Degradation of the adsorbent tenax TA by nitrogen oxides, ozone, hydrogen peroxide, OH radical, and limonene oxidation products

Article abstract of DOI:10.1021/es025680f

The degradation of the adsorbent Tenax TA was studied qualitatively by sampling oxidants common in indoor air followed by thermal desorption and gas chromatography. A total of 25 degradation products were identified. Several degradation products not reported previously were identified: 9 for nitrogen dioxide; 11 for ozone; 2 for hydrogen peroxide; 12 for hydroxyl radical; 1 for ozonelimonene mixtures, but none for nitrogen oxide. Whereas ozone shows a complex degradation of the adsorbent, hydrogen peroxide and limonene-ozone mixtures show few products. Nitrogen dioxide and the hydroxyl radical behave almost identically and produce 2,6-diphenyl-p-benzoquinone as the major degradation product. Reactant specific degradation products were identified for ozone (11) and nitrogen dioxide (1).

Full text of DOI:10.1021/es025680f

Environ. Sci. Technol. 2002, 36, 4121-4126
potential indicators of irritating substances in air. The
Degradation of the Adsorbent Tenax
TA by Nitrogen Oxides, Ozone,
Hydrogen Peroxide, OH Radical, and
Limonene Oxidation Products
application of such indicators, however, assumes either
specificity and/ or a qualitative and quantitative knowledge
of the species capable of forming them.
The purpose of this work was 3-fold: (1) to identify Tenax
degradation products from limonene-O3 mixtures and
determine their oxidant specificity [Limonene was chosen
because of its rapid reaction with O3 and its relatively high
concentration and frequency of occurrence indoors (12).];
(2) to test hydrogen peroxide (H2O2) and the OH radical as
potential products responsible for the degradation caused
by limonene-O3 mixtures; and (3) to scrutinize the degrada-
tion products of Tenax caused by the common oxidants in
indoor air, nitrogen oxide (NO), NO2, and O3.
J A C O B G . K L E N Ø , P E D E R W O L K O F F , *
P E R A . C L A U S E N ,
C O R N E L I U S K . W I L K I N S , A N D
†
T H O R V A L D P E D E R S E N
Department of Indoor Climate, National Institute of
Occupational Health, Lersø Parkallø 105,
DK-2100 Copenhagen Ø, Denmark
Experimental Section
Chem icals. Oxygen (99.999%), NO, NO2, and N2 (O2 + H2O
< 2 ppm) were from Hydrogas, Norway. Acetophenone
(>99.5%), benzoic anhydride (purum), decanal (97%), diphe-
nylethanedione (98%), nonanal (97%), and phthalic anhy-
dride (for syntheses) were from Fluka. Benzaldehyde
(>99.5%), benzoic acid (>99.9%), hydrogen peroxide (35%
in H2O), methanol (>99%), and phenol (>99.5%) were from
Merck. Benzophenone (g99.0%), 2,6-diphenylphenol (98%),
phenyl acetaldehyde (>90%), and phenylglyoxal hydrate
(97%) were from Aldrich. p-Hydroquinone was from Kebolab,
Denmark. Phenylglyoxylic acid (>98%) was from Sigma.
Phenylmaleic anhydride (97%) and 1,2-diphenylethanone
(97%) were from Lancaster Synthesis. 2,6-Diphenyl-p-ben-
zoquinone (DPQ) and 2,6-diphenyl-p-hydroquinone (DPHQ)
were synthesized (10), and 2,4-diphenyl-4-cyclopentene-1,3-
dione was prepared according to Ruggli and Schmidlin (13).
Adsorbent tubes were Perkin-Elmer stainless steel tubes
(length 31/ 2", outer diameter 1/ 4") packed with fresh and
unused 200 ( 1 mg Tenax TA 60-80 mesh (Chrompack, The
Netherlands). Tenax TA will be called Tenax throughout, if
not otherwise stated. Tedlar sample bags were from SKC
Inc., PA, U.S.A.; the volume was 25 L in all experiments.
