PB91-196493
Interaction of Vapour phase Organic Compounds with Indoor Sinks
Acurex Corp., Research Triangle Park, NC
 Prepared for:
 Environmental  Protection Agency,  Research Triangle  Park,  NC
 1991

-------
 AEERL-P-661
I. REPORT NO.
  EPA/600/J-qynfiQ
                                TECHNICAL REPORT DATA
                         (Please rsad fnsimc lions on the reverse before com pie ft
4. TITLE AND SUBTITLE
 The Interaction of Vapor Phase Organic Compounds
 with Indoor Sinks
                                                      5. REPORT DATE
                                                      6. PERFORMING ORGANIZATION CODE
7.AUTHon(s)B.Tichenor (EPA);
 (Univ.  of AK),  L. Sparks (E
 (Acurex)
                          ,_ Z. Guo (A cur ex), J. Dunn
                          EPA),  and M. Mason
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Acurex Corporation
 P. O. Box 13109
 Research Triangle Park, North Carolina  27709
                                                      10. PROGRAM ELEMENT NO.
                                                      11. CONTRACT/GRANT NO,
                                                       68-C2-4701 (Acurex)
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Air and Energy Engineering Research Laboratory
 Research Triangle Park, NC 27711
                                                      13. TYPE OF REPORT MND PERIOD COVERED
                                                       Journal article; 6/89-2/90
                                                      14. SPONSORING AGENCY CODE
                                                        EPA/600/13
is. SUPPLEMENTARY NOTES AEERL project officer is Bruce A.
2991. Indoor Air, 1, 23-35 fl99l).
                                                      Tichenor, MD-54. 919/541-
te. ABSTHACTThe paper reports on the development of sink models based on fundamental
mass transfer theory. The interaction of indoor air pollutants with interior surfaces
(i. e.,  sinks) is a well known, but poorly understood,  phenomenon. Studies have
shown that re-emissions of adsorbed organic vapors can contribute to elevated con-
centrations of organics in indoor environments.  Research is being conducted in
small environmental test chambers to develop data for predicting sink behavior.
The results  of experiments conducted to determine the magnitude and rate of adsorp-
tion and desorption of vapor phase organic compounds for several materials are
presented. Five materials were evaluated: carpet, painted wallboard, ceiling tile,
window glass, and upholstery. Two organic compounds were tested with each mater-
ial:  tetrachloroethylene (a common cleaning  solvent) and ethylbenzene (a common
constituent of petroleum  based solvents widely used in consumer products). Results
of the experimental work show the relevant sink effect parameters for each material
tested and compare the sorptive behavior of  the two organic compounds evaluated.
An indoor air quality model was modified to  incorporate adsorption and desorption
sink rates. The model was used to predict the temporal history of the concentration
of total vapor phase organics in a test house after applying a wood finishing product.
7.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                          b.iDENTIFIERS/OPEN ENDED TERMS
                                                                   c.  COSATI Field/Group
Pollution
Mathematical Models
Mass Trant'er
Organic Compounds
Vapors
Test Chambers
                                          Pollution Control
                                          Stationary Sources
                                          Indoor Air
                                          Sinks
             13B
             12A
             14G
             07C
             07D
             14B
P. DISTRIBUTION STATEMENT

Release to Public
                                          19. SECURITY CLASS (This Report}"
                                          Unclassified
                                          2O. SECURITY CLASS (This page)
                                          Unclassified

-------
                                                                            EPA/600/J-91/069

                                                                                      PB91-196493
                                                                     Indoor Air. I. 23-35 (1991)

                                                      1991 Danish Technical Presi DK-Copenhogen
                                                                        imic
The Interaction of Vap
Compounds with Indoor Sinks
Bruce A. Tiehenor1, Zhishi Guo\ James E. Dunn3, Leslie E. Sparks'
and Mark A. Mason2
  Abstract
  *The interaction of indoor air pollutants with ulterior
  surfaces (i.e., sinks) is a well known, but poorly under-
  stood, phenomenon. Studies have shown that re-emis-
  sions of adsorbed organic vapours can contribute to ele-
  vated concentrations of organic* in indoor environ-
  ments. Research is being conducted in small environ-
  mental test chambers to develop data for predicting
  sinfe behaviour. Tliis paper reports on the development
  of sink models based on fundamental mass transfer
  theory. The results of experiments conducted to deter-
  mine the magnitude and rate of adsorption and desarp-
  tion of vapour phase organic compounds for several
  materials are presented.  Five materials were evaluated:
  carpet, painted waliboard, ceiling tile, window glass,
  and upholstery. Two organic compounds were tested
  with each material: tetrachloroethylene (a common
  cleaning solwni) >'nd ethyibenzene (a common consti-
  tuent ofpetrakum-oased solvents widely used in con-
  sumer products). The results of the experimental work
  are presented showing the relevant sink effect para-
  meter for each material tested and comparing the sorp-
  tax behaomiT of the mo organic compounds evaluat-
  ed. An indoor air quality (IAQ) model was modified
  to incorporate adsorption and desorption sink rates.
  The model WK used to predkt the temporal history of
   Manuscript received: 10 April 1990
   Accepted for publication: 24 September 1990	

