Chemical synthesis of ozone, isotopic labeling, and redistribution

Article abstract of DOI:10.1021/ja980887n

Ozone is formed in good yield by reacting O₂/⁺ salts with water in HF at -78 °C. With isotopically labeled water ¹⁷O-¹⁶O-¹⁶O and ¹⁸O- ¹⁶O-¹⁶O can be prepared, and other isotopomers could be made by starting with ¹⁷O₂/⁺ and ¹⁸O₂/⁺ precursors. ¹⁷O NMR spectroscopy and IR matrix spectroscopy are used for detection and decay of the isotopomers. The ¹⁷O or ¹⁸O label at the terminal position remains at -78 °C, but at higher temperatures scrambling of the atoms is observed. Under strict monomolecular conditions no scrambling is observed till 450 °C, however. Irradiation of ozone in a 15K argon matrix with λ = 254 nm affords scrambling of the label and formation of oxygen. Irradiation with λ > 420 nm initiates only scrambling with little decay. The possible mechanisms of the atomic scrambling are discussed.

Full text of DOI:10.1021/ja980887n

J. Am. Chem. Soc. 1998, 120, 8711-8714
8711
Chemical Synthesis of Ozone, Isotopic Labeling, and Redistribution
Anton Dimitrov,1a Konrad Seppelt,*,1a Dieter Scheffler,1b and Helge Willner†,1b
Contribution from the Institut fu¨r Anorganische und Analytische Chemie der Freie UniVersita¨t, D-14195
Berlin, Germany, and Institut fu¨r Anorganische Chemie der UniVersita¨t,
D-30167 HannoVer, Germany
ReceiVed March 16, 1998
Abstract: Ozone is formed in good yield by reacting O2+ salts with water in HF at -78 °C. With isotopically
labeled water 17O-16O-16O and 18O-16O-16O can be prepared, and other isotopomers could be made by
+
+
starting with 17O2 and 18O2 precursors. 17O NMR spectroscopy and IR matrix spectroscopy are used for
detection and decay of the isotopomers. The 17O or 18O label at the terminal position remains at -78 °C, but
at higher temperatures scrambling of the atoms is observed. Under strict monomolecular conditions no
scrambling is observed till 450 °C, however. Irradiation of ozone in a 15K argon matrix with λ ) 254 nm
affords scrambling of the label and formation of oxygen. Irradiation with λ > 420 nm initiates only scrambling
with little decay. The possible mechanisms of the atomic scrambling are discussed.
Introduction
source reacting with molecular oxygen from another source.
To our knowledge this has never been tried, and the outcome
is questionable in the light of the atomic scrambling mecha-
nisms, which we will discuss below.
Ozone is a fascinating molecule for science ever since it was
isolated for the first time in 1840.2 It took about 10 years to
establish ozone as an allotrope of oxygen. Its importance is
based on many properties, of which only a few are named
here: the occurrence as an allotrope of oxygen under natural
conditions in the upper atmosphere, its molecular structure, its
reactivity toward carbon-carbon double bonds, and lately, its
possible depletion in the atmosphere by man-made gases.
Ozone soon became available in preparative amounts with
the simple process of generating atomic oxygen by the electric
discharge method in the Siemens ozonizator. This method can
be used for preparation of gaseous ozone diluted by oxygen or
for pure ozone as well, and works so simply and effectively
that no other preparation has been needed ever since. Ozone,
however, has been observed in many chemical reactions under
high oxidative conditions, but in no case has a preparative
method been derived from such a reaction.
In the course of our study on heterogeneous reactions between
H2O2/H2SO4 and atmospheric trace gases, we also rediscovered
formation of ozone by fast heating of a H2O2/H2O/H2SO4
mixture (mass ration 4/3/93) with a heat gun up to 250 °C under
dynamic vacuum. Yields are better than 30%. This chemical
formation of ozone by thermal decomposition of peroxomono-
sulfuric acid was first observed in 1902.3
In our investigations of high oxidative materials in liquid HF5
we have found a simple and high yield chemical synthesis of
ozone that can also be used for generation of specifically
isotopically labeled ozone.