Instrum ents. Tenax samples were analyzed by thermal
desorption-gas chromatography. A mass spectrometer (TD-
GC-MS) was used for identification and a flame ionization
detector (TD-GC-FID) for quantification. The TD apparatus
was a Perkin-Elmer ATD 400 with a plug of silanized glass
wool only in a narrow/ wide bore cold trap. The conditions
were as follows: cold trap temperature -30 °C, outlet split
flow 30 mL/ min, desorption flow 46 mL/ min, desorption
time 20 min, transfer line and valve temperature 225 °C and
desorption temperature of the adsorbent tube and the cold
trap 300 °C. The outlet split was 1:16 due to the high
desorption flow in the thermal desorber to get the higher
boiling compounds through the system. Two different gas
chromatographs were used: a Perkin-Elmer Auto system GC
and a Hewlet Packard 5890 series II. Both were equipped
with a 50 m Chrompack CP-Sil 8 CB low-bleed MS (5% phenyl
methyl silicone) analytical column (i.d. 0.25 mm, 0.12 µm
film thickness). The column was directly connected to the
detector and to the thermal desorber. The GC oven program
was 35-280 °C at 5 °C/ min and 16 min at 280 °C. The mass
spectrometers were a Kratos Profile double focusing sector
instrument (scan rate 0.4 per s, mass range m/ z 29-432) and
a Perkin-Elmer Turbomass (scan rate 0.5 per s, mass range
m/ z 30-400). Criteria for positive qualitative identification
was manual inspection of each mass spectrum combined
with library search (Wiley) and retention time identification
with an authentic standard within tR ( 0.03 min.
The degradation of the adsorbent Tenax TA was studied
qualitatively by sampling oxidants common in indoor
air followed by thermal desorption and gas chromatography.
A total of 25 degradation products were identified.
Several degradation products not reported previously
were identified: 9 for nitrogen dioxide; 11 for ozone; 2 for
hydrogen peroxide; 12 for hydroxyl radical; 1 for ozone-
limonene mixtures, but none for nitrogen oxide. Whereas
ozone shows a complex degradation of the adsorbent,
hydrogen peroxide and limonene-ozone mixtures show
few products. Nitrogen dioxide and the hydroxyl radical
behave almost identically and produce 2,6-diphenyl-p-
benzoquinone as the major degradation product. Reactant
specific degradation products were identified for ozone
(11) and nitrogen dioxide (1).
Introduction
Tenax is a synthetic polymer consisting of 2,6-diphenyl-p-
phenylene ether units used for sampling of volatile organic
compounds with boiling points above approximately 50 °C.
It is hydrophobic, thermally stable in the absence of O2, and
has a low chromatographic background (1, 2). However, the
application of adsorbents in general involves the risk of
artifact formation. For example, terpenes may rearrange as
a result of acid catalysis on the adsorbent (3). Ozone (O3)
reacts with adsorbed terpenes to give increased amounts of
oxidation products at the expense of the terpenes. This
oxidative loss is pronounced for e.g. myrcene and limonene,
whereas slowly reacting terpenes are practically unaffected
by O3 (4, 5). Nitrogen oxides and O3 are reported to degrade
Tenax (6-10). Nitrogen dioxide (NO2) mainly forms 2,6-
diphenyl-p-benzoquinone (DPQ) and 2,6-diphenyl-p-hyd-
roquinone (DPHQ), whereas the degradation pattern of O3
is complex and includes the frequently reported benzalde-
hyde and acetophenone (1, 9, 10). Reaction mixtures of
limonene and O3, which contain no residual O3, also degrade
Tenax into DPHQ (10). These mixtures are known to be strong
airway irritants (11). The reactive compounds that degrade
Tenax in these mixtures could reasonably be identical to the
strong irritants. Thus Tenax degradation products could be
* Corresponding author phone: (+45) 39 16 52 72; fax: (+45) 39
16 52 01; e-mail: pwo@ami.dk.
† Department of Chemistry, University of Copenhagen, Univer-
sitetsparken 5, DK-2100 Copenhagen Ø, Denmark.
9
10.1021/es025680f CCC: $22.00
Published on Web 08/24/2002
2002 Am erican Chem ical Society
VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 1 2 1

B: Tenax Exposure to Nitrogen Dioxide. A 150 ppm NO2
certified standard (( 5%, in N2) was diluted 1:109 with
humidified N2 to a final concentration of 1.4 ppm. The
degradation products were quantified by FID, see Table 2.
