 4-38). Using a pulmonary
-------
Table 4-35. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from three
schools3 at three time points (units: jig/kg/day)
Percentiles of the Distribution of Dose Estimates
Child Age Group
4-5 vear olds
Pre-first-year remediation
Post-first-year remediation
Pre-second-year remediation
6- 10 vear olds
Pre-first-year remediation
Post-first-year remediation
Pre-second-year remediation
11 -13 vear olds
Pre-first-year remediation
Post-first-year remediation
Pre-second-year remediation
14 -18 vear olds
Pre-first-year remediation
Post-first-year remediation
Pre-second-year remediation
Mean
0.025
0.007
0.009
0.031
0.008
0.011
0.018
0.005
0.007
0.012
0.004
0.005
SD
0.009
0.002
0.003
0.031
0.003
0.007
0.005
0.002
0.002
0.006
0.002
0.002
p5
0.013
0.004
0.005
0.017
0.005
0.007
0.010
0.003
0.004
0.004
0.001
0.002
p25
0.020
0.006
0.008
0.022
0.006
0.009
0.014
0.004
0.005
0.008
0.002
0.003
p50
0.024
0.007
0.009
0.026
0.007
0.010
0.018
0.005
0.007
0.012
0.004
0.005
p75
0.029
0.008
0.011
0.031
0.009
0.012
0.021
0.006
0.008
0.016
0.005
0.006
p95
0.039
0.011
0.015
0.051
0.012
0.016
0.026
0.008
0.011
0.023
0.006
0.008
p99
0.052
0.012
0.019
0.141
0.017
0.030
0.033
0.010
0.012
0.026
0.007
0.010
"Schools 1,2, and3.
100
— Pre-remediation (1st year)
— Pre-remediation (2styear)
— Post-remediation (1 st Year)
Aroclor 1254 RfD = 0.020 |jg/kg/day
RfD Adjusted for Absorption = 0.017 pg/kg/day
0.01
0.02 0.03
ug/kg/day
0.04
0.05
Figure 4-22. Distributions of estimated absorbed total PCB doses from exposures at three schools for the 6-10 year old
age group at three pre- and post-remediation time points across two years
-------
Table 4-36. Distributions of total absorbed PCB dose estimated by SHEDS based on measurement data from six schools
in this report and from extant measurement data3 (units: jig/kg/day)
Percentiles of the Distribution of Dose Estimates
Child Age Group
6- 10 vearolds
6 schools, this report
Extant data (Appx. D)
11 - 13 vearolds
6 schools, this report
Extant data (Appx. D)
14- 18 vearolds
6 schools, this report
Extant data (Appx. D)
Mean
0.027
0.021
0.015
0.016
0.011
0.011
SD
0.032
0.009
0.005
0.006
0.005
0.005
p5
0.015
0.012
0.009
0.009
0.004
0.004
p25
0.019
0.016
0.012
0.012
0.007
0.008
p50
0.022
0.019
0.015
0.015
0.010
0.011
p75
0.027
0.024
0.018
0.019
0.014
0.014
p95
0.041
0.037
0.022
0.027
0.018
0.020
p99
0.125
0.054
0.028
0.034
0.023
0.028
Extant data and data sources shown in Appendix D.
-------
Table 4-37. Proportion of mean estimated total absorbed PCB dose for 6-10 year olds for inhalation, non-dietary
ingestion, and dermal absorption routes of exposure based on measurements from five schools
Inhalation
%
Non-Dietary
Ingestion
Dermal Absorption
Pre-remediation
30% inhalation absorption
70% inhalation absorption (baseline)
100% inhalation absorption
47
74
78
41
18
15
12
8
6
Post-remediation
30% inhalation absorption
70% inhalation absorption (baseline)
100% inhalation absorption
52
74
79
30
15
13
19
12
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
30003
O-W 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100
Ptercentile range
pathway
Dermal
Non-dietary IG (dust)
Inhalation
Non-dietary IG (residue)
Figure 4-23. Contributions of different exposure routes towards total estimated absorbed PCB doses for the
6-10 year old age group at different percentiles of the total dose estimate based on measurements at six
schools (assuming 70% pulmonary absorption)
-------
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
0_B 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100
Percentile range
pathway
Dermal
Non-dietary IG (dust)
Inhalation
Non-dietary IG (residue)
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
Post-remediation
0_M 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100
Percentile range
pathway
Dermal
Non-dietary IG (dust)
Inhalation
Non-dietary IG (residue)
Figure 4-24. Contributions of different exposure routes towards total estimated absorbed PCB doses for the 6-10 year
old age group at different percentiles of the total dose estimate based on pre-remediation (top) and post-remediation
(bottom) measurements at five schools (assuming 70% pulmonary absorption)
-------
absorption fraction of 30% at the pre-remediation time
point resulted in a 49% decrease in the estimated absorbed
dose at the 50th percentile and a 25% decrease at the 95th
percentile. At the post-remediation time point, using the 30%
pulmonary absorption fraction resulted in a 45% decrease in
the estimated absorbed dose at the 50th percentile and a 23%
decrease at the 95th percentile. The changes were smaller at
the post-remediation time point because the contribution of
inhalation to total exposure was lower due to the decreases in
indoor air concentrations.
Using a pulmonary absorption fraction of 100% at the pre-
remediation time point resulted in an increase in the estimated
absorbed dose of 35% at the 50th percentile and a 23%
increase at the 95th percentile. At the post-remediation time
point using a pulmonary absorption fraction of 100% resulted
in an increase in the estimated absorbed dose of 32% at the
50th percentile and 35% at the 95th percentile. The Public
Health Levels were derived using an assumption of 100%
pulmonary absorption.
Modeled estimates of total PCB absorbed dose resulting from
exposure to PCBs in the school environment was sensitive
to the value used for pulmonary absorption fraction based
on environmental PCB levels for the schools used in this
assessment. While the available literature suggests that
the pulmonary absorbed fraction is likely to be >80%, the
available biomonitoring data do not necessarily support
the high fraction. This is further complicated because the
assumption for SHEDS modeling is that all congeners
in Aroclor 1254 will have the same absorption fraction.
However, the different congeners have different physical
properties that affect both their absorption in pulmonary
tissues as well as the fractions that will be found in vapor and
particle-phase in indoor air and in different human tissues.
