EPA/600/A-92/235
MULTIZONAL MASS BALANCE MODELING OF BENZENE
DISPERSION IN A PRIVATE RESIDENCE
Azzedine Lansari
Environmental Information Technology Services, Computer Sciences Corporation, Research Triangle Park, NC
27709
Andrew B. Lindstrom
Human Exposure and Field Research Division, Atmospheric Research and Exposure Assessment Laboratory,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Brian D. Templeman* and John S. Irwin*
Atmospheric Sciences Modeling Division, Air Resources Laboratory, National Oceanic and Atmospheric
Administration, Research Triangle Park, NC 27711
ABSTRACT
A residence in Roxboro, NC, was found to have its well-water supply contaminated with benzene (» 300 //g/I) and
other organic compounds. The residents of the house do not currently drink the water, but they use it for daily
showers. A study was designed to monitor and model short-term benzene dispersion within the house during and
after a shower.
A multizonal mass balance model, CONTAM88, was used to predict interzonal air flow rates and benzene
concentration distributions within the house. The idealization of the building was created using NBSAVIS, a
preprocessor to CONTAM88. Simulation results showed that the highest concentration occurred in the shower stall.
During the shower, the master bathroom concentration was less than half the shower-stall concentration. Benzene
concentrations in the master bedroom and other rooms were lower. Simulated benzene concentration distributions
showed that benzene from the shower rapidly dispersed in the house, and reached equilibrium in all the rooms in
less than 30 minutes after the shower. These results were supported by SF6 experimental data.
Benzene samples were collected using glass, gas-tight syringes in the shower stall and at various locations in the
house. The average benzene concentration after a 20-minute shower was 978 /jg/m3 in the shower stall, 263 /ig/m3
in the master bathroom, and 70 //g/m3 in the master bedroom. Simulated and average measured benzene
concentrations yielded a similar behavioral trend. It was concluded that multizonal mass balance models may be
useful in designing field study monitoring strategies.
Keyword index: Indoor air, multizonal models, mass balance models, contaminant dispersion, benzene, shower, SF6.
INTRODUCTION
Benzene has the largest production volume of any chemical that has been causally linked to cancer in humans."1
It is a pollutant that is spread in the environment from sources such as tobacco smoke, automobile refueling, and
industrial waste.(:'3-4' Residential use of benzene-contaminated water may result in significant inhalation, ingestion,
and dermal exposures.'51
Previous investigations have shown that trichloroethylene (TCE) contaminated water supply may constitute a
significant point source of human exposure for the bather, and a dispersed source for other inhabitants in the home.
For highly volatile chemicals, inhalation exposures have the potential to be equal or greater than those associated
with direct ingestion of water."¦•7' For households using tap water contaminated with TCE, inhalation exposures in
showers could be as large as or larger than a conservative estimate of ingestion exposure. The assumption that a
70-kg adult consumes 2 liters (1) per day of tap water was considered a conservative estimate of ingestion
exposure.'81
On assignment to the Atmospheric Research and Exposure Assessment Laboratory, U.S. Environmental Protection Agency.