Blank samples were prepared from pure N2. In these samples,
the relative humidity was <1% for practical reasons. There
was no qualitative difference between dry and humid
samples.
Tenax Exposure to Nitrogen Oxide. A 4.90 ppm NO
certified standard (( 2%, in N2) was diluted 1:3.5 with
humidified N2 to 1.4 ppm. The blank samples for NO2 served
as blanks for NO. The same tubes were used for subsequent
NO and NO2 experiments.
C: Tenax Exposure to the OH Radical and Hydrogen
Peroxide. The OH radical was generated by photolysis of
H2O2. N2 was bubbled through an impinger with a 35% H2O2
solution in H2O, and the vapor mixture was led through the
same lamp housing described in section D with the lamp
power turned on. The production of OH radicals was expected
to be substantial considering the high concentration of H2O2
(cf. ref 17) and the significant absorption at 185 nm according
to ref 18. The blank samples for H2O2 served as blanks for
the OH radical.
TABLE 1. Sample Matrix with Experimental Conditions
sample
vol. (L)
no. of
samples
concn
(ppb)
gas type
RH %
a
NO2
NO2
5
9
20
20
5
9
20
144
20
3
1
1
1
3
1
3
1
3
2
3
1
3
2
1
1
5
1
1
1
1400
1400
0
0
52
54
<1
<1
54
<1
43
87
43
87
87
87
51
<1
56
<1
49
53
56
<1
b
blanka
blankb
NOa
1400
NOb
1400
OHa
unknownc
OHb
unknownc
a
H2O2
23000
46000
0
0
50
1000
b
H2O2
336, 532
blanka
blankb
570
397
20
2
20
5
20
20
20
5
a
O3
O3
b
blanka
0
0
blankb
a
lim onene-O3
lim onene-O3
blanka
1700/203d
1700/203d
b
0
0
blankb
a
b
c
d
H2O2 vapor (23 ppm) was generated with the lamp turned
off. Blank samples were prepared as the samples using water
in the impinger instead of the H2O2 solution.
TD-GC-FID. TD-GC-MS. Expected to be negligible. Initial con-
centrations before m ixing.
D: Tenax Exposure to Ozone and Lim onene-Ozone
Mixtures. O3 was generated photochemically in pure O2 (16)
with a mercury lamp (185 and 254 nm) in a thermo-stated
lamp housing controlled by a high performance variable
power supply (11). The VOC generator was turned off (see
below). Humidified N2, O2, and O3 were mixed in a T-union
and introduced into a 103 cm long Teflon tube (i.d. 2.0 mm).
Prior to sampling, the O3 concentration was adjusted by
variation of the lamp current. The O3 concentration was
measured before and after sampling to ensure the stability
of the O3 generation. In all quantitative experiments mixing
ratios of O2 and N2 were 0.21 and 0.79, respectively. Blank
samples were prepared under the same conditions as above,
but with the lamp power turned off. Samples for MS analysis
were made in pure O2 at RH <1% for practical reasons. No
qualitative differences were observed as a result of these
modifications.
The setup was slightly modified, when the limonene-O3
mixture was generated. First, limonene in N2 (16.0 L) was
introduced into a Tedlar bag by use of a VOC generator (Model
350, Analytical Instrument Development, Inc., PA) that was
adjusted to produce 1.7 (0.3 ppm limonene; this was checked
by sampling on Tenax (0.400 L). Second, O3/ O2 (4.2 L) was
introduced into the same Tedlar bag in 4.2 min by adjusting
the O3 generator to produce 203 ( 3 ppb O3 in the final
mixture. Third, the Tedlar bag was mechanically mixed every
5 min in the following 30 min from the onset of introduction
of O3/ O2, whereafter the residual O3 concentration was <5
ppb. Immediately thereafter, the bag air was sampled on
Tenax for 100 min. The mean age of the sampled air was
hence between 30 and 130 min. The concentration of the
hydroxyl radical is considered negligible because of its fast
reaction with excess limonene after complete consumption
of the O3. The blank samples were made from pure N2
introduced into the bag and treated the same way as the
samples.