This complicates both the understanding of pulmonary
exposure and absorption and interpretation in biomonitoring
measurements with regard to distribution and relative storage
in different body tissues versus the relative fractions in
Table 4-38. Sensitivity test results of different pulmonary absorption rates on the distributions of estimated total
absorbed PCB dose for 6-10 year olds based on pre- and post-remediation measurements from five schools3
Percentiles of the Distribution
Model Paramter
Mean
SD
P5
p25
p50
p75 p95 p99
Estimated Absorbed Dose ua/kq/dav
Pre remediation
30% absorption 0.016 0.040
70% absorption (baseline) 0.023 0.015
80% absorption 0.026 0.021
100% absorption 0.032 0.024
0.007
0.014
0.015
0.018
0.009
0.018
0.019
0.024
0.011
0.021
0.023
0.028
0.013
0.025
0.028
0.033
0.028
0.037
0.043
0.046
0.122
0.071
0.106
0.101
Post remediation
30% absorption 0.005 0.004 0.003 0.003 0.004 0.005 0.009 0.019
70% absorption (baseline) 0.008 0.004 0.005 0.006 0.007 0.009 0.012 0.018
80% absorption 0.009 0.005 0.006 0.007 0.008 0.010 0.013 0.033
100% absorption 0.011 0.004 0.007 0.008 0.010 0.012 0.016 0.029
Ratio of Estimated Doses at 30. 80. and 100% to Baseline of 70%
Pre remediation
30% absorption
80% absorption
100% absorption
0.66
1.13
1.35
0.49
1.10
1.31
0.49
1.09
1.34
0.51
1.09
1.35
0.52
1.09
1.31
0.75
1.16
1.23
1.71
1.50
1.43
Post remediation
30% absorption
80% absorption
100% absorption
0.60
1.15
1.31
0.52
1.09
1.30
0.53
1.12
1.32
0.55
1.13
1.32
0.56
1.12
1.29
0.77
1.13
1.35
1.05
1.82
1.58
1 Schools 1, 2, 3, 4, and 5.
-------
circulating blood across the range of congeners. Differences
in congener metabolism are likely to further complicate the
interpretation of blood-based biomonitoring. On the other
hand, congener-specific information can provide information
about the various environmental and dietary sources, as well
as potential toxicity when such information is available.
Dust can be an important source of exposure for children.
Indoor dust samples were not collected from the five
NYC schools. Dust samples were collected from the sixth
school, but since the school had not been routinely cleaned
in the preceding five weeks it is not known if the dust
concentrations were typical for this school. Therefore, in
the SHEDS model analyses estimates of dust concentrations
in each room with an air measurement were made using
partition estimates (see Appendix E). Using this approach,
the estimated concentrations of PCBs in dust were lower
at the post-remediation time point due to the lower air
concentrations. However, there were reasons to assess the
sensitivity of modeled PCB exposures for lower dust and
soil ingestion. First, given that schools are often regularly
cleaned, child exposures to dust in school may be lower
than those in residences. Second, the soil at the five NYC
schools was remediated first by restricted access followed by
removal of soils with > Ippm PCBs. So the post-remediation
exposures to PCBs in soil were likely to be lower.
Sensitivity tests were run by decreasing the dust and soil
concentrations by 70% and by 90% to examine the impact
on SHEDS model absorbed dose estimates. Sensitivity test
results for decreased dust and soil levels are reported in
Table 4-39 for 6-10 year-olds at the post-remediation time
points using measurement information from five schools. For
an assumed 70% reduction in dust and soil concentrations
at the post-remediation time point, there was a 7% decrease
in estimated absorbed dose at the 50th percentile and a
15% decrease at the 95th percentile. For an assumed 90%
reduction in dust and soil levels the decreases in estimated
absorbed doses were 9% and 21% at the 50th and 95th
percentiles, respectively. Although the pre-remediation time
point sensitivity was not examined, the impact would have
been less due to the higher proportion of total exposure from
inhalation at that time point.
4.5.4 Exposure Modeling Uncertainties and Limitations
Models can be useful tools for estimating human exposure
to chemicals in the environment, but it is important to
understand the limitations and uncertainties associated with
model inputs and outputs. Exposure models are designed
to use information about concentrations of chemical in
environmental media, and a person's contact with chemicals
in that environment to estimate the amount of exposure that
may occur. Simple point-estimation models often do not
incorporate variability in environmental levels and human
contact and do not characterize the range of exposures likely
to be encountered by a human population or sub-population.
The SHEDS model incorporates variability in chemical
concentrations and some aspects of human activity (e.g., time
spent in different locations and activities, distributions of
contact rates) in order to estimate distributions of exposure
and absorbed dose. However, there are uncertainties in
some of the assumptions and exposure pathways/scenarios
modeled (e.g., ingestion of caulk was not modeled), the
information available for input into the model, and in some
of the underlying model parameters. Also, while SHEDS
includes sophisticated exposure algorithms, the dose
estimation module in SHEDS is a simple 1-compartment
PK model based on daily absorption rates, and is intended
for screening purposes; it can be linked to PBPK models for
more sophisticated tissue dose modeling if sufficient data are
available (but they are not available at this time for PCBs;
thus, the SHEDS PK model was used in this study).
Table 4-39. Sensitivity test results for post-remediation decreases in dust and soil PCB concentrations on the
distributions of estimated total absorbed PCB dose for 6-10 year olds based on measurements from five schools3
Model Parameter
Mean
SD
Percentiles of the Distribution
p5 p25 p50 p75 p95
p99
Post remediation
Baseline 0.008 0.004
70% reduction in dust/soil 0.007 0.002
90% reduction in dust/soil 0.007 0.001
Estimated Absorbed Dose pg/kg/day
0.005 0.006 0.007
0.005 0.006 0.007
0.005 0.006 0.007
0.009 0.012
0.008 0.010
0.008 0.009
0.018
0.014
0.010
Post remediation
70% reduction in dust/soil 0.89
90% reduction in dust/soil 0.85
Ratio of Estimated Doses at 70 and 90% Reductions to Baseline Estimate
0.92
0.90
0.94
0.92
0.93
0.91
0.91
0.86
0.85
0.79
0.74
0.57
Schools 1, 2, 3, 4, and 5.
-------
While there are uncertainties in the SHEDS absorbed dose
estimates, the probabilistic modeling approach provides
estimates for the range of absorbed doses based on variability
in concentrations and activity. Such information can inform
risk assessments by characterizing not only the average
absorbed dose, but also the upper end of exposures and
absorbed doses in a population. It is also important to
recognize that many of the uncertainties in parameters for
the SHEDS model would also apply to other absorbed dose
estimation approaches, including point-estimation models.
Some of the limitations and uncertainties important for
modeling PCB exposures in school environments are
described below. In many of these areas, uncertainties can be
reduced in the future through collection of additional data or
information.
Estimated Adult PCB Absorbed Dose - Exposure estimation
was not performed for adults, including teachers and staff,
as part of this effort due to the lack of personal activity
data at schools such as those available for children in the
Consolidated Human Activity Database. It is anticipated that
adults would spend more time in school buildings, which
would tend to lead to higher doses, while less contact with
dust and soil and higher relative body masses would lead
to lower doses. The absorbed doses experienced by adults
in school environments with PCB levels found in the six
schools may be similar to those estimated for the 14-18
year old age group, but good adult activity information would
need to be applied in the model to determine if that is the
case.
