EPA/600/A-92/259
ATTACHMENT 1.
Household Exposures to Benzene from Showering with		
Gasoline-Contaminated Ground
By: Andrew B. Lindstrom', V. Ross Highsmith1, Timothy J. Buckley', William J. Pate2,
Larry C. Michael3, and R. Mark Johnson4
1 U.S. EPA, Atmospheric Research and Exposure Assessment Laboratory, RTP, NC 27711
1 NC Department of Environment, Health, and Natural Res., P.O. Box 27687, Raleigh, NC 27611
i Research Triangle Institute, P.O. Box 12194, RTP, NC 27709-2194
4 Acurex Corporation, P.O. Box 13109, RTP, NC 27709
ABSTRACT
In a private residence using benzene-contaminated ground water (« 300 fig ft), a series of
experiments were performed to assess the benzene exposures that occur in the shower stall, bathroom,
master bedroom, and living room as a result of a single 20 minute shower. Sampling methodologies
used in this assessment included: fixed site Summa™-polished canisters and Tenax GC® cartridges;
personal Tenax GC® devices; and, grab samples collected with glass gas-tight syringes. Integrated
Summa™ and Tenax GC® samples were collected from the target microenvironments over 20, 60, and
240 minute periods. These results are contrasted with the long-term personal samples (6 h) and grab
samples that were collected at 0, 10, 18, 20, 25, 25.S, and 30 minutes. Results indicate that maximum
short-term benzene concentrations occurred in the shower stall (758 - 1673 /ig/m3) and bathroom (366 -
498 ng/m3). The total dermal and inhalation dose resulting from a single 20 minute shower was
estimated to be approximately 281 fig. This is 2 - 3.5 times higher than the mean inhalation dose
received during concurrent 6 h occupation of the house. The benzene dose relating to a single shower
and continuous occupancy of the residence was shown to be approximately 697 fig/day, with the shower
accounting for 40 % of the daily total (24 % dermal and 16 % inhalation), and the remaining 60 %
relating to respiration in the house for the balance of the day.
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's
peer review and administrative review policies and approved for presentation and publication. Mention
of trade names or commercial products does not constitute endorsement or recommendation for use.
INTRODUCTION
Use of volatile organic compound (VOC) contaminated ground water for ordinary domestic
purposes can lead to oral, dermal, and inhalation exposures. Several recent studies have demonstrated
that the inhalation route may be as important as, or more important than, direct ingestion of
contaminated water1-2. Studies have also demonstrated that residential water use can lead to significant
airborne exposures in areas remote from the water use zone3. Much of this preceding work concerning
the water-to-air transfer of pollutants and resulting exposures has been conducted with radon4,
trichloroethylene', and chloroform5. This study characterizes household benzene exposures that can
occur with the use of gasoline contaminated water.
The Environmental Protection Agency has estimated that between 100,000 - 400,000 of the 3 -
5 million underground storage tanks used in the U.S. for underground liquid petroleum or chemical
storage may have been leaking at one time or another during their lifetime6. This presents ground water
contamination problems which can in turn lead to oral, dermal, and inhalation exposures resulting from
normal domestic water use. This investigation was conducted in a single family residence known to be
using benzene-contaminated groundwater. Chemical analyses performed by North Carolina Department
of Environment, Health, and Natural Resources indicate that the contaminant is a petroleum product
(unpublished data). Specific objectives of this investigation include: (1) assess shower related exposures
that occur in various parts of the house as a result of a single 20 minute shower; (2) examine the
l

-------
relationships between contaminant levels monitored using Tenax-GC®, Summa™ polished canisters, and
glass, gas-tight syringes; and, (3) support concurrent modeling and human exposure biomarker studies.
