Environ. Sci. Technol. 2001, 35, 2758-2764
Indoor Chemistry: Ozone and
Volatile Organic Compounds Found
in Tobacco Smoke
RICHARD J. SHAUGHNESSY,†
T. J. MCDANIELS,‡ AND
C H A R L E S J . W E S C H L E R * ,§
Department of Chemical Engineering, University of Tulsa,
600 S. College Avenue, Tulsa, Oklahoma 74104,
R. W. Holland Engineering, P.O. Box 472336,
Tulsa, Oklahoma 74147, and Department of Environmental
and Community Medicine University of Medicine & Dentistry
of New Jersey-Robert Wood Johnson Medical School,
170 Frelinghuysen Road, Piscataway, New Jersey 08854
The deliberate generation of ozone in indoor settings has
been promoted as a method to reduce the concentration
of indoor pollutants. The present study examines the effect
of ozone on a subset of volatile organic compounds
(VOCs) found in tobacco smoke. The decays of these
compounds were measured in a static room-sized chamber:
(1) in the absence of ozone, (2) in the presence of
moderate ozone concentrations (<0.115 ppm), and (3) in
the presence of high ozone concentrations (<1.4 ppm). At
moderate ozone concentrations there was little effect on
the monitored VOCs. At high ozone concentrations there was
a small, unanticipated reduction in the concentration of
some of the saturated VOCs, apparently caused by OH radicals
produced as a consequence of the ozone/alkene reactions.
There was also a much larger reduction in the concentrations of those compounds with unsaturated carbon bonds.
However, this reduction was largely matched by an
increase in the concentration of a number of aldehydes.
Some of these aldehydes are more potent irritants than their
precursors. Furthermore, even a relatively small ventilation
rate (∼0.1 h-1) would produce a greater reduction in
the monitored VOCs than that produced by a moderate
amount of ozone.
Introduction
In recent years the use of ozone, intentionally produced by
an ozone generator, has been promoted as a means to
“reduce” the concentration of indoor pollutants. However,
reactions between ozone and certain unsaturated organic
compounds in indoor environments have been shown to
generate products that are more reactive and/or irritating
than their precursors (1-10). This paper examines the effect
of ozone on the concentration of volatile organic contaminants from a specific sourcessmoking. The experiments have
been conducted in a room-sized environmental chamber
intended to represent a simple indoor environment.
There are literally hundreds of different volatile organic
compounds (VOCs) present in environmental tobacco smoke
* Corresponding author phone: (732)235-4114; fax: (732)445-0116;
e-mail: [email protected].
† University of Tulsa.
‡ R. W. Holland Engineering.
§ University of Medicine & Dentistry of New Jersey-Robert Wood
Johnson Medical School.
2758
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 13, 2001
FIGURE 1. Constituents of environmental tobacco smoke monitored
in this study. In Group I, separate structures are shown for the
meta- (m-) and para- (p-) xylene isomers. However, the concentrations of these two compounds are reported together as m,p-xylene
since they are difficult to separate chromatographically. See text
for Group assignments.
(ETS). In this study we have chosen to monitor 18 of these
compounds. They have been chosen because they are some
of the more abundant organic constituents of ETS and
because they are representative of the different types of
compounds known to be present. Based on their expected
behavior in the presence of ozone, we have assigned these
VOCs to one of three groups: Group I s compounds that
neither react with ozone at a significant rate nor are ozone
reaction products; Group II s compounds that react with
ozone at a rate comparable to or faster than typical air
exchange rates; Group III s compounds that do not react
with ozone at a significant rate but are among the products
of the reaction of ozone with other compounds found in
tobacco smoke. Figure 1 lists these compounds by group
and displays their structures. In addition to the separately
monitored VOCs, we have also monitored an undifferentiated
signal produced by the sum of the VOCs present in the
chamber.
Experimental Section
Chamber. Experiments were conducted in a modified
Association of Home Appliance Manufacturers (AHAM) AC-1
chamber (11). The chamber dimensions are 3.66 × 2.83 ×
10.1021/es001896a CCC: $20.00
 2001 American Chemical Society
Published on Web 06/02/2001
2.41m for a total volume of 25 m3. The interior surfaces consist
of wallboard with linoleum flooring per the AHAM specifications. The chamber is flushed using a single pass ventilation
system with a 150 m3/min fan drawing air from the room.