Experimental Section
Reagents. O2+AsF6- and O2+SbF6- are prepared by UV irradiation
of O2/F2/AsF5 or O2/F2/SbF5 gaseous mixtures in large glass containers,
following literature procedures.6 HF is purified by 2-fold distillation
in a stainless steel vacuum line and poly-perfluoroethylene-perfluo-
rovinyl ether (PFA) container. No requirement for extreme drying of
the HF is needed, since water will take part in the reaction anyway.
But natural water in traces will influence the isotopic yield if H217O
and H218O are used as reactants.
Spectroscopy. 17O NMR spectra were taken of a JEOL JNM-LA
400 instrument at 54.10 MHz with H2O as external standard. The
reactions are performed in PFA tubes equipped with stainless steel
valves that are connected to a stainless steel vacuum line. Matrix
isolation studies are performed on an equipment described previously.7
Synthesis of Ozone. Caution: Pure ozone can detonate. A 2180
mg (9.58 mmol) sample of O2+AsF6- was weighed into a 20 mL PFA
tube, and 10 mL of HF was condensed into it. A 450-500 mg (∼25
mmol) sample of H216O, H217O, or H218O was added by syringe at -78
°C. Under stirring the dioxygenyl salts dissolve, the solution turns
Lately the surprising reaction
deep blue, and another colorless precipitate is found, which is identified
- 8
by its Raman spectrum as H3O+AsF6
. In cases when H2O excess is
FeIII(porphyrin)NO2 + O2 f FeII(porphyrin)NO + O3 (1)
used, which is advisable for completing the reaction, the remaining
nonvolatile material is partly liquid at room temperature, possibly due
to formation of H5O2+AsF6-. Its 19F NMR spectrum in CH3CN clearly
shows the quartet fine structure of AsF6-, because of the coupling of
the 19F nuclei with the spin 3/2 75As atom. The ozone dissolved in HF
allows a chemical preparation in solution and for immediate
further reaction.4 In principle this reaction could be used for
the preparation of isotopically labeled ozone also.
Isotopically labeled ozone requires a preparation from two
different oxygen precursors. One might conceive a preparation
by reacting electrically generated atomic oxygen from one
(5) Drews, T.; Seppelt, K. J. Angew. Chem. 1997, 109, 264; Angew.
Chem., Int. Ed. Engl. 1997, 36, 273.
(6) McKee, D. E.; Bartlett, N. Inorg. Chem. 1973, 12, 2738.
(7) Jacobs, J.; Kronberg, M.; Mu¨ller, H. S. P.; Willner, H. J. Am. Chem.
Soc. 1994, 116, 1106.
(8) Christe, K. O.; Schack, C. J.; Wilson, R. D. Inorg. Chem. 1975, 14,
2224. Christe, K. O.; Charpin, P.; Soulie, E.; Bougon, R.; Fawcett, J.; Russel,
D. R. Inorg. Chem. 1984, 23, 3756.
(1) (a) Freie Universita¨t Berlin. (b) Universita¨t Hannover.
(2) Scho¨nbein, C. F. Pogg. Am. 1840, 50, 616.
(3) W. Steadel, Z. Anorg. Chem. 1902, 14, 642.
(4) Castro, E. J. Am. Chem. Soc. 1996, 118, 3984.
S0002-7863(98)00887-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/15/1998

8712 J. Am. Chem. Soc., Vol. 120, No. 34, 1998
DimitroV et al.
is pumped from -78 °C via PFA tubing into a -196 °C cold trap,
containing a few 100 mg of powdered, dry NaF for absorbance of
residual HF. Oxygen is pumped away under these conditions. The
yield was determined by oxidation of a KI solution and back-titration
of the liberated iodine with a standardized solution of thiosulfate.9 The
yield turned out to be 83%, based on O2+AsF6. It is desired not to
keep the ozone for prolonged times in the HF solution since under
-
these conditions decomposition is observed. O2+SbF6- or O2+Sb2F11
can be used similarly. Ozone was also prepared by the discharge
method in a closed system,10 both with natural oxygen, with oxygen
enriched to 35% 17O and with oxygen enriched to 50% 18O.