   1  U.S. Environmental Protection Agency
     Air and Energy Engineering Research Laboratory
     Indoor Air Branch (MD-54)
     Research Triangle Park, NC 27711, USA
   '  Acurex Corporation
     Environmental Systems Division
     Research Triangle Park, NC 27709, USA
   '  University of Arkansas
     Department of Mathematical Sciences
     Fayeiteville, AR 72701, USA
                                               the concentration of total vapour phase organia in a
                                               test house after application of a wood finishing pro-
                                               duct. The predicted results are presented and compared
                                               to measured values. Suggestions for further research on
                                               indoor sinks are presented.
                                               KEYWORDS:
                                               Indoor sinks, Vapour phase organics, Adsorption,
                                               Desorption, Mass transfer, IAQ models
                                               Introduction
                                               The interaction of indoor air pollutants with
                                               interior surfaces (i.e., sinks) is a well known
                                               phenomenon. While it  is possible  that the
                                               mass transfer mechanism could be absorp-
                                               tion (i.e., for a water soluble compound to a
                                               moist surface),  this paper assumes  that the
                                               mechanism is adsorption. Researchers have
                                               developed models and data to deal with the
                                               adsorption of particles and NOj to a variety
                                               of indoor surfaces (Nazaroff and Cass,  1989;
                                               Yanagisawa et  al., 19S7) and, for NO2, the
                                               generation of NO from the surface has also
                                               been reported (Billick and Nagda, 1987; Spi-
                                               cer et aL3 1989). Berglund et al. (1988) pres-
                                               ented an extensive review of the literature on
                                               adsorption and desorption of vapour  phase
                                               organic compounds in indoor environments,
                                                including data on a ventilation system heat
                                                exchanger. The same authors reported that
                                                organic  compounds  adsorbed on  building
                                                materials in a  7-year-old preschool were re-
                                                emitted  (i.e.,  desorbed) when  placed in  a
                                                small environmental test chamber ventilated
IfTBli MtitorizM H raaraduet md nil tMi
 nMrt PtmltiKNi 'ir htOar ra pradiKllon
 Btiri ht MHiimd Iran DM copyright owner.

-------
24  8. Tichenor el a!.; VOC Interactions with Indoor Sinks
 with clean air (Berglund et al., 1989). Zsol-
 nay el al. (1987) reported on the transport of
 Hndane vapours between  treated  wood  and
 cotton cloth. Gebefugi and Korte (1989) ex-
 tended this research to include other fungi-
 cides and  additional textile materials. The
 Danish Town Hall Study provided empirical
 evidence that re-emission  of  indoor pollu-
 taots, including vapour phase organies, from
 fleecy (i.e., large surface area) materials may
 contribute  to the  Sick Building  Syndrome
 (Nielsen, 1987; Nielser., 1988; Valbjorn  and
 Skov,  1987). Seifen and Schmahl (1987) re-
 ported on research to determine the adsorp-
 tion and desorption behaviour of several or-
 ganic compounds and several surface mater-
 ials, but concluded "... no rule can be de-
 disced to predict the sorption behaviour of
 different VOC  with  different  materials".
 Dunn and Tichenor (1988) presented data
 showing the effect of sinks in small environ-
 mental test  chambers used to  determine
 emission rates of organic compounds for la-
 tex caulk and moth repellant, and they illu-
 strated the use of adsorption and desorption
 rate constants in mass balance models- Ti-
 chenor et al. (1988) reported on studies in an
 IAQ test house where para-dkhlorobenzene
 was measured in significant concentrations
 several days after the source (moth repellant)
 was removed. Measurements in the same test
 house, taken 12  months  after  additional
 moth repellant emission  tests, showed con-
 centrations  of  para-dichlorobenzene  well
 above the outdoor concentration,  indicating
 the  re-emission   of para-dkhlorobenzene
 previously adsorbed on interior surfaces.
  In view of the significant impact of sinks
 on indoor concentrations of vapour phase or-
 ganic compounds,  research is being conduc-
 ted to develop data for predicting sink beha-
 viour.  This paper reports  on  the develop-
 ment of sink models and the  results of ex-
periments conducted to determine the mag-
 nitude and  rate of adsorption and desorption
of vapour phase organic compounds for sev-
eral materials. In  addition, an IAQ model
was modified to incorporate adsorption and
desorption sink rates, and  the model was
used to predict the temporal history of the
conce  .ration of total vapour phase organics
in a test house after application of a  wood
finishing product. The concentrations pre-
dicted by the model are presented and com-
pared to measured values. Finally, sugges-
tions for further research on indoor sinks are
presented.


Theory
Adsorption Isotherms
The study  of adsorptive/desorptive  beha-
viour of gas molecules with solid surfaces has
resulted in  several well known  adsorption
isotherm  equations  (Daniels and Alberty,
1961).
  The Langmuir isotherm assumes a mono-
layer of molecules on a homogeneous surface
with all adsorption sites mutually indepen-
dent and identical (i.e., all sites require the
same energy to adsorb molecules). At equi-
librium, the Langmuir isotherm can be rep-
resented by:
k,Q(l-8) = k/6
(1)
where ka is the adsorption rate constant; Ce is
the equilibrium concentration of die adsor-
bate in the gas phase; k/ is a desorption con-
stant; and 8 is the proportion of available ad-
sorption sites occupied. For the low vapour
concentrations of single compounds used in
the experiments  discussed below,  it can be
assumed  that the occupied sites are a very
small fraction of the available sites (ie., 8
< < 1). Thus, equation (1) can be simplified
to a linear form:
     = k/0
(2)
Since the equilibrium mass per unit area on
the sink, M*, is directly proportional to 8, it
follows that:

k,Cc = k,]Ms                           (3)

where kd is the desorption rate constant. For

-------
                                          B. Tichenor el ol.: VOC Interactions wilh Indoor Sinks  25
Moss {*}
to chamber
Chamber
k'J