Inlet air passes through a 60% efficient prefilter, a carbon
filtration glide-pak, and finally a HEPA filter. During the
experiments the chamber was operated in static mode; that
is, there was no forced ventilation. The chamber air exchange
rate was measured on three occasions using standard carbon
dioxide decay techniques. The results were 0.018, 0.019, and
0.018 air changes/h (h-1). A ceiling fan in the chamber was
operated at low speed during testing to prevent stratification.
Initial mixing of the smoke was facilitated by the use of an
oscillating table fan positioned on the floor.
Ozone Generator. A commercially available appliance
was used to generate ozone. This device produces a corona
discharge on wire mesh screens connected to a high-voltage
power supply. It is equipped with a fan that pulls air from
the back of the unit, across the wire mesh screens, and out
the front of the unit. This fan was operated at a high setting
(0.17 m3/s) for all tests. The device has a dial that alters the
voltage to the wire mesh screens. The dial is circumscribed
by markings indicating recommended settings based on the
floor area of the room in which the device operates. For
purposes of these tests, the device was operated in two
modes: “room setting”, the manufacturer’s recommended
setting based on the floor area of the chamber, and “high
setting”, the dial at its maximum position. Ozone concentrations increased exponentially over the duration of a 5-h
experiment; at the “high setting” ozone levels within the
chamber were approximately 0.14 ppm after 1 h, in the range
of 0.30-0.32 ppm after 2.5 h and 1.2-1.4 ppm after 5 h. The
lower ozone concentrations produced at the “room setting”
proved difficult to monitor due to interference from the
tobacco smoke. Previous chamber tests, conducted in the
absence of ETS but under otherwise identical conditions,
resulted in ozone levels of 0.09-0.115 ppm after 5 h of
operation. In the presence of tobacco smoke, the ozone
concentrations in the chamber at the “room setting” would
be smaller due to reactions with ETS constituents.
Measurements. VOC samples were collected on multisorbent tubes (Tenax TA, Ambersorb XE-340, and carbon)
and analyzed by thermal desorption GC/MS following U.S.
EPA Method TO-1 (12). Both positive and negative artifacts
have been reported when collecting compounds on Tenax
in the presence of ozone (13-16). In initial experiments an
ozone scrubber was placed upstream of the multisorbent
tube. However, the scrubber removed a fraction of the higher
molecular weight VOCs. Consequently, scrubbers were not
used in the VOC measurements that are reported in this
study. To examine positive artifacts produced by collecting
VOCs in the presence of ozone, two experiments were
conducted at the high ozone setting but in the absence of
ETS. Of the 18 VOCs monitored in this study, three displayed
a positive artifactsbenzaldehyde, nonanal, and decanal. The
increases were roughly comparable to those observed when
smoke was present and the ozone generator operated at a
“high setting”. However, it should be noted that in the absence
of smoke, the final ozone concentration at the high setting
was almost four times greater than when smoke was present.
Given these interacting factors, no attempt was made to
correct for this positive artifact; the increases reported for
benzaldehyde, nonanal, and decanal in the Results and
Discussion section should be considered upper limits. A
negative sampling artifact may have occurred for one or more
of the Group II compounds. That is, some of the Group II
compounds may have undergone further reaction with ozone
after they had been captured by the sorbent. Kinetic analyses
suggest that such negative artifacts are small (see below).
Nonetheless, concentrations reported for Group II com-
pounds should be considered lower limits; their actual gasphase concentrations may have been larger than those
reported in the Results and Discussion section. The analytical
precision for individual compounds in replicate samples was
between 10% and 15%. Exceptions were isoprene (20%) and
hexanal, nonanal, and decanal (30-40%).
Aldehyde samples were concurrently collected on silica
cartridges impregnated with 2,4-dinitrophenylhydrazine.
Copper tubing, internally coated with potassium iodide (36
cm × 0.64 cm o.d.), was used upstream of the sampler to
remove ozone from the sample-gas stream and limit formaldehyde artifacts known to occur when using this method
in the presence of ozone (17). Aldehyde samples were
analyzed for formaldehyde and acetaldehyde by HPLC
following U.S. EPA Method TO-11 (18). Authentic standards
were used for both compound identification and calibrations.
The analytical precision for individual compounds in replicate
samples was typically 10% or better.