The adsorbent tubes were cleaned in a stream of 90 mL/
min pure N2 at 300 °C for 180 min and 340 °C for 30 min
using a sample tube conditioning apparatus (TC-20, Markes
International, England). The purity of each tube was verified
with TD-GC-FID prior to use and accepted if the area of
unknown peaks were less than 100 µV‚s, which corresponds
to ca. 1 ng of toluene. The performance of the TD-GC-FID
system was validated with high and low controls (14). The
peak area of the degradation products in µV‚s was integrated
automatically or manually. Peak areas of degradation prod-
ucts were taken as a measure of the relative abundance (FID
peak area percent of all identified degradation products for
each oxidant) and reported as a mean of at least three
samples. The sampling flow on Tenax was 0.200 L/ min, and
the sampling volumes for all experiments are listed in Table
1.
A photometric analyzer (API model 400, San Diego, CA)
interfaced to a PC was used to monitor O3; the analyzer was
6-point calibrated with an external O3 source. A chemilu-
minescence analyzer for O3 (AID 560, Analytical Instruments
Development, Inc.) was used to measure O3 in the limonene-
O3 mixture. Similarly, a chemiluminescence analyzer for NO
and NO2 (API model 200A, San Diego, CA) was used. The
sampling flows of the analyzers were ca. 800 and 500 mL/
min, respectively. H2O2 was sampled (ca. 5 L) in duplicate
on dedicated sampling tubes and analyzed by HPLC (15).
A: Sam pling of Reactive Gases on Tenax. The experi-
mental setup in Figure 1 shows the generation of reactive
gases in sections B (NO or NO2), C (OH or H2O2), D (O3 or
O3-limonene), and a common section A, where individual
gases were humidified and pumped through an adsorbent
tube with Tenax. Combining a dry and a wet N2 stream
controlled the relative humidity. All gas flows were adjusted
by flow controllers and measured before and after sampling
by a flow meter (Gillian Instrument Corporation NJ). The
humidified gas was mixed in a T-union and introduced into
a 103 cm long Teflon tube (i.d. 2.0 mm). The gas was sampled
on adsorbent tubes at 200 ( 3 mL/ min using Alpha-1 pumps
(Ametek, PA). The residual flow was led to exhaust. The
relative humidity was measured by a Testo 601 hygrometer
(Testoterm GmbH & Co., Germany). The temperature was
ca. 21 °C and the relative humidity ca. 50% in all quantitative
experiments, see Table 1 for additional data.
Results and Discussion
The TD-GC-FID chromatograms of the cleaned adsorbent
tubes showed no degradation products but 2,6-diphen-
ylphenol (U), see Figure 2f. Traces of more volatile degrada-
tion products such as phenol, benzaldehyde, and acetophe-
none are often encountered on cleaned Tenax tubes (2).
However, in this study an empty cold trap and the split flow
9
4 1 2 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

FIGURE 1. Experimental setup.
in the TD system was 1:16 in favor of the higher boiling
compounds. 2,6-Diphenylphenol appeared as a trace in all
blanks. Tenax is formed from oxidative coupling of this
compound, and its presence is presumably due to unpoly-
merized material (6). No other products in Table 2 were
detected on the blanks. Several new degradation products
have been identified in this study compared to a previous
one (10). The major cause is a better chromatographic
resolution by the use of a narrower GC column resulting in
the separation of coeluting GC peaks, a more sensitive MS
unit, and the extensive use of authentic standards.
The diphenyl hydroquinoid structure of Tenax is reflected
in the structures of most degradation products, which include
aromatic aldehydes, ketones, phenols, acids, acid anhydrides,
pyranes, quinones, and hydroquinones. Hitherto not reported
degradation products were identified for NO2: 9, O3: 11,
H2O2: 2, OH radical: 12, and limonene-O3 mixture: 1, but
none was detected for NO. Their identities and relative
abundances are shown in Table 2. The aldehydes I and K
were not quantified, because they are not considered
degradation products, since they have no resemblance to
the Tenax structure. Their origin is unclear, one source could
be certain impurities retained in the Tenax, despite cleaning.
Acetic acid (B) was not found as a NO2-Tenax degradation
product at low concentrations. Formic acid (A) was detected
with MS, but not with FID, because of the low detector
response of this compound. All compounds were identified
by MS in addition to GC retention time criteria, except for
T, W, and Z (Table 2). A few minor GC peaks related to
degradation with retention times greater than 45 min remain
unidentified.
