Secondary ozonides of substituted cyclohexenes: A new class of pollutants characterized by collision-induced dissociation mass spectrometry using negative chemical ionization

Article abstract of DOI:10.1016/j.chemosphere.2007.09.018

Recent studies indicate that secondary ozonides of cyclic alkenes are formed in atmospheric reactions and may be relatively stable. The secondary ozonides (SOZs) of cyclohexene (1), 1-methylcyclohexene (2), 4-isopropyl-1-methylcyclohexene (3) and 4-isopropenyl-1-methylcyclohexene (limonene) (4) have been characterized by rapid gas chromatography electron ionization (EI), negative and positive chemical ionization (CI: ammonia, isobutane and methane) and collision-induced dissociation (CID) mass spectrometry. Both EI and positive CI spectra were found unsuitable for reproducible analysis. However, negative CI showed stable (M-H)^- ions with minor fragmentation. CID of the (M-H)^- ions resulted in simple and reproducible fragmentation patterns for all four SOZs with loss of m/z 18, 44 and 60, tentatively assigned as H₂O, CO₂ and C₂H₄O₂ or CO₃, respectively. Thus, negative CI-MS-MS in combination with rapid gas chromatography is the preferred method for identification of secondary ozonides of cyclohexenes.

Full text of DOI:10.1016/j.chemosphere.2007.09.018

Available online at www.sciencedirect.com
Chemosphere 70 (2008) 2032–2038
www.elsevier.com/locate/chemosphere
Secondary ozonides of substituted cyclohexenes: A new class
of pollutants characterized by collision-induced dissociation
mass spectrometry using negative chemical ionization
Asger W. Nørgaard, Jacob K. Nøjgaard, Per A. Clausen, Peder Wolkoff *
Indoor Environment Group, National Research Centre for the Working Environment, Lersø Parkall e´ 105, DK-2100 Copenhagen Ø, Denmark
Received 28 June 2007; received in revised form 10 September 2007; accepted 11 September 2007
Available online 26 October 2007
Abstract
Recent studies indicate that secondary ozonides of cyclic alkenes are formed in atmospheric reactions and may be relatively stable.
The secondary ozonides (SOZs) of cyclohexene (1), 1-methylcyclohexene (2), 4-isopropyl-1-methylcyclohexene (3) and 4-isopropenyl-1-
methylcyclohexene (limonene) (4) have been characterized by rapid gas chromatography electron ionization (EI), negative and positive
chemical ionization (CI: ammonia, isobutane and methane) and collision-induced dissociation (CID) mass spectrometry. Both EI and
À
positive CI spectra were found unsuitable for reproducible analysis. However, negative CI showed stable (MÀH) ions with minor frag-
À
mentation. CID of the (MÀH) ions resulted in simple and reproducible fragmentation patterns for all four SOZs with loss of m/z 18, 44
and 60, tentatively assigned as H O, CO and C H O or CO , respectively. Thus, negative CI-MS–MS in combination with rapid gas
2
2
2
4
2
3
chromatography is the preferred method for identification of secondary ozonides of cyclohexenes.
2007 Elsevier Ltd. All rights reserved.
ꢀ
Keywords: Indoor air; Mass spectrometry; Secondary ozonides; Substituted cyclohexenes
1
. Introduction
been identified, because of their instability. Thus, a number
of relatively unstable oxidation products have been identi-
fied using ‘‘soft’’ sampling and chromatographic techniques
for identification (Li et al., 2002; Docherty et al., 2004).
Recently, the secondary endo-ozonide of limonene has been
identified by use of pressurized liquid extraction and cold-
on-column gas chromatography as an abundant oxidation
product (Nørgaard et al., 2006), trapped from the gas phase
ozonolysis of limonene, a common VOC in indoor environ-
ments (e.g. J a¨ rnstr o¨ m et al., 2006). Higher concentration of
alkenes can build up in indoor environments (Singer et al.,
2006), as opposed to outdoors, and result in the formation
of secondary ozonides (SOZs), e.g. from ozonolysis of terp-
enes. This has been substantiated in recent studies about
cyclic alkenes (Aschmann et al., 2003; Nøjgaard et al.,
Although substantial research has been carried out on
indoor air pollutants and their potential health impact,
there are still many unknowns (Carslaw and Wolkoff,
2
006). Most research has focussed on single compounds
present indoors and relatively little attention has been paid
to the significance of the reactions between them, which are
likely to occur (Weschler et al., 2006; Wolkoff et al., 2006).
