ARTICLE IN PRESS
Atmospheric Environment 40 (2006) 6696–6710
www.elsevier.com/locate/atmosenv
Indoor secondary pollutants from cleaning product and air
freshener use in the presence of ozone
Brett C. Singera,b,, Beverly K. Colemanb,c, Hugo Destaillatsb, Alfred T. Hodgsonb,
Melissa M. Lundena, Charles J. Weschlerd,e, William W Nazaroffb,c
a
Atmospheric Sciences Department, Environmental Energy Technologies Division, Lawrence Berkeley National Laboratory,
Berkeley, CA 94720, USA
b
Indoor Environment Department, Environmental Energy Technologies Division, Lawrence Berkeley National Laboratory,
Berkeley, CA 94720, USA
c
Department of Civil and Environmental Engineering, University of California, Berkeley, CA 94720-1710, USA
d
Environmental and Occupational Health Sciences Institute, University of Medicine and Dentistry of New Jersey & Rutgers University,
Piscataway, NJ 08854, USA
e
International Centre for Indoor Environment and Energy, Technical University of Denmark, Lyngby DK-2800, Denmark
Received 8 March 2006; received in revised form 6 June 2006; accepted 7 June 2006
Abstract
This study investigated the formation of secondary pollutants resulting from household product use in the presence of
ozone. Experiments were conducted in a 50-m3 chamber simulating a residential room. The chamber was operated at
conditions relevant to US residences in polluted areas during warm-weather seasons: an air exchange rate of 1.0 h1 and an
inlet ozone concentration of approximately 120 ppb, when included. Three products were used in separate experiments. An
orange oil-based degreaser and a pine oil-based general-purpose cleaner were used for surface cleaning applications.
A plug-in scented-oil air freshener (AFR) was operated for several days. Cleaning products were applied realistically with
quantities scaled to simulate residential use rates. Concentrations of organic gases and secondary organic aerosol from the
terpene-containing consumer products were measured with and without ozone introduction. In the absence of reactive
chemicals, the chamber ozone level was approximately 60 ppb. Ozone was substantially consumed following cleaning
product use, mainly by homogeneous reaction. For the AFR, ozone consumption was weaker and heterogeneous reaction
with sorbed AFR-constituent VOCs was of similar magnitude to homogeneous reaction with continuously emitted
constituents. Formaldehyde generation resulted from product use with ozone present, increasing indoor levels by the order
of 10 ppb. Cleaning product use in the presence of ozone generated substantial fine particle concentrations (more than
100 mg m3) in some experiments. Ozone consumption and elevated hydroxyl radical concentrations persisted for 10–12 h
following brief cleaning events, indicating that secondary pollutant production can persist for extended periods.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Air quality; Formaldehyde; Indoor air chemistry; Secondary organic aerosol; Terpenes
Corresponding author. Atmospheric Sciences and Indoor Environment Departments, Environmental Energy Technologies Division,
Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA. Tel.: +1 510 486 4779; fax: +1 510 486 5928.
E-mail address: [email protected] (B.C. Singer).
1352-2310/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.atmosenv.2006.06.005
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
1. Introduction
Many consumer cleaning products and air
fresheners contain terpenoids and related compounds as active ingredients or fragrances that
volatilize during product application and use
(Singer et al., 2006). Some of these compounds
react rapidly with ozone to form secondary pollutants (Liu et al., 2004; Nazaroff and Weschler, 2004;
Sarwar et al., 2004; Wolkoff et al., 1998).
Ozone enters indoor environments with ventilation air and typically is present indoors at levels that
are 20–70% of concurrent outdoor levels (Weschler,
2000). Ozone also may be introduced by indoor
sources including devices designed to generate
ozone (Boeniger, 1995), certain air cleaners (Britigan et al., 2006; Niu et al., 2001; Phillips et al.,
1999), and some photocopiers and printers (Lee
et al., 2001; Leovic et al., 1996).
Ozone–terpenoid reactions produce carbonyls
such as formaldehyde and acetaldehyde, organic
acids, hydrogen peroxide, secondary organic aerosol, and hydroxyl (OH) radical (Nazaroff and
Weschler, 2004). Subsequent reactions of OH with
volatile organic compounds (VOCs) can generate
additional products. Information about many of the
individual terpenoid oxidation products is limited.
However, the mixture of reaction products appears
to have significant irritant properties (Wolkoff
et al., 2006).
Secondary pollutant formation from reactions
involving ozone and terpenoid constituents of
consumer products has been studied in a reactor
tube (Wolkoff et al., 2000), ventilation ducts (Fick
et al., 2005), small Teflon-lined chambers (Destaillats et al., 2006; Wainman et al., 2000), room-sized
stainless-steel chambers (Fan et al., 2003; Liu et al.,
2004; Sarwar et al., 2004), unoccupied offices
(Weschler and Shields, 1997a, 1999, 2003), and
residences (Hubbard et al., 2005; Long et al., 2000).
Many of these studies used pure compounds such as
d-limonene. A few studies used actual consumer
products (Destaillats et al., 2006; Liu et al., 2004;
Long et al., 2000; Sarwar et al., 2004). Information
has been generated regarding formation of reaction
products (Li et al., 2002; Rohr et al., 2003; Weschler
and Shields, 1999), reactant consumption and
product yields (Destaillats et al., 2006), secondary
pollutant levels in residences (Long et al., 2000) and
chambers (Liu et al., 2004; Sarwar et al., 2004;
Weschler and Shields, 1999), and associations of
secondary pollutants with specific consumer pro-
6697
ducts (Destaillats et al., 2006; Sarwar et al., 2004).
Data have also been generated for development and
validation of mathematical models (Liu et al., 2004;
Sarwar et al., 2003). Despite this progress, important questions remain unanswered about the quantity of secondary pollutants formed during typical
residential use of cleaning products and air fresheners and about the relevance of laboratory results
for elucidating potential human exposures in actual
indoor environments.
In the current study, we aimed to partially bridge
the gap between bench-scale and building-scale
environments by conducting experiments in a
simulated residential room. Three consumer products—a general-purpose cleaner, a degreaser, and
a plug-in scented-oil air freshener (AFR)—were
applied in realistic manners in a chamber constructed with standard building materials. For each
product, experiments were executed in the absence
of ozone and with ozone introduced into the
chamber air supply. Ozone interactions with these
products were studied to contribute to the following
objectives: (1) broadly characterize ozone–terpenoid
reactions for a range of consumer products and
compounds, (2) quantify the formation of very
volatile carbonyl reaction products, (3) estimate the
levels of OH formed from ozone–terpenoid chemistry, and (4) quantify the formation of secondary
organic aerosol.
2. Methods
2.1. Chamber, products, and application protocols
Reactions between ozone (O3) and consumer
products containing terpenes and related compounds were studied in a 50-m3 chamber designed
to simulate a residential room. The chamber,
products, and most experimental protocols have
been described previously (Destaillats et al., 2006;
Singer et al., 2006). Key points are summarized
here.
The chamber is finished with painted gypsum
wallboard with sheet aluminum on the floor. The
floor was partially covered with noncontiguous 3.9
and 7.0 m2 sections of vinyl tile flooring. A table
with laminate top (1.16 m2) was present. Supply air
was drawn from outdoors and directed through a
bed of activated carbon to remove ambient VOCs
and O3. The chamber was ventilated at 1 h1 at a
positive pressure of 5 Pa relative to the building.
A household oscillating fan, set to low or medium
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
6698
speed, was operated during experiments to promote
mixing. The air-exchange rate (AER) was measured
during each experiment by monitoring the decay of
injected SF6 using a photoacoustic infrared analyzer
(Model 1302, Brüel & Kjær). A high-frequency
corona discharge O3 generator (OzoneLab GE30/
FM100R, Yanco Industries, Ltd.) was connected to
the supply air 1 m before the chamber inlet. In
designated experiments, the generator provided O3
at 120 ppb (measured) in the supply air. Chamber
air temperature and relative humidity were monitored at two locations (HOBO H8 Pro, Onset
Computer Corp.).
