Synthesis of hydrogen polyoxides H2O4 and H 2O3 and their characterization by raman spectroscopy

Article abstract of DOI:10.1002/ejic.201100767

Hydrogen tetroxide (H₂O₄) and hydrogen trioxide (H₂O₃) have been prepared in relatively large quantities as components of peroxy radical condensates. The polyoxides were characterized by the Raman spectra of the oxygen framework. Lines at 500, 756, and 878 cm ^-1 correspond to skeletal oscillations of H₂O₃ in the condensate (OOO bend, asymmetric OO stretch, symmetric OO stretch). Lines at 449, 589, 624, 827, and 865 cm^-1 match the skeletal vibrations of H₂O₄ (OOO bend 1, OOO bend 2, center OO stretch, asymmetric OO stretch, symmetric OO stretch). The line of the O₄ torsion vibration of hydrogen tetroxide was not observed experimentally. The assignment was confirmed by B3LYP/6-31+G(d,p) calculations of the vibrational spectra. According to the information available in the literature, this work is the only report of the preparation of H₂O₄ in significant concentrations and its Raman spectrum. For the first time convincing evidence is provided for the existence of hydrogen tetroxide (in low-temperature solids). Copyright

Full text of DOI:10.1002/ejic.201100767

FULL PAPER
DOI: 10.1002/ejic.201100767
Synthesis of Hydrogen Polyoxides H2O4 and H2O3 and Their Characterization
by Raman Spectroscopy
Alexander V. Levanov,*[a] Dmitri V. Sakharov,[b] Anna V. Dashkova,[a] Ewald E. Antipenko,[a]
and Valeri V. Lunin[a]
Dedicated to the memory of Professor N. I. Kobozev (1903–1974)
Keywords: Hydrogen tetroxide / Hydrogen trioxide / Raman spectroscopy / Oxygen / Radical reactions
Hydrogen tetroxide (H2O4) and hydrogen trioxide (H2O3)
OO stretch, symmetric OO stretch). The line of the O4 torsion
have been prepared in relatively large quantities as compo- vibration of hydrogen tetroxide was not observed experimen-
nents of peroxy radical condensates. The polyoxides were
characterized by the Raman spectra of the oxygen frame-
work. Lines at 500, 756, and 878 cm–1 correspond to skeletal
tally. The assignment was confirmed by B3LYP/6-31+G(d,p)
calculations of the vibrational spectra. According to the infor-
mation available in the literature, this work is the only report
oscillations of H2O3 in the condensate (OOO bend, asymmet- of the preparation of H2O4 in significant concentrations and
ric OO stretch, symmetric OO stretch). Lines at 449, 589, 624,
827, and 865 cm–1 match the skeletal vibrations of H2O4
(OOO bend 1, OOO bend 2, center OO stretch, asymmetric
its Raman spectrum. For the first time convincing evidence
is provided for the existence of hydrogen tetroxide (in low-
temperature solids).
peroxide and ozone[2] as well as in the reaction of ozone
with a number of organic compounds.[3] There is evidence
that H2O3 may play an important role in immune reactions
in living organisms.[4]
Introduction
Hydrogen is the most widespread element in the universe,
and oxygen is the same on Earth. However, there are only a
few compounds of these two elements: H2O, H2O2, unstable
radicals OH, HO2, and HO3, and hydrogen polyoxides
H2O3 and H2O4. The polyoxides are of interest from the
viewpoint of chemical structure theory as molecules with
catenated oxygen chains and higher members of the homo-
logical series starting from H2O and H2O2. Furthermore,
Questions such as “how many members are there in the
homological series H2O, H2O2, ..., H2On?”[5] and “what are
the properties of polyoxides H2On (n Ͼ 2) if they exist?”[6]
have occupied the minds of chemists since the end of the
19th century. H2O3 was first mentioned in 1880 by Ber-
thelot as a possible intermediate in the decomposition of
hydrogen peroxide.[7] H2O4 was proposed in 1895 by Mend-