The sum of volatile organic compounds present in the
chamber air was monitored by a total-ion-current (TIC)
procedure described in detail by Hodgson (19). In the present
study the TIC chromatograms were integrated to include
compounds with volatilities between those of 2-methylbutane
and n-hexadecane. A mixture of 12 aliphatic and aromatic
hydrocarbons, similar to that employed by Hodgson (19),
was used for calibration. The results were expressed as mass/
unit volume hydrocarbon equivalents.
A Dasibi Environmental Corp., Inc. model 1003-PC
ultraviolet photometric ozone monitor was used to measure
ozone concentrations (wavelength, 254 nm; range 0.001-5
ppm; precision ( 3% of the reading). Three point calibrations
were performed each week of testing.
Procedure. For each set of experiments, the test chamber
was sealed, and the filtration/air-recirculation loop was
turned on for a minimum of 60 min. The filtration/air
recirculation system was then shut off, and the smoke of
four unfiltered research cigarettes (1R4 from the Tobacco
and Health Research Institute, University of Kentucky) was
delivered to the chamber. This was accomplished by passing
filtered air through the lit cigarette into a glass container and
finally into the chamber through stainless steel tubing. After
a brief mixing interval, “initial” VOC and aldehyde samples
were collected. The ozone-generating device was then turned
on (except for “no ozone” experiments); this point was
considered time zero. Additional samples were collected at
1 h, 2.5 h, and 5.0 h after time zero. A total of nine experiments
were conducted: three no ozone experiments, three experiments at moderate ozone concentrations, and three experiments at high ozone concentrations. In several experiments
background VOC samples were collected prior to smoke
injection. For the 18 compounds monitored in this study,
background levels were negligible compared with the levels
measured following the injection of smoke.
Results and Discussion
The initial concentrations of the 18 constituents of ETS
monitored in this study vary greatly; the levels range from
lows of 2.5 and 3.0 µg/m3 for decanal and nonanal to highs
of 360 and 650 µg/m3 for isoprene and acetaldehyde. The
average initial concentrations of the monitored VOCs are
shown in Table 1. This table also summarizes changes, as
percent of initial concentration, of the monitored VOCs in
the presence of different ozone concentrations both 2.5 and
5 h after the experiments were initiated.
No Ozone. As can be seen in Table 1, even when the
ozone generator is not operating, most of the Group I and
II compounds display a decrease in concentration over the
course of the experiments (isoprene and 2,5-dimethyfuran
are the exceptions to this statement). Some of this decrease
is due to the chamber’s modest air exchange rate (0.018 air
VOL. 35, NO. 13, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2759
TABLE 1. Summary of Initial Concentrations and Changes in Those Concentrations for the Monitored VOCs at Different Ozone
Levels
% change after 2.5 hb
initial concna (µg/m3)
no ozone
mod. ozone
benzene
toluene
m,p-xylene
o-xylene
ethylbenzene
pyridine
50
100
40
8
16.5
40
0%
-5%
-5%
+5%
-5%
-30%
isoprene
2,5-dimethylfuran
pyrrole
styrene
2-furancarboxaldehyde
d-limonene
360
20
70
25
80
110
acetaldehyde
formaldehyde
hexanal
benzaldehyde
nonanal
decanal
650
20
3.5
6
3
2.5
compound
a
Average of nine sets of data.
b
% change after 5 hc
high ozone
no ozone
mod. ozone
high ozone
Group I
-10%
-10%
-10%
-10%
-15%
-45%
-10%
-15%
-20%
-20%
-15%
-20%
-10%
-10%
-15%
--10%
-45%
-10%
-10%
-10%
-15%
-15%
-60%
-25%
-25%
-35%
-30%
-30%
-45%
+30%
+20%
-20%
-5%
-20%
Group II
-15%
-30%
-35%
-20%
-30%
-60%
-90%
-97%
-80%
-35%
+5%
-25%
-35%
-10%
-40%
-20%
-45%
-45%
-25%
-35%
-98%
-96%
-98%
-97%
-70%
-5%
-25%
-97%
-15%
-35%
-98%
-10%
+130%
+100%
+20%
+60%
+210%
Group III
-10%
+260%
+80%
+30%
+100%
+390%
+15%
+1400%
+600%
+280%
+660%
+460%
-20%
+270%
+300%
+10%
+160%
+250%
-15%
+390%
+140%
+30%
+90%
+480%
-5.0%
+2500%
+470%
+390%
+170%
+400%
Average of three sets of data; (5%. c Average of two sets of data; (5%.