Three widely available consumer products containing O3-reactive chemicals were employed: an
orange oil-based degreaser (OOD) packaged as an
aerosol foam; a general-purpose pine oil-based
cleaner (POC) packaged as a concentrated liquid;
and an AFR. POC and AFR are products GPC-1
and AFR-1 in Singer et al. (2006). Ozone-initiated
chemistry involving the same three products has
been investigated in a bench-scale (200-L), Teflonlined chamber (Destaillats et al., 2006).
Table 1 summarizes the experimental design and
the measured environmental parameters (AER,
temperature, RH). In Exps. A–C, OOD was sprayed
onto a 0.11-m2 section of sheet aluminum to
simulate cleaning of a cooktop. After 1 min, the
excess was wiped away with a paper towel and
the towel and product container were removed from
the chamber. In Exps. D–H, about 4 L of a solution
of 1 part POC to 16 parts water was applied by
sponge mop to the 3.9-m2 section of vinyl flooring
using a previously described protocol (Singer et al.,
2006). The AFR was plugged into an electrical
outlet approximately 2 days before the start of Exp.
J. The device was initially set to low (1) then
switched to high (3) at the end of Exp. J and
remained at this setting through Exp. K. The AFR
container was weighed at several points including at
the start and completion of each experiment. The
measured product volatilization rate did not vary
with device setting in the manner expected (Table 1).
OOD and POC experiments started with product
application (t ¼ 0). For each product, one or more
experiments were conducted with O3 and two
control experiments were conducted without introduction of O3. The POC+O3 experiment was
conducted in triplicate (Exps. F–H). Exp. I entailed
floor mopping with water in the presence of O3, but
Table 1
Summary of experiments
Producta
Amt.b
Exp.
O3c
Start date (2005)
Start time
AER (h1)d
Temp. (1C)e
RH (%)f
OOD
OOD
OOD
POC
POC
POC
POC
POC
H2O/Mop
AFR (1)
AFR (1)
AFR (3)
AFR (3)
6.7 g
3.7 g
3.7 g
52 g
52 g
51 g
50 g
50 g
—
45 mg h1
43 mg h1
31 mg h1
29 mg h1
A
B
C
D
E
F
G
H
I
J
J
K
K
No
No
Yes
No
No
Yes
Yes
Yes
Yes
No
Yes
No
Yes
5 June
9 June
12 June
3 June
21 June
7 June
14 June
23 June
2 June
26 June
26 June
27 June
27 June
11:10
10:00
10:30
11:00
9:30
9:40
9:10
10:00
10:32
7:10
12:10
12:00
17:00
0.99
1.03
1.04
1.08
1.01
1.16
0.99
1.00
1.03
0.99
1.00
1.01
0.95
22.2
23.2
22.4
22.7
23.2
23.9
22.2
21.8
22.8
20.8
21.3
21.1
21.4
36
56
45
46
48
36
48
53
44
53
54
53
55
a
OOD ¼ orange oil degreaser: sprayed onto 0.11 m2 sheet aluminum, surface wiped dry with paper towels after 1 min, towel and
product removed (2 min procedure). POC ¼ pine oil cleaner: applied in dilute solution to 3.9 m2 of vinyl flooring using sponge mop
(7 min procedure). AFR ¼ scented oil air freshener: set to ‘‘low’’ and plugged into electrical outlet for 2 days before Exp. J; setting
switched to ‘‘high’’ for Exp. K.
b
Amount of product dispensed. For OOD and AFR, the vast majority of dispensed product was released to chamber air. For POC,
much of the dispensed product remained in solution and was removed from the chamber when mopping was completed.
c
‘‘Yes’’ ¼ ozone at 114–120 ppb in supply air; ‘‘No’’ ¼ no ozone in supply or chamber air.
d
Air exchange rate determined from measured decay of injected SF6. Uncertainty estimated at 70.02–0.05 h1 based on standard
deviation of n ¼ 3 determinations in each of Exps. C, E, and H.
e
Mean temperature over 12 h for OOD, POC, and mop w/H2O; mean over 5 h for each phase of Exps. J–K. Standard deviations were
p0.3 1C except for Exps. B (1.0), E (1.1), and F (1.4).
f
Mean RH over same periods shown for temperature. Standard deviations were p3% RH for all but Exp. B (5%).
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
without POC. The floor was wet and dry mopped
with tap water after each POC experiment to
remove residue before the next experiment. When
used, O3 was introduced to the chamber at least 12 h
prior to the start of OOD and POC experiments. In
the AFR experiments, VOC concentrations were
nominally at steady state when the O3 was
introduced. Each AFR experiment included measurement of pollutant concentrations over multiple
hours before and after introduction of O3.
2.2. Air quality measurements
Chamber air was sampled to quantify concentrations of O3, specific VOCs, very volatile carbonyl
compounds, and size-resolved particles. Analytical
methods are summarized below.
Ozone was measured continuously with a UV
analyzer (Model 400, Advanced Pollution Instrumentation, Inc.) calibrated to a primary standard.
VOCs were collected on sorbent tubes containing
Tenaxs-TA or Tenax backed by a carbonaceous
6699
sorbent. Tubes were thermally desorbed then
analyzed by gas chromatography with mass selective detection (TD-GC/MS) (Singer et al., 2004).
VOCs were sampled over integrated periods (0–30,
30–90, 90–240, and 240–720 min) during all POC
experiments and with greater time resolution in
some POC and all OOD and AFR experiments (B
and C, E–H, and J and K). Integrated samples were
collected at 3.3 mL min1. Time-resolved samples
were collected at 15–110 mL min1 over periods of
1–20 min using peristaltic pumps (Cole-Parmer),
with precision of better than 1%. Flow rates were
measured during sampling. Samples were collected
in duplicate with a subset analyzed to assess
analytical precision. VOC analytes are listed in
Tables 2 and 3.
VOCs were collected without ozone scrubbers.
This likely introduced a negative bias for some
measurements since terpenoids collected on Tenax
may be degraded by ozone in sample air (Calogirou
et al., 1996). Degradation increases with ozone
concentration, sampling time, and the bimolecular
Table 2
Time-averaged concentrations (ppb) of POC constituentsa
Analyte
CAS #
With ozoneb (Exps. F–H)
No ozone (Exps. D and E)
0–30 min
30–90 min
1.5–4 h
4–12 h
0–30 min
30–90 min
Terpene HCs
a-Pinene
Camphene
a-Phellandrenec
a-Terpinene
d-Limonene
g-Terpinene
Terpinolene
80-56-8
79-92-5
99-83-2
99-86-5
5989-27-5
99-85-4
586-62-9
12.5
12.1
9.3
19.9
166
18.0
129
7.1
6.3
4.6
8.8
82
8.9
61
2.2
1.8
1.1
1.3
23
2.5
11.8
0.2
0.2
0.2
0.2
4.8
0.5
2.9
12.2
12.2
8.7
3.9
159
17.5
134
6.0
6.0
2.6
ndb
66b
6.7b
33b
Terpene alcohols
1-Terpineolc
b-Terpineolc
4-Terpineolc
a-Terpineol
g-Terpineol
586-82-3
138-87-4
562-74-3
98-55-5
586-81-2
1
10.6
7.8
103
14.1
14.5
5.2
3.7
55
6.5
6.0
2.3
1.7
26
2.5
1.8
0.9
0.6
10.7
0.9
33
10.5
8.7
105
13.3
14.7
5.1
3.5
53b
4.4b
Other VOCs
p-Cymene
Eucalyptol
99-87-6
470-82-6
15.1
39
7.3
17.5
2.2
3.5
0.6
0.5
14.7
39
7.3
17.3
a
Mean of two experiments without ozone, three experiments with ozone. Compounds listed by retention time (RT) within group (Singer
et al., 2006).
b
For the most reactive terpenes (a-terpinene, terpinolene, d-limonene, a-terpineol and g-terpineol) concentrations reported for the
30–90 min ‘‘with O3’’ category may by biased low owing to degradation of these compounds on Tenax samplers exposed to ozone.
c
Quantified by total ion current (TIC) based on d-limonene response. a-Phellandrene identity confirmed with pure standard. Terpineols
tentatively identified by matching mass spectra to NIST database. Uncertainty in TIC quantitation estimated as 730% or less. Also
tentatively identified were terpene HCs eluting at 26.3, 28.2, and 29.4 min; 0–30 min concentrations of these compounds were estimated by
TIC to be 4–8 ppb.