METHODS
A series of experiments were conducted in a 290 m2 (3100 ft2) north central North Carolina
single family residence from June 11-13, 1991. The residential water was supplied from a single 30
m deep well located on the homeowners' property. Contamination was first discovered in 1986 when
an unusual odor was noticed by the residents of the house. After a water test showed benzene at 425
Hg/t, the homeowners stopped drinking the water and installed a small (15 x 9 cm) charcoal filtration
device in the main water supply line. The residents continued to use the water for bathing and washing.
From 1986 - 1991 benzene was measured in the water despite the filter (and perhaps because of varying
filter efficiencies) at concentrations ranging between 33 and 673 ngIt.
Identical experiments were conducted on three consecutive days following an established
protocol. The experiments involved an individual taking a 20 minute shower with the bathroom door
closed, allowing that individual five minutes to dry off and dress, and then opening the bathroom door
and allowing him to leave. This individual then participated in blood and breath sampling in support
of the biomarker segment of this study7. Whole air samples were collected in Summa™ polished
canisters8 using conventional mass flow control devices in bathroom, bedroom, living room, and in the
ambient air. To save space and avoid electrical hazards, canister samples were collected using flow
restrictors in the shower and bathroom. Tenax GC® samples were also collected in the shower,
bathroom, and living room8. The integrated samples were collected over 20, 60, and 240 minute
periods, all beginning as the shower was turned on and the bathroom door was closed. Glass, gas-tight
syringe samples were simultaneously collected from the shower, bathroom, bedroom, and living room
at 0, 10, 18, 20, 25, 25.5, and 30 minutes, and at additional times in selected areas until the end of
each study period. The members of the sampling team that were initially stationed in the bathroom and
the living room wore personal Tenax GC® sampling devices to assess their exposures. Water samples
were collected for VOC analysis at the shower head from a preaerator bypass valve, and at drain level
at the beginning and end of each shower period to provide a measure of water-to-air transfer efficiency.
Water temperature and flow rate were also measured at the beginning and end of each shower period.
The entire experiment was reviewed and approved by an appropriate Human Subjects Review
Committee. The glass, gas-tight syringe samples were analyzed on-site using a Photovac 10S50 portable
gas chromatograph. Samples (1 ml) were injected onto a CPSIL-5 (Photovac, Inc.) capillary column
operated at 40 °C with 10 ml/min zero air carrier gas flow. Canister samples were analyzed on an
LKB 2091 magnetic sector GC/MS/COMP system operated in the multiple ion detection mode. Tenax
GC® and water samples were analyzed on an HP5988A quadrupole GC/MS/COMP system operated in
multiple ion detection mode. The water samples were analyzed by purging 5 ml aliquots of each sample
onto Tenax GC® cartridges with 480 ml helium (40 ml/min for 12 min) followed by thermal desorption.
RESULTS/DISCUSSION
Three laboratory control canisters had mean benzene recoveries of 96 % with a relative standard
deviation (RSD) of 2 %, and six field control Tenax GC® cartridges had mean recoveries of 128 % (18
% RSD). Eleven syringe field controls had mean benzene recoveries of 97 % (9 % RSD). Three field
control water samples had mean benzene recoveries of 72 % (15 % RSD). Field blank canister, Tenax
GC®, syringe, and water samples were all reported to be at or below their levels of detection (40 ^g/m3,
0.4 /xg/m3, 0.16^g/m3, and 0.6/zg/f, respectively).
Analysis of blind performance evaluation Tenax GC® and Summa™ canister samples resulted in
benzene recoveries ranging from 36 - 64 %, and 73 & 76 %, respectively. The lower than normal
2

-------
spike sample recoveries most likely resulted from the wide calibration range required to support
expected sample loadings. Blank Tenax GC* and Summa™ canister samples reported below their limit
of detection (0.4 and 40.0 jtg/m3, respectively). Excellent agreement was observed in daily duplicate
60 minute Summa™ canister and Tenax GC® samples, with a median relative difference of 4.2 % (range
0 - 10.3 %). Analysis of duplicate syringe samples showed a median relative difference of 20.2 %
(range 3.6 - 40.7 %).