Nitrogen Dioxide and Nitrogen Oxide. The product
formation in a gas sample of 1.4 ppm NO2 is considerably
higher than typically encountered in ambient air (see Figure
2a). This was necessary in order to identify and properly
quantify the degradation products formed in trace amounts.
The degradation of Tenax by NO2 is dominated by DPQ and
DPHQ. The two compounds are interconverted in the
analytical system, and they are considered a single compound
of 98% abundance (10). Nine compounds have not previously
been reported, see Table 2. They each make up less than
0.5% of the total abundance. NO2 levels are highly variable,
9
VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 1 2 3

TABLE 2. Relative Abundance of Tenax Degradation Products (FID Peak Area Percent) from Sampling of NO, NO , O, HO , OH
2
3
2 2
Radical, and Limonene-O Mixtures
3
.
tR (min)
reactant gases/degradation productsa
NO
NO
OH
H O
O
3
limonene-O
3
identificationc tR/MSe
2
2
2
3.46
4.01
A: form ic acid
B: acetic acid
C: benzaldehyde
D: phenol
E: benzonitrile
F: phenyl acetaldehyde
G: acetophenone
H: phenylglyoxal
I: nonanal
nqb,f
nqb,f
0.4b
nqb,f
0.4b
4b
nqb,f
1b
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+/+
+ /+
+/+
+/+
/+
13.17
13.73
13.98
15.92
16.68
16.95
17.74
20.66
21.02
23.39
23.55
24.80
30.98
33.49
35.97
37.89
40.05
40.78
46.65
47.19
nqb,f 11
5
0.06b
0.1b,d
0.8b
0.9d
0.1b
1b
8
2b,d
nqf
46
J benzoic acid
K: decanal
0.5b
6b
nqb,f
nqf
nqb,d,f
nqb,d,f
1b
L: p-hydroquinone
M: phenyl glyoxylic acid
N: phthalic anhydride
O: phenylm aleic anhydride
P: benzophenone
Q: 1,2-diphenylethanone
R: diphenylethanedione
S: benzoic anhydride
T: diphenyl propanetrione
U: 2,6-diphenylphenol
V: 2,4-diphenyl-4-cyclopentene-
1,3-dione
W: unknown arom atic ketone
X: 2,6-diphenyl-p-benzoquinone
(DPQ)
Y: 2,6-diphenyl-p-hydroquinone (DPHQ)
Z: 2,6-diphenyl-4H-pyran-4-one
0.02b
0.3
0.4b
0.4b
5
0.6b,d
0.5b
0.4b,d
0.1d
0.5d
3
0.04b
0.1
nqf 2b
nqb,f
13b
+/+
+/+
2d
48.42
49.71
5d
93
5
83b
2b
nqd,f
+/+
53.40
53.05
8b
87
+/+
/+
nqd,f
a
The relative abundance (FID peak area % of all identified degradation products) as an average of n ) 3 sam ples. See Table 1 for additional
b
c
d
e
data. Hitherto not reported degradation product. Identification by MS and retention tim e tR. Specific degradation product. + ) positive
identification according to criterion. f nq ) not quantified, see text.
though, and between 5 and 500 ppb are encountered in major
cities (11, 19). The analyst should be aware of these products.
Since benzonitrile is formed by NO2 only, the N is likely to
be transferred from NO2 and not from impurities in the
polymer. Benzonitrile is specific for NO2 and could serve as
a quantitative marker. TD-GC analysis of Tenax degradation
products from air samples could become an alternative to
chemiluminescence measurement of NO2. This was previ-
ously suggested by Clausen and Wolkoff, which demonstrated
that the formation of [DPQ + DPHQ] could be calibrated in
dry air (10).
Identical conditions and same concentrations of NO and
NO2, respectively, resulted in a [DPQ + DPHQ] production
50 times less for NO than for NO2. Either NO has a low reaction
rate for degradation of Tenax, or traces of NO2 are present
in the system from oxidation of NO. Hence, the degradation
of Tenax by NO is negligible.
Hydroxyl Radical. The OH radical produced no specific
degradation products. Its chromatogram resembled that of
NO2-Tenax (see Figure 2b). NO2 produces benzonitrile and
1,2-diphenylethanone in small amounts (see Table 2), but
otherwise the degradation products from the OH radical and
NO2 are similar. The two species are both radicals, and it
appears that they share the same degradation routes.