Levels ofPCBs in school dust - Interior dust samples were
not collected as part of the NYC remedial pilot investigation.
Dust can be an important source of exposure through
inhalation, non-dietary ingestion, and dermal contact. PCB
concentration data were not available for dust. For the
purposes of this modeling effort, the dust concentrations
were estimated for each room with an indoor air PCB
measurement using an estimated solid/air partition coefficient
(see Appendix E). Wipe sample data were not used as the
surrogate for dust because the wipes likely contained some
distribution of dust-bound and surface-residue PCBs, but that
distribution cannot be defined from the measurement. Also,
the ug/100 cm2 units for wipes cannot be simply translated to
the ug/kg units for dust. The uncertainty in concentrations of
PCBs in dust can be reduced by collection of dust samples;
protocols for future sampling should including bulk dust
sample collection and should include both concentration and
loading where possible.
Building ventilation conditions - Air samples were
collected at multiple locations (including classrooms,
gymnasiums, cafeterias, transitional spaces) at several time
points at multiple schools under different conditions. Air
concentrations of indoor pollutants are strongly impacted
by ventilation rates in a building or in a room. Actual rates
of ventilation with outdoor air and with air from adjoining
spaces are difficult to measure in individual rooms in older
buildings. While the air PCB measurements certainly
incorporated some level of variability in ventilation effects,
it is not possible to quantitatively characterize the impact
of ventilation on air concentrations from the available data.
Exposures (and doses) might be substantially different under
different ventilation conditions. Doubling the outdoor air
ventilation rate to a room would result in an approximately
50% decrease in indoor air PCB concentrations if all other
factors were unchanged, while reducing the outdoor air
ventilation rate by half would approximately double indoor
air PCB levels. Uncertainties due to ventilation effects can
be reduced by collection of baseline data on ventilation and,
where successive measurements are performed, making
those measurements under similar ventilation conditions.
However, it will remain difficult to accurately assess air flows
between a room and other adjacent spaces in older buildings
that may also contain PCBs, limiting the ability to fully
account for ventilation impacts on PCBs in indoor air.
Temperature conditions - Both laboratory and building
studies have demonstrated that PCB emissions increase
with increasing temperature, and that temperature can affect
indoor air PCB concentrations (Guo et al, 2011; Macintosh
et al., 2012). The measurements used in this study were
based primarily on measurements made during the summer
when ambient temperatures are highest. Indoor air levels
and estimated exposures may be somewhat lower during
colder months. There was some information to suggest a
relationship between indoor air PCB concentrations and
ambient temperature with one of the NYC schools (NYC
SCA, 2012). However, Hazrati and Harrad (2006) showed
no seasonal variations in indoor PCB concentrations as, in
effect, any increased volatilization in summer may be offset
by increased ventilation.
Dermal contact and non-dietary ingestion rates - Dermal
contact rates with potentially contaminated surfaces have not
been directly assessed for children in school environments.
Likewise, non-dietary ingestion rates of PCBs have not been
directly characterized for children in school environments.
Thus, values from the literature based on other studies were
used as model inputs. These values are the best available
information at this time.
Pulmonary absorption of PCBs - Limited information is
available to determine the pulmonary absorption of PCBs
through the lungs; thus, a value of 70% absorption was
assumed for this purpose and sensitivity tests were performed
using values of 30%, 80%, and 100% absorption to examine
the impact on estimates of absorbed dose. Because a
majority of the modeled absorbed dose resulted from
the inhalation pathway, the value selected for pulmonary
absorption can have an important impact on absorbed dose
estimates. Sensitivity analyses indicated that the median
absorbed dose estimate for 6-10 year-old children would be
49% lower assuming pulmonary absorption of 30%, and 35%
higher assuming pulmonary absorption of 100%.
Dermal absorption of PCBs - Some animal and human
cadaver skin absorption data are available for selected
Aroclors. However, dermal absorption may be affected by a
-------
number of factors including skin conditions; dermal loading
rates; and how much of the PCBs are bound to soil, dust,
or caulk particles. There remains uncertainty in absorption
rates in natural environments under different conditions.
The default values from the literature are the best available
information for estimating dermal absorption at this time.
Estimation of dose asAroclor 1254 - Most of the available
absorption data is based on Aroclor 1254, and the
environmental data generated at five of the schools was only
available as Aroclor measurements (primarily reported as
Aroclor 1254 with an altered Aroclor pattern). Thus, the
SHEDS results are based on modeling total PCBs as Aroclor
1254. This is appropriate since the RfD of 0.020 ug/kg/d is
based on Aroclor 1254 as well. However, it is likely that the
actual congener concentrations in environmental media did
not exactly match an Aroclor 1254 pattern. The analytical
results suggested the mixture in school air was similar
to a modified Aroclor 1254 pattern. Congener-specific
results showed that the pattern in air was weighted more
towards the more volatile congeners, but not as much as
would be predicted from vapor emissions alone, suggesting
inhalation of dust with a congener pattern similar to Aroclor
1254. The absorption rates for all congeners may not be
well characterized when testing is done with Aroclor 1254
mixtures. Norstrom et al. (2010) showed predictions of
exposure to PCBs in ambient air and predicted inhalation
leads to greater exposures of lower-chlorinated congeners
as compared to dietary intake of more highly chlorinated
congeners.
Comparisons of dose estimates within and between schools -
The strength of the SHEDS model is its ability to estimate the
distribution of exposures in a population or sub-population
that incorporates variability in chemical concentrations
in multiple media and variability in human activities that
contribute to exposure. Its ability to characterize exposure
distributions is improved as more data become available
for more locations and/or scenarios, and it relies on
relatively large sample sizes to generate useful estimates for
relevant populations. In this study, SHEDS modeling has
been performed using measurements at up to six schools,
providing some range of variability in concentrations that
may, or may not represent well the larger universe of older
school buildings with PCB sources. Due to the relatively
small number of measurements at each school, SHEDS was
not applied to estimate exposures on a school-by-school
basis. It is clear from the air measurements of total PCBs
that differences in exposures among individuals might be
expected both between schools, and also within schools.
SHEDS incorporates this variability in its distributional
estimates, but is not well suited for comparison of exposure
levels across different rooms in a school or between schools
given the small number of schools and small number of
measurements within each school.
Dietary and residential exposure to PCBs - SHEDS
modeling estimates in this report are limited to estimates of
absorbed doses (and exposure pathway analysis) resulting
from school environments. A more complete model
assessment would include the contribution from dietary
sources as well as contributions from residential and outdoor
exposures away from school. The evaluation of dietary
exposures is important because dietary intake is often
characterized as the primary route of exposure to PCBs in the
general population. Summaries of estimated children's PCB
dietary intake based on FDA Total Dietary Study data for the
years 1991 through 1997 show considerable variability across
quarters and an average intake of 0.008 ug/kg/day, with
the most recent 1997 estimate of 0.003 ug/kg/day for 6 and
10-year old children (ATSDR 2000). These values can be
compared to the median estimates of 0.020 ug/kg/day (pre-
remediation) and 0.008 ug/kg/day (post-remediation) time
points for 6 - 10 year-old children based on the measurement
results for five schools. It is possible that PCB levels in
foods have continued to decrease since 1997. Given the
small number of samples with measurable PCB results in the
most recent Total Diet Study, it is not clear whether there are
sufficient U.S. data to support development of distributional
parameters for SHEDS modeling of total dietary exposure.