Waterborne benzene concentrations from the preshower head samples ranged from 185 - 367
pg/t (N = 5, mean = 292 ng/t,) while drain level samples ranged from below the detectable limit (0.6
fig/1) to 198 iig/i. This results in a mean water-to-air transfer efficiency of 0.88 (range 0.73 - >
0.99). Analysis of the syringe sample benzene concentrations suggests a wave of benzene moving from
the shower stall to the rest of the house over approximately 60 minutes. Figure 1 is a plot of benzene
levels during the June 12 experiment. Peak benzene concentrations in the shower stall were collected
in the 18 minute sample on June 11th and 12th (758 and 832 jtg/m3 respectively), and in the 20 minute
sample on June 13th (1673 jtg/m3). Benzene concentrations in the bathroom tended to increase for the
first ten minutes and then remain uniformly elevated for the duration of each shower period. Peak
bathroom concentrations were collected in the 10 minute sample on the 11th (366 ftg/m3), in the 25 min
sample on the 12th (371 ^g/m3), and in the 25 min sample on the 13th (498 ng/m3). Maximum
bedroom benzene levels occurred immediately as, or shortly after, the bathroom door was opened, with
peaks at 30 min on the 11th (81 /ig/m3), at 25.5 min on the 12th (146 /*g/m3), and at 30 min on the 13th
(125 jxg/m3). The highest benzene concentrations in the living room were found still later, with peak
levels at 36 min on the 11th (40 jig/m3), 70 minutes on the 12th (62 /ig/m3), and at 48 min on the 13th
(54 fig/m3).
Overall, the collocated integrated Summa™ and Tenax GC* samples were in good general
agreement. The relationship between the two sampling methods can be described by the linear model:
[Tenax GC*] = 0.89 x [Summa™] + 41.36 /ig/m3	(1)
with an r = 0.95 at a P < 0.001 (Figure 2).
This regression line is not significantly different from the line of one-to-one correspondence (F = 2.82,
P = 0.0908). Although the Tenax GC® benzene concentrations were typically higher than the Summa™
canisters (mean difference 21.9 fig/m3, standard error of the difference = 12.5)., a paired T-test shows
the two sampling methods were not significantly different (N = 17, T = -1.75, P = 0.0984).
The personal exposures of the two monitoring team members, assessed using personal Tenax
GC® monitors, were examined and compared with the microenvironmental results. For the first 30
minutes of each experiment one individual was based in the bathroom and the other in the living room.
After the shower period was completed, however, these individuals were free to move to other parts
of the house to assist in sample processing. Their daily personal inhalation dose can be calculated using
the following equation5:
D, = CA x MV x T x F	(2)
Where D, = inhalation absorbed dose (jig); CA = concentration of benzene in the air 0*g/m3); MV =
minute ventilation rate (0.014 m3/min); T = duration of exposure (min); F = fraction of benzene
absorbed (70 %, based on 100 % absorption of alveolar ventilation volume)9. Using this model, we
can estimate that the benzene dose for the individual in the bathroom ranged from 97.7 to 184 fig (mean
= 133.3 /xg). Although the first 20 minutes of exposure in the bathroom corresponds to only 6 % of
3

-------
the total experimental period (6 h), the results from the fixed-site 20 minute bathroom Tenax GC®
samples indicate this period accounted for between 18 - 38 % of this individual's daily absorbed dose.
Slightly reduced exposures are estimated for the individual in the living room (53.0 to 111 fig, mean
= 78.9 fig). Site specific mean 4 h doses (based on mean Tenax GC* + Sum ma™ canister data) were
calculated to be 122 fig for the bathroom, 100 fig for the bedroom, and 79.4 fig for the living room.