-------
In 1985, a residence in Roxboro, NC, was identified as using ground water contaminated with benzene, xylene, and
other organic compounds. The benzene contamination has been characterized by measurements of 7 figl\, 32 fig/l,
and 445 txgll in 1986, 1989, and 1990, respectively. The homeowners have continued to reside in the house and
use the water for all their normal purposes except drinking and cooking. In 1991, the U.S. Environmental
Protection Agency (EPA) conducted a series of tests to assess shower-related exposures that occur throughout the
house during and after a single 20-minute shower, to determine the relationship between various monitoring
techniques and to assess the usefulness of multizonal mass balance models during experimental study designs.
STUDY OBJECTIVES
The objectives of this study are to (1) investigate the possibility of using multizonal mass balance models to predict
locations where benzene concentrations are significantly different from background concentrations in order to
optimize sampling times and locations during a field study design, and (2) test the model performance in a well-
defined microenvironment.
THE MULTIZONAL MASS BALANCE MODEL, CONTAM88
The multizonal mass balance model used in this investigation is the National Institute of Standards and Technology
(N1ST) model, NBSAVIS/CONTAM88191 developed for EPA, to simulate transient contaminant concentration
distribution in buildings. The model is based on the element-assembly approach, which assumes that a building can
be represented as a combination of well-mixed zones linked by flow and kinetic elements (contaminant mass
transport and decay). CONTAM88 solves a set of mass balance and flow equations. The mathematical formulation
of the contaminant concentration is:
[W]C * [M]f = G
dt
where; C	=	vector containing the discrete concentration values
| W]	=	system mass transport matrix containing flow rate data
[MJ	=	system matrix containing mass (volume) data
G	=	system generation vector containing kinetics data.
NBSAVIS is a preprocessor to CONTAM88 that allows the idealization of the building through the generation of
a file thai describes the building configuration, including indoor and outdoor contaminant sources. Data input to
NBSAVIS are controlled by a series of screen-fill subroutines, which allow the user to specify interior and exterior
wall types, interior and exterior doors, windows, open passageways, filters and fans, room descriptions, and HVAC
system descriptions.
RESIDENCE DESCRIPTION AND IDEALIZATION
The private residence, which is located in a rural area in Roxboro, NC, is a single-story house. The house has
three bedrooms, a bathroom, a family room, a laundry room, and an open area that consists of a living room,
kitchen, and dining room (Figure 1). The master bedroom area includes a bathroom with a separate shower. The
house also has a full basement, an attic, and a carport. The residents of the house get their water from a nearby
well, located south of the residence.
The NBSAVIS preprocessor was used to build the idealization of the house. The parameters of the house that were
measured to run NBSAVIS are as follows:
Physical dimensions (including all windows, doors, and other openings),
HVAC system output, including locations of all vents and the associated air flow rates,
Contaminant source information (name, molecular weight, emission rate),
Source locations (outside or inside, particular rooms of the house), and
Local meteorological conditions (temperature, wind speed, and wind direction).
L.

-------
RESULTS AND DISCUSSION
Air flow rates from all the vents of the HVAC system were measured using an Omega HH-30 vane anemometer;
the HVAC return flow rate (Table 1) was measured using a Shortridge Instruments Flow Hood. Constant
meteorological conditions were assumed, because the duration of the simulations did not exceed 4 hours. The
estimated local meteorological conditions were: 2 m/s wind speed, 220° wind direction, and 25 °C temperature.
In the first stage of the study, a 15-minute shower was simulated—with water temperature of approximately 40 °C,
at a flow rate of 10 1/minute. The contaminant, benzene, was modeled as a point source located in the shower stall.
The most recent benzene-in-water concentration (445 /xg/1 measured in the house in 1990) and a 61 % transfer
efficiency of TCE from shower water-to-air (from McKone and Knezovich'8') were used to estimate the benzene
emission rate—45 /xg/s. Also, sulfur hexafluoride (SF6) was released in the shower for 15 minutes and its dispersion
was monitored throughout the residence. Syringe samplers were placed in all rooms of the house (one sampler per
room, in a location not exposed to direct air flow from the vents) to monitor concentration gradients.
Each room in the house was considered as one zone, except the master bathroom which was considered as two zones
because it has a separate shower stall. During the entire testing period, the HVAC system was running (only fan
on). The ceiling fans were also running at their lowest speed to allow constant contaminant mixing without
disturbing interzonal air flows. Using the above conditions, benzene dispersion throughout the house during and
after the shower was simulated.
Simulation results for the first stage of this study showed that the highest benzene concentrations occurred in the
shower and master bathroom, then in decreasing concentrations, in the master bedroom and hallway. The living
room, dining room, kitchen, and family room had lower concentrations. These results were supported by the SF6
experimental data (Figure 2). Modeled benzene concentration distributions showed that benzene rapidly dispersed
in the house, and all rooms in the house reached equilibrium within 30 minutes after the shower (Figures 3a and
3b). Therefore, a total sampling time of about 50 minutes may be chosen. After that time, simulated concentrations
of about 50 /xg/m3 were found in the house.
In the second stage of the study, the living room, dining room, kitchen, and family room were considered as one
well-mixed zone. Furthermore, the total simulation time was 50 minutes. The shower was run for 20 minutes with
the bathroom door closed. After the shower, the shower-stall door was open and the bathroom door was kept closed
for 5 minutes to allow for the individual to dry off and get dressed. After that time, the shower-stall and bathroom
doors were opened. The average measured shower water flow rate was about 6.3 I/min, and the average waterborne
benzene concentration from the pre-shower head samples was 292 /xg/1. Waterborne benzene concentrations from
the pre-shower head samples and the drain-level samples were measured and used to calculated the water-to-air
transfer efficiency. The average calculated benzene transfer efficiency was 88% yielding a benzene emission rate
of 27.5 /xg/s. Using the above conditions, benzene dispersion throughout the house during and after the shower was
simulated. Simulation results of benzene dispersion showed a benzene concentration of 625 /xg/m3 in the shower
stall after a 20-minute shower, 278 /xg/m3 in the master bathroom, and 148 /xg/m3 in the master bedroom. The rest
of the rooms in the house had concentrations of less than 40 /xg/m3.
Benzene concentration levels were measured during a 3-day, 3-shower period (i.e., 1 shower each day).00' Glass,
gas-tight syringe samplers were placed in the shower stall, bathroom, master bedroom, and living room. The total
sampling time was 120 minutes. After 20 minutes, the shower-stall concentration reached an average value of
978 /xg/m3 (standard deviation, SD, equal to 514 /xg/m3), the master bathroom concentration reached 263 /xg/m3
(SD = 64 /xg/m3), the master bedroom concentration reached 70 /xg/m3 (SD = 14 /xg/m3), and the living room
concentration reached 40 /xg/m3 (SD = 16 /xg/m3). Figures 4 and 5 show the simulated and measured benzene
concentrations in the shower stall and the master bathroom, respectively.
During the shower, there was significant variability in the data, which may be due to incomplete mixing, dynamic
variation in the benzene-in-water concentration, and experimental errors. The benzene concentration during the third
shower was much higher than during the first two showers. This difference may be due to variability in water flow,
as well as sampling inaccuracies due to incomplete mixing. Differences between simulated and measured
concentrations may be due to model limitations. For instance, the assumption of a well-mixed zone may be too
3