In addition, there may be segments of the population with
above average dietary intakes, particularly for high fish
consumers. Residential indoor air levels have been found
in a range of a few to about 14 ng/m3, much lower than
the levels found in schools with PCB sources. Outdoor air
levels are typically lower than those in indoor residential
air. Additional time and effort are needed to examine the
extant PCB dietary data (including recent Total Diet Study
data from the FDA) as well as residential and ambient air
data to determine their suitability for incorporation into the
SHEDS model. However, the current extant data available
for assessing background exposures is limited and may not be
sufficiently robust for probabilistic estimates in the general
population. Initial work is focusing on dietary intake of
PCBs from fish and seafood.
The factors discussed above, as well as other model inputs,
contribute to uncertainty in modeled exposure and absorbed
dose estimates resulting from PCBs in school environments.
Sensitivity testing for two important parameters helps
define some of the range of uncertainty. Uncertainty around
SHEDS absorbed dose estimation distributions could be
better characterized given sufficient data. However, there
is still insufficient school measurement data for a full
uncertainty characterization for SHEDS PCB modeling
of school exposures. Collection of additional data or
information is likely to help reduce uncertainties and allow
for better uncertainty characterizations in the future.
-------
-------
5.0
Conclusions
5.1 Sources of PCBs in School Buildings
PCB-containing caulk is a primary source of PCBs in and
around school buildings. PCB emissions from caulk can
potentially result in concentrations from hundreds to over
a thousand nanograms per cubic meter in indoor air. PCBs
from exterior caulks around windows and ventilation intakes
can lead to elevated concentrations in indoor spaces. PCBs
in exterior caulk are likely to enter the soil near school
buildings. Caulk containing PCBs was found to be mostly
intact and still somewhat flexible, but visual examination
alone may not be adequate for determining if PCBs are
present and testing is needed to determine if caulk or other
sealants in a building contain PCBs.
PCB-containing fluorescent light ballasts remain in use in
some older school buildings and are a primary source of
PCBs. Emissions from intact ballast capacitors can lead to
the presence of PCBs in school environments. PCB residues
from previously failed ballast capacitors may remain in
fixtures even if the ballast is replaced. Leaking or bursting
capacitors are likely to substantially elevate PCB levels in
indoor environments when they fail. Because these ballasts
have exceeded their expected operational lifetimes, failure
and possible leakage will continue and is likely to increase
for ballasts remaining in place.
Several paint samples had total PCB concentrations above
100 ppm, up to 718 ppm. PCBs were used as plasticizers
or flame retardants in some paints, so it is possible that
these paints may have incorporated PCB when they were
originally applied. Thus, it is possible that paints could
be primary sources of PCBs in buildings based on our
definition. Although they were not encountered in our study,
window glazing and ceiling tile surface coatings containing
PCBs have been reported in school buildings and would be
considered primary sources.
Other primary sources of PCBs may have been used in school
buildings but are no longer present today. For example,
carbonless copy paper and PCB-containing capacitors in
early computer video display terminals may have been used
in school buildings. The potential impact of previously
removed sources on current PCB levels in building
environments cannot be easily determined.
Many of the building and furnishing materials in schools with
apparent primary PCB sources contain PCB levels in the 4
to 100 ppm range. It appears likely that these materials have
absorbed PCBs that have been emitted from primary sources.
While primary sources remain in buildings these other
materials are likely to be in quasi-dynamic equilibrium when
temperature and ventilation conditions remain relatively
constant. However, when primary sources are removed
these materials may serve as secondary sources for emissions
of PCBs into air in the building. Paints may be the most
significant secondary sources given their large surface areas
and relatively high PCB concentrations, but other materials
may be important as well. Following mitigation of primary
sources it may, in some cases, be necessary to consider
mitigation actions for secondary sources.
5.2 School Environment PCB Levels and
Exposures
PCBs are present in indoor air, dust, and on surfaces in
school buildings with PCB-containing source materials, and
are likely to be present in the soil near buildings with exterior
PCB-containing caulk. Building occupants are exposed to
PCBs through contact with these environmental media.
Estimated average total absorbed doses that could occur
from the PCBs in school buildings with environmental levels
that were found in these six schools were near the reference
dose levels for Aroclor 1254 (0.020 ug/kg/day). Because the
Aroclor 1254 reference dose is based on an oral applied dose,
a more direct comparison with the SHEDS absorbed dose
might be the RfD adjusted by a gastrointestinal absorption
factor. Using the 85% absorption factor applied in SHEDS,
the adjusted RfD would be 0.017 ug/kg/d. Over 50% of
the estimated distribution of absorbed doses exceeded the
adjusted reference dose level for the two younger age groups.
These estimates do not include the additional background
exposures to PCBs that occur outside of the school
environment, including exposures from dietary intake and
inhalation of PCBs in outdoor and indoor air at non-school
locations.
Pre-remediation PCB concentrations in indoor air were found
to exceed EPA's 2009 public health guidance levels (ranging
from 70 to 600 ng/m3 depending on age) in many of the
rooms at the six schools that were evaluated. Inhalation was
estimated to be responsible for over 70% of the exposure
that could occur in buildings with environmental levels of
PCBs that were found in these six schools. Mitigation efforts
that focus on reducing indoor air PCB concentrations are
likely to have the greatest impact on reducing exposures,
although cleaning to reduce dust levels will also have an
impact. Based on information from the five New York City
schools, it appears that mitigation efforts can be successful
in substantially reducing indoor air concentrations and
exposures to PCBs.
-------
5.3 Complexity of PCBs in School Buildings
PCBs in school buildings present a complex problem from
exposure assessment, risk assessment, and mitigation
decision-making perspectives. Different aspects of this
complexity are summarized below.
There may be multiple primary sources of PCBs in school
buildings. Numerous different kinds of caulks and sealants
may be present and widespread across many building
locations and they must be sampled to determine whether
they contain PCBs. Fluorescent light ballasts containing
PCBs may be present and light fixtures may be contaminated
with residues from ballasts that have previously failed.
PCBs are semi-volatile organic chemicals with a wide range
of vapor pressures that will vaporize from primary sources
and can be transported throughout indoor and outdoor
environments. They are absorbed by dust and soil which can
result in additional transport and exposure.
PCBs absorb into numerous materials in a building resulting
in a reservoir that remains even after primary sources are
removed or otherwise mitigated. These secondary sources
may result in continuing exposures following removal or
remediation of primary sources.