eleev as a higher member of the H2On homological series
by analogy with the corresponding sulfur compounds H2S,
H2S2, H2S3, and H2S4.[8] A detailed review of the subject
prior to 1968 has been written by Venugopalan and Jones.[9]
The experimental investigation of H2O3 and H2O4 is
closely connected with research on the so-called peroxy rad-
ical condensate. This condensate is a glassy solid, obtained
on a cold surface (80 K) by condensation of water, hydro-
gen peroxide vapor or H2/O2 mixtures dissociated in a low-
pressure electrical discharge and by low-temperature inter-
action of hydrogen atoms with oxygen or with ozone. Upon
heating above 150 K, the condensate decomposes vigor-
ously to yield large amounts of gaseous oxygen and concen-
trated hydrogen peroxide solution. The formation of hydro-
gen peroxide from a low-temperature reaction of hydrogen
atoms with molecular oxygen was detected by Boehm and
[1]
H2O3 and H2O4 are intermediates of the important radi-
·
·
cal/radical reactions HO2 + OH· Ǟ H2O + O2 and HO2
+
·
HO2 Ǟ H2O2 + O2. These reactions appear almost every-
where on Earth and are of great importance in atmospheric
chemistry, combustion and flame chemistry, oxidation reac-
tions in living organisms, aqueous radiochemical processes,
and aqueous ozone chain reactions. It was found that H2O3
is formed during the thermal reaction between hydrogen
[a] Department of Chemistry, M. V. Lomonosov Moscow State
Unversity
Leninskiye Gory 1, Building 3, 119991 Moscow, Russia
Fax: +7-495-939-4575
E-mail: levanov@kge.msu.ru
[b] A. V. Topchiev Institute of Petrochemical Synthesis, Russian
Academy of Sciences
Leninsky Prospect 29, 119991 Moscow, Russia
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201100767 or from
the author.
5144
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2011, 5144–5150

Hydrogen Polyoxides H2O4 and H2O3
Bonhoeffer in 1926.[10] Hydrogen peroxide was found as a ganic compounds[3] or hydrogen peroxide[2] in organic sol-
1
product of the low-temperature condensation of electrodis- vents. H2O3 has been detected by H and 17O NMR spec-
sociated water vapor by Lavin and Stewart in 1929,[11] and troscopy.
of an electrodissociated gas mixture of 2H2 + O2 by Brewer
Giguère et al. did not succeed in observing the true set of
and Westhaver in 1930.[12] In 1932, Geib and Harteck dis- HOOOOH oxygen framework fundamental vibrations for
covered that on decomposition of the condensate obtained H2O4, which is discussed below. Subsequent attempts to ob-
from the low-temperature reaction of H + O2, not only was tain H2O4 in a low-temperature matrix and to characterize
hydrogen peroxide produced but also large amounts of gas- the polyoxide by IR spectroscopy were not successful. Tso
eous oxygen.[13]
and Lee[23] tried to prepare HOOOOH by photolysis of
In 1940, Ohara suggested that oxygen evolution and hy- formaldehyde or glyoxal in an O2 matrix at 15 K. The pri-
drogen peroxide formation upon warming the condensates, mary products of the photooxidation are two HO2· radicals
obtained from dissociated water vapor and by H + O2 reac- and one or two CO molecules. In the secondary reaction,
tion, were due to the decomposition of hydrogen tetroxide, the HO2· radicals do not give HOOOOH but a cyclic dimer
H2O4 Ǟ H2O2 + O2, which is stable only at low tempera- (HO2)2, which complexes with CO as determined by IR
tures.[14] For these condensates the molar ratio of O2/H2O2 spectroscopy.[23] This conclusion was confirmed theoretic-
in the decomposition products was less than unity. This ally by Schaefer et al.[24] Thus, the existence of HOOOOH
could be interpreted by assuming that the condensates con- in low-temperature solids was thrown to doubt.
tained H2O4 as well as H2O2 and H2O.