There is an increasing concern about the health impact of
oxidation products from reaction between ozone and chem-
ically reactive volatile organic compounds (VOCs) that are
emitted from a number of building materials and consumer
products (e.g. Nazaroff and Weschler, 2004). There are indi-
cations that some of the causative species have hitherto not
2
007). Such SOZs are expected to degrade at elevated
temperature (cf. Tobias et al., 2000; Nørgaard et al.,
006), thus warranting a versatile analytical method for
*
Corresponding author. Tel.: +45 39165272; fax: +45 39165201.
2
E-mail address: pwo@nrcwe.dk (P. Wolkoff).
0
045-6535/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2007.09.018

A.W. Nørgaard et al. / Chemosphere 70 (2008) 2032–2038
2033
O
O
O
O
2.2. Instrumentation
O
O
O
O
Three different mass spectrometers (I–III) were used in
this study. MS I: Perkin Elmer GC–MS (GC: Autosystem
XL/MS: TurboMass); MS II: Varian GC–MS/MS (GC:
CP-3800/MS: 1200L). Both instruments were equipped
with auto sampler and on-column injector, a medium polar
deactivated pre-column 2.5 m · 0.53 mm i.d. (Varian) and
a 10 m · 0.53 mm i.d. Varian VF-5ms Rapid Low Bleed/
MS (0.25 lm film thickness) column. GC oven program
O
O
O
O
1
2
3
4
À1
was: 50 ꢁC hold 1 min, ramp 1: 5 ꢁC min to 80 ꢁC, ramp
À1
2
5
4
: 25 ꢁC min to 250 ꢁC hold 2 min for SOZ analysis and
À1
0 ꢁC hold 1 min, ramp 1: 5 ꢁC min to 80 ꢁC, ramp 2:
À1
5 ꢁC min to 250 ꢁC hold 10 min for analysis of TPPO.
Helium was used as carrier gas; the pressure was adjusted
5
6
7
8
À1
to be equal to a column flow of 1.5 ml min
.
MS III: Kratos Profile double focusing magnetic sector
instrument equipped with a HP 5890 GC and on-column
injector for manual injection. The chromatographic system
consisted of a medium polar retention gap (Varian deacti-
Fig. 1. Secondary ozonides (SOZs) of cyclohexene (1), 1-methylcyclohex-
ene (2), 1-methyl-4-isopropylcyclohexene (3), limonene (4) and their
corresponding cyclohexenes (5–8).
vated pre-column 2.5 m · 0.53 mm i.d.) and
a
6
0 m · 0.32 mm i.d. Chrompack CP-Sil 19 CB low bleed/
future indoor campaigns, because they are of concern
regarding sensory irritation of the eyes and upper airways,
including inflammation in the lower airways (Cvitas et al.,
MS (0.25 lm film thickness) column. The GC oven pro-
À1
gram was: 30 ꢁC hold 2 min, ramp: 4 ꢁC min to 250 ꢁC
hold 5 min. Helium was used as carrier gas; column flow:
2
005; Wolkoff et al., 2006).
À1
1
.5 ml min . Injection volumes were 0.5 ll. The MS trans-
The purpose of this study was to identify the best analyt-
fer lines and MS source temperatures were kept at 120 ꢁC.
Ionization energy of 70 eV was applied in EI mode. For CI
experiments ammonia, isobutane and methane were used
as reagent gases at pressures between 1 and 5 torr and ion-
ization energy of 30 eV (MS I and III) and 150 eV (MS II).