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
6700
Table 3
VOC concentrations and reactions with ozone in first AFR experiment (J)
Analyte
Ozone-reactive VOCs
d-Limonene
Dihydromyrcenol
Linalool
Linalyl acetatee
b-Citronellol
a-Citral
Other VOCs
Benzyl acetate
Bornyl acetate
CAS #
Steady VOC
concentration
without O3 (ppb)a
Homogeneous
consumption of
VOC by O3 (h1)b
Steady VOC
concentration with
O3 (ppb)c
Homogeneous
consumption of O3
by VOC (h1)d
5989-27-5
18479-58-8
78-70-6
115-95-7
7540-51-4
141-27-5
2.770.3
11.271.2
7.370.9
3.670.3
1.7870.18
0.4570.05
0.88
0.005
1.85
1.75
0.99
1.43
10.7
2.6
1.32
0.89
0.03
0.00
0.10
0.05
0.02
f
f
f
140-11-4
76-49-3
16.771.6
4.670.5
—
—
15.570.2
4.370.1
—
—
a
VOC concentrations in chamber air (mean71 std. dev.) based on n ¼ 4 measurements over period 0–5 h before ozone was introduced
into the supply air at 114 ppb.
b
Consumption of VOC by homogeneous gas-phase reaction under steady-state conditions; calculated using published bimolecular
reaction rates (see text) and steady O3 concentration measured 3–5 h after O3 introduced to chamber.
c
Steady-state concentrations with O3 calculated for ozone-reactive VOCs and measured for other VOCs. Calculation based on VOC
emission from air freshener, removal by ventilation and loss by homogeneous reaction. Emission rates were calculated from steady
chamber air concentrations before ozone added, and thus include sorption effects. Measurements based on n ¼ 3 samples collected 3–5 h
after ozone was introduced.
d
Consumption of O3 calculated using expected VOC concentrations and published bimolecular reaction rates.
e
Quantified by total ion current (TIC) based on linalool response; identity confirmed by pure standard. Uncertainty in TIC quantitation
estimated as 730% or less.
f
Value could not be calculated because bimolecular reaction rate is not available.
reaction rate of the terpenoid with O3. A MnO2coated ozone scrubber has been found to reduce
ozone-degradation of some terpenes but retained
several compounds, notably linalool. Several scrubber materials have been shown to be effective at
removing ozone at 73–78 ppb without interfering
with analysis of ppb levels of five terpene hydrocarbons and five oxidation products, but no terpene
alcohols were examined (Fick et al., 2001). Scrubbers were not used in the current study owing to
concerns that they retain (through sorption) the
terpene alcohols in POC and the terpene alcohol,
aldehyde, and ester constituents of AFR. Instead,
when possible, samples were collected at low flow
rates and over short durations to reduce ozoneinduced degradation. The effect of sampling in the
presence of ozone was estimated for several VOCs
based on Calogirou et al. (1996). Fig. 2 of that paper
addressed the effect of sampling time for a fixed
ozone concentration of 120 ppb and Fig. 1 provided
data on recoveries at various ozone concentrations
relative to 120 ppb. Overall recovery was estimated
as the product of these two factors. Figs. 1 and 2 of
the current study show the estimated effect of
ozone-induced degradation on sampled VOCs.
Carbonyl compounds were collected on coated
silica cartridges (P/N 047205, Waters Corp.); during
experiments with O3, each was preceded by an O3
scrubber (WAT054420, Waters Corp.). Cartridges
were extracted with 2 mL acetonitrile. Extracts were
analyzed by HPLC with UV diode-array detection
at 360 nm, following ASTM Method D 5197.
Derivatives were quantified to determine concentrations of formaldehyde, acetaldehyde, and acetone.
Size-resolved particle number concentrations
were quantified using an optical particle counter
(OPC) (Lasair 1003, Particle Measuring Systems,
Inc.). The OPC was placed outside the chamber,
and sample air was drawn from the chamber at
0.03 L min1 through 1.4 m of 1.7-mm ID copper
tubing. The nominal size bins of the OPC are based
on the instrument’s response to polystyrene latex
(PSL) calibration aerosol. The secondary organic
aerosol generated in our experiments has different
optical properties than PSL, necessitating adjustment of the bins. Based on data collected in
similar experiments (Destaillats et al., 2006), we
used OPC bin boundaries reported by Hand and
Kreidenweis (2001) that were determined by calibration of the same instrument model with oleic
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
limonene concentration (ppb)
10000
Exp. B: No ozone
Exp. C: With ozone
1000
100
10
1
0.1
0
2
4
6
8
time (h)
10
12
Fig. 1. Limonene concentrations in chamber air following
application of the orange oil degreaser (OOD). Data have been
adjusted to account for ozone-induced degradation of limonene
on Tenax; the symbol represents the corrected value and the Tbar extends down to the measured value. Adjustments based on
Figs. 1 and 2 of Calogirou et al. (1996).
acid. This provided bin lower-size limits (aerodynamic diameters) of 0.15, 0.24, 0.36, 0.47, 0.62, and
0.89 mm, respectively, for the first six bins; negligible
counts were recorded in the two largest bins.
The particle volume distribution was determined
by multiplying the number concentration by
(p/6 GMD3), where GMD is the geometric mean
diameter of the bin. The particle mass distribution
was estimated, assuming a particle density of
1 g cm3. This method may underestimate the true
mass concentration as organic aerosol density has
been estimated to be approximately 1.2 g cm3
(Turpin and Lim, 2001). The total particle mass
concentration is described as PM1.1 based on the
upper boundary of 1.1 mm for the largest size bin
considered.
Hydroxyl radical concentrations were determined
by an indirect method (Weschler and Shields,
1997a). Diffusion vials containing 1,3,5-trimethylbenzene (TMB) and perchloroethylene (PCE) were
1000
1000
100
Exp. E: No ozone
Exp. G:With ozone
Exp. H: With ozone
10
1
d-Limonene
concentration (ppb)
concentration (ppb)
Eucalyptol
0.1
100
Exp. E: No ozone
Exp. G:With ozone
Exp. H: With ozone
10
1
0.1
0
2
4
6
8
time (h)
10
12
0
1000
2
4
6
8
time (h)
10
12
1000
100
Exp. E: No ozone
Exp. G:With ozone
Exp. H: With ozone
10
1
0.1
alpha-Terpineol
concentration (ppb)
Terpinolene
concentration (ppb)
6701
100
Exp. E: No ozone
Exp. G:With ozone
Exp. H: With ozone
10
1
0.1
0
2
4
6
8
time (h)
10
12
0
2
4
6
8
time (h)
10
12
Fig. 2. Constituent concentrations in chamber air following floor mopping with a dilute solution of the pine oil cleaner (POC). Data for
Exps. G and H have been adjusted to account for ozone-induced degradation of analytes on Tenax; the symbol represents the corrected
value and the T-bar extends down to the measured value. Adjustments based on Figs. 1 and 2 of Calogirou et al. (1996).