In this paper we describe the synthesis of condensates
Kobozev et al.[15] were the first to synthesize peroxy radi- containing both H2O4 and H2O3 in significant amounts.
cal condensates by reaction of hydrogen atoms with liquid The polyoxides, H2O4 in particular, were characterized by
ozone. The molar ratio of O2/H2O2 in its decomposition Raman spectra of their oxygen frameworks, and the experi-
products was equal to unity, which is distinct from the con- mental spectra are in good agreement with the results of
densates obtained by other methods, and this ratio was later quantum chemical calculations.
confirmed by others.[16] Magnetic susceptibility measure-
ments of the condensate showed that it did not contain oc-
Results and Discussion
cluded oxygen; oxygen was only formed during decomposi-
tion at elevated temperatures.[17] These results were taken
The peroxy radical condensates were obtained by low-
as evidence that the condensate synthesized by H + O3(l) temperature condensation of O2 and H2 gas mixtures disso-
contained only H2O4 and water. ciated in a low-pressure microwave discharge. Raman spec-
However, subsequent investigations of the condensates tra of the condensates show new lines (solid arrows), which
obtained from electrodissociated hydrogen/oxygen mixtures do not arise from H2O or H2O2 (Figures 1 and 2). As the
have shown that, under certain conditions, the molar ratio oxygen content increases in the initial mixture, the new lines
of O2/H2O2 in their decomposition products can be as high become more pronounced (Figure 2). In addition, signals
as 10:1 and even higher.[18] These facts show that the con- arising from ozone (706, 1039, and 1106 cm–1) and molecu-
densate could contain H2O3; on warming, H2O3 decom- lar oxygen (1552 cm–1) appear.
poses according to the reaction H2O3 Ǟ H2O + O2. All the
Other lines in the spectra of peroxy radical condensates
information on the subject leads to the conclusion that the have analogs in the spectrum of hydrogen peroxide (Fig-
condensates contain not only H2O4 but also H2O3, and ure 1).[25] The intense broad band with a maximum at
H2O3 can be the predominant component.
3180 cm–1 corresponds to the stretching oscillations of OH
Vibrational spectroscopy is the most suitable method for groups bound with hydrogen bonds. The bending oscilla-
detection of the active components in the condensates and tions of HOO give rise to the broad signal with a maximum
their structure determination. Some new bands in the IR at 1448–1468 cm–1. The combination of HOO bending
spectra of condensates were first observed by Yagodovskaya oscillations reveals itself as a peak with a maximum at
and Nekrasov,[19] which were tentatively assigned to H2O4 2840 cm–1. In the condensate prepared by codeposition of
and H2O3. The first convincing evidence for the presence of nondissociated H2O2 and H2O vapors, these two HOO
H2O3 in the peroxy radical condensates was found by Gig- bending lines have frequencies of 1425 and 2830 cm–1,
uère et al.[20,21] They reported fundamental vibrations of respectively. In pure crystalline H2O2, the frequencies are
the oxygen framework in H2O3 by studying IR and Raman 1385, 1420 and 2760, 2835 cm–1. The broad unresolved
spectra of the condensates.
shoulder at 40–380 cm–1 is in the same region where lattice
Detailed spectroscopic characterization of H2O3 has vibration lines are observed in the spectrum of crystalline
been performed only recently. Engdahl and Nelander ob- H2O2. The broad weak signals of fused quartz (from the
served all the fundamental vibrations of HOOOH isolated reactor) may appear in the region 40–550 cm–1.
in an Ar matrix, and their experimental results agree very
An intense peak in the neighborhood of 880 cm–1 repre-
well with the theory.[22] Suma et al. determined the precise sents the convolution of three signals (Supporting Infor-
geometrical structure of H2O3 in the gas phase using ad- mation). This shows itself particularly well in oxygen-rich gas
vanced Fourier transform microwave spectroscopy tech- mixtures where a shoulder at ca. 865 cm–1 can be observed.
niques.[6] Considerable progress has also been made in
The shape of the peak near the maximum led us to sus-
studying the H2O3 formation in reactions of ozone with or- pect that there are several signals with heights positioned
Eur. J. Inorg. Chem. 2011, 5144–5150
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5145

A. V. Levanov et al.
FULL PAPER
The new lines can be divided into two sets: 500, 756, and
878 cm–1 and 449, 589, 624, 827, and 865 cm–1. As the oxy-
gen content increases in the initial mixture, the line inten-
sities of the second set grow more rapidly than those of the
first (Figure 2). Within both sets, the intensities change in
the same manner.
In order to elucidate whether the new lines correspond
to normal vibrations, which include only O atoms or both
O and H atoms, the spectra of condensates obtained from
O2 + H2 were compared to those obtained from O2 + D2O
gas mixtures (Figure 3). The substitution of deuterium for
hydrogen does not cause any appreciable shift of the new
lines. Therefore, we conclude that the new lines are due to
normal vibrations incorporating O atoms only.