FIGURE 2. Initial and final concentrations for the monitored constituents of environmental tobacco smoke. Ozone generator off.
changes/h). Theoretically, this air exchange rate should
produce about a 5% decrease after 2.5 h and a 10% decrease
after 5 h. However, some of the Group I and II compounds
decrease by percentages greater than these values during
the no ozone experiments. The larger decreases are due, in
part, to sorption on chamber surfaces. The results in Table
1 suggest that, among the Group I and II compounds, sorption
is greatest for pyridine, pyrrole, and 2-furancarboxaldehyde.
These are the most polar compounds in Groups I and II, and,
2760
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 13, 2001
hence, these compounds are expected to have the greatest
affinity for the chamber’s surfaces.
Figure 2 shows the initial (t ) 0) and final (t ) 5 h) mass
concentrations for the 18 monitored VOCs when there is no
ozone in the chamber (i.e., ozone generator not operating).
The figure is a composite, summarizing the results from
multiple “no ozone” experiments. With the exception of
isoprene, the small decreases noted above are apparent for
the first 12 compounds in Figure 2. In the case of the last six
FIGURE 3. Initial and final concentrations for the monitored constituents of environmental tobacco smoke. Ozone generator at moderate
setting.
compounds, the Group III compounds, there is a decrease
in the concentration of acetaldehyde, little change in the
concentration of benzaldehyde, and increases in the concentrations of formaldehyde, hexanal, nonanal, and decanal.
The most likely cause of these increases is autoxidation (O2)
of unsaturated hydrocarbons, including fatty acids, known
to be present in ETS (20, 21). Note that hexanal, nonanal,
and decanal had initial concentrations less than 5 µg/m3, so
that it does not require much of an absolute increase in
concentration to produce a large relative increase. A small
amount of acetaldehyde is also likely produced by autoxidation, but the small amount produced is overwhelmed by
the amount lost to surface sorption and air exchange. The
latter losses are large, in an absolute sense, since acetaldehyde
has a large initial concentration.
Moderate Ozone. Figure 3 is analogous to Figure 2 but
shows composite results when the ozone levels are moderate.
When the ozone-generating device is operated at a moderate
setting, the reduction in the concentration of Group I
compounds is nearly identical to that observed when the
ozone generator is not operating (Figures 2 and 3, Table 1).
As noted above, the decay is primarily a combination of
sorption on the chamber surfaces and air exchange. In the
case of the Group II compounds, their decrease is somewhat
greater in the “moderate ozone” experiments than in the no
ozone experiments; this is true both 2.5 and 5 h into the test.
Comparing Figures 2 and 3 (t ) 5 h), the larger decreases are
most apparent for isoprene and d-limonene. The effect of
moderate ozone on the concentration of Group III compounds is also small but in the opposite direction from that
observed for Group II compounds. Formaldehyde and
benzaldehyde show distinct increases in their concentrations
compared with the no ozone conditions. The small increases
in the concentrations of these compounds are matched by
small decreases in the concentrations of isoprene,
d-limonene, and styrene at moderate ozone levels. Acetaldehyde’s concentration at 5 h has not decreased as much in
the presence of moderate ozone as in the no ozone case. Its
concentration at 5 h is actually 30 µg/m3 larger with moderate
ozone than with no ozone, consistent with increased
acetaldehyde production. None of the Group II compounds
are expected to produce acetaldehyde upon reaction with
ozone; other unsaturated organic constituents of the ETS
must serve as precursors for the acetaldehyde.
High Ozone. Figure 4 shows composite results for
concentrations of the 18 monitored VOCs when the ozone
levels are high. Under these conditions, the decreases
observed in the concentrations of the Group II compounds
and the increases in the concentrations of the Group III
compounds are quite large. Even the compounds in Group
I display a slightly larger decrease in concentration in the
presence of high ozone levels than in the presence of either
no ozone or moderate ozone (see Table 1). Based on the very
slow rate at which these compounds react with ozone (see
Kinetics subsection and Table 2), the decreases observed in
the Group I compounds cannot result from the direct reaction
with ozone. Instead, we hypothesize that the reductions are
due to reactions with hydroxyl radicals produced as a
consequence of ozone reacting with Group II compounds
and other unsaturated compounds present in ETS.