ARTICLE IN PRESS
6702
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
placed inside the chamber below the air inlet.
Concentrations of TMB, but not PCE, are reduced
by reaction with OH. Neither compound is depleted
substantially by reaction with O3. TMB and PCE
concentrations were quantified before and during
experiments with O3 and during Exp. E without O3.
TMB and PCE were quantified using the VOC
methods described previously, in some cases from
the same samples. Separate TMB/PCE samples
were collected at higher volume to facilitate
quantitation of these compounds during the early
part of cleaning product experiments, when their
concentrations were much lower than product
constituent VOCs. OH was estimated from the
measured decrement in the TMB/PCE ratio (R) as
shown in Eq. (1). Here, l is the AER, k the
bimolecular OH-TMB reaction rate constant
(1.44 ppb1 s1) (NIST, 2000), R(0) the ratio measured prior to t ¼ 0, i.e. when no OH was present,
and R(t) the ratios measured at time t during an
experiment.
l
Rð0Þ
½OH ¼
1 .
(1)
k
RðtÞ
from the substrate. Mop application of POC yielded
peak levels of 170–200 ppb d-limonene, 70–200 ppb
terpinolene, and 110–130 ppb a-terpineol measured
at 10 min in Exps. E–H (Fig. 2). Unsaturated
terpenoids totaled 480–630 ppb during the first
30 min following POC use in Exps. D–H (Table 2).
With no O3 present, most POC constituents initially
declined at close to the measured AER. Observed
deviations from first-order decay of POC constituents and limonene from OOD are consistent with
reversible sorption (Singer et al., 2004).
The most abundant terpenoids in AFR—
d-limonene, dihydromyrcenol, b-citronellol, linalool, and linalyl acetate—were each present at
steady levels of 1–11 ppb prior to introduction of
O3 (Table 3). Combined, unsaturated terpenoids
were present at levels of 27 and 17 ppb for Exps. J
and K.
Mopping of the floor with water only (Exp. I)
produced no change in background VOC levels in
air.
Hydroxyl radical concentrations were estimated
using the TMB/PCE ratios rather than TMB
concentrations directly because variations related
to sampling (e.g., pump flow rates), analysis (e.g.,
thermal desorption efficiency), and temperaturedependent emissions are similar for TMB and PCE.
Uncertainty in the calculated OH concentration is
proportional to uncertainty in the TMB decrement
([R(0)/R(t)]1).
The presence of ozone altered the concentration
profiles of several terpenoids, whereas the profiles of
saturated VOCs were largely unaffected. The initial
(0–1 h) first-order decay rates for a-terpinene,
terpinolene, g-terpineol, d-limonene and a few other
compounds were higher with O3 than without O3.
Attributing this difference to reactive decay,
VOC–O3 reaction rates were estimated from POC
experiments by comparing initial decay rates
calculated for Exps. G and H with those calculated
for Exp. E. First-order decay rates for reactive
compounds also were compared to the mean of nonreacting compounds in the same experiments to
provide a second estimate of the decay owing to
reaction. Bimolecular reaction rate constants were
estimated using the mean O3 concentration over the
10–60 min interval. Reaction rate constants determined this way were within a factor of 2–3 of
published values (NIST, 2000) for d-limonene,
terpinolene, a-phellandrene, and g-terpinene.
Among compounds without published reaction
rates, g-terpineol and an unspecified terpene hydrocarbon eluting at 28.2 min were estimated to react
with O3 at rates approximately 2–3 and 5–6 times,
respectively, as fast as d-limonene. a-Terpineol
initially did not decline more rapidly when O3 was
present, contrary to expectations based on its
published O3 reaction rate (Wells, 2005a). Delayed
3. Results and discussion
3.1. VOC constituents and concentrations
POC and AFR each emitted mixtures of VOCs,
whereas OOD only emitted d-limonene. POC
emissions were dominated by terpene hydrocarbons
and alcohols (Table 2), whereas AFR also emitted
substantial quantities of terpene aldehydes and
esters as well as relatively unreactive saturated
VOCs (Table 3).
VOC profiles following application of OOD and
POC are shown in Figs. 1 and 2. Fig. 1 demonstrates that spray application of OOD produced
peak d-limonene concentrations in chamber air of
950 ppb in Exp. B and 1400 ppb in Exp. C. In
Exp. B (no O3), d-limonene persisted at this level for
approximately 90 min, suggesting ongoing emissions
3.2. Effect of ozone on VOC concentrations
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
faster reaction rate with O3. Concentration patterns
of p-cymene, eucalyptol, and even some terpenes
(e.g., a-pinene and camphene) were largely unaffected by the presence of ozone.
In the AFR experiments, VOC emissions from
the source continued at approximately steady levels
while ozone was introduced. This is confirmed by
the small measured change in steady-state concentrations of benzyl acetate and bornyl acetate after
ozone was added (Table 3). Expected steady-state
concentrations of ozone-reactive VOCs in the
presence of ozone were calculated by mass balance,
using the mass emission rate inferred before ozone
was added (adjusted for the change in AFR
volatilization as shown in the second column of
Table 1), and considering removal by ventilation
and reaction. Removal by reaction was calculated
using the measured ozone concentration and
published bimolecular reaction rates for terpene–
ozone reactions. Table 3 indicates that the 47 ppb of
ozone in chamber air at steady state (Exp. J) is
140
140
120
120
ozone concentration (ppb)
ozone concentration (ppb)
emission or sorption processes for terpene alcohols
(Singer et al., 2006) may have masked a-terpineol
decay owing to reaction. Destaillats et al. (2006)
reported a similar deviation in the reactivity
of a-terpineol in bench-scale experiments. Since
g-terpineol likely sorbs at a rate that is similar to
a-terpineol, the reaction rate for the g isomer likely
is faster than estimated above. Degradation on
Tenax samplers is of minimal concern for these data
owing to the high concentrations of terpene
compounds (Fig. 2 and Table 2), low ozone levels
(Fig. 3) and short (1–2 min) sampling intervals
during this period.
Ozone continued to influence concentration
patterns over the 12-h time scale of each experiment
with OOD and POC. The addition of O3 extended
the period over which d-limonene decayed in
roughly a first-order process to about 8 h for OOD
(Fig. 1) and 6 h for POC (Fig. 2). A similar pattern
was observed for a-terpineol and terpinolene; the
latter decayed faster (Fig. 2), consistent with its
100
80
60
40
20
OOD
0
-2
0
2
100
80
60
40
20
POC
4
6
time (h)
8
10
-2
12
0
2
4
6
time (h)
(b)
8
10
12
140
140
AFR
ozone in supply air
ozone concentration (ppb)
ozone concentration (ppb)
ozone in supply air
0
(a)
120
100
80
60
ozone in chamber
40
20
Mop with H2O
120
100
80
modeled, no AFR
60
40
20
0
0
-2
(c)
6703
0
2
4
6
time (h)
8
10
-1
12
(d)
0
1
2
3
4
time (h)
5
6
7
Fig. 3. Ozone concentration profiles: (a) Exp. C: OOD sprayed at t ¼ 0; (b) Exp. H: mopping with POC started at t ¼ 0; (c) Exp. I:
mopping with water only; (d) Exp. K: AFR plugged-in for 424 h, then ozone added to chamber air supply starting at t ¼ 0. Data at top of
each panel were measured in supply air stream. Data in bottom half of each panel were measured in chamber.
ARTICLE IN PRESS
6704
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
expected to reduce gas-phase concentrations of
d-limonene and b-citronellol by about half, and
concentrations of linalool and linalyl acetate by
about two-thirds.