Over 120 different PCB congeners have been measured in
indoor air. These different congeners have a wide range of
physical properties.
Ventilation with outdoor air is an important factor in the
indoor air PCB concentrations that will result from source
emissions. However, ventilation in older school buildings is
often poor, difficult to assess, and difficult to improve.
An illustration of complexity of the situation that could be
faced in a school classroom with different sources of PCBs
is shown in Figure 5-1. Multiple primary sources of PCBs
may be emitting PCBs into the air, onto surfaces, into dust
and soil. Some of the PCBs are absorbed into other building
materials that serve both as sources and sinks for PCBs.
Ventilation occurs both from exchange with outdoor air and
from exchange with air in other building spaces. PCBs can
be carried between these school spaces. Finally, it is likely
that there may be over 100 different PCB congeners present,
with a range of vapor pressures and other physical and
chemical properties that affect transport and absorption.
5.4 Study Limitations
There are important limitations and uncertainties in the
information included in this report. Key limitations and
uncertainties are summarized below.
PCB measurement results were available from only six
schools. It is not known if these results are representative
of older schools nationwide, both in terms of the presence
of PCB-containing materials and components and the
environmental concentrations measured in and around the
school buildings.
Ventilation
V Primary PCB
Source
Secondary PCB
Sources and Sinks
Dust/Soil
Figure 5-1. Illustration of the complexity of PCBs in school buildings
Materials and components containing PCBs were likely to
have been used in buildings other than schools. This report
does not address whether and to what extent PCBs may be
a potential problem in other types of buildings, and if so,
whether environmental concentrations and exposures are
likely to be similar.
PCB emissions from materials and light ballasts were not
directly measured at the six schools. Modeled emission
estimates and the resulting predictions of indoor air
concentrations have considerable uncertainties. Emission
parameters are likely to vary across different materials and
for different temperature and ventilation conditions. Two
different types of chambers were used to test caulk and
light ballasts emissions, possibly impacting comparability.
Emissions from light ballasts are likely to vary depending on
the lighting fixture design and the condition of the ballast and
capacitor. Emissions from light fixtures contaminated from
previously leaking or failed ballasts could not be evaluated.
Attributing the relative impact of PCB emissions from caulk
and light ballasts on PCB levels in the schools was difficult
because both sources were present in most buildings, and the
Aroclor mixture used in every light ballast was not identified.
Several paint samples were found to have several hundred
ppm of PCBs, and it is not clear whether these contained
PCBs when installed and might be considered primary
sources.
There is uncertainty in modeled estimates of PCB exposures
due to uncertainties in key model parameters. In particular,
there is limited information for pulmonary absorption
fraction from the range of PCB congeners in vapor and
particle-bound forms. There is also uncertainty in total PCB
exposures because of the lack of robust data for background
exposures from diet and other non-school sources.
-------
6.0
References
Allen JR, Norback DH, Hsu 1C. (1974). Tissue modifications in monkeys as related to absorption, distribution, and excretion of
polychlorinated biphenyls. Arch Environ Contain Toxicol, 2:86-95.
ASTM International (2006), D5116-06 Standard Guide for Small-Scale Environmental Chamber Determinations of Organic
Emissions from Indoor Materials/Products, ASTM International, West Conshohocken, PA.
ASTM (2006). ASTM Method D6661-01 - Standard Practice for Field Collection of Organic Compounds from Surfaces Using
Wipe Sampling. ASTM International.
ASTM (2007). ASTM Method D6245-07 - Standard Guide for Using Indoor Carbon Dioxide Concentrations to Evaluate
Indoor Air Quality and Ventilation. ASTM International.
ATSDR (2000). Toxicological Profile for Polychlorinated Biphenyls (PCBs). U.S. Department of Health and Human Services,
Public Health Service, Agency for Toxic Substances and Disease Registry, http://www.atsdr.cdc.gov/toxprofiles/tpl7.pdf.
November, 2000.
ATSDR (2011). Addendum to the Toxicological Profile for Polychlorinated Biphenyls. U.S. Department of Health and Human
Services, Public Health Service, Agency for Toxic Substances and Disease Registry, http://www.atsdr.cdc.gov/toxprofiles/
pcbs_addendum.pdf April, 2011.
Bartlett KH, Marinez M, and Bert J. (2004). Modeling of Occupant-Generated CO2 Dynamics in Naturally Ventilated
Classrooms. Journal of Occupational and Environmental Hygiene, 1:139-148.
Bleavins MR, Breslin WJ, Aulerich RJ, and Ringer RK. (1984). Placenta! and mammary transfer of a polychlorinated biphenyl
mixture (Aroclor 1254) in the European ferret (mustela putorius furo). Environ Toxicol Chem, 3:637-644.
Coghlan KM, Chang MP, Jessup DS, Fragala MA, McCrillis K, and Lockhart TM. (2002). Characterization of Polychlorinated
Biphenyls in Building Materials and Exposures in the Indoor Environment. Proceedings: Indoor Air 2002. The 9th International
Conference on Indoor Air Quality and Climate, Monterey, CA, USA, June 30-July 5, 2002.
Currado GM and Harrad S. (1998). Comparison of Polychlorinated Biphenyl Concentrations in Indoor and Outdoor Air and
the Potential Significance of Inhalation as a Human Exposure Pathway. Environ Sci Technol, 32, 3043-3047.
Delle Site A. (1997). The Vapor Pressure of Environmentally Significant Chemicals: AReview of Methods and Data at Ambient
Temperature. JPhys Chem Re/Data, 26(1): 157-193.
Dilberto JJ, Jackson JA, Birnbaum LS. (1996). Comparison of 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) Disposition
Following Pulmonary, Oral, Dermal, and Parenteral Exposure to Rats. Toxicol and Applied Pharmacol, 138:158-168.
Erickson MD and Kaley IIRG. (2011). Applications of polychlorinated biphenyls, Environmental Science and Pollution
Research, 18(2): 135-151.
Erickson MD. (1997). Analytical Chemistry of PCBs, Second Edition. CRC Press, Boca Raton, FL, USA.
Falconer RL, and Bidleman TF. (1994). Vapor Pressures and Predicted Particle/Gas Distributions of Polychlorinated Biphenyl
Congeners as Functions of Temperature and Ortho-Chlorine Substitution. Atmos Environ, 28(3):547-554.
Frame GM, Cochran JW, and Bowadt SS. (1996). Complete PCB Congener Distributions for 17 Aroclor Mixtures Determined
by 3 HRGC Systems Optimized for Comprehensive, Quantitative, Congener-Specific Analysis. J High Resol Chromatogr,
19:657-668.
Frederiksen M, Meyer HW, Ebbehoj NE, Gunnarsen L. (2012). Polychlorinated biphenyls (PCBs) in indoor air originating
from sealants in contaminated and uncontaminated apartments in the same housing estate. Chemosphere, 89(4):473-479.