Figure 1. Raman spectra of peroxy radical condensates (a and b)
in comparison with Raman spectra of condensates of H2O2
(70 mol-%) in water (c), solid crystalline hydrogen peroxide (d), and
solid water (e). New lines are marked with solid arrows. The peroxy
radical condensate was synthesized from a mixture of 86vol.-% O2
and 14vol.-% H2. The spectra are normalized with reference to the
height of the peak at ca. 880 cm–1, which is taken as 100% (a, b,
c, e) or 200% (d).
Figure 3. Spectra of the peroxy radical condensates synthesized
from oxygen/deuterium (a, b) and oxygen/hydrogen (c, d) systems.
Experimental conditions: (a, b) initial gas mixture 70vol.-% O2 +
30vol.-% D2O [gas flow rate 3.0 L/h (STP)]; (c, d) initial gas mix-
ture 86vol.-% O2 + 14vol.-% H2. The spectra are normalized with
reference to the height of the peak at ca. 880 cm–1.
Thus, the new lines result from compounds of oxygen
and hydrogen other than H2O, H2O2, O2, and O3. Free radi-
·
cals, HO2 in particular, should be virtually “invisible” in
the Raman spectra as their concentration in the conden-
sates is low (less than 0.5wt.-%).[26] (HO2)2 and isomeric
hydrogen peroxide H2OO can also be excluded from con-
sideration.[27] It is likely that the new lines arise from H2O4
and H2O3.
In order to assign the new lines correctly, quantum chem-
ical calculations for the vibrational spectra of hydrogen per-
oxide (HOOH) and polyoxide (HOOOH, HOOOOH) mo-
lecules were performed. The density functional method
B3LYP/6-31+G(d,p) was employed (Supporting Infor-
mation).[28] The calculations were carried out for isolated
molecules in an ideal gaseous state. The experimental Ra-
man spectra were taken from glassy solid condensates with
Figure 2. Raman spectra of peroxy radical condensates synthesized
from O2/H2 mixtures of different compositions. O2 content in the
initial gas mixture in vol.-%: 98.3 (a), 96.6 (b), 86 (c), 80 (d), 70
(e), 60 (f), 50 (g), 33 (h); gas flow rate in L/h (STP, standard tem-
perature and pressure, 20 °C and 1 atm): 3.2 (a, c–h), 1.6 (b). The
spectra are normalized with reference to the height of the peak at
ca. 880 cm–1.
very close to 880 cm–1. Numerical deconvolution disclosed strong hydrogen bonding. Evidently, the matrix exerts a
that the peak is a superposition of three lines with maxima strong influence on oscillations including hydrogen atoms
at 865, 878, and 881 cm–1 (Supporting Information). The and much less influence on oxygen framework oscillations.
line at 881 cm–1 is caused by OO stretching vibrations of For this reason, comparison of the calculated and experi-
H2O2. The lines at 865 and 878 cm–1 are new and do not mental frequencies makes sense only for oxygen framework
have analogs in the H2O2 spectrum.
oscillations. The calculated frequencies were scaled in a way
5146
www.eurjic.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2011, 5144–5150

Hydrogen Polyoxides H2O4 and H2O3
that the calculated frequency of the H2O2 OO stretching 880 cm–1 is superimposed by the lines of the symmetric O–
vibration (unscaled value 921 cm–1) was equal to its experi-
O
stretching vibrations of H2O2 (881 cm–1), H2O3
mental value (881 cm–1) in solid hydrogen peroxide.