Sink
(x-y-w)
f.
(w)

                                                              Fig. 1 T\ist chamber schernolsc
                                                              with sink: mass balance equa-
                                                              tions
               Exit (y)
   x = Mass to  chamber  over time  t {dx/dt = R(t)]
   y = Mass from chamber over time t [dy/dt =kz (x-y-w)}
  w = Moss in the sink [dw/dt = k3(x—y—w)—k4w]
a sink with an area of A, the mass balance
equation is:

                                      (4)
Also, an equilibrium constant, k,., can be de-
fined as:
          = MJCC
(5)
The Freundlich isotherm, originally  deve-
loped as an empirical fit  to non-Langmuir
isotherms, was later derived by assuming an
exponential  distribution of  adsorption site
energies. The Freundlich isotherm can be
represented by:
 m = kC
(6)
 where m is the mass adsorbed per unit of ad-
 sorbent and k and a are empirical constants.
   The  BET (Brunauer-Emmett-Teller) iso-
 therm provides for multilayer adsorption.
 For the low  vapour concentrations encoun-
 tered in indoor environments, it is assumed
 that the available adsorption sites will not be
 covered, so the BET isotherm is probably
 not appropriate.

 Application to Chamber Testing
 Dunn and Tichenor (1988) proposed a set of
 ordinary  differential equations to describe
 the mass balance m a test chamber using the
 schematic shown in Figure 1:
      dx/dt = R(t)

      dy/dl = k2(x-y-w)
      dw/dt = k3(x-y-w) -
                                     (7)
                                     (8)
                                     (9)
where x is the mass that enters the chamber
over time t;  y is  the  mass that exits the
chamber over time  t; w is the mass in the
sinks R(0 is the *nPut rate to *e chamber; k2
is the air exchange rate, N; k3 is the rate con-
stant to the sink; and kj is the rate constant
from the sink.
  Expressing the sink term in a form consis-
tent with the linear  form of Langmuir's
equation gives:
dw/dt = k,CV-k4MA                (10)
where  the concentration in the chamber, C
= (x-y-w)/V; V is the chamber volume; M is
the mass per  unit area in the sink; and A is
the sink area, (Note that w = MA.)
  At equilibrium, dw/dt = 0, Thus,
k3CV =  k
-------
26  B. Tichenor et a!.: VOC Interactions with Indoor Sinks
Substitution of equations (7), (8), and (10)
and replacing ka and k4 yield:

VdC/dt = RdHsfCV-k.CA + kjMA    (13)

or
dC/dt =
   R(t)/V-NC-k4CA/V + k.,M A/V      (14)

Substituting L = A/V in equation (14) gives:

dC/dt =  P.(t)/V - NC - kaCL + kdML  (15)

Appropriate units for equation (15) are:

C   = concentration, mg/mj
t    = time, h
R(t) = emission rate of source, mg/h
V   = chamber volume, m3
N   = air exchange rate, h"1
ka   = adsorption rate, m/h
L   = ratio of sink area to chamber volume,
      m"1
kd   = desorption rate, h'
M   = mass per unit area in sink, mg/m2

Substituting for k3 and  ki  in equation (Ib;
yields:
AdM/dt = Ak,C-AkdM

or simply

dM/dt = kaC
(16)
(17)
Thus, if Langmuir adsorption processes are
assumed, equations (15) aiid (17) describe
the rate of change of concentration and mass
in the sink, respectively, for small chambers.


Experimental Methods
Experiments to determine adsorption and
desorption rates for vapour  phase  organic
compounds were  conducted in small (53
litre) environmental test chambers (Tichenor
et al.31988; Tichenor, 1989).
Test Materials and Compounds
Five sink materials were evaluated: (a) carpet
- nylon fibre pile bound with a styrenebma-
diene latex to a jute  backing; (b) wailboard
(gypsum drywaU) - finished with an interior
latex paint;  (c) ceiling tile -  1.5 cm thick
composite material  (mineral  fibre,  starch,
clay, gypsum perlite paper fibre, silicate); (d)
window glass; and (e) upholstery - throw pil-
low  with  50% polyester/50% cotton fabric
covering a polyester fibre fill. Each material
was tested with two organic compounds: (a)
tetrachloroethylene (perchloroethylene) - a
common cleaning solvent and (b) ethylben-
zene - a common constituent of petroleum-
based solvents widely used in consumer pro-
ducts.