Results
Ozone is chemically generated by the reaction
2O2 AsF6+ + 3H2O f O3 + O2 + 2H3O+AsF6
(2)
+
-
The ozone can be easily identified by the known characteristics
such as color, smell, vapor pressure, and, occasionally and
involuntarily, its explosiveness. For the latter reason the
preparation of only small amounts of undiluted ozone is strongly
recommended. Ozone can be safely dissolved in inert solvents
at low temperature like CFCl3 and CH2Cl2, or it can safely be
absorbed on silica gel at -78 °C for further chemical use.11
The major advance of the chemical preparation presented here
over the electric discharge method is of course the possibility
of generating certain specifically labeled ozone. Here we
present 16O-16O-17O and 16O-16O-18O by reaction of 16O2
+
with 17O and 18O enriched water. With 17O and 18O enriched
+
O2 the isotopomers 17O-17O-16O and 18O-18O-16O could
be made, and if both components are enriched four more
isotopomers are within preparative reach. Isotopomers not
accessible by this route are those which contain a central oxygen
isotope different from both terminal oxygen isotopes, such as
16O-17O-16O.
Figure 1. (a) 17O NMR spectrum of ozone with natural abundance of
16O, 17O, and 18O, relative to H217O. δ ) 1586.1, 1021.9 ppm. Solution
in CFCl3, -78 °C. (b) 17O NMR spectrum of ozone, made by the
discharge method from oxygen containing 40% 16O, 35% 17O, and 25%
18O. Inserts show isotopic splitting and 17O-17O coupling: A, 17O-
16O-xO, δ ) 1587.1 ppm; B, 17O-18O-xO, δ ) 1585.5 ppm; C, 17O-
17O-xO, δ ) 1586.3 ppm, partially resolved six equidistant lines due
In the case of 17O labeling the corresponding NMR spetros-
copy is the analytical tool of choice. In Figure 1 are shown
the 17O NMR spectra of normal ozone, such as 17O enriched
ozone generated by the electric discharge method, and of the
17O-16O-16O isotopomer generated by the chemical method.
Our measurement shows a two-line spectrum, as has been
measured once before, and the chemical shift data agree quite
well.12 The assignment of the two lines with quite extreme
chemical shifts to the two different oxygen environments was
done correctly in the original paper but was reversed later for
unknown reasons. The correct assignment is obvious from the
intensity ratio and the fine structure of the two signals (see
below). Thus the terminal oxygen atoms have an even more
extreme shift than the central oxygen atom.
5
to I ) /2, J ) 101.5 Hz; D, 16/18O-17O-16/18O, δ ) 1021.9 ppm; E,
17O-17O-16/18O, δ ) 1021.9 ppm, partially resolved six equidistant
lines. (c) 17O NMR spectrum of 17O-16O-16O prepared by the chemical
+
method from O2 and H217O.
O2+ on H2O, resulting in O-O-OH2+, which can be oxidized
+
further by another O2 to protonated ozone (reaction 3). This
mechanism is less likely than the following one, since H2O is
present in anhydrous HF solely as H3O+, which has very little
nucleophilicity. We therefore assume that at first a redox
reaction occurs between O2+ and H2O, giving H2O+. After loss
of H+, the OH radical combines with a second O2+ to protonated
ozone (reaction 4).
In the 17O/18O enriched but statistically labeled ozone, isotopic
shifts as well as some 17O-17O coupling is partially resolved
(Figure 1b). Figure 1c shows the 16O-16O-17O spectra, made
from H217O (10% enrichment). The intensity ratio of the
terminal versus the central oxygen labeling is found to be 60:1.
The theoretical expectation would be 270:1. The difference can
be caused by traces of normal water in the HF and/or some
scrambling of the label during the reaction.