Based on the measured net reductions (“high ozone” case
corrected for the natural decay measured in the no ozone
case) in the concentrations of benzene, toluene, m,p-xylene,
o-xylene, and ethylbenzene and the reported rate constants
for the reactions between these compounds and the hydroxyl
radical, we have calculated the necessary hydroxyl radical
concentration to effect such a reduction during the high
ozone experiments. The result is (9 ( 8) × 10-5 ppb OH. The
production of the hydroxyl radical in indoor settings as a
consequence of ozone/alkene chemistry has been discussed
in recent studies (4, 5). The magnitude of this calculated
hydroxyl concentration for the high ozone chamber experiments is consistent with that expected based on the estimated
sources and sinks of OH during the high ozone experiments
VOL. 35, NO. 13, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2761
FIGURE 4. Initial and final concentrations for the monitored constituents of environmental tobacco smoke. Ozone generator at high setting.
TABLE 2. Rate Constants Reported in the Literature for the Reaction of Ozone with Group I, II, and III Compounds at 25 °Ci
compound
rate constants
(cm3 molecule-1 s-1)
group I compounds
isoprene
2,5-dimethylfuran
pyrrole
styrene
2-furancarboxaldehyde
d-limonene
Group III compounds
<10-20 c
1.2 × 10-17 d
(4-20) × 10-17 e
1.6 × 10-17 f
2.2 × 10-17 g
(0.5-1) × 10-17 e
2.1 × 10-16 h
<10-20 c
calculated % change after 2.5 h
high
mod.
ozoneb
ozonea
0%
8%
49%
10%
14%
5%
75%
0%
0%
65%
100%
75%
86%
48%
100%
0%
calculated % change after 5 h
mod.
high
ozonea
ozoneb
0%
15%
74%
19%
25%
9%
94%
0%
1%
88%
100%
94%
98%
73%
100%
1%
equivalent ventilation
rate at 0.05 ppm ozone
(h-1)
<0.00004
0.05
0.2-0.9
0.07
0.10
0.02-0.04
0.9
<0.00004
a Assuming average [O ] ) 0.03 ppm. b Assuming average [O ] ) 0.40 ppm. c References 22 and 23. d Reference 24. e Reference 25. f Reference
3
3
26. g Reference 27. h Reference 28. i Calculated percent changes, based on these rate constants, and ventilation rates required to produce equivalent
changes also tabulated.
(on the order of 10-4 ppb using an approach similar to that
discussed in ref 4).
Note that the pyridine reduction is apparently smaller at
the “high setting” than at the “room setting”; at the high
setting polar ozone/alkene reaction products (aldehydes,
ketones and organic acids) may be displacing pyridine from
sorption sites on the chamber surfaces, but additional
experiments would be necessary to test this hypothesis.
Turning to the Group II compounds, when the ozone
generating device is operated at the high setting, within 2.5
h the initial concentration is reduced by 90% or more for
2,5-dimethylfuran, pyrrole, and d-limonene (see Table 1).
By 5 h all of the compounds in Group II, with the exception
of 2-furancarboxaldehyde, have been reduced in concentration by more than 95%. The percent changes in the
concentrations of Group III compounds are also large but
opposite from those observed for the Group II compounds.
The formaldehyde concentration has increased by 1400% at
2.5 h and by 2500% at 5 h. Formaldehyde is one of the products
directly formed by the reaction of ozone with isoprene,
styrene, and d-limonene. At the high setting, the amount of
formaldehyde produced (500 µg/m3or 405 ppb) is greater
2762
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 13, 2001
than the net reduction in the concentration of isoprene,
styrene, and d-limonene (485 µg/m3or 152 ppb), suggesting
that other ETS constituents with terminal double bonds also
react with ozone to produce formaldehyde. Benzaldehyde is
a direct product of the ozone/styrene reaction. The increase
in the concentration of benzaldehyde at the high setting (23
µg/m3or 5.3 ppb) is matched by the decrease in the
concentration of styrene (24 µg/m3or 5.6 ppb). Acetaldehyde
is the only Group III compound that does not display a sharp
increase at high ozone. However, acetaldehyde’s concentration is, on average, 100 µg/m3 greater at t ) 5 h in the high
ozone experiments than in the no ozone experiments.