Typical scan ranges were m/z 35–400 in EI mode and m/z
ical conditions for SOZs of cyclic alkenes (see Fig. 1) by
combined gas chromatography and mass spectrometry as
an alternative to spectroscopic methods (FT-IR and
NMR), which have previously been used for SOZ identifi-
cation (Griesbaum et al., 1996; Neeb et al., 1996; Winter-
halter et al., 2000). Characterization of the mass spectral
fragmentation patterns has been carried out for some SOZs
of alkyl and alkenyl substituted cyclohexenes under various
conditions of ionization including electron ionization (EI)
and positive chemical ionization (CI) using ammonia, iso-
butane and methane as reagent gas and negative CI using
isobutane and methane.
60–400 in CI mode. Precursor and product ion scans were
carried out on MS II. For CI-MS–MS experiments only
methane and isobutane could be used as reagents gasses.
Argon was used as collision gas at pressures of 1–2 mtorr
and the collision energy (CE) was usually ramped
(2–24 eV) during most MS–MS experiments.
2.3. Synthesis and purification of secondary ozonides of
cyclic alkenes
2
. Experimental
2
.1. Chemicals
The SOZs were synthesized by cryo-ozonolysis of the
alkenes (500 ll) dissolved in 100 ml pentane in a 200 ml
impinger until a pale blue colour appeared (Griesbaum
Cyclohexene (5) (puriss > 99.5%), 1-methylcyclohexene
(
(
6) (puriss > 99%), 4-isopropyl-1-methylcyclohexene (7)
purum > 97%), R-(+)-limonene (8) (purum > 96%), tri-
et al., 1996). O was generated photochemically in pure
O with a mercury lamp (185 and 254 nm) in a thermostatic
2
3
phenyl phosphine (TPP) (puriss > 98.5%) and triphenyl
phosphine oxide (TPPO) (purum > 98%) were purchased
from Fluka. The purity of all alkenes was checked by
GC–MS analysis prior to ozonolysis. Dichloromethane
lamp housing controlled by a high performance variable
power supply (Clausen et al., 2001).
The reaction mixture was purged with N to remove
2
excess O and to reduce the volume to about 2 ml. Subse-
3
(
DCM) (>99%) and methanol (99.8%) were obtained from
Merck. Pentane (98%) was purchased from Aldrich. O2
>99.9999%) from Hydro Gas (Norway) was used to gen-
quently, the SOZs were purified by the use of preparative
normal phase LC-UV. The system consisted of a Perkin
Elmer series 410 LC-pump, a Waters model 440 UV-detec-
tor and a 50 cm stainless steel column (i.d. 3.9 mm) packed
with silica gel (Merck type 9385, 230–400 Mesh). The reac-
tion mixture (2 ml) was, by loop injection, introduced into
(
erate O . Methane (quality 5.5), isobutane (quality 3.5)
and ammonia (quality 4.8) for CI-MS experiments were
obtained from AGA (Denmark).
3

2034
A.W. Nørgaard et al. / Chemosphere 70 (2008) 2032–2038
Fig. 2. EI mass spectra of secondary ozonides (1–4). A proposed fragmentation pathway is shown for SOZ 2 in Fig. 3.

A.W. Nørgaard et al. / Chemosphere 70 (2008) 2032–2038
2035
the chromatographic system. First, excess alkene was
eluted with 95% pentane and 5% DCM), subsequently,
the SOZ was eluded with 100% DCM followed by other
oxidation products. The purified SOZs, in DCM, were
stored at 5 ꢁC until use.
ters (Budzikiewicz et al., 1967). EI mass spectral data for
the SOZs 1–4 are presented in Fig. 2. SOZ 1 and SOZ 2
show molecular ions M at m/z 130 and 144, respectively.
A molecular ion of low abundance could in some experi-
ments be observed at m/z 186 in the case of SOZ 3, while
the molecular ion of SOZ 4 was absent. SOZ 1 and SOZ 2
+
Å
2
.4. Reaction with triphenyl phosphine
show loss of molecular oxygen (O ) directly from the molec-
2
ular ion, which was confirmed by MS–MS experiments. O2
loss from the molecular ion has also been reported in a pre-
vious study of the fragmentation patterns of some second-
ary ozonides (Castonguay et al., 1969). The subsequent
0
.5 M TPP in DCM (50 ll) was added to a purified solu-
tion of SOZ (approx. 250 ll). The solutions were stored at
room temperature for at least 48 h before analysis. The oxi-
dation of TPP to TPPO by SOZs is well known (e.g. Gries-
baum et al., 1996; Fleming et al., 1997). The samples were
analyzed by rapid GC–MS and the identity of TPPO was
confirmed by comparison with an authentic standard.