However, it is considered unlikely that the OH radical in
ambient or indoor air will contribute significantly to Tenax
degradation, because of its low concentrations indoors (20).
Hydrogen Peroxide. Only small amounts of benzaldehyde
and 2,6-diphenylphenol appear in the chromatogram (see
Figure 2c), although H2O2 was expected to degrade Tenax.
It is an important oxidant in the atmosphere along with O3
and the OH radical. The concentration of hydrogen peroxide
was 23 ppm, much higher than in ambient air, where levels
around 1 ppb or lower have been reported (e.g. refs 21-23).
Sources of atmospheric H2O2 are the ozonolysis of alkenes
(24) and the bimolecular self-reaction of hydroperoxy radicals
(25), though concentrations appear to be in the low ppb
range (26). Organic hydroperoxides are also formed in the
ozonolysis of alkenes and in the reaction of the hydroperoxy
radical with organic peroxy radicals (27, 28). Like H2O2 they
are expected not to degrade Tenax directly. However, organic
hydroperoxides are better adsorbed than hydrogen peroxide
and may thermally decompose prior to or during desorption.
If the decomposition products of hydroperoxides are radicals,
formation of DPQ and DPHQ are expected as for NO2 and
the OH radical.
Ozone. The chromatogram is dominated by benzalde-
hyde, phenol, acetophenone, benzoic acid, phenylmaleic
anhydride, DPHQ, see Table 2 and Figure 2d. p-Hydro-
quinone and phenyl glyoxylic acid were not observed at low
concentrations of O3. Formic acid, acetic acid, phenylglyoxal,
p-hydroquinone, phenylglyoxylic acid, phthalic anhydride,
benzophenone, 1,2-diphenylethanone, diphenylethane-
dione, DPQ (in trace amounts), and DPHQ have not been
previously reported as O3-Tenax degradation products.
Compound Z appeared as a tiny shoulder on the DPHQ peak,
and it was not possible to quantify it accurately. Mass spectra
of all degradation products were available, except compound
W. The spectrum of H has been reported (29). The negligible
molecular ion at m/ z 134 is presumably due to the weak
RC(O)-CHO bond and a highly stable RC)O ion (m/ z 105).
Eleven degradation products were found to be specific
for O3, and most of them were formed in small amounts
while H, V, and W were formed in measurable amounts. W
has the characteristics of an aromatic ketone but remains
unidentified (30). 2,4-Diphenyl-4-cyclopentene-1,3-dione
and compound W appear promising as O3 markers, since
they are both specific and abundant. In addition, their low
volatility reduces the risk of peak overlap with volatile organic
compounds during sampling of ambient or indoor air (e.g.,
F and H). The application of SIM MS would enhance the
response, which is convenient regarding the sample volume.
9
4 1 2 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

FIGURE 2. FID chromatograms of Tenax degradation products formed by (a) NO ; (b) OH; (c) H O ; (d) O ; (e) O /limonene reaction products,
2
2
2
3
3
and (f) FID chromatogram of a cleaned Tenax adsorbent tube.
Modification of the desorber conditions would probably not
enhance the detector response, because of their low volatility.
The application of several degradation products using
multivariable statistics is a possibility.
even though the OH radical can produce DPHQ. The OH
radical is highly reactive and is probably consumed im-
mediately by the excess limonene. It is more likely that as
a result of thermal decomposition before or during the
thermal desorption of Tenax, other reaction products either
produce the OH radical or other radicals. Possibly, organic
hydroperoxides, which would be stronger adsorbed to Tenax
than H2O2, could decompose and degrade the adsorbent.
Based on the above findings, it appears promising to
develop specific calibration procedures for both NO2 and
O3, so the residual degradation of e.g. O3-terpene mixtures
may be estimated. Since the contribution of DPQ/ DPHQ
from the OH radical and hydrogen peroxide is probably
negligible, calibration procedures for NO2 and O3 would
suffice.