+
+O2
O2+ + H2O f •+OOOH2
+8 •O-O-OH
-H
+8
-O2
+8 O3 (3)
+8 O3 (4)
-H
OOOH+
We can only speculate on the mechanism of the ozone
formation, since no intermediates in the overall reactions have
been observed. One possibility is the electrophilic attack of
+
+O2
O2+ + H2O f O2 + H2O+•
+8 OH
-H
8
(9) Boelter, E. D.; Putnam, G. L., Lash, E. J. Anal. Chem. 1950, 22,
1533.
(10) Griggs, M.; Kaye, S. ReV. Sci. Instrum. 1968, 39, 1685.
(11) Hesse, G.; Bayer, E. Z. Naturforsch. B 1964, 19, 875.
(12) McFarlane, W.; McFarlane, H. C. E. Oxygen. In Multinuclear NMR;
Mason, Joan Ed.; Plenum Press: New York, 1987; Chapter 14, p 406.
-H
OOOH+
IR spectroscopy of matrix isolated ozone is the second method
we have used for analysis. Here we have chosen the 18O isotope

Chemical Synthesis of O3, Isotopic Labeling, and Redistribution
J. Am. Chem. Soc., Vol. 120, No. 34, 1998 8713
Table 1. Vibrational Wavenumbers (cm-1) for IR Band Positions (main matrix side) of 16O, 18O Isotopomers of Ozone Isolated in Neon and
Argon (in parentheses) Matrix and Their Assignments
assign acc
C2ν, Cs sym
16O3
Ia
18O16O16O
Ia
1.0
0.3
15
14)
0.3
0.9
9
6)
18O16O18O
16O18O16O
18O18O16O
18O3
3041.0
2782.8
2109.7
(2108.4
1796.0
1722.8
1104.3
(1104.5
1039.6
(1039.8
699.8
1.4
0.2
14
13)
0.2
0.7
1.6
0.7)
2994.7
2746.1
2089.1
(2087.6
1767.0
1693.4
1091.4
(1090b
1025.6
(1025.8
683.4
2975.4
2715.6
2059.0
(2057.5)
1733.1
1668.1
1073.5c
-
1016.7
(1016.7)
667.0c
(670.9)
2947.5
2701.0
2048.2
(2047.4)
1759.8
1683.0
1075.5c
-
2899.0
2677.8
2026.1
(2024.7)
1730.4
1653.3
1061.4
(1060.4)
991.6
2878.9
2631.9
1994.9
(1993.5)
1697.0
1628c
1042.7
(1042.7)
982.2
3ν3, 3ν2
2ν3 + ν2
3ν3, 2ν2
ν1 + ν2
ν2 + ν3
ν1(a1), ν1(a)
100
100)
7
100
100)
9
1006.0
(1006.3)
692.2
ν3(b1), ν2(a)
ν2(a1), ν3(a)
(991.6)
676.3
(680.2)
(982.5)
660.3
(664.0)
(703.6
7)
(687.2
11)
(696.2)
a Integrated relative intensities I(ν3) ) 100. b Broad band. c (0.5 cm-1
.
Table 2. Vibrational Wavenumbers (cm-1) of Isotopomers of O2
Isolated in an Argon Matrix
16O2
16O18O
18O2
monomer
dimer
1550.0
1544.9
1506.2
1501.6
1461.5
1457.2
Figure 3. IR spectra of ozone enriched in 18O isolated in an argon
matrix. Main matrix sites are of 16O3(1) and 16O16O18O (2), besides
18O16O18O (3), 16O18O16O (4), 16O18O18O (5), and 18O3 (6). Lower trace
without and upper trace with UV photolysis λ > 420 nm. Arrows in
the upper trace indicate increase of 16O-18O-16O and decrease of 16O-
16O-18O during photolysis, with little formation of O2.