(878 cm–1), and H2O4 (865 cm–1). The lines of other oxygen
A comparison of the experimental condensate spectrum framework oscillations are clearly separated. The calculated
and the calculated Raman spectra of HOOH, HOOOH, spectra fit those obtained experimentally fairly well (Fig-
and HOOOOH is given in Figure 4 and Table 1. The first ure 4). The difference in line positions is not principal, be-
set of new lines is consistent with the three skeletal vi- cause the strong intermolecular interactions in the conden-
brations of hydrogen trioxide. The second set corresponds sate were not taken into account in the calculations.
to the vibrations of the hydrogen tetroxide oxygen frame-
Denis and Ornellas[29] have calculated the vibrational
work. However, the second set comprises only five lines spectra of H2O3 and H2O4 at higher levels of theory
when there are six skeletal vibrations of H2O4. The line aris- [CCSD(T)/cc-pVTZ and B3LYP/6-311+G(3df,2p)], but Ra-
ing from the O4 torsion vibration does not manifest itself man intensities have not been reported. The CCSD(T)
in the experiment. This was expected, because – according methodology gives the frequencies closest to the experimen-
to the calculations – this line and the lines of the other tal values obtained in this work. The frequencies obtained
torsion oscillations of H2O2, H2O3, and H2O4 are of low by B3LYP methods are systematically higher than those ob-
Raman intensity and fall in the region of the rather strong served experimentally, which is due to methodological
broad unresolved signal (“shoulder”) at 50–400 cm–1. This reasons, but the performance of these methods is still good.
shoulder is present in all of our spectra and obscures the Our frequencies (unscaled) calculated with the 6-31+G(d,p)
O4 torsion line.
basis set are closer to the experimental results than those
obtained with the 6-311+G(3df,2p) basis set.[29]
In addition to this work, the vibrational spectrum of
H2O3 and H2O4 in a peroxy radical condensate was re-
ported by Giguère et al.[20,21] (see Table 1 for a comparison
of the frequencies). For H2O3, the asymmetric OO stretch-
ing (755 cm–1) and OOO bending (500 cm–1) frequencies co-
incide with our values. However, the shoulder signal at ca.
855 cm–1, which Giguère et al. assigned to the H2O3 sym-
metric OO stretching vibration, is in fact due to the analo-
gous vibration of H2O4. Its frequency is equal to 865 cm–1
in our experiments. The line of the symmetric OO stretching
vibration of H2O3 is at 878 cm–1 and is visually almost in-
distinguishable from that of H2O2 at 881 cm–1. The lines at
865 and 878 cm–1 were identified by the numerical decon-
volution of the peak at ca. 880 cm–1 as mentioned above.
Giguère et al.[20,21] observed with certainty only two oxy-
gen framework lines for H2O4 with frequencies of 820–830
and 450 cm–1 in their experiment. This is because their con-
densates contained less than 0.6% of this polyoxide.[30]
Other frequencies were assumed based on rather unclear
experimental signals and by analogy with the vibrational
spectra of the homological series of hydrogen polyoxides
and polysulfides. As a result three (out of six) frequencies
assigned by Giguère et al.[21] to H2O4 skeletal oscillations
disagree with our new experimental and theoretical values
and with the theoretical frequencies of Denis and Or-
nellas.[29] We were able to prepare condensates that con-
tained up to 20% of H2O4, which allowed us to characterize
its vibrational spectrum in greater detail.
Figure 4. Experimental Raman spectrum of the peroxy radical con-
densate and calculated Raman spectra of hydrogen tetroxide, hy-
drogen trioxide, and hydrogen peroxide in the range of the oxygen
framework vibrations. The spectra are normalized with reference
to the height of the peak at ca. 880 cm–1. The condensate was syn-
thesized from the initial gas mixture 96.6vol.-% O2 + 3.4vol.-% H2
[gas flow rate 1.6 L/h (STP)].
Estimates of the composition of the peroxy radical con-
densates are presented in Figure 5 (see Exp. Sect. for the
composition determination procedure). At an oxygen con-
tent of 60–70%, the prevailing condensate components are
H2O2 and H2O3. The largest amount of H2O3 in the con-
densate was 30%, which was attained when the initial gas
On the whole, the experimental spectrum in the region mixture contained 70% O2. With an increased oxygen ratio
350–1000 cm–1 represents the superposition of the lines of in the initial gases, the H2O2 and H2O3 content decreased
the oxygen framework oscillations of HOOH, HOOOH, and that of H2O and H2O4 increased. The condensate pre-
and HOOOOH. The most intense signal in the vicinity of pared from 98% O2 and 2% H2 had the highest concentra-
Eur. J. Inorg. Chem. 2011, 5144–5150
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5147

A. V. Levanov et al.
FULL PAPER
Table 1. Frequencies of fundamental vibrations of H2O3 and H2O4 oxygen frameworks [cm–1].