3.3. Ozone concentrations and reactions
Ozone concentration profiles are shown in Fig. 3
for selected experiments. Prior to the introduction
of cleaning products, O3 concentrations in the
chamber were steady at about one-half the level in
the supply air. This relationship, which pertained
throughout the study, was used to calculate, by
material balance, O3 deposition and decomposition
on surfaces in the chamber. The 1.0 h1 O3 loss rate
on surfaces is at the low end of those measured in
residences (Lee et al., 1999), but similar to that
measured in a telephone switching office (Weschler
et al., 1994) and consistent with values expected for
an unfurnished room.
Ozone levels dropped rapidly with application of
the cleaning products. Concentrations decreased to
the minimum observed levels within 10 min of OOD
use in Exp. C (Fig. 3(a)) and within 13 min of POC
use in Exp. H (Fig. 3(b)). Minimum O3 concentrations were steady at 4 ppb from 10 to 60 min after
OOD use in Exp. C and at 6 ppb from 10 to
40 min after POC use in Exp. H. In Exps. F and G,
minimum concentrations were 7.5 ppb and persisted
from 20 to 40 min. Ozone concentrations increased
gradually beginning about 1 h after the cleaning
activity. Concentrations remained below the steady
state level for 10–12 h following product application, indicating continued consumption by reaction
with residual cleaning product constituents. In Exp.
I, O3 concentrations in chamber air started to
decline during the hour before the floor was mopped
with water and reached a low of about 44 ppb at
12 min after mopping (see Fig. 3(c)). No similar
decline in O3 was observed prior to the start of
mopping in Exps. F–H despite similar preparation
activities. The sharper decline following the start of
Exp. I may in part be related to an increase in the
ozone deposition and decomposition rate caused
by increased chamber RH (Grøntoft et al., 2004)
associated with mopping.
The relationship between the quasi-steady O3
concentration in the chamber during the first hour
and the concentration in the supply air was used to
calculate, by material balance, the rate of O3
consumption by reaction following product use.
Steady concentrations of 4, 6, and 7.5 ppb corre-
spond to total O3 consumption rates of 30, 20, and
16 h1. Since uptake on room surfaces (1 h1) is
much slower, most O3 consumption is attributable
to reaction with cleaning product constituents. VOC
concentrations were combined with known or
estimated O3 reaction rates to calculate O3 consumption attributable to gas-phase reactions and to
predict steady O3 concentrations. For example, in
Exp. C (OOD), the average gas-phase concentration
of d-limonene over the 10–60 min interval was
1070 ppb. Combining this with the d-limonene–O3
reaction rate constant of 5.2 106 ppb1 s1
(NIST, 2000) and the AER of 1 h1 yields a
predicted O3 consumption rate (reaction+deposition) of 21 h1 and a steady concentration of
5.5 ppb. A similar calculation for Exp. H (POC)
yields estimates of O3 consumption rates by the five
predominant terpenoids, based on their mean
concentrations over the 10–40 min interval: terpinolene, 17 h1; a-terpinene, 3.6 h1; d-limonene,
3.0 h1; a-terpineol, 2.1 h1; and a-phellandrene,
2.1 h1. An O3 concentration of 4 ppb was predicted
based on the total consumption rate of 28 h1.
Thus, measured O3 concentrations were consistent
with predictions considering loss to be dominated
by homogeneous reactions.
Over the course of the first hour following
cleaning product application, the total mass of O3
consumed was 18–19 mg, which includes most of
the 6.5 mg present initially and about 95% of the
13 mg entering with supply air over this period.
In the AFR experiments, O3 was introduced into
the chamber already containing reactive and nonreactive VOCs. AFR use in the absence of O3
produced reactive VOC concentrations in air that
were 1.5–2 orders of magnitude lower than in
experiments with cleaning products. The AFR
source continued to emit throughout the measurement periods. In Exps. J and K (Exp. K shown in
Fig. 3(d)) after approximately 3 h, the chamber O3
concentrations were steady at 81% and 87% of the
57 ppb expected for the inlet O3 concentrations of
114–115 ppb. The total ozone loss rates from
chamber air were determined by material balance
to be 2.5 h1 for Exp. J and 2.3 h1 for Exp. K.
These rates include air-exchange (1 h1), heterogeneous decomposition and homogeneous reaction.
The O3 consumption and secondary pollutant
formation potential of the AFR product has three
possible components: (1) constituents already present in the air when O3 is introduced, (2) continuously emitted constituents, and (3) a reservoir of
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
constituents sorbed to material surfaces. Compounds already in the air when ozone is introduced
should impact ozone consumption only over the
time scale of air exchange, i.e. for a few hours at
most. Steady-state levels depend on the other
two factors. Their relative importance was assessed
by calculating O3 consumption associated with
homogeneous reactions with VOCs emitted by
AFR. This analysis used the calculated steady-state
concentrations of VOCs in the presence of ozone
(Table 3, column 5) and published or estimated
rate constants for reaction with O3. The following
published rate constants were used: d-limonene,
5.2 106 ppb1 s1 (NIST, 2000); linalool,
1.1 105 ppb1 s1 (NIST, 2000); b-citronellol,
5.9 106 ppb1 s1 (Ham et al., 2006). Based on
the results of Destaillats et al. (2006), we estimated
an ozone reaction rate for linalyl acetate that is
twice the ozone reaction rate for d-limonene. The
reaction rate between ozone and dihydromyrcenol
was estimated at 3 108 ppb1 s1 (Wells, 2005b).
The results of this calculation, shown in Table 3 for
Exp. J, suggest that homogeneous reaction with
identified AFR constituents accounted for O3
consumption rates of 0.2 h1 in Exp. J and 0.1 h1
in Exp. K. The remaining ozone consumption,
1.3 h1 in Exp. J and 1.2 h1 in Exp. K, is assumed
to result from heterogeneous reactions. We suspect
that these rates are higher than the surface decomposition rate without AFR (1.0 h1) because of
ozone reaction with AFR constituents sorbed to
room materials. Note that this result was obtained
6705
for an unfurnished chamber; residential rooms
contain larger quantities of material surfaces and
more plush materials that would be expected to
increase the concentration of sorbed VOC mass.
Heterogeneous reactions between ozone and AFR
constituents could thus be more important in
residences than in our chamber experiments.
3.4. OH radical concentrations and reactions
OH concentrations calculated from the measured
TMB/PCE ratios are summarized in Table 4. The
results indicate significant OH concentrations persisting for 10–12 h following a single cleaning event;
this finding is consistent with the observation that
measurable ozone consumption persists over the
same period. Indeed, in the case of OOD, the OH
concentrations were higher during the periods 2–12
or 6–10 h after the cleaning activity than during the
first 2 h. Although not necessarily intuitive, these
results are consistent with predictions from a model
that captures the essential features of the system
(Destaillats et al., 2006). d-Limonene, through
reaction with O3, is both a source of OH and at
the same time a sink for OH. As the limonene
concentration in the OOD experiment decreases,
the O3 concentration increases. The rate of OH
production does not decline as fast as the rate of
OH removal, with the net effect that the OH
concentration is expected to peak several hours
after the cleaning event ends. The OH concentrations estimated during constant AFR operation
Table 4
Measured TMB/PCE ratios and calculated OH concentrations for selected experiments
Product
Exps.
Time (h)a
n
TMB/PCEb (ppb/ppb)
OH, calc.b,c
(105 molecules cm3)
OOD
C
POC
G and H
AFR
J and K
to0
0–2
2–12
6–10
to0
1–6
8–10
12
to0
1–5
4
6
9
5
4
12
4
3
15
12
1.73570.019
1.70370.008
1.61070.012
1.59070.016
1.69570.014
1.56970.005
1.59870.003
1.65370.016
1.73670.003
1.66170.006
—
1.070.6
3.970.7
4.670.8
—
3.970.5
2.970.4
1.270.6
—
2.270.2
a
Hours since start of experiment. For OOD and POC, experiment started (t ¼ 0) with use of product in chamber already containing
ozone; for AFR, experiment started with introduction of ozone into chamber already containing AFR constituents. Time intervals for
analysis based on availability of data and trends observed in time-resolved measurements.
b
Mean7standard error. TMB concentrations were 0.7 ppb.
c
Calculated from TMB/PCE ratios.