Test Procedures
A sample of the sink material was placed in a
test chamber supplied with clean, humidified
air (45% RH, 23 °C) at a nominal rate of 1 air
change per hour (ACH). At the start of each
test, a portion of the inlet flow was replaced
by  flow from  a  constant temperature oven
that held a  permeation tube containing the
test compound. The  test compound was ad-
ded to the chamber inlet flow at a constant
rate for 48 hours, then the flow from the per-
meation oven was shut off and replaced by
flow from the clean air system.
  The concentration of the test compound
was monitored at the outlet of the chamber
by  gas chromatography. Compounds  were
identified  by retention time and quantified
by  electronic  integration   of  detector re-
sponse. Tetrachloroethylene was  quantified
using an electron capture  detector (ECD);
ethylbenzene was quantified using a flame
ionization detector (FID). Samples were col-
lected using gaslight syringes, automated gas
sampling valves equipped with fixed volume
sample loops, or  solid sorbent  sampling
tubes containing graphitized carbon adsor-
bents.  Syringe and sampling  loop samples
were injected directly to the packed column
of a gas chromatograph equipped with both

-------
                                          B. Tichcnor el ol: VOC Interactions with Indoor Sinks  27


*E
j[
c

b
u
u

0
O




9-
E -
7-
6-

5-
4 •


3 -


2-

1 -
0 •
Adsorption

icol "No Sink" Curve










sorpllan
v, 	
1 _*_ ' , — i r-f* ran 1i
                                                             Fig. 2 Chamber test (etbyiben-
                                                             zeno/eorpcl): adsorption and
                                                             desorplion represented by
                                                             crosshatciied areas
                20
40   15     60
  Time (hrs)
FID and ECD- Sorbent samples were therm-
ally  desorbed first to a cryogenic trap and
then to a gas chromatograph equipped with a
fused silica column and an FID-
  Samples were collected as frequently as
possible for the first few hours in order to
define the rising portion of the concentration
vs. time curve {see Figure 2). Once the con-
centration reached an apparent equilibrium
value, the sampling frequency was decreased.
Forty-eight hours after the  start of the test,
the injection of the organic compound was
terminated. At this point, the sampling fre-
quency was again increased. Sampling conti-
nued until  the  concentration reached the
quantification limit of the method.

Test Conditions
A wide variety of test conditions was used.
Initial testing was conducted on carpet with
tetrachloroethylene to investigate the effect
of concentration  and temperature on  the
sorption behaviour. Subsequent tests with
the  other sink materials and with ethylben-
zene %vere conducted under fewer conditions.
Tests  to  determine the sink behaviour of
empty chambers were also carried out Table 1
summarizes the conditions for all the tests.
AH  tests were conducted at a nominal 1 ACH
 and a  water vapour content  equivalent to
 45% relative humidity at 23 °C [Note that te-
                  trachloroethylene is designated as PCE (per-
                  chloroethylene) to  avoid  confusion with
                  TCE (trichloroethylene).]


                  Data Analysis
                  Data obtained from  the small chamber tests
                  were analysed assuming a Langmuir type ad-
                  sorption. These analyses yielded values for:
                  Mc, the mass per unit area on the sink at
                  equilibrium; ka, the adsorption rate constant;
                  kd, the desorption rate constant; and k,., the
                  equilibrium constant.
                  Determination of
                  Given the experimental procedures described
                  above, the mass in the sink is assumed to be
                  at equilibrium at the time (Q the flow of the
                  test compound is stopped (see Figure 2). The
                  total mass of the test compound exiting the
                  chamber from time i^ to the ?nd of the test is
                  the sum of (a) the mass of -ie compound in
                  the chamber air at time t> and (b) the mass of
                  the compound emitted from the sink. The
                  total mass exiting the chamber from t, to the
                  end of the test  is  obtained from  the  area
                  under  the curve (Ac) described by the  data
                  points multiplied by the air exchange  rate
                  (N) and the chamber volume (V).  In  this
                  case, the AC was calculated using the trape-
                  zoid rule. The mass of the compound in the

-------
28  B, Tichenor ct at.: VOC Inter actions with Indoor Sinks
Table 1 Tesl conditions for sink experiments
Material

Carpet


Wallboard

Ceiling
tile
Window
glass
Upholster}'
(pillow)
Empty
chamber

Tetrachloroeihylene
chamber air is the
don at equilibrium
lumck f\f\ tt^tlc*
Compound

PCE1
PCE
Ethylbenzene
PCE
Elhylbenzene
PCE
Elhylbenzene
PCE
Elhylbenzene
PCE
Ethylbenzene
PCE
PCE
Elhylbenzene
= Perchloroethylene (PCE)
Concentration
range (rog/m")
6-50
14-15
8-9
19-26
il-13
23-28
10-11
19
10-11
28-29
8-11
5-38
12-13
10-11

Temperature
<°C)
23
35
2?
23
23
23
23
23
23
23
23
23
35
23

No. of
tests
10
2
2
2
2
2
2
1
2
2
2
4
2
2

product of the concentra- where
(Ce) and the chamber vo

fN
— V" ~*~
icl +kVHiN + k_I +
^a1-* *SJ/ 11* «Vjij
. V.N. va
*SJ.T'»*"NJJ /">| \
    = (A^NV -
(18)
where M,. is the equilibrium mass per unit
area (mg/m2) in the sink at time t^.
Determination of k,, and kd
Equations (15) and (17) can be solved analy-
tically for certain simple cases. If the portion
of the concentration vs. time curve from I, to
the end of the curve is considered., the initial
conditions are: t = 0, QO) = C« and  M(0)
 = M. = C^ka/kd) [from equation (3)J. For
these conditions the solutions for equations
(15) and (17) are:
C(t) =
M(t) =
(20)
Values for k, and k,t were obtained by fitting
equation (19), where Q is a known experi-
mental value, to the chamber concentration
vs. time data from t, to the end of the test
using a nonlinear regression curve fit rou-
tine; in this case, NUN by SAS (1985). Fig-
ures 3 and 4 show the data and the fitted
Langtnuir adsorption curve for  two  of the
experimental  runs  (ethylbenzene/wallboard
and tetrachloroethylene/carpet); the theoreti-
cal "no-sink" curve, defined by C(t)=CceMl,
is also shown.


