Published on Web 10/06/2005
The Rotational Spectrum and Structure of HOOOH
Kohsuke Suma, Yoshihiro Sumiyoshi, and Yasuki Endo*
Department of Basic Science, Graduate School of Arts and Sciences, The UniVersity of Tokyo, Komaba, Meguro-ku,
Tokyo 153-8902, Japan
Received August 25, 2005; E-mail: endo@bunshi.c.u-tokyo.ac.jp
Hydrogen oxides, H2On with n ) 1 (i.e., water) and n ) 2 (i.e.,
hydrogen peroxide), are familiar species in various fields of natural
science. However, extended species (n > 2), namely, hydrogen
polyoxides, have been believed to be quite unstable due to the
repulsion between the lone-pair electrons on the oxygen atoms.
Thus, the questions, “how long can you make an oxygen chain?1”
and “if such oxygen chain molecules exist, what kind of roles do
these molecules play?” have long been fascinating chemists even
from the19th century.2 Recently, we have reported a microwave
observation of a radical species with three oxygen atoms, HO3,
and determined its precise molecular structure. HO3 is regarded as
a weak adduct of HO-O2 and is expected to be a sink of the
atmospheric OH radical.3 On the other hand, no detection of a closed
shell species, H2O3, in the gas phase has been reported yet. Its
existence and stability in the gas phase, as well as its precise
molecular structure, are still open questions. Recent theoretical
studies predict its existence in the gas phase. For example, McKay
et al. predicted that the oxygen chain could be extended infinitely
in a very low temperature gas.1 Several experiments on H2O3 in
the condensed phase have meanwhile been reported. Infrared
absorption spectra in a solid state4 and the Ar matrix5 and NMR
spectra in several organic solutions6 have shown its existence in
respective media. These reports indicate that H2O3 is far more stable
than was previously believed. Plesnicˇar et al. have recently reviewed
these experiments.6
Wentworth et al. reported that all antibodies were capable of
catalyzing the oxidation of water by singlet oxygen (1∆ O2) to
generate H2O2 and probably O3, as well, and postulated that these
antibodies carry the reaction through H2O3 as a key intermediate.7
Furthermore, H2O3 is expected to be involved in oxidation processes
that span atmospheric, environmental, and biological systems.6 It
is thus expected that precise rotational transitions could be used to
probe H2O3 and to elucidate its potential roles in these chemical
systems, if they are observed.
In the present study, we were able to detect H2O3 by Fourier
transform microwave (FTMW) spectroscopy8 and FTMW-mm-
wave double resonance9 and triple resonance spectroscopy, provid-
ing sufficient data to precisely determine its ground-state geometry.
H2O3 was produced in a pulsed discharge nozzle (PDN) by
discharging a gas mixture of O2 (10%) and Ar, passed through a
reservoir filled with a 30% hydrogen peroxide solution. Discharged
products were expanded as a supersonic jet into a FTMW
spectrometer cavity. Since H2O3 is a singlet molecule without any
hyperfine splittings and only one component of the dipole moment,
µb, has a nonzero value in the trans form (C2 symmetry), the
observable transition by our FTMW spectrometer, whose frequency
coverage is 4-40 GHz, is limited to only one, 303 - 212, at about
21 GHz, which was predicted by an ab initio calculation.10-12
Therefore, we performed double resonance spectroscopy by moni-
toring all possible candidate lines observed in the region of 21 (
2 GHz, to check which of the lines belongs to H2O3. The lines to
be checked by double resonance spectroscopy were selected
beforehand to a few lines by checking if they show a diamagnetic
behavior, are the discharge products, need both hydrogen peroxide
and O2, and do not depend on a particular buffer gas, Ar or Ne.
When one of the observed lines at 20 838.6 MHz was used as a
monitor for 303 - 212, double resonance signals at the predicted
regions, 45 (312 - 303), 79 (212 - 101), and 186 (321 - 212) GHz,
were observed. For the confirmation of the spectral carrier, a newly
developed triple resonance method was also applied to observe the
transition, 110 - 101, at 42 GHz. Details of this technique will be
reported elsewhere. In total, five rotational transitions, as shown
in Supporting Information, were observed. Neither a-type nor c-type
transitions were observed around the predicted frequencies, indicat-
ing that the dipole moment components, µa and µc, are 0 or very
small. These results support that the spectral carrier is H2O3 in the
trans form. We could not observe transitions of the cis isomer (Cs
symmetry) despite careful scans with similar experimental condi-
tions in sufficiently wide frequency regions. The transition at
20 838.6 MHz was also observed slightly weakly by discharging a
mixture gas of H2O/O2/Ar. Rotational transitions of D2O3 were thus
searched for by using the D2O/O2/Ar mixture gas. Three transitions
listed in the table in Supporting Information were observed by
FTMW spectroscopy. Using these transitions as monitors, two more
transitions were observed by the double resonance technique. In
total, five rotational transitions were observed for D2O3. Although
the dipole moments, µb, of H2O3 and D2O3 are almost identicals
about 1 D predicted by B3LYP/cc-pVQZ calculations10ssignals
of D2O3 were much weaker than expected from those of H2O3. By
irradiating the microwave parallel to the supersonic jet beam,13 small
hyperfine splittings due to the electric quadrupole moment of
deuterons were resolved. Since D2O3 has two equivalent deuterons,
which obey Bose-Einstein statistics, the rotational energy levels
consist of ortho (Itotal ) 0, 2) and para (Itotal ) 1) states. The 110-
101 transition (para) is well reproduced by the electric quadrupole
coupling constants obtained by the ab initio calculation.10
However, extra lines were observed for two other transitions 303-
12 and 211-202, as shown in the figures in Supporting Information.
2
They were observed for both ortho and para states and could not
be explained by changing the values of the electric quadrupole
coupling constants within a reasonable range. On the other hand,
no such doubling was observed for H2O3. At present, we were not
able to explain the origin of these splittings. They could be
considered to be the splittings caused by the tunneling between
the two chiral isomers through the torsional motions of OH moieties.
However, these splittings are estimated to be extremely small by
the simple WKB (Wentzel-Kramers-Brillouin) approximation
along the one-dimensional torsion vibration coordinate using a
potential energy surface obtained by the B3LYP/cc-pVTZ calcula-
tion.10 H2O3 should be considered to be a chiral molecule at the
time scale of our experimental apparatus, and no tunneling splittings
should be observed. Although parity-violation effect may also
9
14998
J. AM. CHEM. SOC. 2005, 127, 14998-14999
10.1021/ja0556530 CCC: $30.25 © 2005 American Chemical Society