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
were similar to those determined during OOD and
POC use.
In bench-scale experiments with the same products, OH concentrations varied over a wider range
but were centered at the same order of magnitude as
reported here (Destaillats et al., 2006). The reported
OH concentrations also are similar to those
measured in an office in which O3 and d-limonene
were added to achieve steady-state levels of 110
and 60 ppb, respectively (Weschler and Shields,
1997a). In actual indoor environments, OH levels
may be lower because of additional scavenging by
background VOCs.
3.5. Secondary pollutants: very volatile carbonyls
Production of formaldehyde was associated
with the use of each product in the presence of
O3 (Table 5). Relative to experiments without
O3, cleaning product use in the presence of O3
led to increased formaldehyde concentrations of
9–16 ppb and 5–10 ppb over the periods of 0–4 h
and 4–12 h after cleaning. Introduction of O3 during
AFR use resulted in a 6 ppb increase in formaldehyde (2–5 h after introduction of O3). Here, the
condition appeared to approximate steady state,
based on stable concentrations of ozone and
reactive VOCs. Acetone was produced during the
use of POC and AFR with O3. For POC, the
increase over the 0–4 h period was 29 ppb. There
was no clear indication of acetaldehyde production
in any experiment. The production of formaldehyde
and acetone, but not acetaldehyde, is consistent
with the location of the unsaturated carbon–carbon
bonds in the reactive VOCs constituents of these
products.
3.6. Secondary pollutants: particle number and mass
concentrations
Shortly after the introduction of OOD and
POC, particle number concentration started to
increase in several size ranges, with the largest
increase occurring in the smallest measured range
(0.15–0.24 mm). For OOD, particle growth over
time shifted through progressively larger size
ranges in a particle growth wave similar to those
reported previously (Sarwar et al., 2004). POC use
produced substantial particle growth only up to the
0.24–0.36 mm size range. Particle number concentrations were elevated for 8–10 h for OOD and 6 h
for POC.
The evolution of calculated fine particle mass
concentration is presented in Fig. 4 for three
300
PM1.1 concentration (µg m-3)
6706
Exp. C: OOD
Exp. G: POC
Exp. J: AFR
250
200
150
100
50
0
-2
0
2
4
6
time (h)
8
10
12
Fig. 4. Particle mass concentrations resulting from use of
consumer products in presence of ozone; calculated from
measurements of the size-resolved particle number concentration.
Table 5
Measured concentrations (ppb) of very volatile carbonylsa
Product (Exps.)
Ozone
Samples
Time (h)b
Formaldehyde
Acetaldehyde
Acetone
OOD (A and B)
OOD (C)
OOD (C)
POC (D and E)
POC (F–H)
POC (F–H)
Mop w/H2O (I)
AFR (J and K)
AFR (J and K)
No
Yes
Yes
No
Yes
Yes
Yes
No
Yes
n¼6
n¼2
n¼2
n¼8
n¼6
n¼6
n¼4
n¼4
n¼4
0–12
0–4
4–12
0–12
0–4
4–12
0–12
(3)–0
2–5
8.272.2
23.7 (1.3)
17.9 (0.2)
7.371.0
16.071.3
12.072.8
9.870.6
5.470.2
11.271.2
1.070.2
2.2 (0.4)
1.9 (0.1)
1.370.4
1.770.5
1.970.5
2.970.2
0.570.4
0.770.4
1.770.7
3.9 (0.4)
2.5 (0.1)
2.170.5
3171
12.479.8
1.870.4
0.770.8
14.274.1
Multiple experiments under same condition averaged.
a
Values are mean71 standard deviation for n42, mean (absolute deviation ) for n ¼ 2.
b
Time is relative to introduction of product for OOD and POC and introduction of ozone for AFR.
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
6707
Table 6
Secondary fine particles measured during use of consumer products with and without presence of ozone
Product
Exp.
Ozone
Maximum number (cm3)
Maximum PM1.1a (mg m3)
Mean PM1.1a,b (mg m3)
OOD
OOD
OOD
POC
POC
POC
POC
POC
H2O, Mop
AFR
AFR
AFR
AFR
A
B
C
D
E
F
G
H
I
J
J
K
K
No
No
Yes
No
No
Yes
Yes
Yes
Yes
No
Yes
No
Yes
240
520
44,000
620
540
38,000
35,000
36,000
840
290
1550
230
750
2.1
3.3
280
4.5
2.9
138
131
132
5.1
1.8
6.9
1.6
4.0
1.0
1.6
89
3.5
2.4
36
31
34
3.5
1.4
4.8
1.2
3.0
Mass of particles with aerodynamic diameters o1.1 mm. Mass concentrations calculated from size-resolved particle number
concentrations (see text for details).
b
For OOD and POC mean values calculated for 12 h starting with product use (Exps. A–H). For AFR, mean values for ‘‘No’’ ozone
condition calculated for 5 h before ozone introduced; mean values for ‘‘Yes’’ ozone condition calculated for 5 h after start of ozone.
a
experiments. For each cleaning product experiment,
mean PM1.1 concentrations were calculated for 12-h
periods starting with product use; for AFR experiments, mean PM1.1 concentrations were calculated
at steady levels before and after O3 addition
(Table 6). Peak PM1.1 concentrations were estimated to be about 275 mg m3 for OOD and
135 mg m3 for POC. In contrast to the cleaning
products, use of the AFR in the presence of O3
produced much smaller quantities of PM1.1
(2–5 mg m3 increases, Table 6). Particle concentrations measured in these experiments are generally
similar to those reported for experiments conducted
in other large chambers and in indoor environments
(Fan et al., 2003; Long et al., 2000; Sarwar et al.,
2004).
3.7. Relevance to residential scenarios and exposures
The protocols and conditions employed in this
study were designed to be relevant to product use
under
typical
residential
conditions.
The
114–120 ppb O3 in the chamber inlet air, which is
analogous to outdoor air entering a residence, is
within a factor of two of outdoor levels commonly
occurring during warm, sunny conditions in US
urban areas (EPA, 2004). The steady-state level of
O3 in the chamber without added terpenoids is
within the range of reported indoor values (Weschler, 2000), albeit at the higher end of this range. At
lower but still substantial outdoor ozone concentra-
tions (e.g., 40–80 ppb), the basic chemistry observed
in this study would still occur. When ozone is the
limiting reagent (e.g., just after use of products
similar to OOD or POC), lower ozone levels will
produce lower concentrations of secondary pollutants. Lower ozone levels will have less impact for
secondary pollutants associated with the AFR since
ozone is not limiting in this case.
The use of a chamber finished with painted
wallboard allowed for sorption to a material that
is ubiquitous in US indoor environments. Owing to
the use of appropriately scaled cleaning surface area
to room volume ratios and realistic application
protocols, concentrations measured at the start of
OOD and POC experiments are expected to be
relevant to product use in residences. The AER of
1 h1 corresponds to the 80th percentile of an
empirical distribution reported for US detached
residences across all seasons (Murray and Burmaster, 1995); use of a higher than average value
may be especially appropriate for cleaning events
during which windows may be opened. Use of a
single air freshener in the 50-m3 chamber approximates scenarios in which multiple plug-in devices
are used in a larger residence or a single device is
used in an isolated room or a small apartment. At
this moderate rate of air freshener use, although
airborne concentrations of reactive terpenoids were
1.5–2 orders of magnitude lower than those resulting from cleaning product use, reactive chemistry
was still measurable.