-------
simplistic. Also, the assumptions and parameter estimation used in the idealization of the house may constitute a
significant source of uncertainty. Overall, the model did well in predicting the zones of significantly different
concentrations and the time necessary for the contaminant to reach equilibrium throughout the house.
CONCLUSIONS
In the first stage, CONTAM88 was used to plan the study design. SF6 was used to measure flow rates within the
house. Modeled benzene concentrations in the shower were more than twice the master bathroom's concentration
during the shower. After the shower and opening the shower door, benzene quickly dispersed in the house.
Concentration equilibrium was reached within 30 minutes. This result suggests that a total sampling period of less
than 50 minutes would be appropriate for this type of study. Simulation results also showed that the living room,
dining room, kitchen, and family room have similar concentrations. Therefore, they were grouped into one zone.
The SF6 experimental analysis yielded similar results.
In the second stage of the study, benzene concentrations were simulated and measured in the shower, master
bathroom, master bedroom, and living room. The average measured shower benzene concentrations were about
40% higher than the simulated ones; the simulated master bathroom concentrations were about 6% higher than the
measured ones, and the simulated master bedroom concentrations were about 100% higher than the average
measured one. The simulated concentrations in the rest of the rooms were about 20% lower than the measured
ones. Therefore, CONTAM88 may only be used to simulate broad trends of concentration distribution throughout
the house. Using CONTAM88 for the exposure assessment suggested that a 1-hour sampling time should be
appropriate for a 20-minute shower. The model also helped in deciding the rooms in which to locate the samplers,
to monitor benzene concentration distribution. Simulation results will hopefully help investigators plan field studies
and minimize the cost of the studies.
ACKNOWLEDGEMENTS
The authors wish to acknowledge and thank Mark Johnson and David Proffitt for the air exchange and SF6 data, used during the preliminary
stage of the study. The authors also thank Larry Michael for the benzene data. This work was sponsored by the Indoor Air Research Section,
U.S. Environmental Protection Agency.
DISCLAIMER
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's peer 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.
REFERENCES
1.	EPA. 1984. National emission standards for hazardous air pol!utants:Regulation of benzene. Federal Regisler. 49(110~):23.478-23.495.
2.	Fishbein, L. An overview of environmental and toxicological aspects of aromatic hydrocarbons. I. Benzene. Sci. Total Environ.. 40. 189-
218. 1984.
3.	International Agency for Research on Cancer (IARC). IARC Monograph on the Evaluation of the Carcinogenic Risk of Chemicals and
Dyestuffs. Volume 29, pp. 93-148. Lyon, France. 1982.
4.	Webster, R.C., Maibach, H.I., Gruenke, L.D., and Craig, J.C. Benzene levels in ambient air and breath of smokers and nonsmokers in
urban and pristine environments. I. Toxicol. Environ. Health. 18:567-573. 1986.
5.	Shehata. A. T. A multi-route exposure assessment of chemically contaminated drinking water. Toxicology and Industrial Health.
l(4):277-298. 1985.
6.	Andelman, J. B., A. Couch, and W. W. Thurston. Inhalation exposures in indoor air to irichloroethylene from shower water.
Environmental Epidemiology, pp. 201-213. 1991.
7.	Andelman, J. B. Inhalation exposure in the home to volatile organic contaminants of drinking water. Science Total Enviro.. 47:443-460.
1985.
8.	McKone, T.E., and J.P. Knezovich. The transfer of irichloroethylene (TCE) from a shower to indoor air: experimental measurements
and their implication. J. Air & Waste Mgmt. Assoc.. 41:832-837. 1991.
9.	Grot, R.A. User's Manual NBSAVIS/CONTAM88. A user interface for air movement and contaminant dispersal analysis in multizone
buildings. National Institute of Standards and Technology, Gaithersburg, MD. 1991.
10.	Michael, L. C. VOC support to the Roxboro, NC, benzene investigation. Research Triangle Institute. RTI/4657-07A/02F. 1991.