Results
The experimental results, including  values
for Q, Ma ka kd, and k« are presented in Ta-
bles 2-5. The relative standard deviations of
the estimates for ka and kd ranged from 7%
to 20%, with an average of 14%. Table 2 pro-
vides the results for carpet tested with tetra-

-------
                                                   B. Tkhenor cl ol.: VOC Interactions with Indoor Sinks
                                                  Data
                                           ongmur Adsorption

                                           No Sink" Curv«
Fig, 3 Chamber lesf (ethjrlben-
icnc/wollboordj: Langrnuir and
"no sink" predictions
                               B     !0    12
                                 Time 
-------
30   B. Tichenor ci ol.: VOC Inleioctions wilh Indoor Sinks
Table 3 Experimental results - letrochloroelhytcne ond woltboard, ceiling file, and pillow
Sink
material
Wallboard

Ceiling lile
Pillow

ACH
0.982
0.994
1.03
1.03
0.923
0,925
Q
(nig/m1)
19.2
25.8
22.6
28.4
28.4
28.6
Mt
(rng/ni1)
2.67
3.79
3.68
4.61
632
8.17
k,
(m/h)
0.22
0.20
0.09
0,11
0,02
0.03
kd

1.6
1,4
0.54
0.69
0.09
0.1 1
k,
(m)
0.14
0-15
0.16
O.i6
0.22
0.29
 1) The sample area (A) for the wallboard and ceiling tile was 0.14 m>; for the pillow, 0.266 m-'.
 2) The effective chamber volume (V) for the wallboard and ceiling die tests was 0.051 m1; for the pillow,
    0.046 m1.
 Table 4 Experimental results - elhylbenzene and carpet, wollboard, ceiling tile, ond pillow
Sink
material
Carpet
Wallboard
M
Ceiling tile
Pillow
ACH
0.945
0.982
0.975
0.949
0.971
0.988
0,956
0.991
c,
(mg/m5)
8.2
8.6
11.4
127
9.8
11.2
7.6
10.7
M.
(mgtar)
7.79
8.29
3.40
3.83
4.30
4.26
1.73
3.41
k,
(w/h)
0.07
0.08
0.58
032
0.25
0.23
0.004
0-005
K,
(h-)
0.08
0.08
1.9
1.1
0.58
0.60
0.018
0.015
k.
(m)
0.95
0.96
0.30
0.30
0.44
0.38
0.23
0.32
  f)° The sample area (A) for the carpet, wallboard, and ceiling tile was 014 m'; for the pillow^,0.266 m!-
  2) The effective chamber volume (V) for (he carpet, wallboard, and ceiling ate iests was 0.051 m'; for the
Notes:
   Th'
   Thi
   pillow, 0.046m'.
  Table 5 Average adsorption and desorp»bn rate constants and equilibrium constants - Langmuir iso« ->rrn model

PCE1

Elhylbenzene

Rate
constant
k,(m/h)
kd (h"1)
Mm)
k,(m/h)
kj (h •')
Mm)
Carpet
0.13
0.13
0.97
0.08
0.08
0.95
Carpet
(35 °C)
0.30
0.59
0.52
—
—
Wallboard
0.21
1.5
0.14
0.45
1.5
0.30
Ceiling
tile
0.10
0.61
0.16
0.24
0.59
0,41
Pillow
0.03
0.10
0,25
0.004
0.016
0.27
  •Teirachloroeihykne = Pcrchloroethylene (PCE)

  Note: All lests were conducted at 23 °C, except where noted.

-------
                                             B. Tithenor ct ol,: VOC Interactions with Indoor Sinks  31
     60
  „  40
  n

  I
  £  30
     20
     10
                                                                Fig. 5 Cor relation between M.
                                                                ond C«: teiracWotocthylcnc/
                                                                carpel
                10
                        20
                                 30
                                                   50
                                                            60
 ehloroeihylene; Table 3 presents the results
 for the remainder of the tetrachloroethylene
 experiments; Table 4 provides information
 on the experiment  conducted with ethyl-
 benzene; Table 5 presents a summary of ihe
 Langmuir rate  constanis  for  all the tests.
 Note that the tests conducted on the empty
 chamber  and on  window glass showed no
 sink  effect, so no  results are presented for
 these experiments.
 Discussion
 Applicability of Langmuir Isotherm
 Figure 5 presents a least squares plot of M«.
 vs. Cc for  tetrachloroethylene and carpet at
 23 °C, The plot shows a linear relationship
 with a slope of 0.97 and an r2 correlation of
 0.97. This  linearity confirms the applicabil-
 ity of the Langmuir equilibrium at low con-
 centrations represented by equations (3) and
 (5) [i.e., M^kAijQ^ktCJ, with the slope
 of the  line equal  to the average k,. value for
 the 10 tests (see Table 5).
   Fitting the kinetic data (i.e., C vs. t) using
 equations (19) and  (21) to obtain k, and kd
 values based on a Langmuir sorption process
 shows  a good fit for ceiling tile and wall-
 board (see Figure 3). On the other hand, the
fit of the data for the pillow and carpet daia
indicates that  ihe  desorption  process  for
 these materials deviates from the  Langmuir
 assumptions (see Figure 4). Thus, it appears
 that  Langmuir kinetics apply to relatively
 flat, smooth surfaces (e.g.( ceiling tile, wall-
 hoard), but may be inappropriate for more
 complex surfaces  (e.g., carpet and upholst-
 ery).  In spite of the relatively poor fit of the
 kinetic data for carpel and pillow, it is felt
 that the values of ka, kj, and kc for these ma-
 terials can be used to compare their sorption
 behaviour with other materials and to com-
 pare the sink effects of different compounds.