ARTICLE IN PRESS
6708
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
The increase in formaldehyde concentrations
resulting from consumer product use may be
evaluated in the context of exposure guidelines.
The State of California has established non-cancer
reference exposure limits (RELs) of 68 ppb for acute
(1 h) and 2.2 ppb for chronic (10 yr or more)
exposures to formaldehyde (OEHHA, 2000), and
an interim 8-h REL of 27 ppb based on the acute
value (CARB, 2005). California’s no significant risk
level (o1:100,000) for cancer is a 70-year intake rate
of 40 mg day1 (OEHHA, 2005). Formaldehyde
increments of 6–12 ppb averaged over 12 h following
use of a single cleaning product constitute about
20–40% of the 8-h REL that was set to protect
against irritancy. When this is combined with other
indoor sources and formaldehyde transported from
outdoors, such increments may increase the frequency or extent of exceedences of the 8-h REL.
Chronic exposures scale with the frequency of
activity. A single event at the formaldehyde levels
found in our study would contribute roughly
20–50% of the weekly average exposure or intake
allowable under California’s non-cancer and cancer
chronic exposure guidelines. More frequent cleaning
activity—e.g., by professional house cleaners and
fastidious homemakers—could lead to exposures
that exceed the chronic levels on a weekly averaged
basis. On the other hand, the levels of secondary
pollutants from ozone–terpene chemistry are expected to correlate with indoor ozone levels,
whether from ambient air or indoor sources. A
thorough exposure assessment must consider the
influence of the spatial variability, and diurnal and
seasonal cycles of ambient ozone on secondary
pollutant formation.
In contrast to episodic use of cleaning products,
AFR use represents a continuous emissions source
and therefore presents a different set of exposure
considerations. The incremental 6 ppb of formaldehyde measured during steady use of the air freshener
in the presence of 50 ppb of residual ozone is about
25% of California’s 8-h REL and exceeds the 2 ppb
REL for chronic exposure.
Table 6 shows that use of terpenoid-containing
cleaning products in the presence of ozone can cause
substantial increases in the concentrations of fine
particles. Epidemiological studies have shown associations between increases in ambient fine-particle
concentrations and mortality and morbidity (Pope
et al., 2002). However, the relative toxicity of
particles measured at outdoor monitoring stations
versus that of secondary organic aerosol derived
from ozone–terpenoid reactions has yet to be
determined.
4. Conclusions
This study adds to an emerging body of evidence
that use of terpenoid-containing cleaning products
or air fresheners, combined with indoor ozone,
produces substantial levels of secondary air pollutants to which occupants may be exposed. Specifically, both formaldehyde and fine particulate mass
are generated in quantities that may result in
exposures under some circumstances that are
substantial in relation to health-based standards
and guidelines. Also, chemistry involving both
ozone and OH generates ‘‘stealth’’ products that
cannot be measured with the methods employed
here, but which are suspected to cause sensory
irritation (Weschler and Shields, 1997b; Weschler,
2004; Wolkoff et al., 1997) and may have other
adverse effects.
Sorption of reactive terpenoids onto material
surfaces can delay their removal by ventilation and
extend the time over which they are available for
indoor reaction with ozone. Thus, exposures to
secondary products can persist for hours following a
cleaning event. Air fresheners emit terpenoids at
rates that produce substantially lower concentrations than those that occur after use of a cleaning
product containing these compounds. However,
since plug-in air fresheners emit terpenoids continuously, their use may cause more chronic
exposures to secondary pollutants. Sorption of
AFR constituents to material surfaces appears to
cause heterogeneous reactions that accounted for
half or more of the additional ozone reactivity in
our chamber experiments and could be even more
important in fully furnished indoor environments.
The results of these experiments and our related
work (Destaillats et al., 2006; Nazaroff and
Weschler, 2004; Singer et al., 2006) should be
considered in the context of the hygienic and
psychological benefits of cleaning activities and the
perceptual benefits that some experience from airfreshener use. Our results suggest that consumers
should be cognizant of exposures to primary and
secondary air pollutants that can result from
product use and consider appropriate opportunities
for mitigation. Exposures to secondary pollutants
formed by reactions of ozone with terpenoid
constituents of consumer products can be reduced
by means of the following simple measures: (1) use
ARTICLE IN PRESS
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
products with lower concentrations of ozonereactive constituents, (2) use products in dilute form
whenever appropriate, (3) ensure adequate ventilation during and for several hours following cleaning, (4) clean during periods of low occupancy and
allow adequate time before the space is occupied by
sensitive individuals, (5) rinse surfaces with water
following product use, (6) promptly remove cleaning supplies (e.g., paper towels, sponges, mops)
from the occupied area, and (7) rinse sponges and
mops before storing. Also, use of indoor ozone
generators or ionizing air cleaners should be
avoided during and following use of cleaning
products and in the presence of air fresheners
containing ozone-reactive constituents.
Acknowledgments
The authors gratefully acknowledge technical
assistance provided by Toshifumi Hotchi and Doug
Sullivan of LBNL. We thank Rich Sextro of LBNL,
and Dorothy Shimer with colleagues from CARB
for their comments on the draft manuscript.
This work was funded by the California Air
Resources Board Contract no. 01-336. The statements and conclusions in this report are those of the
researchers and not necessarily those of the California ARB. The mention of commercial products,
their source, or their use in connection with material
reported herein is not to be construed as actual or
implied endorsement of such products. All work at
LBNL was conducted under US DOE Contract no.
DE-AC02-05CH11231.
References
Boeniger, M.F., 1995. Use of ozone generating devices to
improve indoor air-quality. American Industrial Hygiene
Association Journal 56, 590–598.
Britigan, N., Alshawa, A., Nizkorodov, S.A., 2006. Quantification of ozone levels in indoor environments generated by
ionization and ozonolysis air purifiers. Journal of the Air &
Waste Management Association 56, 601–610.
Calogirou, A., Larsen, B.R., Brussol, C., Duane, M., Kotzias, D.,
1996. Decomposition of terpenes by ozone during sampling
on Tenax. Analytical Chemistry 68, 1499–1508.
CARB, 2005. Indoor Air Pollution in California—Report to the
California Legislature. Principal authors: Shimer, D., Phillips,
T.J., Jenkins, P.L., California Air Resources Board,
July 2005. /http://www.arb.ca.gov/research/indoor/ab1173/
finalreport.htmS.
Destaillats, H., Lunden, M.M., Singer, B.C., Coleman, B.K.,
Hodgson, A.T., Weschler, C.J., Nazaroff, W.W., 2006.
Indoor secondary pollutants from household product emis-
6709
sions in the presence of ozone. A bench-scale chamber study.
Environmental Science & Technology, 40, 4421–4428.
EPA, 2004. The ozone report—measuring progress through 2003.
Report No. EPA 454/K-04-001, US Environmental Protection Agency. /http://www.epa.gov/airtrends/ozone.htmlS.
Fan, Z.H., Lioy, P., Weschler, C., Fiedler, N., Kipen, H., Zhang,
J.F., 2003. Ozone-initiated reactions with mixtures of volatile
organic compounds under simulated indoor conditions.
Environmental Science & Technology 37, 1811–1821.
Fick, J., Pommer, L., Andersson, B., Nilsson, C., 2001. Ozone
removal in the sampling of parts per billion levels of terpenoid
compounds: an evaluation of different scrubber materials.
Environmental Science & Technology 35, 1458–1462.
Fick, J., Pommer, L., Astrand, A., Ostin, R., Nilsson, C.,
Andersson, B., 2005. Ozonolysis of monoterpenes in mechanical ventilation systems. Atmospheric Environment 39,
6315–6325.
Grøntoft, T., Henriksen, J.F., Seip, H.M., 2004. The humidity
dependence of ozone deposition onto a variety of building
surfaces. Atmospheric Environment 38, 59–68.