fragmentations after the loss of O are similar to those
2
reported for the 1,2-epoxides of cyclohexenes 5 and 6,
respectively (Budzikiewicz et al., 1967; Strong et al.,
1969), see Fig. 3 for proposed fragmentation. SOZ 3 and
SOZ 4 show minor losses of 32 Da and 34 Da, respectively,
which are tentatively assigned as O and H O ; the latter
3
. Results and discussion
2
2
2
has previously been confirmed for SOZ 4 by accurate mass
determination of ion m/z 150 (Nørgaard et al., 2006). The
overall fragmentation pathways for SOZ 3 and SOZ 4 after
the loss of O or H O appear to mimic the mass spectra of
3
.1. EI-MS
Multiple EI experiments were performed on the three dif-
2
2
2
ferent instruments and slightly different spectra were
achieved. Such differences have been reported for analysis
of alicyclic epoxides when using different mass spectrome-
their corresponding epoxides; this is shown, for example, by
comparison of the mass spectrum of limonene epoxides
(NIST Mass spectral database v2.0). Loss of O from SOZs
2
O
O
O
O
O
O
-
O2
m/z 144
O
O
m/z 112
O
methyl shift
H- shift
ring contraction
O
HCO
ring contraction
COCH3
O+
O+
-
CHO
CH3
-
CH3
O
-
C H
5 9
-
COCH3
m/z 55
m/z 69
+
CH CO
3
m/z 43
m/z 83
m/z 97
m/z 69
Fig. 3. Proposed EI fragmentation pathway for the SOZ of 1-methylcyclohexene (according to Budzikiewicz et al., 1967; Strong et al., 1969).

2036
A.W. Nørgaard et al. / Chemosphere 70 (2008) 2032–2038
Fig. 4. CID mass spectra of secondary ozonides in negative chemical ionization (methane).
1
–3 and loss of both O and H O from SOZ 4 have been
sampling Townsend discharge mass spectrometry (ASTDI-
MS) (Nøjgaard et al., 2007).
2
2
2
reported in a previous study of SOZs by use of atmospheric

A.W. Nørgaard et al. / Chemosphere 70 (2008) 2032–2038
2037
3
.2. CI-MS, positive mode
tively assigned as H O, CO , and C H O or CO , respec-
2 2 2 4 2 3
tively. In addition, SOZ 1 shows loss of 28 and 32
+
(
M+H) ions were observed in CI-MS in positive mode
tentatively assigned as CO and O , respectively.
2
+
À
for all SOZs. (M+H) ions were observed by use of meth-
ane or isobutane as reagent gas. Ammonia as reagent gas
produced (M+NH ) ions. The relative abundance (r.a.)
The high stability of the (MÀH) ions of all four SOZs
indicate the presence of an acidic hydrogen atom, presum-
ably the ring bound H. This stability indicates that the neg-
ative charge is delocalized in the trioxolane ring.
+
4
+
+
of the (M+H) and (M+NH4) ions ranged from 6% to
5% for all reagent gasses.
The appearance of the (M+NH ) ions and the absence
5
+
4
+
4. Conclusions
of (M+H) ions with ammonia as reagent gas indicates
that the proton affinities (PA) of the SOZs are above
GC–EI-MS will in some cases be insufficient for identi-
fication of the SOZs, because of the low abundance or
absence of their molecular ions for SOZs 3–4. This may
explain why SOZs have not been detected previously in
atmospheric samples. In addition, EI is further invalidated
by substantial variation of the fragmentation patterns that
depends on the instrument used. Positive chemical ioniza-
tion results in extensive fragmentation even when using a
soft reagent gas such as ammonia. Thus, positive CI pro-
vides much information; however, interpretation is ham-
pered by poor reproducibility. The simple and predictable
fragmentation patterns in negative CI-MS–MS mode
makes this method ideal for identification of the examined
SOZs.