Fischer RC, Wittlinger R, and Ballschmiter K. (1992). Retention-index Based Vapor Pressure Estimation for
Polychlorobiphenyl (PCB) by Gas Chromatography. Fresenius' Journal of Analytical Chemistry, 342(4) :421-425.
Foreman, W. T and Bidleman, T F. (1985). Vapor pressure estimates of individual polychlorinated biphenyls and commercial
fluids using gas chromatographic retention data. Journal of Chromatography A, 330:203-216.
Gabio T, Piechotowski I, Wallenhorst T, Klett M, Cott L, Friebel P, Link B, Schwenk M. (2000). PCB-Blood Levels in
Teachers, Working in PCB-Contaminated Schools. Chemosphere, 40:1055-1062.
-------
Glen G., Zartarian V, Smith L, and Xue J. (2012). The Stochastic Human Exposure and Dose Simulation model for
multimedia, multipathway chemicals (SHEDS-Multimedia): residential module. SHEDS-Residential Version 4 Draft Technical
Manual. U.S. EPA National Exposure Research Laboratory. May 21, 2012. http://www.epa. gov/heasd/products/sheds_
multimedia/files/SHEDSResidential_TechManual_2012.pdf
Godwin C, and Batterman S. (2007). Indoor Air Quality in Michigan Schools. Indoor Air, 17:109-121. http://www.epa.gov/
heasd/products/sheds_multimedia/files/SHEDS_Residentialv4_Techmanual_06-16-2010.Final.pdf. June 16, 2010.
Guo Z, Liu X, Krebs KA, Stinson RA, Nardin JA, Pope RH, and Roache NF. (2011). Laboratory Study of Polychlorinated
Biphenyl (PCB) Contamination and Mitigation in Buildings Part 1. Emissions from Selected Primary Sources. National Risk
Management Research Laboratory, U.S. EPA. EPA/600/R-11/156, October 2011.
Guo Z, Liu X, Krebs KA, Stinson RA, Nardin JA, Pope RH, and Roache NF. (2012). Laboratory Study of Polychlorinated
Biphenyl (PCB) Contamination and Mitigation in Buildings Part 2. Understanding the Secondary Sources. National Risk
Management Research Laboratory, U.S. EPA. EPA/600/R-12/156A, January 2012.
Guo Z. (2002). Review of indoor emission source models — Part 2. Parameter estimation, Environmental Pollution,
120:551-564.
Guo Z. (2000). Simulation tool kit for indoor air quality and inhalation exposure (IAQX) Version 1.0 User's Guide, U.S.
Environmental Protection Agency, National Risk Management Research Laboratory, Research Triangle Park, NC, Report
No. EPA-600/R-00-094, 76 pp. http://www.epa.gov/nrmrl/appcd/mmd/iaq.html
HansenLG. (1999). The ortho side of PCBs: Occurrence and disposition. Kluwer Academic Publishers. Boston, MA.
Harrad S, Goosey E, Desborough J, Abdallah MA, Roosens L, Covaci A. (2010). Dust for U.K. Primary School Classrooms
and Daycare Centers: The Significance of Dust as a Pathway of Exposure of Young U.K. Children to Brominated Flame
Retardants and Polychlorinated Biphenyls. Environ Sci Technol, 44, 4198-4202.
Harrad S, Ibarra C, Robson M, Melymuk L, Zhang X, Diamond M, and Douwes J. (2009). Polychlorinated Biphenyls
in Domestic Dust from Canada, New Zealand, United Kingdom and United States: Implications for Human Exposure.
Chemosphere, 76:232-238.
Hazrati S, and Harrad S. (2006). Causes and Variability in Concentrations of Polychlorinated Biphenyls and Polybrominated
Diphenyl Ethers in Indoor Air. Environ Sci Technol, 40:7584-7589.
Heinzow B, Mohr S, Ostendorp G, Kerst M, Korner W. (2007). PCB and Dioxin-like PCB in Indoor Air of Public Buildings
Contaminated with Different PCB Sources - Deriving Toxicity Equivalent Concentrations from Standard PCB Congeners.
Chemosphere, 67:1746-1753.
Herrick RF, Meeker JD, Altshl L. (2011). Serum PCB Levels and Congener Profiles Among Teachers in PCB-Containing
Schools: A Pilot Study. Environ Health, 10:56.
Herrick RF, Lefkowitz DJ, and Weymouth GA. (2007). Soil Contamination from PCB-Containing Buildings. Environ Health
Perspect, 115(2): 173-175.
Herrick RF, McClean MD, Meeker JD, Baxter LK, and Weymouht GA. (2004). An Unrecognized Source of PCB
Contamination in Schools and Other Buildings. Environ Health Perspect, 112(10): 1051-1053.
Holmes KK, Shirai JH, Richter KY, Kissel JC (1999). Field Measurement of Dermal Soil Loadings in Occupational and
Recreational Activities. Environmental Research Section A, 80:148-157.
Holmes DA, Harrison BK, and Dolfing H. (1993). Estimation of Gibbs Free Energies of Formation for Polychlorinated
Biphenyls. Environ Sci Technol, 27:725-731.
Hore P, Zartarian V, Xue J, Ozkaynak H, Wang SW, Yang YC, Chu PL, Sheldon L, Robson M, Needham L, Barr D, Freeman
N, Georgopoulos P, Lioy PJ. (2006). Children's Residential Exposure to Chlorpyrifos: Application of CPPAES Field
Measurements of Chlorpyrifos and TCPy within MENTOR/SHEDS-Pesticides Model. Sci Total Environ, 366:525-537.
Hosomi, M. (2005). Volatilization of PCBs from PCB-containing Ballast in Fluorescent Lamp and Indoor PCB pollution:
Odor of PCBs. J. Japan Association on Odor Environment, 36(6): 323-330.
Hu X, Adamcakova-Dodd A, Lehmler H, Hu D, Kania Korwel I, Hornbuckle KC, Thorne PS. (2010). Time Course of
Congener Uptake and Elimination in Rats after Short-Term Inhalation Exposure to Airborne Polychlorinated Biphenyl (PCB)
Mixture. Environ Sci Technol, 44:6893-6900.
Hummel AA, Braun KO, Fehrenbacher MC. (1996). Evaporation of a liquid in a flowing airstream, Am IndHygAssoc J,
57:519-525.
Kissel JC, Richter KY, Fenske RA. (1996). Factors Affecting Soil Adherence to Skin in Hand-Press Trials. Bull. Environ.
Contam. Toxicol, 56:722-728.
-------
Kohler M, Tremp J, Zennegg M, Seller C, Minder-Kohler S, Beck M, Lienemann P, Wegmann L, and Schmid P. (2005). Joint
Sealants: An Overlooked Diffuse Source of Polychlorinated Biphenyls in Buildings. Environ Sci Technol, 39:1967-1973.