Ham, J.E., Proper, S.P., Wells, J.R., 2006. Gas-phase chemistry
of citronellol with ozone and OH radical: rate constants and
products. Atmospheric Environment 40, 726–735.
Hand, J.L., Kreidenweis, S.M., 2001. A new technique for
obtaining aerosol size distributions with applications to
estimates of aerosol properties. CIRA Technical Paper No.
0737-5352-49, Colorado State University, Fort Collins, CO.
Hubbard, H.F., Coleman, B.K., Sarwar, G., Corsi, R.L., 2005.
The effects of an ozone generating air purifier on indoor
secondary particles in three residential dwellings. Indoor Air
15, 432–444.
Lee, K., Vallarino, J., Dumyahn, T., Ozkaynak, H., Spengler,
J.D., 1999. Ozone decay rates in residences. Journal of the Air
& Waste Management Association 49, 1238–1244.
Lee, S.C., Lam, S., Fai, H.K., 2001. Characterization of VOCs,
ozone, and PM10 emissions from office equipment in an
environmental chamber. Building and Environment 36,
837–842.
Leovic, K.W., Sheldon, L.S., Whitaker, D.A., Hetes, R.G.,
Calcagni, J.A., Baskir, J.N., 1996. Measurement of indoor air
emissions from dry-process photocopy machines. Journal of
the Air & Waste Management Association 46, 821–829.
Li, T.H., Turpin, B.J., Shields, H.C., Weschler, C.J., 2002. Indoor
hydrogen peroxide derived from ozone/d-limonene reactions.
Environmental Science & Technology 36, 3295–3302.
Liu, X.Y., Mason, M., Krebs, K., Sparks, L., 2004. Full-scale
chamber investigation and simulation of air freshener
emissions in the presence of ozone. Environmental Science
& Technology 38, 2802–2812.
Long, C.M., Suh, H.H., Koutrakis, P., 2000. Characterization of
indoor particle sources using continuous mass and size
monitors. Journal of the Air & Waste Management Association 50, 1236–1250.
Murray, D.M., Burmaster, D.E., 1995. Residential air exchange
rates in the United States: empirical and estimated parametric
distributions by season and climatic region. Risk Analysis 15,
455–469.
Nazaroff, W.W., Weschler, C.J., 2004. Cleaning products and air
fresheners: exposure to primary and secondary air pollutants.
Atmospheric Environment 38, 2841–2865.
NIST, 2000. Chemical Kinetics Database. Standard Reference
Database 17, Version 7.0, Release 1.3. National Institute of
ARTICLE IN PRESS
6710
B.C. Singer et al. / Atmospheric Environment 40 (2006) 6696–6710
Standards and Technology. /http://kinetics.nist.gov/index.
phpS.
Niu, J.L., Tung, T.C.W., Burnett, J., 2001. Ozone emission rate
testing and ranking method using environmental chamber.
Atmospheric Environment 35, 2143–2151.
OEHHA, 2000. Air toxics hot spots risk assessment guidelines.
Office of Environmental Health Hazard Assessment. /http://
www.oehha.org/air/hot_spots/index.htmlS.
OEHHA, 2005. Proposition 65 safe harbor levels: no significant
risk levels for carcinogens and maximum allowable dose
levels for chemicals causing reproductive toxicity. Reproductive and Cancer Hazard Assessment Branch, Office of
Environmental Health Hazard Assessment, California EPA,
August 2005.
Phillips, T.J., Bloudoff, D.P., Jenkins, P.L., Stroud, K.R., 1999.
Ozone emissions from a ‘‘personal air purifier’’. Journal of
Exposure Analysis and Environmental Epidemiology 9,
594–601.
Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D.,
Ito, K., Thurston, G.D., 2002. Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air
pollution. Journal of the American Medical Association 287,
1132–1141.
Rohr, A.C., Weschler, C.J., Koutrakis, P., Spengler, J.D., 2003.
Generation and quantification of ultrafine particles through
terpene/ozone reaction in a chamber setting. Aerosol Science
and Technology 37, 65–78.
Sarwar, G., Corsi, R., Allen, D., Weschler, C., 2003. The
significance of secondary organic aerosol formation and
growth in buildings: experimental and computational evidence. Atmospheric Environment 37, 1365–1381.
Sarwar, G., Olson, D.A., Corsi, R.L., Weschler, C.J., 2004.
Indoor fine particles: the role of terpene emissions from
consumer products. Journal of the Air & Waste Management
Association 54, 367–377.
Singer, B.C., Revzan, K.L., Hotchi, T., Hodgson, A.T., Brown,
N.J., 2004. Sorption of organic gases in a furnished room.
Atmospheric Environment 38, 2483–2494.
Singer, B.C., Destaillats, H., Hodgson, A.T., Nazaroff, W.W.,
2006. Cleaning products and air fresheners: emissions and
resulting concentrations of glycol ethers and terpenoids.
Indoor Air 16, 179–191.
Turpin, B.J., Lim, H.J., 2001. Species contribution to PM2.5 mass
concentrations: revisiting common assumptions for estimating organic mass. Aerosol Science and Technology 35,
602–610.
Wainman, T., Zhang, J.F., Weschler, C.J., Lioy, P.J., 2000.
Ozone and limonene in indoor air: a source of submicron
particle exposure. Environmental Health Perspectives 108,
1139–1145.
Wells, J.R., 2005a. Gas-phase chemistry of alpha-terpineol with
ozone and OH radical: rate constants and products.
Environmental Science & Technology 39, 6937–6943.
Wells, J.R., 2005b. Personal Communication. Exposure
Assessment Branch, Health Effects Laboratory Division,
National Institute of Occupational Safety and Health,
Morgantown, WV.
Weschler, C.J., 2000. Ozone in indoor environments: concentration and chemistry. Indoor Air 10, 269–288.
Weschler, C.J., 2004. New directions: ozone-initiated reaction
products indoors may be more harmful than ozone itself.
Atmospheric Environment 38, 5715–5716.
Weschler, C.J., Shields, H.C., 1997a. Measurements of the
hydroxyl radical in a manipulated but realistic indoor
environment. Environmental Science & Technology 31,
3719–3722.
Weschler, C.J., Shields, H.C., 1997b. Potential reactions among
indoor pollutants. Atmospheric Environment 31, 3487–3495.
Weschler, C.J., Shields, H.C., 1999. Indoor ozone/terpene
reactions as a source of indoor particles. Atmospheric
Environment 33, 2301–2312.
Weschler, C.J., Shields, H.C., 2003. Experiments probing the
influence of air exchange rates on secondary organic aerosols
derived from indoor chemistry. Atmospheric Environment 37,
5621–5631.
Weschler, C.J., Shields, H.C., Naik, D.V., 1994. Indoor chemistry
involving O3, NO, and NO2 as evidenced by 14 months of
measurements at a site in southern California. Environmental
Science & Technology 28, 2120–2132.
Wolkoff, P., Clausen, P.A., Jensen, B., Nielsen, G.D., Wilkins,
C.K., 1997. Are we measuring the relevant indoor pollutants?
Indoor Air 7, 92–106.
Wolkoff, P., Schneider, T., Kildeso, J., Degerth, R., Jaroszewski,
M., Schunk, H., 1998. Risk in cleaning: chemical and physical
exposure. Science of the Total Environment 215, 135–156.
Wolkoff, P., Clausen, P.A., Wilkins, C.K., Nielsen, G.D., 2000.
Formation of strong airway irritants in terpene/ozone
mixtures. Indoor Air 10, 82–91.
Wolkoff, P., Wilkins, C.K., Clausen, P.A., Nielsen, G.D., 2006.
Organic compounds in office environments—sensory irritation, odor, measurements and the role of reactive chemistry.
Indoor Air 16, 7–19.