Degradation Mechanism s. The mechanisms for the
formation of the degradation products are unknown. Neher
and Jones hypothesized that the formation of DPQ from
Tenax GC (a less pure polymer) was due to chain scission of
terminal groups of the polymer primarily by NO in stack
gases (6). Our results on Tenax, however, indicate DPQ to be
produced predominantly by NO2, which usually also are
Lim onene-Ozone Mixtures. The limonene-O3 mixtures,
with no residual O3, produced DPHQ (major) and 2,6-
diphenylphenol (minor), see Figure 2e. No other degradation
products could be detected with certainty. However, the yield
of DPHQ was lower than in the other types of experiments.
The chromatogram resembles that of NO2 and the OH radical,
where DPQ is the major product. The major degradation
product is considered the same in all three experiments,
because under the analytical conditions DPQ and DPHQ
cannot be differentiated. In fact, NO2, the OH radical, and
limonene-O3 mixtures degrade Tenax roughly in the same
way, since the degradation products formed in trace amounts
are not normally detected.
Since NO2 and O3 were absent in the mixture, other
reaction products including secondary products as particles
must be responsible for the degradation of Tenax and
formation of DPHQ. It is unlikely that the OH radical
produced from the ozonolysis of limonene is responsible,
9
VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 1 2 5

In Identification and Analysis of Organic Pollutants in Air; Keith
L. H., Ed.; Butterworth: London, 1984; pp 371-387.
(2) Harper, M. J. Chromatogr. A 2000, 885, 129-151.
(3) Stro¨ mvalll, A.-M.; Petersson, G. J. Chromatogr. 1992, 589, 385-
389.
(4) Calogirou A.; Larsen B. R.; Kotzias D. Atmos. Environ. 1999, 33,
1423-1439.
(5) Grosjean, E.; Grosjean, D. J. Atmos. Chem. 1999, 32, 205-232.
(6) Neher, M. B.; Jones, P. W. Anal. Chem. 1977, 49, 512-513.
(7) Pellizzari, E. D.; Demian, B.; Krost, K. J. Anal. Chem. 1984, 56,
793-798.
(8) Walling, J. F.; Bumgarner, J. E.; Driscol, D. J.; Morris, C. M.;
Riley, A. E.; Wright, L. H. Atmos. Environ. 1986, 20, 51-57.
(9) Cao, X.; Hewitt, C. N. Environ. Sci. Technol. 1994, 28, 757-762.
(10) Clausen, P. A.; Wolkoff, P. Atmos. Environ. 1997, 31, 715-725.
(11) Wolkoff, P.; Clausen, P. A.; Wilkins, C. K.; Nielsen, G. D. Indoor
Air 2000, 10, 82-91.
(12) Weschler, C. J. Indoor Air 2000, 10, 269-288.
(13) Ruggli, P.; Schmidlin, J. Helv. Chim. Acta 1944, 27, 499-502.
(14) Wolkoff, P. Atmos. Environ. 1998, 32, 2659-2668.
(15) Effkemann, S.; Brødsgaard, S.; Mortensen, P.; Linde, S.-A.; Karst,
U. J. Chromatogr. 1999, 855, 551-561.
(16) Dohan, J. M.; Masschelein, W. J. Ozone: Sci., Eng. 1987, 9, 315-
334.
(17) Nozie´re, B.; Barnes, I.; Becker, K.-H. J. Geophys. Res. 1999, 104-
(D19), 23,645-23,656.
(18) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.;
Molina, M. J. Chemical kinetics and Photochemical data for use
in stratospheric modelling; Evaluation No. 12; Jet Propulsion
Laboratory: Pasadena, CA, January 15, 1997.
(19) Weschler, C. J.; Shields, H. Environ. Sci. Technol. 1994, 28, 2120.
(20) Weschler, C. J.; Shields, H. C. Environ. Sci. Technol. 1997, 31,
3719-3722.
(21) Lee, J. H.; Leahy, D. F.; Tang, I. N.; Newman, L. J. Geophys. Res.
1993, 98, D2, 2911-2915.
(22) Jackson, A. V.; Hewitt, C. N. Atmos. Environ. 1996, 30, 819-830.
(23) Riedel, K.; Weller, R.; Schrems, O.; Ko¨ nig-Langlo, G. Atmos.
Environ. 2000, 34, 5225-5234.
(24) Sauer, F.; Scha¨fer, S.; Neeb, P.; Horie, O.; Moortgart, G. K. Atmos.
Environ. 1999, 33, 229-241.