The individual 20 minute showers had inhalation doses ranging from 79.6 - 105 fig (mean =
95.9 fig). Adding the average dose absorbed in the bathroom during the 5.5 minutes following the
shower (using the overall 20 & 25 minute mean syringe level of 318 fig/irr5) gives a total average
shower-related inhalation dose of 113 fig.
The average dermal dose resulting from the 20 min showers was calculated to be 168 fig using
the following equation10:
Dd = CW x SA x EF x K„ x T x U (3)
Where: Dd = dermal dose (fig); CW = concentration of benzene in water (292 fig/f) ; SA = surface
area of the 6'4" male volunteer (20,900 cm2); EF = exposure factor, % body surface exposed (75 %);
K,, = benzene dermal permeability constant (0.11 cm/hr)n; T = time in hrs (0.33); U = units
conversion (lf/1000 cm3). This result is consistent with a previous study of chloroform exposures in
showers which demonstrates that dermal and inhalation doses are roughly equivalent5. The total mean
shower-related dose (inhalation + dermal) was calculated to be 281 fig - an exposure that is 2 - 3.5
times higher than the mean 6 h exposures of the other members of the sampling crew. Using the results
of this experiment as a model, we can calculate a worst case exposure scenario for the residents of this
house by adding the shower-related total to an additional 23.6 hours of inhalation exposure at the
average 4 hour living room concentration (30 fig/m3 x 0.014 m3/min x 1415 min x 0.7 = 416 fig). This
leads to a total benzene dose of approximately 697 fig/day, with the shower accounting for 40% of the
daily total (24 % dermal and 16 % inhalation), and the remaining 60 % from continuous respiration in
the house.
Research has demonstrated that the principal nonoccupational sources of benzene exposure
include active smoking, automotive and travel related activities, passive smoking, and the use of certain
household products and building materials12. Because contaminated water was continuously present in
the house during the study (e.g., from toilets, drain traps, and dripping faucets) it is impossible to
determine how much of the benzene measured was directly related to the ground water or to the other
potential residential sources. To put the potential exposures received in this house into perspective, it
is useful to compare the benzene concentrations measured during this study with those measured in other
exposure assessment studies. For example, in a study of the air pollutants associated with
environmental tobacco smoke13, the median benzene level in 200 nonsmoking homes was found to be
about 7 fig/m3. In 300 homes with one or more smokers, the study showed that median benzene
concentrations were significantly higher at 10.5 fig/m3. The four hour long-term average concentration
in this test house (30 fig/m3) was four times higher than these nonsmoking houses and almost three times
higher than the homes with smokers. In a study of pollutant concentrations in commuter vehicles in
Los Angeles14, mean summertime benzene concentrations were found to be 31 fig/m3 - approximately
equal to the 4 h average level in this test residence.
CONCLUSIONS
The results of this study suggest the potential for elevated dermal and inhalation related benzene
exposures resulting from residential use of gasoline-contaminated ground water. Because only one
4

-------
residence was sampled however, the results are not necessarily representative of other homes with
ground water contamination problems. The total dermal and inhalation exposure resulting from a single
20 minute shower (281 ng) was estimated to be from 2 - 3.5 times higher than inhalation exposures that
occur during concurrent 6 h occupation of the house. The absorbed dose relating to a single 20 minute
shower and continuous occupancy of the residence was shown to be approximately 697 ugldzy, with
the shower accounting for 40 % of the daily total (24 % dermal and 16 % inhalation), and the
remaining 60 % relating to respiration in the house for the balance of the day. This study confirms the
observations made in similar studies regarding the importance of dermal and inhalation exposures during
a shower and inhalation exposures in areas far removed from the water use zone.
Differences in benzene concentration measured with Summa™ polished canisters and collocated
Tenax GC® cartridges were not found to be statistically significant. Samples collected with glass, gas-
tight syringes demonstrated a pulse of benzene moving from the shower stall through the rest of the
house over the course of approximately 60 minutes. This sampling method may be useful in
determining instantaneous contaminant concentrations in multiple locations in future studies.