Figure 2. IR spectra of ozone enriched in 18O isolated in an argon
matrix. Main matrix sites are of 16O3 (1) and 16O16O18O (2), besides
18O16O18O (3), 16O18O16O (4), 16O18O18O (5), and 18O3 (6). Lower trace
without and upper trace with UV photolysis λ ) 254 nm. Arrows in
the upper trace indicate increase of 16O-18O-16O and decrease of 16O-
16O-18O during photolysis, along with formation of O2.
with the mechanism of scrambling of the atoms within the
molecule. For an intramolecular exchange of the atoms within
one ozone molecule two mechanisms are conceivable: (1)
closing of the molecule to a three-membered-ring species and
opening at any of the three bonds of the regular triangle; and
(2) elimination of one terminal oxygen atom that would remain
in the environment of the residual O2 species, and recombination
of these two entities in a statistical manner.
In this connection it is noteworthy that the existence of a
regular, triangular ozone has been predicted by ab initio
calculations many times as being the second minimum on the
potential hyperface, with an energy =30 kcal/mol higher than
normal ozone.16-22 A barrier of 55 kcal is calculated between
the normal and triangular ozone. This transition state with C2V
symmetry and a bond angle of R = 83° is extremely close in
energy and geometry to the minimum of an excited state.
Historically the D3h ring structure of ozone as the ground state
has already been rejected by L. Pauling following guesses on
the heats of formation.23 But it took the measurement of the
for a larger band separation than would be achieved with 17O.
Also 18O is much less expensive than 17O. For matters of
unambiguous assignments vibrational data are collected on all
16O-18O ozone isotopomers, see Table 1. These data complete
(and in some part correct) previous measurements.13,14
The lower traces of Figures 2 and 3 show matrix IR spectra
of isotopic mixtures of ozone consisting mainly of 16O3 and
16O16O18O. It is interesting to note that the intensity of the ν1
band of 16O16O18O strongly increases in comparison to that of
16O3 due to the break of symmetry.
Labeled ozone should be of use for mechanistic studies, e.g.,
for the investigation of the reaction mechanism of the ozoniza-
tion of olefines (Crige´e reaction.15). Here we will only deal
(13) Green, D. W.; Ervin, K. M. J. Mol. Spectrosc. 1981, 88, 51.
(14) Anmin, T.; Ruifeng, L.; Zemin, S.; Xuefeng, Z.; Guosen, Y. Int. J.
Quantum Chem. 1989, 36, 765. Flaud, J.-M.; Camy-Peyret, C.; Malathy
Davy, V.; Rinsland, C. P.; Smith, M. A. H. J. Mol. Spectrosc. 1987, 122,
221-228. Perkin, A.; Vasserot, A.-M.; Flaud, J.-M.; Camy-Peyret, C.;
Rinsland, C. P.; Smith, M. A. H.; Malathy Davy, V. J. Mol. Spectrosc.
1990, 143, 311-317. Flaud, J. M.; Bourgeois, M. T.; Barbe, A.; Plateaux,
J. J.; Camy-Peyret, C. J. Mol. Spectrosc. 1994, 165, 464-469. Bourgeois,
M. T.; Flaud, J. M.; Barbe, A.; Regalia, L.; Camy-Peyret, C.; Plateaux, J.
J. J. Mol. Spectrosc. 1995, 172, 552-558.
(16) Wright, J. S. Can. J. Chem. 1973, 51, 139.
(17) Hay, P. J.; Dumming, T. H.; Goddard, W. A., III J. Chem. Phys.
1975, 62, 3912; Chem. Phys. Lett. 1973, 23, 457.
(18) Shih, S.; Buonker, R. J.; Peyerimhoff, S. D. Chem. Phys. Lett. 1974,
28, 463.
(19) Burton, P. G.; Harcey, M. Dawn, Nature 1977, 266, 826.
(20) Hay, P. J.; Dumming, T. H. J. Chem. Phys. 1977, 67, 2290.
(21) Jones, R. O. J. Chem. Phys. 1985, 82, 325.
(22) Xantheas, S. S.; Atakity, G. J.; Ellert, S. T.; Ruedenberg, K. J. Chem.
Phys. 1991, 94, 8054.
(15) Kuczkowski, R. L. Acc. Chem. Res. 1983, 16, 42.
(23) Pauling, L. J. Am. Chem. Soc. 1932, 54, 3570.