(25) Finlayson-Pitts, B.J; Pitts, J. N. Chemistry of the Upper and Lower
Atmosphere; Academic Press: San Diego, 2000.
(26) Li, T.-S.; Turpin. B. J.; Shields, H. C.; Weschler, C. J. Environ. Sci.
Technol. 2002, 36, in press.
present in minor amounts in stack gases. Therefore, it is
reasonable to assume that NO2 is responsible for the DPQ
production. Other degradation products may arise from
degradation of the polymer or other molecules formed during
depolymerization. Roberts et al. (1) and Pellizari et al. (7)
have observed a decreasing yield in benzaldehyde, phenol,
and acetophenone after repeated O3 exposure of Tenax GC.
It was hypothesized that the adsorbent surface was depleted
in readily degradable oligomers. However, another explana-
tion could be degradable impurities not removed by thermal
cleaning of the Tenax.
The observed degradation products are not necessarily
primary in nature but may arise from thermal breakdown of
unknown degradation products. This could be studied by
solvent extraction of Tenax exposed to O3 followed by NMR,
FTIR, or another cold analytical technique, as described by
Neher and Jones (6). Clausen and Wolkoff (10) found that
DPQ and DPHQ were not thermally degraded to other Tenax
degradation products. O3 could add to the double bonds of
DPQ, but this reaction should be much slower than its
reaction with terpenes, since the carbonyl groups are electron
withdrawing (31). In a series of experiments, crystalline DPQ
and DPHQ was exposed to high concentrations of O3 for 1
h and dissolved in methanol. TD-GC-MS analysis revealed
none of the observed degradation products (10). However,
oxidation of DPQ on the Tenax surface cannot be excluded.
It appears that several mechanisms are involved in the
breakdown of Tenax. It is interesting that the NO2 and OH
radicals behave almost identical with respect to degradation
products but not NO. If this is general for reactive radicals,
the formation of the 11 specific O3-Tenax products suggests
degradation routes other than simple radical mechanisms.
Degradation of adsorbed terpenes on Tenax in the
presence of O3 (4) and NO2 (32) may also be explained by
secondary oxidation by DPQ formed on the Tenax surface,
possibly during thermal desorption. This is particularly
relevant in the case of NO2, because its reaction rate with
alkenes (in the gas phase) are orders of magnitude lower
than those of O3 (33).
The present study warrants an evaluation of O3-Tenax
degradation products as a method for quantitative NO2 and
O3 analysis. Also, the influence of residual mixtures of O3 and
NO2 needs to be studied.
(27) Neeb, P.; Sauer, F.; Horie, O.; Moortgart, G. K. Atmos. Environ.
1997, 31, 10, 1417-1423.
(28) Tobias, H. J.; Ziemann, P. J. Environ. Sci. Technol. 2000, 34,
2105-2115.
(29) Andresen, B. D.; Biemann, K. J. Label. Comps. Radiopharmaceut.
1978, 15, 479-488.
(30) Clausen, P. A. Unpublished FT-IR and MS data, 2001.
(31) Paquette, L. A. Encyclopedia of reagents for organic chemistry
synthesis; John Wiley & Sons Ltd.: England, 1995; Vol. 6, pp
3837-3843.
(32) Pommer, L.; Fick, J.; Andersson, B.; Nilsson, C. Atmos. Environ.
2002, 36, 1443-1452.
(33) Atkinson, R.; Aschmann, A. M.; Winer, A. M.; Pitts, J. N. Int. J.
Chem. Kinet. 1984, 16, 697-706.
Acknowledgments
The study was partially supported by Center for Indoor Air
Research (Linthicum, MD). We thank Dr. S. Hammerum,
Department of Chemistry, University of Copenhagen for a
sample of 2,4-diphenyl-4-cyclopentene-1,3-dione. We are
grateful for the excellent work by Ms. Vivi Hansen and Mr.
Kjeld Larsen.
Literature Cited
(1) Roberts, J. M.; Fehsenfeld, F. C.; Albritton, D. L.; Sievers, R. E.
Sampling and analysis of monoterpene hydrocarbons in the
atmosphere with Tenax gas chromatographic porous polymer.
Received for review March 27, 2002. Revised manuscript
received July 12, 2002. Accepted July 17, 2002.
ES025680F
9
4 1 2 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002