7
87 kJ/mol, which is the energy required for formation of
+
the (M+NH4) adduct ion (Keough and DeStefano,
981; Westmore and Alauddin, 1986) and below 854 kJ/
1
mol (the PA for ammonia).
All SOZs showed a high degree of fragmentation in all
experiments, and their CI spectra were complex and diffi-
cult to reproduce on the different instruments; thus indicat-
ing the instability of the (M+H) ions. Common fragment
ions in the positive CI spectra of all SOZs correspond to
fragment losses with the mass of 16, 17, 18, 28, 32, 34, 36
and 44 Da, which tentatively are assigned as O, OH,
H O, CO, O , H O , 2H O and CO , respectively. All
SOZs eliminate OH, H O with methane as reagent gas,
+
Å
Å
2
2
2
2
2
2
Å
2
2
Å
while elimination of O, H O and CO was observed with
2
2
2
isobutane as reagent gas. In the case of NH all the SOZs
3
Å
eliminate O; in addition, SOZs 2–4 eliminate two fragment
Acknowledgements
+
ions at 33 and 60 Da less than (M+NH ) .
4
The positive CI-MS–MS experiments with methane and
isobutane as reagent gas confirmed some of the losses men-
The work was in part supported by Philip Morris Inter-
national. We thank Mr. K. Larsen for excellent technical
assistance.
tioned above. In general, collision-induced dissociation
+
(
CID) of the (M+H) ions resulted in extensive fragmenta-
tion for all four SOZs even at a CE as low as 2 eV. For
References
example, the most abundant peaks for SOZs 1–3 originate
Å
from the elimination of OH and H O, while SOZ 4 elimi-
2
Aschmann, S.M., Tuazon, E.C., Arey, J., Atkinson, R., 2003. Products of
Å
nates OH, H O and H O . Although the positive CI spectra
2
2
2
the gas-phase reaction of O with cyclohexene. J. Phys. Chem. A 107,
3
show unique losses, they cannot be used for characteriza-
tion due to their large variability between instruments.
2247–2255.
Budzikiewicz, H., Djerassi, C., Williams, D.T., 1967. Mass Spectrometry
of Organic Compounds. Holden-Day Inc., San Francisco.
Carslaw, N., Wolkoff, P., 2006. Guest editorial. Indoor Air 16, 4–6.
Castonguay, J., Bertrand, M., Carles, J., Flisz a´ r, S., Rousseau, Y., 1969.
Mass spectrometry of ozonides. Can. J. Chem. 47, 919–934.
Clausen, P.A., Wilkins, C.K., Wolkoff, P., Nielsen, G.D., 2001. Chemical
and biological evaluation of a reaction mixture of R-(+)-limonene/
ozone: formation of strong airway irritants. Environ. Int. 26, 511–
3
.3. CI-MS, negative mode
All SOZs show base peak (MÀH)À ions and negligible
fragmentation is observed by use of methane and isobutane
as reagent gasses. SOZ 1 shows consistently elimination of
1
5
22.
8 Da (ꢀ1% r.a.), assigned as H O loss, while SOZs 2–4 all
2
Cvitas, T., Klasinc, L., Kezele, N., McGlynn, S.P., Pryor, W.A., 2005.
New directions: How dangerous is ozone? Atmos. Environ. 39, 4607–
4608.
Docherty, K.S., Kumboonlert, K., Lee, I.J., Ziemann, P.J., 2004. Gas
chromatography of trimethylsilyl derivates of a-methoxyalkyl hydro-
À
show an abundant (MÀ15) ion, see Fig. 4. These
À
(
MÀ15) ions were examined by various MS–MS experi-
ments, however, no precursor ions could be identified;
probably, these ions originate from the loss of a methyl
peroxides formed in alkene-O
15.
Fleming, F.F., Huang, A., Sharief, V.A., Pu, Y., 1997. Unsaturated
nitriles: domino ozonolysis–aldol synthesis of highly reactive
3
reactions. J. Chromatogr. 1029, 205–
group during the ionization process. The fact that no
2
À
(
MÀ15) ion was observed for SOZ 1 adds further sup-
port, since this SOZ contains no methyl group.
a
À
oxonitriles. J. Org. Chem. 62, 3036–3037.