Approximate normal
mode[a]
This work,
Giguère et
Engdahl and This work,
Nelander,[d,e] calculated
experimental B3LYP/6-
31+G(d,p),
This work,
calculated
B3LYP/6-
31+G(d,p),
unscaled
Denis and Or-
Denis and Or-
nellas,[29] calcu- nellas,[29] cal-
lated, B3LYP/6- culated,
experimental al.,[b,c] ex-
perimental
311+G(3df,2p)
CCSD(T)/cc-
pVTZ
scaled
HOOOH
Symm OO stretch
Asym OO stretch
OOO bend
878
756
500
ca. 855
755
500
821.0
776.3
509.1
878
760
486
917
795
508
925
795
516
879
784
508
HOOOOH
Symm OO stretch
Asym OO stretch
Center OO stretch
OOO bend 1
OOO bend 2
O4 torsion
865
827
624
589
449
–
ca. 855
ca. 823
ca. 764
(435)
450
ca. 98
873
814
594
567
462
162
912
851
621
593
483
169
922
847
629
607
479
172
882
833
672
601
470
166
[a] The descriptions of H2O4 normal modes in this work and in refs.[21,29] are somewhat different. [b] The values are from the closing
work,[21] which is based on the experimental results from refs.[20] [c] One should consult the paper by Giguère at al.[21] with regard to
obtaining and assigning the frequencies. For H2O4, only the peaks at 450 and ca. 823 cm–1 were clearly observed. [d] Experimental
values from ref.[22] The authors[22] observed the complete set of nine HOOOH fundamental oscillations, and not just oxygen framework
fundamentals. [e] The difference between the frequencies reported in this work and by Engdahl and Nelander[22] is most likely to result
from the different environments in the two cases. In this work H2O3 was synthesized as a component of peroxy radical condensates,
which consist of H2O, H2O2, H2O3, and H2O4 in comparable quantities, with strong intermolecular interactions between the constituents.
Engdahl and Nelander prepared H2O3 isolated in an Ar matrix with virtually no intermolecular interactions.
tion of H2O4 (20%), although its main component was H2O H2O3 in the condensate (OOO bend, asymmetric OO
(55%). The absolute quantity of condensate in our experi- stretch, symmetric OO stretch). The lines at 449, 589, 624,
ments was about 20 mmol. Therefore, condensates with rel- 827, and 865 cm–1 match the skeletal vibrations of H2O4
atively high contents of H2O3 and H2O4 have been pre- (OOO bend 1, OOO bend 2, center OO stretch, asymmetric
pared. This made it possible to perform an experimental OO stretch, symmetric OO stretch). The O4 torsion vi-
determination of the important properties of the polyox- bration of hydrogen tetroxide was not observed experimen-
ides, such as their heat of decomposition.
tally. According to the data available in the literature, this
work is the first to report the preparation of H2O4 in a
significant concentration and describe its Raman spectrum.
Experimental Section
General: Peroxy radical condensates were synthesized from oxygen/
hydrogen gas mixtures of different compositions and from D2O
vapor/O2 mixtures, dissociated at low pressure (0.85–1.0 Torr) by
electrodeless microwave discharge. The experimental setup was es-
sentially a conventional discharge-flow system with a low-tempera-
ture reactor-cryostat filled with liquid nitrogen. The condensate
was formed on the cryostat “finger”. Immediately after the synthe-
sis, the condensate was investigated in situ by Raman spectroscopy.
The condensate was decomposed by heating, with the formation of
gaseous O2 and liquid H2O2 solution in water. The oxygen evolved
was measured manometrically. The H2O2 solution was weighed, the
amount of H2O2 was determined by titration with KMnO4, and
the amount of water by difference. The condensate composition
was determined from the amounts of the decomposition products
and the relative intensities of the H2O3 and H2O4 peaks in the
Raman spectra. A detailed description of the procedures is given
below.
Figure 5. Composition of the peroxy radical condensate vs. O2 con-
tent in the initial gas mixture O2/H2.
Experimental Setup and Procedure: The low-temperature reactor is
shown in Figure 6. The outer part of the reactor was made of fused
quartz of optical quality. The inner cryostat (“finger”), was made
of glass, and its bottom was a polished stainless steel conical cap.