-------
I
ROXOOnOHOUStSfUUY POUU'AKT EXN4.Nl
B.. I ^
I
JUr
UO • Mttlt'SW
fi • RHwn
S • S»M(f
« • WleJow
v • Vent
Figure 1. Roxboro house. Pollutant benzene.
WCASURED SF6 CONCENTRATIONS
(11 mm shower)
Moath
Mbcd
Kitchen
homily
Time (rvn)
Figure 2. Measured SF6 concentration distribution
within the house. During and after a 15 min
shower.

-------
Table i. Air flow measurements from
the Roxboro house."
Room-vent f
Air flow
(m'/min) j
Kitchen
3.77 1
Laundry room
3.10 1
Dining room-1
2.78
Dining roon-2
1.96
Family room-1
2.96
Family room-2
4.02
Bedroom 1-1
1.31
Bedroom 1-2
0.81
Bedroom 2
1.32
Bathroom
4.28
Master bedroom-1
3.02
Master bedroom-2
3.22
Master bathroom
7.14
Living room-l
3.00
Living roon-2
3.29
Basement
5.92
Return
44.53
ROOBOOQ MOUSE STWY - POU.UTAWT BEKZEKC
W.V V	V «	v	w
Matter
w.V
8«*
J
MB
Ob
LM»q
Rmm

V
Bcri'MMI
1
BrArvaai
t
Rmm
Giefct i
Oialfl|
Him
\
Uw^ty

H* r
C*P«rt
T
i»r
-21' T
R	•Rtlim
%	- SW~rf
V	•Whiilw
v	- Vcwt
Figure 1. Schematic of the Roxboro house.
MEASURED SF6 CONCENTRATIONS
(15 min shower)
Mbath
Mbed
Hall
Living
Kitchen
Fomily
5 10 15 20 25 JO 35 40 45 50 55 60
Time (min)
figure 2. Measured SF6 concentration distribution
within the house. During and after a 15-min.
shower.