Kinetics. The results presented in Table 1 and Figures
2-4 are anticipated, based on the rate constants reported in
the literature (or estimated) for gas-phase reactions between
ozone and the 18 monitored ETS constituents (22-28). Table
2 summarizes such rate constants. Ozone is known to react
with the Group I and III compounds very slowly (secondorder rate constants smaller than 10-20 cm3 molecule-1 s-1).
In the presence of 0.050 ppm of ozone, these compounds
have half-lives longer than 16 000 h (95 weeks). Hence,
reactions between ozone and Group I and III compounds
are too slow to be of any practical consequence in indoor
settings.
Table 2 also lists the calculated percent changes in the
concentrations of the monitored VOCs after 2.5 and 5 h at
both moderate and high ozone settings. The calculated
reductions compare reasonably well with the measured
reductions (correcting the values in Table 1 for the percent
change observed in the absence of ozone). Of the Group II
compounds, d-limonene reacts with ozone the fastest, and,
indeed, it displays the largest reductions when the moderate
and high ozone results are compared to the no ozone results
(see Table 1 and Figures 2-4). Of the Group II compounds,
2-furancarboxaldehyde reacts with ozone the slowest, and
the percent reductions at moderate and high ozone levels
reflect this. (Rate constants have not been reported for the
reaction of ozone with either 2-furancarboxaldehyde or 2,5dimethylfuran. However, based on the reported ozone/furan
rate constant (25), and the properties of the substituent
groups, rate constants have been estimated.) The other Group
II compounds react with ozone at rates between that of
d-limonene and 2-furancarboxaldehyde, and the measured
results reported in Table 1 are consistent with this. Note that
a large fraction of the reduction observed for pyrrole in the
moderate ozone experiments is due to some process other
than homogeneous reaction with ozone; sorption to chamber
surfaces may play a role. It should also be noted that
heterogeneous oxidation of selected adsorbed compounds
may enhance the overall destruction rate over that predicted
based solely on homogeneous rate constants. However, the
reasonable agreement between the calculated homogeneous
changes reported in Table 2 and the measured changes in
Table 1 indicates that heterogeneous processes do not
contribute order of magnitude increases in the oxidative
effects of ozone.
The final column in Table 2 shows the equivalent
ventilation rates that would be needed to remove a given
compound at the same rate as 0.050 ppm of ozone. The
equivalent ventilation rates have been calculated assuming
that the ozone concentration is significantly larger than that
of the ETS constituent and calculating a pseudo-first-order
rate constant (in units of h-1) at a concentration of 0.050
ppm ozone. A ventilation rate as small as 0.1 h-1 removes the
monitored VOCs, with the exception of d-limonene and 2,5dimethylfuran, at a rate faster than 0.050 ppm of ozone.
Ozone is a very powerful oxidant, but it reacts with
saturated organic compounds, such as those found in Groups
I and III, at a very slow rate. For a homogeneous reaction to
alter the concentration of indoor pollutants, it must occur
at a rate that is comparable to, or faster, than the air exchange
rate (29). The smaller the air exchange rate, the more time
available for homogeneous chemistry to occur. During the
ozone/ETS experiments the chamber was deliberately ventilated at a very small rate (0.018 h-1) to provide conditions
that favored ozone reactions with ETS constituents. In typical
indoor settings, the air exchange rate is significantly larger
than 0.018 h-1 (30), and, consequently, there is less time
than in the chamber experiments for ozone to react with the
constituents of ETS. In other words, under normally occurring
indoor conditions ozone would have even less of an effect
on the concentrations of organic compounds in ETS than
that shown in Table 1 and Figures 2-4.
Σ ETS Constituents. As noted in the Experimental Section,
in each experiment the total concentration of ETS constituents with volatilities between those of 2-methylbutane and
n-hexadecane was monitored using a total ion current (TIC)
procedure. However, this measurement fails to measure
formaldehyde and acetaldehyde, two potentially important
species. Hence, we have added the concentration of the latter
two compounds to the TIC measurement and termed the
resulting value “Σ ETS”. Figure 5 presents results for the Σ
FIGURE 5. Σ ETS concentrations (TIC signal plus formaldehyde and
acetaldehyde concentrations) at t ) 2.5 or t ) 5 h, normalized by
respective Σ ETS concentration at t ) 0 h. Results shown for
experiments with no ozone (ozone generator off), moderate ozone
(ozone generator on at a moderate setting), and high ozone (ozone
generator on at a high setting).