REFERENCES
1.	N J. Giardino, E. Gumerman, N.A. Esmen, I.B Andelman, C.R. Wilkes, tod M J. Small, "Real-Time air measurements of trichloroethylene
in dome«lic bithroomi tiling contaminated water," in Proceeding of ihc 5th International Conference on Indoor Air Quality and Climate: Indoor
Air '90. Vol 2. International Conference on Indoor Air Quality and Climate, Ottawa, 1990, pp 707 - 712.
2.	T.E. McKone, "Human expoaure to volatile organic compounds in household lap water: the indm-vr inhalation pathway.* Environ. Sci Technol.
21: 1194(1987).
3.	C.R. Wilket, MJ. Small, J.B. Andelman, and J. Manhall, "Air quality model for volatile constituent* from indoor usee of water," in
Proceeding! of the 5th International Conference on Indoor Air finality and Climate: Indoor Air '90. Vol 2. International Conference on Indoor
Air Quality and Climate, Ottawa, 1990, pp 783 - 788.
4.	H.M. Prichard, and T.F. Ceaell, "An estimate of population exposures due to radon in public water supplies in the area of Houston, Texas,'
Health Phv». 41: 599 (19811.
5.	C.P. Weiael, PJ. Lioy, and W.K. Jo, "Exposure to volatile organic compoundsresulting from ahowering with chlorinated water," in Proceedings
of the 5lh International Conference on Indoor Air Quality and Climate: Indoor Air '90. Vol 2. International Conference on Indoor Air Quality
and Climate, Ottawa, 1990, pp 495 - 500.
6.	U.S. Environmental Protection Agency: "Underground storage tank corrective action technologies.* EPA/625/6-87-015, Washington, D.C.
(1987).
7.	T. J. Buckley, A.B. Lindnrom, V.R. High smith, W. Becbtold, E. Pellizzari, L. Sheldon, and K. Thomas, "The time-coune and sensitivity of
muconic acid ai a biomarker for human environmental exposure to benzene," These Proceedings.
8.	W.T. Winberry, L. Forehand, N.T. Murphy, A. Ceroli, B. Fbinney and A. Evans, Compendium of Methods for the Determination of Air
Pollutants in Indoor Air. EPA-600/S4-90/010, U.S. Environmental Protection Agency, Reaearch Triangle Park, 1990.
9.	A.C. Guyton, Basic Human Physiology: Normal Function and Mechanism of Disease." 2nd ed., W.B. Saunders Company, Philadelphia, 1977
pp 393-438.
10.	H.S. Brown, D.R. Bishop, and C.A. Rowan, "The role of akin absorption as a route of exposure for volatile organic compounds (VOCs) in
drinking water," Am. J. Public Health 74(5):479-484 (1985).
11.	I.H. Blank and DJ. McAulifTe,	nf K>irrAtw thmtiyh human itin." J. Invest. Dermatol 85:522-526(1985).
12.	L. Wallace, "Major aourcea of Exposure to benzene and other volatile organic chemicals," Risk Anal. 10:59-64 (1990).
13.	L. Wallace, E. Pelizzari, T. Hartwell, K. Pcrritt, and R. Ziegenfus, 'Exposures to benzene and other volatile organic compounds from active
and passive smoking," Arch. Environ. Health 42: 272-279 (1987).
14.
SCAQMD, "In-vehicle air toxics characterization study in the south coast air basin," South Coast Air Quality Management District, El Monte,
CA, May, 1989.

-------
Figure 1. Portable GC Data
June 12,1991
1,000
Shower
End
600
CO
O) 600
Bathroom Door Open
400
200

o
18 20
25.5 30
0
10
48
Time (minutes)
Shower Bathroom Bedroom Living Room
—A			G	
Figure 2. Tenax/Summa Comparison
1,000
CO
E 500
"o>
3 300
-------