-------
BENZENE SIMULATED CONCENTRATIONS
(t 5 mm tsho*er, *5 ug/s emission rote)
—	Shower
•— M&ath
Wbed
Hal!
-	- Uv_Dm
1000
900
800
700
600
500
400
300
200
100
20 25 30
Time (min)
Figure 3a. Simulated benzene concentration
distribution during and after a 15-min. shower, for
the rooms of highest concentration levels.
BENZENE SIMULATED CONCENTRATIONS
(15 min shower, 45 ug/s emission rcte)
liv_Din
Laundry
Bcsement
10 15 20 25 30 35 40 45 50
Time (min)
SIMUIATE0 /MEASURED BENZENE CONCENTRATIONS
(20 min shower, 27 ug/s emission rate)
1 SCO
£ 1400
X
Q>
3 1200
800
o S00
400
2QQ
0
0
5
5 20 25 30 35 40 45 50
0
Time (min)
Figure 4. Simulated (S) and measured (M) benzene
transient concentration in the shower stall, during a
3-day, 3-shower experiment.
Figure 3b. Simulated benzene concentration
distribution during and after a 15-min. shower, for
the rest of the house.
SIMULATED/MEASURED BENZENE CONCENTRATIONS
(20 min shower, 27 uq/s emission rate)
mo
*
	 S-Mboth

• M-Mfcothl
450

» M-Mfccth2
400

* W-Mbctn3

350
• . ¦


¦

300
¦ r
	 ¦

250
-	" \ *

200
: X *\ •
/ ~ \

150
¦ / X *

100
" / r

50

• I
n
L i i . i . i — >-.&-» i ¦ .
, ,
0 5 10 15 20 25 30 35 • 40 45 50
Time (min)
Figure 5. Simulated (S) and measured (M) benzene
transient concentration in the master bathroom,
during a 3-day, 3-shower experiment.
7

-------
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-92/235
2.
3.
4. TITLE AND SUBTITLE
Multizonal mass balance modeling of benzene
dispersion in a private residence.
5,REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTBQRCS)
Azzedine Lansari, CSC
Andrew Lindstrom, HERB/AREAL
Brian Templeman, HEMB/AREAL
John S. Irwin, HEMB/AREAL
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Computer Science Corporation
Research Triangle Park, NC 27711
10.PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
No. 68-WO-0043
12. SPONSORING AGENCY NAME AND ADDRESS
Atmospheric Research and Exposure Assessment
Laboratory - - RTP
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
1*. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
Proceedings of the 1992 U.S. EPA/A&WMA Symposium on Measurement
of Toxic and Related Air Pollutants, Durham, NC. May 1992.
16. abstract
A residence in Roxboro, NC, was found to have its well-water supply contaminated with Benzene (*= 300^1) and other organic compounds. The
residents of the house do not currently drink the water, but they use it for daily showers. A study was designed to monitor and model short-term
benzene dispersion within the house during and after a shower.
A multizonal mass balance model, CONTAM88, was used to predict interzonal air flow rates and benzene concentration distributions within the
house. The idealization of the building was created using NBSAVIS, a preprocessor to CONTAM88. Simulation results showed that the highest
concentration occurred in the shower stall. During the shower, the master bathroom concentration was less than half the shower-stall concentration.
Benzene concentrations in the master bedroom and other rooms were lower. Simulated benzene concentration distributions showed that benzene from
the shower rapidly dispersed in the house, and reached equilibrium in all the rooms in less than 30 minutes after the shower. These results were
supported by SF6 experimental data.
Benzene samples were collected using glass, gas-tight syringes in the shower Mall and at various locations in the house. The average benzene
concentrations after a 20-minute shower were 978 n%!ve? in the shower stall, 263 ^g/m3 in the master bathroom, and 70 jig/m3 in the master
bedroom. Simulated and average measured benzene concentration " :lded a similar behavioral trend. It was concluded that multizonal mass balance
models may be useful in designing field study monitoring strsteg. .
17. KEY WORDS AND DOCUMENT ANALYSIS
». DESCRDTOI
(..IDENTIFIERS/ OPEN ENDED TERMS
c.COSATI



18. DISTRIBUTION STATEMENT
Release to Public reproduced by
U.S DEPARTMENT OF COMMERCE
« NATIONAL TECHNICAL INFORMATION SERVICE
SPRINGFIELD. VA 22161
' 1 »
19. SECURITY CLASS (TO. R«>c«)
UNCLASSIFIED
21 .NO. OF PACES
,8
». SECURITY CLASS (Tbii hte)
UNCLASSIFIED
22. PRICE .

-------