8714 J. Am. Chem. Soc., Vol. 120, No. 34, 1998
DimitroV et al.
electron diffraction24 and the microwave spectrum25 to establish
the ground-state structure as it is known until today.
that a monomolecular atomic exchange does not take place
thermally, leaving no room for speculation of a transient
triangular ozone species under these conditions.
We found that the terminal oxygen label scrambles to the
central position under various conditions, e.g., if gaseous ozone
is kept in an inert vessel (Kel-F) at room temperature and
approximately 1 bar of pressure for 1 day. Simultaneously
strong decomposition into oxygen is observed. We can show
that there is an intermolecular atomic exchange taking place
by mixing 17O2 and 16O3 under similar conditions and keeping
the mixture at room temperature for several days. In the ozone
remaining after 1 day we could qualitatively observe 17O
enrichment in all atomic positions (reaction 5). Alternatively
we have observed IR spectra of 18O3 and 16O2 (1:1, 200 mbar)
and find no scrambling within 2 h at room temperature. This,
however, changes when the mixture is heated to 100 °C for a
few minutes. It is noteworthy that intermolecular exchange is
not observed between different isotopomers of ozone under
similar conditions (6).
Irradiation of ozone in the matrix at 14 K has been done by
light of λ ) 254 and > 420 nm. The high-energy light produces
decomposition into oxygen and scrambling of the terminal 18
O
label in the remaining ozone, see Figure 2. The low-energy
light produces very little oxygen, but nevertheless scrambling
of the label occurs, see Figure 3.
Upon irradiation with λ ) 254 nm generation of O(1D) and
O2(1∆g) occurs.26 Because under matrix condition 1O2 is more
long living than 1O atoms, the latter changes into 3O, which is
prohibited from recombining back to ozone, and therefor has
time to escape from the matrix cage. Reaction of diffusing O
atoms with isolated ozone molecules will finally form two O2
in one matrix cage. As evidence of this we observe the splitting
of the vibrational bands of 16/18O2 oxygen isotopomers into two
lines, which we assign to a single oxygen molecule in a matrix
cage, and to two such molecules, see Table 2.
days, r.t.
17(18)
16O3 + 17(18)O2
8
O3 + 16O2
(5)
(6)
Ozone absorbs also in the visible region (Chappius band),
and in our matrix experiment with less energetic irradiation of
hours, 70 °C
3
3
λ > 420 nm only O2 and O are formed, so that (almost)
quantitative recombination can occur.
days, r.t.
N
16O3 + 18O3
16,16,18O3 etc.
While this discussion is based on the assumption that the
transient intermediate of the ozone decomposition is a van der
Waals molecule O2‚‚‚O, which is also supported by ab initio
calculations,22 the possibility of a triangular regular ozone as
the transient cannot be excluded totally. Short-time spectros-
copy is needed to address this important question.
Labeled ozone pumped through quartz tubes of 30 cm length
at 1 and 10-2 mbar results, beginning at 350 °C, in scrambling
of the label along with strong decomposition, as evidenced by
17O NMR spectroscopy.
To arrive at clearly monomolecular conditions ozone consist-
ing mainly of 16O3 and 16O-16O-18O mixed with a large excess
of argon (1:500) is pumped through a quartz tube at 10-4 mbar
of partial pressure. Not until 450 °C (the maximum temperature
possible) have we observed any scrambling of the label, as
evidenced by IR matrix spectroscopy. Therefore we conclude
Acknowledgment. We thank the Fonds der Chemischen
Industrie and the European Union (EU contract LAMOCS
ENV4-CT 95-0046) for financial support of this study.
JA980887N
(24) Shand, W.; Spurr, R. A. J. Am. Chem. Soc. 1943, 65, 179.
(25) Trambarulo, R.; Guosh, S. N.; Burrns, C. A.; Gordy, W. J. Chem.
Phys. 1995, 21, 851; Phys. ReV. 1993, 91, 222.
(26) International geophysics series: Warneck, P. Chemistry of the
natural atmosphere; Academic Press: London, 1987; Vol. 41.