ETS for each of the cases examinedsno ozone, moderate
ozone, and high ozone. The plotted values have been
normalized by dividing the Σ ETS concentration at time )
“t” by the Σ ETS concentration at time ) 0. In all nine
experiments the Σ ETS concentration was measured at t )
0 and t ) 2.5 h. In six of the nine experiments the Σ ETS
concentration was measured at t ) 5 h. To put Figure 5 in
perspective relative to Figures 2 - 4, the average initial
concentration of the Σ ETS for the nine experiments was
4260 µg/m3 compared to an initial concentration of 1600
µg/m3 for the sum of the compounds that make up Groups
I, II, and III. That is, the compounds in Groups I, II, and III
account for 40% of the compounds measured in the Σ ETS
concentration.
In the no ozone experiments, the Σ ETS concentration is
approximately 95% of its initial value after 2.5 h and 75% of
its initial value after 5 h (due to surface losses and air
exchange). In the moderate ozone experiments, the Σ ETS
concentration is approximately 75% of its initial value after
2.5 h and 65% at the 5-h point. In the high ozone experiments,
the Σ ETS concentration is about 85% of its initial value at
2.5 h and 65% of its initial value at 5 h. Hence, when the
changes in Σ ETS concentrations in the presence of moderate
or high ozone concentrations are compared with those in
the presence of no ozone, taking into account the analytical
precision, it is apparent that there is little difference among
the three scenarios. This is an extension of the results obtained
by monitoring the Group I, II, and III constituents of ETS.
Ozone only reacts with a subset of the organic compounds
that contribute to the Σ ETS signal; however, other compounds are generated as a consequence of the reactions with
ozone. Furthermore, not all of the products of the ozone/
VOC chemistry will contribute to the Σ ETS signal. Indeed,
if all of the products could be accounted for, one would
actually expect the concentration of organic compounds in
the chamber air to be greater when ozone is present than
when it is absent. This last statement derives from the fact
that, in general, the reaction of ozone with an unsaturated
organic compound produces more than one product.
However, some of these products will wind up in the
condensed phase (secondary organic aerosols or adsorbed/
absorbed on chamber surfaces). Other products such as
certain organic acids, although in the gas phase, are not
readily detected by the analytical procedures employed in
this study (the analytical methods are more sensitive to
nonpolar compounds than to oxidized species (19)).
Odor. It is interesting to note that some of the more
odorous constituents of ETS are among the compounds that
VOL. 35, NO. 13, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2763
react with moderate or high levels of ozone. The smell of ETS
is a composite, representing the net contribution of numerous
odorous compounds. To the extent that the concentrations
of some of these compounds are altered by ozone, ozone is
expected to alter the smell of ETS. This does not mean that
the resulting mixture is less irritating or harmful; it simply
has an odor that may be less offensive to some individuals.
Irritation. At elevated concentrations, ozone is a known
sensory irritant. At settings that result in low ozone concentrations, the ozone generating device has little effect on
the concentrations of the organic compounds present in
tobacco smoke. At settings that result in high ozone
concentrations, the ozone generating device only reduces
the concentration of a small subset of the organics present
in tobacco smokesprimarily those compounds with unsaturated carbon-carbon bonds. However, the reduction in
the concentration of these unsaturated hydrocarbons is
countered by an increase in the concentration of a number
of more highly oxidized compounds (6, 16, 23). Given their
increased water solubility, many of the oxidation products
are anticipated to be stronger mucous membrane irritants
than their precursors. Indeed, recent experiments using a
mouse bioassay support this conjecture (8, 9).
Acknowledgments
The authors would like to thank Al Hodgson of Berkeley
Analytical Association for his support, patience, and expertise
throughout this study. Parts of this work were presented at
Indoor Air 99, the 8th International Conference on Indoor Air
Quality and Climate, Edinburgh, Scotland.
Literature Cited
(1) Weschler, C. J.; Hodgson, A. T.; Wooley, J. D. Environ. Sci.
Technol. 1992, 26, 2371-2377.