 Sink Strength
 The capability of a sink material  to adsorb
 indoor air pollutants is represented by k,.; the
 higher the kc value, the greater the mass ad-
 sorbed on  the sink at a given concentration
 (i.e., the stronger  the sink). An analysis oi
 variance (AOV) of the k,. values in Tables 2 -
 4  shows  a significant  (p  = 0.06) COM-
 POUND  X  MATERIAL interaction (i.e.,
 the differences in  the kc  values among  the
 materials depend on the compound). Carpet
 was a significantly stronger sink (p <  0.01)
 than the other three materials for both com-
 pounds, while no significant differences in k«.
were detectable (p  = 0.01) among  the other
 three materials  for either compound. As sta-
ted previously, glass was not a sink for either
compound.

-------
32  B. Tichenor et ol. • VOC Interactions with Indoor Sinks
  The sink strengths of the two compounds
tested arc also different. Ceiling tile adsorbed
significantly  more (p <  0.01) cihylbeii/cne
than ictrachloroethyicnc. A similar trend (p
= 0.05) was  evident for wallboard. For car-
pet and pillow, no differences (p =  0.79 and
p - 0.80, respectively) in sink strength were
observed between the two compounds.
 Adsorption and Desorption Rates
 Values ."or ka and kd were analysed to deter-
 mine differences in adsorption and dcsorp-
 tion rates for the materials and compounds
 tested. An AOV  indicated a significant MA-
 TERIAL X COMPOUND interaction for k,
 (p < 0.01), but not for kd (p = 0.99). Other
 findings include:

 -  wallboard and ceiling tile adsorbed ethyl-
   benzene at significantly higher rales (p <
   0.01) than carpel and pillow. Elhylbenzcne
   adsorption by wallboard was faster than by
   ceiling tile;  the adsorption rates for carpel
   and pillow were not significandy different;
 -  wallboard adsorbed tetrachloroethylene at
   significantly higher rates (p <  0.01) than
   the pillow, with carpet and ceiling tile rep-
   resenting intermediate, but indistinguish-
   able, values;
 -  ethylbenzene  adsorbed  at a  significantly
   higher rate than  tetrachloroethylene on
   the wallboard (p <  0.01) and ceiling die
   (p <  0.01).  There were no significant dif-
   ferences in the adsorption rates for the two
   compounds for carpet (p  = 0.22) and pil-
   low (p = 0.68);
 - wallboard desorbed at a significandy high-
   er rate (p < 0.01) than the other three ma-
   terials. Ceiling tile desorbed at a signific-
   antly higher rate (p  < 0.01) than carpet
   and pillow,  which are indistinguishable;
 -  there  was no significant  difference (p =
    0.56) in the desorpdon rates for ethylben-
    zene and tetrachloroethylene for any of the
    sink materials.
Effect of Temperature
Only limited data were collected  to investi-
gate the effect of temperature. Based on ther-
modynamics, both the adsorption and  de-
sorption rates should be higher at elevated
temperatures. The values for k, and kj for the
tetrachloroethylene/carpet  tests conducted at
35 °C indeed show significantly higher values
(p  < 0.001) than those observed at  23 °C.
The significantly lower value of kc at 35 °C
occurred because the desorption rate was ele-
vated more than the adsorption rate.


Practical Implications
The data presented herein, while  limited, do
confirm that common  indoor materials ad-
sorb and subsequently re-emit vapour phase
organic  compounds.  The type  of material
and compound affects the rates of adsorption
and desorption as well as  the amount of ma-
terial adsorbed.
   Recent experiments  conducted at  an in-
door air quality (1AQ) test house (Jackson et
al, 198?) provide further confirmation of the
 importance of indoor  sinks. A  wood stain
 was applied to a bare oak floor, and the con-
centration  of  total vapour  phase organics
 (based  on sampling with Tenax/charcoal and
 analysis with gas chromatograph/fiame ioni-
 zadon  detector) was followed over  time as
 shown in Figure 6.
    An IAQ model (Sparks, 1988)  was  used to
 evaluate the data. The emission  rate  of total
 organics from the stain was calculated from:
 ER,01 =
(22)
 where, ERIO, is the total emission rate (mg/
 m'-h); EI^ is the emission rate during the
 application of the product (mg/rn2-h); ER» is
 the initial emission rate of the drying phase
 (mg/rn2 h);  k is the decay constant for the
 emission rate during drying (h"1); and t is the
 time (h). The following  values were used:
 ER1W = 15,000 mg/mMt for the first 0.4
 hours; ERo =  17,000 mg/m2-h;  k = 0.4 tv1.