Kohler M, Zennegg, and Waeber. (2002). Coplanar Polychlorinated Bipheynyls (PCBs) in Indoor Air. Environ Sci Technol,
36:4735-4740.
Liebl B, Schettgen T, Kerscher G, Eroding H, Otto A, Angerer J, Drexler H. (2004). Evidence for Increased Internal Exposure
to Lower Chlorinated Polychlorinated Biphenyls (PCB) in Pupils Attending a Contaminated School. IntJHyg Environ Health,
207:315-324.
Macintosh DL, Minegishi T, Fragal MA, Allen JG, Coghlan KM, Stewart JH, McCarthy JF. (2012). Mitigation of Building-
related Polychlorinated Biphenyls in Indoor Air of a School. Environ Health, 11:24, 1-10.
McCurdy T, Glen G, Smith L, and Lakkadi Y. (2000). The National Exposure Research Laboratory's Consolidated Human
Activity Database. J Expos Anal Environ Epidem, 10:566-578.
Martinez A, Erdman NR, Rodenburg ZL, Eastling PM, Hornbuckle KC. (2012). Spatial distribution of chlordanes and PCB
congeners in soil in Cedar Rapids, IA USA. Environ Pollut, 161:222-228.
Natural Resources Canada, Lighting Reference Guide - Fluorescent Lamp Ballasts, http://oee.nrcan.gc.ca/publications/
equipment/lighting/9953. Accessed July 2012.
Nazaroff WW, Bhangar S., Mullen NA, Hering SV, and Kreisberg NM. (2010). Ultrafine Particle Concentrations in
Schoolrooms and Homes, Final Report Contract No 05-305. California Air Resources Board, Sacramento, CA.
Nessel CS, Amoruso MA, Umbreit TH, Meeker RJ, Gallo MA. (1992). Transpulmonary Uptake and Bioavailability of 2,3,7,8-
TCDD from Respirable Soil Particles. Chemosphere, 25(l-2):29-32.
NIOSH (1975). Current Intelligence Bulletin 7. Polychlorinated (PCBs). National Institute for Occupational Safety and
Health. Novembers, 1975. http://www.cdc. gov/niosh/docs/1970/78127_7.html. Accessed July 2012.
Norstrom K, Czub G, McLachlan MS, Hu D, Thorne PS, Hornbuckle KC. (2010). External exposure and bioaccumulation of
PCBs in humans living in a contaminated urban environment. Environ Int, 36(8):855-861.
NYC SCA (2010). Remedial Investigation Plan for the New York City School Construction Authority Pilot Study to Address
PCB Caulk in New York City School Buildings. EPA Consent Agreement and Final Order Docket Number: TSCA-02-
2010-9201. TRC Engineers Inc. and the New York City School Construction Authority. July, 2010. http://www.nycsca.org/
Community /Programs/EPA-NYC-PCB/PCBDocs/EPAApproved RemdialInvestigationWorkPlan.pdf
NYC SCA (2011). Interim Remedial Investigation Report for the New York City School Construction Authority Pilot Study to
Address PCB Caulk in New York City School Buildings. TRC Engineers, Inc. June 2011. http://www.nycsca.org/Community/
Programs/EPA-NYC-PCB/PCBDocs/EPApprlnterRemedlnvestRep.pdf
NYC SCA (2012). Final Remedial Investigation Report for the New York City School Construction Authority Pilot Study to
Address PCB Caulk in New York City School Buildings. TRC Engineers, Inc.
Ozkaynak H, Xue J, Zartarian VG, Glen G, Smith L. Modeled Estimates of Soil and Dust Ingestion Rates for Children (2011).
Risk Anal, 31(4):592-608.
Persily AK. (1997). Evaluating Building IAQ and Ventilation with Indoor Carbon Dioxide. ASHRAE Transactions, 103(2): 193-
204.
Philips Lighting (undated). Ballast lifetime calculations, Philips Lighting Technical Note TN005.
http://lighting.philips.com/gl_en/global_sites/fluo-gear/dimming/download/pdf/technical-notes/tn005.pdf. Accessed March,
2011.
Price HA, Welch RL. (1972). Occurrence of poly chlorinated biphenyls in humans. Environ Health Perspect, 1:73-78.
Priha E, Hellman S, Sorvari J. (2005). PCB Contamination from Polysulphide Sealants in Residential Areas - Exposure and
Risk Assessment. Chemosphere, 59, 537-543.
Qian K, ZhangY, Little JC, Wang X. (2007). Dimensionless correlations to predict VOC emissions from dry building materials.
Atmos Environ, 41:352-359.
Rudel RA, Seryak LM, and Brody JG. (2008). PCB-Containing Wood Floor Finish is a Likely Source of Elevated PCBs in
Resident's Blood, Household Air and Dust: A Case Study of Exposure. Environ Health, 7(2).
Scheff PA, Paulius VK, Huang SW, and Conroy LM. (2000). Indoor Air Quality in a Middle School, Part I: Use of CO2 as a
Tracer for Effective Ventilation. App Occ and Environ Hygiene, 15(ll):824-834.
Schlummer M, Moser GA, McLachlan MS. (1998). Digestive tract absorption of PCDD/Fs, PCBs, and HCB in humans: Mass
balances and mechanistic considerations. ToxicolAppl Pharmacol, 152:128-137.
-------
Staiff D, Quinby GE, Spencer DL, Starr Jr. HG. (1974). Polychlorinated biphenyl emission from fluorescent lamp ballasts.
Bulletin of Environmental Contamination and Toxicology, 12(4):455-463.
Sullivan, D. (2008). Polychlorinated Biphenyls (PCBs) and Indoor Air: Source Investigation and Remdial Approach for
a Public School Building in New Bedford, Massachusetts, USA. 28th International Symposium on Halogenated Organic
Pollutants (POPs), Birmingham, England, UK, August 17-22, 2008.
TRC. (2006). New Bedford High School Indoor Polychlorinated Biphenyls Sampling. Report of Findings. Prepared by the
TRC Environmental Coroporation. Prepared for the City of New Bedford, MA. November 17, 2006.
TRC. (2008). New Bedford High School Indoor Polychlorinated Biphenyls Source/Sink Sampling Program. Report of
Findings. Prepared by the TRC Environmental Coroporation. Prepared for the City of New Bedford, MA. October 2008.
TRC. (2009). New Bedford High School Indoor Polychlorinated Biphenyls Quasi-Random Bulk Material Sampling Program.
Report of Findings Addendum. Prepared by the TRC Environmental Coroporation. Prepared for the City of New Bedford, MA.
May 2009.
U.S. EPA (2010). Recommended Toxicity Equivalence Factors (TEFs) for Human Health Risk Assessments of
2,3,7,8-Tetrachlorodibenzo-p-dioxin and Dioxin-like Compounds. EPA/100/R 10/005, Risk Assessment Forum, Office of the
Science Advisor. December 2010.