(2) Zhang, J.; Wilson, W. E.; Lioy, P. J. Environ. Sci. Technol. 1994,
28, 1975-1982.
(3) Boeniger, M. F. Am. Indust. Hyg. Assn. J. 1995, 56, 590-598.
(4) Weschler, C. J.; Shields, H. C. Environ. Sci. Technol. 1996, 30,
3250-3258.
(5) Weschler, C. J.; Shields, H. C. Environ. Sci. Technol. 1997, 31,
3719-3722.
(6) Weschler, C. J.; Shields, H. C. Atmos. Environ. 1997, 31, 34873495.
(7) Weschler, C. J.; Shields, H. C. Atmos. Environ. 1999, 33, 23072318.
(8) Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Hougaard, K. S.; Nielsen,
G. D. Atmos. Environ. 1999, 33, 693-698.
(9) Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Nielsen, G. D. Indoor
Air 2000, 10, 82-91.
2764
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 13, 2001
(10) Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. Environ. Health
Perspectives 2000, 108, 1139-1145.
(11) Association of Home Appliance Manufacturers AC-1. AHAM
Standard Method for Measuring Performance of Portable
Household Electric Cord-Connected Room Air Cleaners (AC-1);
Developed by IIT Research Institute: Chicago, IL, 1988.
(12) 12.Method TO-1, Revision 1.0: Method For The Determination
Of Volatile Organic Compounds in Ambient Air Using Tenax
Adsorption and Gas Chromatography/Mass Spectrometry (GC/
MS); U.S. Environmental Protection Agency, Center for Environmental Research Information, Office of Research and
Development: 1984.
(13) Pellizzari, E.; Demian, B.; Krost, K. Anal. Chem. 1984, 56, 793798.
(14) Pellizzari, E. D.; Krost, K. J. Anal. Chem. 1984, 56, 1813-1819.
(15) Calogirou, A.; Larsen, B. R.; Brussol, C.; Duane, M.; Kotzias, D.
Anal. Chem. 1996, 68, 1499-1506.
(16) Morrison, G. C.; Ph.D. Dissertation, University of California at
Berkeley, 1999.
(17) Kleindienst, T. E.; Corse, E. W.; Blanchard, F. T.; Lonneman, W.
A. Environ. Sci. Technol. 1998, 32, 124-130.
(18) Compendium of Methods for the Determination of Toxic Organic
Compounds in Ambient Air, 2nd ed.; Compendium Method
TO-11A, Determination of Formaldehyde in Ambient Air Using
Adsorbent Cartridge Followed by High Performance Liquid
Chromatography (HPLC); EPA/625/R-96/010b; U.S. Environmental Protection Agency, Center for Environmental Research
Information, Office of Research and Development: 1999.
(19) Hodgson, A. Indoor Air 1995, 5, 247-257.
(20) Johnstone, R. A. W.; Plimmer, J. R. Chem. Rev. 1959, 59, 885936.
(21) Stedman, R. L. Chem. Rev. 1968, 68, 153-207.
(22) NIST Standard Reference Database 17. NIST Chemical Kinetic
Database, Version 6.1; U.S. Department of Commerce: 1994.
(23) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and
Lower Atmosphere, Academic Press: San Diego, 2000.
(24) Grosjean, E.; Grosjean, D. Int. J. Chem. Kinet. 1996, 28, 911918.
(25) Atkinson, R.; Aschmann, S. M.; Carter, W. P. L. Int. J. Chem.
Kinet. 1983, 15, 51.
(26) Atkinson, R.; Aschmann, S. M.; Winer, A. M.; Carter, W. P. L.
Atmos. Environ. 1984, 18, 2105.
(27) Atkinson, R.; Aschmann, S. M.; Fitz, D. R.; Winer, A. M.; Pitts,
J. N., Jr. Int. J. Chem. Kinet. 1982, 14, 13.
(28) Shu, Y.; Atkinson, R. Int. J. Chem. Kinet. 1994, 26, 1193-1205.
(29) Weschler, C. J.; Shields, H. C. Indoor Air 2000, 10, 92-100.
(30) ASHRAE. 1997 ASHRAE Handbook: Fundamentals; American
Society of Heating, Refrigerating and Air Conditioning Engineers: Atlanta, GA, 1997.
Received for review November 22, 2000. Revised manuscript
received April 17, 2001. Accepted April 24, 2001.
ES001896A