-------
                                           B.lkhenor et ol.: VOC Interactions with indoor Sinks  33
    1000
     100
  f
"5 A
                                  fig. 6IAQ test house experiment
                                  [wood slain): model predictions
                 too
200        300
  Time (hrs)
The value for ER^p was developed from on-
going experiments; the values for ER,, and k
were obtained from chamber tests (Tichenor
and Quo, 1988).
   Figure 6 shows the results of the  model
runs for three cases; (a) assuming no sink ef-
fect; (b) assuming Laiipmuir type sinks; and
(c)  assuming  re-emitting  sinks  are non-
Langmuir (noted as  "Best  Fit"). Note that
the prediction assuming "no sink" fails to ac-
count for the increased concentration due to
the re-emitting sinks  after the first few hours
of the test.
   The curve labelled "Langmuir sink" was
 predicted by the model with the sink beha-
 viour described by equation (17) and using
 the  chamber derived values for  k,  and kd
 based  on  the ethylbenzene tests. The  test
 house carpet values  (ka  =  0.08 m/h; kd =
 0.08 h'1) were taken directly from the cham-
 ber data. Single values for all wall and ceiling
 surfaces (k, = 0.35 m/h; kd =  1.0 h'1) were
 based on averages of the wallboard and ceil-
 ing die chamber data. The  "Langmuir sink"
 curve does indicate re-emissions from the
 sinks, but  ii does not adequately predict the
 time history of these re-emissions. The fail-
 ure of the Langmuir isotherm data to accura-
 tely predict the sink behaviour could be
 caused by several factors:
                  -  the ethylbenzene data do not properly ac-
                    count for the sink behaviour of the com-
                    plex  petroleum-based  solvent  emitted
                    from the stain;
                  -  the test house sink surfaces are  different
                    and more complex than the surfaces evalu-
                    ated in the chambers; and
                  -  the desorption behaviour is not adequately
                    defined by a Langmuir process. (Note the
                    previous discussion of the poor fit of the
                    chamber kinetic data for carpet and  pil-
                    low.)

                  The curve labelled "Best Fit" was predicted
                  by the model assuming all sink surfaces be-
                  haved according to a Langmuir adsorption
                  process and a non-Langmuir desorption pro-
                  cess. The rate of change of mass in the sink
                  was described by:
                                                       (23)
                   where, k,  and A  have been previously de-
                   fined; C, is the concentration at time t, (mg/
                   m3);kd" is the desorption rate constant; M, is
                   the mass in the sink at time t (mg/m2); and n
                   is an  empirical constant. Note that the de-
                   sorpdon portion of this equation is  analo-
                   gous  to a Freundlich isotherm  described
                   previously [see equation (6)]. The following

-------
34  B. Tichenor ei ol.: VOC Inferddions with Indoor Sinks
values, obtained from analyses of the cham-
ber data for ethylbenzene/carpetj were used
in the IAQ model:  ka  = 0.1 m/h;  k/ =
0.004; and n = 1.5. The fit of the model (see
the "Best Fit" line) is quite good for the sink
behaviour described  by equation (23).  The
theoretical  basis for  this equation  is weak,
and farther research is required to refine this
approach.


Conclusions
Experiments conducted to examine the beha-
viour of indoor sinks vis-a-vis vapour phase
organic compounds  have clearly shown  ad-
sorption to and re-emission from common
indoor  materials  for two typical indoor air
pollutants.  Of the five materials tested, only
window glass had no measurable sink effect.
Both the type of material and the type of or-
ganic compound  affect  the rates of adsorp-
tion and desorpuon,  as well as the equilibri-
um  mass in the  sink. Higher temperatures
increase both the adsorption and desorption
rate constants. Fundamental  Langmuir  ad-
sorption theory appears  adequate to describe
the behaviour of smooth materials, such as
wallboard and ceiling die. The desorption
kinetics of rough, complex materials, such as
carpet and  upholstery, appear  to be governed
by non-Langmuir  processes. Experiments
conducted  in an IAQ test house confirm the
importance of sinks in controlling the levels
of vapour phase organics over extended per-
iods. If appropriate adjustments are made to
account for non-Langmuir desorption pro-
cesses,  IAQ models  can adequately  predict
the  impact of sinks on the  concentration/
time history in full scale (i.e., test house) en-
vironments.


Research Needs
Further research is needed to increase under-
standing of the behaviour  of indoor sinks
with respect to vapour phase organic com-
pounds:

- examination of the equilibrium and kine-
  tic behaviour of additional sink materials
  and additional  organic  compounds, in-
  cluding  mixtures  of  vapour  phase orga-
  nics;
- evaluation of the behaviour of sinks under
  a wider range of conditions (i.e., tempera-
  tt  -, humidity, concentration);
- examination of the behaviour of sinks ex-
  posed to  dynamic,  non-steady  source
  emissions;
- development of sink models that account
  for  non-Langmuir desorption  processes.
  Such models should be capable of predict-
  ing re-emissions from sinks over long per-
  iods (e.g., for weeks or months);
- incorporation of sink models  into IAQ
  models and validation in full-scale indoor
  environments.  Methods  for simplifying
  complex, multiple sinks  and for  dealing
  with vapour phase organic  mixtures are
  needed.
A preliminary version of this paper was presented
at ttie Fifth International Conference on Indoor
Air Qtiality and Comfort, Toronto, Canada, July
1990.