Griesbaum, K., Hilß, M., Bosch, J., 1996. Ozonides of mono-, bi- and
tricyclic terpenes. Tetrahedron 52, 14813–14826.
J a¨ rnstr o¨ m, H., Saarela, K., Kalliokoski, P., Pasanen, A.L., 2006.
Reference values for indoor air pollutant concentrations in new,
residential buildings in Finland. Atmos. Environ. 40, 7178–7191.
The (MÀH) ions of all SOZs appear to be relatively
stable, and significant fragmentation was first observed at
CE of at least 8 eV. Results from such negative CI-MS–
MS experiments are presented in Fig. 4. All SOZs eliminate
the following fragments: 18, 44 and 60, which are tenta-

2038
A.W. Nørgaard et al. / Chemosphere 70 (2008) 2032–2038
Keough, T., DeStefano, A.J., 1981. Factors affecting reactivity in
ammonia chemical ionization mass spectrometry. Org. Mass Spec-
trom. 16, 527–533.
Li, T.H., Turpin, B.J., Shields, H.C., Weschler, C.J., 2002. Indoor
hydrogen peroxide derived from ozone/D-limonene reactions. Environ.
Sci. Technol. 36, 3295–3302.
concentrations of glycol ethers and terpenoids. Indoor Air 16,
179–191.
Strong, M.K., Brown, P., Djerassi, C., 1969. Mass spectrometry in
structural and stereochemical problems – CLXXVIII: the electron-
impact promoted fragmentation of 1,2-cyclohexene oxide. Org. Mass
Spectrom. 2, 1201–1214.
Nazaroff, W.W., Weschler, C.J., 2004. Cleaning products and air
fresheners: exposure to primary and secondary air pollutants. Atmos.
Environ. 38, 2841–2865.
Tobias, H.J., Docherty, K.S., Beving, D.E., Ziemann, P.J., 2000. Thermal
desorption mass spectrometric analysis of organic aerosol formed from
3
reactions of 1-tetradecene and O in the presence of alcohols and
Neeb, P., Horie, O., Moortgat, G.K., 1996. Formation of secondary
ozonides in the gas-phase ozonolysis of simple alkenes. Tetrahedron
Lett. 37, 9297–9300.
Nøjgaard, J.K., Nørgaard, A.W., Wolkoff, P., 2007. Secondary ozonides
of endo-cyclic alkenes analyzed by atmospheric sampling Townsend
discharge ionization mass spectrometry. Int. J. Mass Spectrom. 263,
carboxylic acids. Environ. Sci. Technol. 34, 2105–2115.
Weschler, C.J., Wells, J.R., Poppendieck, D., Hubbard, H., Pearce, T.A.,
2006. Workgroup report: indoor chemistry and health. Environ.
Health Perspect. 114, 442–446.
Westmore, J.B., Alauddin, M.M., 1986. Ammonia chemical ionization
mass spectrometry. Mass Spectrom. Rev. 5, 381–465.
8
8–93.
Winterhalter, R., Neeb, P., Grossman, D., Kollhof, A., Horie, O.,
Moortgat, G.K., 2000. Products and mechanism of the gas phase
reaction of ozone with b-pinene. J. Atmos. Chem. 35, 165–197.
Wolkoff, P., Wilkins, C.K., Clausen, P.A., Nielsen, G.D., 2006. Organic
Nørgaard, A.W., Nøjgaard, J.K., Larsen, K., Sporring, S., Wilkins, C.K.,
Clausen, P.A., Wolkoff, P., 2006. Secondary limonene endo-ozonide: a
major product from gas-phase ozonolysis of R-(+)-limonene at
ambient temperature. Atmos. Environ. 40, 3460–3466.
compounds in office environments
– sensory irritation, odor,
Singer, B.C., Destaillats, H., Hodgson, A.T., Nazaroff, W.W., 2006.
Cleaning products and air fresheners: emissions and resulting
measurements and the role of reactive chemistry. Indoor Air 16,
7–19.