U.S. EPA. 2008. Standard Operating Procedure for Sampling Porous Surfaces for Polychlorinated Biphenyls (PCBs). EIA
Porous Sampling 1, Revision 3, July 2008.
U.S. EPANERL (2010). AResearch Study to Investigate PCBs in School Buildings, EPA600R-10/074, National Exposure
Research Laboratory, http://www.epa.gov/pcbsincaulk/research-plan.pdf.
U.S. EPANRMRL (2010). PCBs in Caulk: Evaluation of Mitigation Methods and Source Characterization, http://www.epa.
gov/pcbsincaulk/research-plan-draft.pdf.
U.S. EPA (1976). PCBs in the United States Industrial Use and Environmental Distribution. Contract 68-01-3259 Task 1 Final
Report. U.S. EPA Office of Toxic Substances, Washington D.C. EPA/560/6-76-005, February 25, 1976.
Volckens J, and Leith D. (2003) Partitioning theory for resporitory deposition of semivolatile aerosols. Ann Occup Hyg,
47(2): 157-164.
Weschler CJ, Nazaroff WW. (2010). SVOC partitioning between the gas phase and settled dust indoors. Atmos Environ,
44:3609-3620.
Wester RC, Maibach HI, Bucks DA, McMaster J, Mobayen M, Sarason R, and Moore A. (1990). Percutaneous absorption and
skin decontamination of PCBs: In vitro studies with human skin and in vivo studies in the Rhesus monkey. J Toxicol Environ
Health, 31:235-246.
Wester RC, Maibach HI, Sedik L, Melendres J, and Wade M. (1993). Percutaneous absorption of PCBs from soil: In vivo
rhesus monkey, in vitro human skin, and binding to powdered human stratum corneum. J Toxicol Environ Health, 39(3):375-
382.
WHO (2006). Van den Berg M, BirnbaumLS, DenisonM, DeVitoM, Farland W, Feeley M, Fiedler H, HakanssonH,
Hanberg A, Haws L, Rose M, Safe S, Schrenk D, Tohyama C, Tritscher A, Tuomisto J, Tysklind M, Walker N, Peterson RE.
Review - The 2005 World Health Organization Reevaluation of Human and Mammalian Toxic Equivalency Factors for Dioxins
and Dioxin-like Compounds. Toxicol Sci, 93(2):223-241.
Xue J, McCurdy T, Spengler J, and Ozkaynak H. (2004). Understanding variability in time spent in selected locations for
7-12-year old children J. Expos. Anal. Environ. Epidem. 14: 221-233.
Zartarian V G, Xue J, Glen G, Smith L, Tulve NS, and Tornero-Velez, R. (2012). Quantifying Children's Aggregate (Dietary
and Residential) Exposure and Dose to Permethin: Application and Evaluation of EPA's Probabilistic SHED-Multimedia
Model. J Expos Sci Environ Epidem, 22(3):267-273.
Zhang X, Diamond M, Robson M, Harrad S. (2011), Sources, emissions, and fate of polybrominated diphenyl ethers and
polychlorinated biphenyls indoors in Toronto, Canada. Environ Sci Technol, 45:3268-3274.
-------
Appendix A.
PCB Congener Information
Table A-l. Information for the 209 PCB congeners
Congener
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Chlorine
Positions
2
3
4
2-2
23
2-3
24
2-4
25
26
3-3
34
3-4
35
4-4
23-2
24-2
25-2
26-2
23-3
234
23-4
235
236
24-3
25-3
26-3
24-4
245
246
Dioxin- Vapor Vapor
No. of Like Pressure Pressure
Chlorines Congener (torr) Ref.ae
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2.07E-03
2.72E-03
1.22E-03
1.33E-03
1.46E-03
1.19E-03
1.53E-03
2.60E-03
6.85E-04
5.56E-04
6.24E-04
9.44E-04
5.82E-04
4.96E-04
5.82E-04
6.38E-04
1.04E-03
2.27E-04
2.21 E-04
1.97E-04
3.35E-04
5.31 E-04
2.79E-04
2.92E-04
5.31 E-04
2.43E-04
3.48E-04
6.38E-04
a
b
b
b
b
b
b
b
c
b
c
c
b
b
b
b
b
b
b
b
b
b
b
b
b
b
d
b
Weight
%of
Congener
in Aroclor
1242'
0.34
0.02
0.11
2.71
0.11
1.24
0.18
6.48
0.4
0.14
0.04
0.17
1.95
3.44
3.29
9.14
0.84
0.77
3.08
0.01
0.13
0.61
1.38
0.44
7.31
0.08
Weight
% of Weight Weight
Congener % of % of
in Aroclor Homolog Homolog
1254' inA1242 inA1254
0.47 0
0.06 13.42 0.24
0.02
0.13
0.03
0.09 48.02 1.24
0.08
0.25
0.04
0.03
0.19
-------
Table A-l. Information for the 209 PCB congeners (continued)
Congener
Number
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
Chlorine
Positions
25-4
26-4
34-2
35-2
34-3
35-3
34-4
345
35-4
23-23
234-2
23-24
235-2
23-25
236-2
23-26
24-24
245-2
24-25
246-2
24-26
25-25
25-26
26-26
234-3
23-34
235-3
23-35
236-3
234-4
2345
2346
235-4
236-4
2356
24-34
245-3
24-35
Dioxin- Vapor Vapor
No. of Like Pressure Pressure
Chlorines Congener (torr) Ref.ae
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
2.60E-04
4.42E-04
2.21 E-04
3.35E-04
9.66E-05
1.43E-04
8.61 E-05
1.52E-04
9.02E-05
9.02E-05
1.01 E-04
1.43E-04
1.14E-04
2.21 E-04
1.16E-04
1.22E-04
1.27E-04
1.36E-04
1.97E-04
1.50E-04
2.60E-04
3.88E-04
4.35E-05
3.51 E-05
8.45E-06
3.43E-05
5.43E-05
1.06E-04
1.14E-04
4.42E-05
5.31 E-05
b
b
b
b
b
b
b
c
b
b
b
b
b
b
a
b
b
b
b
b
b
d
c
b
e
b
b
b
b
b
b
Weight
%of
Congener
in Aroclor
1242'
7.82
2.05
5.35
0.02
0.07
2.19
0.79
0.69
1.25
0.19
3.63
0.91
0.38
0.92
1.17
2.6
0.23
3.47
0.71
0.11
1.85
0.32
1.17
0.11
1.68
3.38
0.17
Weight
% of Weight Weight
Congener % of % of
in Aroclor Homolog Homolog
1254' inA1242 inA1254
0.28
0.05
0.16
0.07
0.12 32.71 16.4
0.01
0.15
2.31
0.05
0.14
0.12
1.1
5.38
0.12
0.55
0.02
0.18
0.02
0.59
1.01
------- |