-------
                                                      B. Tichenor el at.: VOC Interactions with Indoor Sinks   35
References

Bcrgiund, B., Johansson, I., and LindvaH, T. (1988) "Ad-
  sorption and desorption of organic compounds in in-
  door materials", Proceedings of Heallhy Buildings '88,
  Stockholm,  Swedish Council for Building Research,
  Vol. 3, pp. 299-309.
Bergtod, B., Johansson, I., and Lindvall, T. (1989) "Vo-
  latile organic compounds from building materials in a
  simulated chamber  study", Environment  International,
  15,385-388.
Billick, I. and Nagda, N. (1987) "Reaction decay of ni-
  trogen dioxide". In: Proceedings of Indoor Air '87, Ber-
  lin, Institute for Water, Soil, and Air Hygiene, Vol. 1,
  pp. 311-315.
Daniels, E and Alberty, R. (196!) Physical Chemistry, 2nd
  ed., New York, John Wiley & Sons.
Dunn, J. and Tichenor,  B.  (1988) "Compensating for
  sink effects in emissions test chambers by mathemati-
  cal modeling", Atmospheric Environment, 22,885-894.
Gebefugi, I. and Korte, F. (1989) Indoor Accumulation of
  Semivolatiles. Paper presented at 82nd Annual Meet-
  ing of Air and Waste Management  Association, Paper
  No. 89-86.1, Anaheim, CA.
Jackson,  M., Clayton, R., Stephenson, E.,  Guyton, W.,
  and Bunch, J. (1987) "EPA's indoor air quality test
  house, 1. Baseline studies". In: Proceedings of the 198?
  EPAIAPCA Symposium on Measurement of Toxic and
  Related Air  Pollutants,  EPA-600/9-87-010  (NTIS
  PB88-113402), 104-108, Research Triangle Park, NC.
Nazaroff, W. and Cass, G. (1989) "Mass-transport aspects
  of pollutant removal at  indoor surfaces", Environment
  Imermlional, 15,567-584.
Nielsen,  E (1987) "Potential pollutants - their import-
  ance to the Sick Building Syndrome - and their re-
  lease mechanism".  In:  Proceedings of Indoor Air '87,
  Berlin, Institute for Water, Soil, and Air Hygiene, Vol.
  2, pp. 598-602.
Nielsen, E (1988) "The importance of building materials
  and building construction to the Sick Building Syn-
  drome", Proceedings of Healthy Buildings 'SS, Stock-
  holm, Swedish Council for Building Research, Vol. 3,
  pp. 391-399.
SAS  Institute, Inc. (1985) SAS User's Guide: Statistics,
   Version 5 Edition, Gary, NC.
Seifert, B. and Schmahl, H. (1987) "Quantification of
  soiption effects for selected organic substances pres-
  ent in indoor air". In: Proceedings of Indoor Air  '87,
  Berlin, Institute for Water, Soil, and Air Hygiene, Vol.
  1, pp. 252-256.
Sparks. L. (1988) Indoor Air Quality Model - Version  1.0,
  EPA-600/8-88-097a (NTIS PB89-133607), U.S. EPA,
  Air and Energy  Engineering Research Laboratory,
  Research Triangle Park, NC.
Spicer, C, Coutant, R.» Ward, G., Joseph, D., Gaynor, A.,
  and BilKck, I. (1989) "Rates and mechanisms of NO*
  removal from indoor air by residential materials", En-
  vironment International! 15,643-654.
Tichenor, B. (1989) Indoor Air Sources: Using Small En-
  vironmental Test Chambers to Charaaaize Organic Emis-
  sions from Indoor Miaeriah and  Products, EPA-69Q/8-
  89-074 (NTIS PB90-H0131), U.S. EPA, Air and En-
  ergy Engineering Research Laboratory, Research Tri-
  angle Park, NC.
Tichenor, B. and Guo, Z. (1988) "The effect of ventila-
  tion on emission rates of wood finishing materials",
  Proceedings of Heallhy Buildings 'SS Stockholm, Swed-
  ish Council for Building Research, Vol.  3, pp. 423-
  432.
Tichenor, B.,  Sparks, L.,  White, J., and Jackson, M.
  (1988) Evaluating Sources of Indoor Air Pollution. Paper
  presented at 81st Annual Meeting of the Air  Pollution
  Control Association, Dallas, TX,  (Paper  No.  88-
  110.2).
Valbjom, O. and Skov, P. (1987) "Influence of indoor cli-
  mate on the Sick Building Syndrome prevalence". In:
  Proceedings  of Indoor Air 87,  Berlin, Institute for
  Water, Soil, and Air Hygiene, Vol. 2, pp. 593-597.
Yanagisawa, Y, Fasano, A., Spengler, J., Ryan, E,  and
  Billick, I. (1987) "Removal of nitrogen dioxide by in-
  terior materials". In: Proceedings of Indoor Air '87, Ber-
  lin, Institute for Water, Soil, and Air Hygiene, Vol. 1,
  pp. 376-380.
Zsolnay, A., Gebefiigi, I., and Korte, E (1987)  "The use
  of factorial design to estimate  synergistic effects in the
  environment". In: Proceedings of Indoor Air '87, Ber-
  lin, Institute for Water, Soil, and Air Hygiene, Vol. 1,
  pp. 265-269.

-------