The cap was soldered to the Kovar junction welded with glass. The
cryostat was filled with liquid nitrogen at 77 K. The parts of the
reactor were connected by ground conical joints. Apiezon grease
Conclusions
Hydrogen polyoxides were prepared in relatively large
quantities as components of peroxy radical condensates.
The polyoxides were characterized by the vibrational Ra-
man spectra of their oxygen frameworks. The lines at 500,
756, and 878 cm–1 correspond to skeletal oscillations of M was used on all the joints and stopcocks of the setup.
5148
www.eurjic.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2011, 5144–5150

Hydrogen Polyoxides H2O4 and H2O3
were measured at one definite point on the finger cap surface, lo-
cated at the half-height of the cone. Numerical deconvolution of
peaks in the Raman spectra was performed with the PeakFit 4.12
program.[32]
Analysis of Decomposition Products: After Raman investigation, li-
quid nitrogen was removed from the cryostat and the reactor
heated up to room temperature. The condensate decomposed to O2
and aqueous hydrogen peroxide solution. The oxygen evolved was
measured by its pressure in a known volume. Before the decomposi-
tion, a known amount of dry gaseous oxygen was admitted into
the setup. This approach ensured that the evaporation of H2O and
H2O2 was negligible. The total pressure was recorded over the en-
tire course of the decomposition with the electronic pressure gauge
until the setup reached room temperature. The pressure of the oxy-
gen evolved was deduced from the total pressure vs. time curve.
At high oxygen content in the initial gas mixture, the condensates
contained quantities of ozone and occluded molecular oxygen. As
a rule, these quantities were negligible compared to the main com-
ponents. Only when the oxygen content was more than 95% did
they become appreciable. On warming the condensate, ozone and
occluded molecular oxygen were the first to be released into the
gas phase, and the molecular oxygen formed in the breakdown of
H2O3 and H2O4 was subsequently evolved. This allowed us to sepa-
rate these gases with some degree of certainty. In this work the
composition of condensates is given without the content of ozone
and occluded molecular oxygen.
Figure 6. Reactor for the synthesis of peroxy radical condensates
and their investigation by Raman spectroscopy.
The liquid decomposition products (H2O2 and H2O) were collected
in a special trap and remained on the reactor walls. The trap was
weighed, the amount of H2O2 in the trap was determined by ti-
tration with potassium permanganate according to a standard tech-
nique,[33] and the amount of H2O was determined by difference.
The reactor walls were rinsed with distilled water, and the amount
of H2O2 on the walls was also determined by titration. The amount
of H2O on the walls was estimated from the ratio of H2O2/H2O in
the trap. This procedure enabled all the condensate decomposition
products to be determined.
The setup was pumped with Vacfox VC-75 and GLD-201 vacuum
pumps. The input gases (very high purity grade) were taken from
cylinders with pressure regulators and conducted to the setup
through calibrated capillaries. The gas flow rates were regulated by
varying pressure at the capillary inlets. The setup was furnished
with a thermostatted vaporizer for admission of H2O or D2O va-
pors into the reactor. The overall gas flow rate was 3.2 L/h (STP)
in nearly all experiments, unless otherwise noted.
Determination of Peroxy Radical Condensate Composition: The
main components of the peroxy radical condensates were H2O,
H2O2, H2O3, and H2O4. Their content in the condensates was
evaluated from the amounts of condensate decomposition products
(O2, H2O, H2O2) and the relative intensities of the H2O3 and H2O4
peaks at 500 and 449 cm–1 in the Raman spectra of the condensates.
The following relationships hold true:
The pressure during the synthesis was controlled by a differential
manometer filled with vacuum pump oil (density 0.88 g/cm3), and
took values between 0.85 and 1.0 Torr. The pressure during the
condensate decomposition was measured with an AIR-20/M2 elec-
tronic pressure gauge connected to a computer.
The discharge was sustained by a Luch-11 (2450 MHz) medical
microwave generator, whose nominal output power was 12 W in
O2/H2 mixtures and 42 W in the O2/D2O mixture. The design of
the microwave discharge cavity was similar to that described by
Broida et al.[31] (type 5).
n(H2O) + n(H2O3) = N
,
H2O
n(H2O2) + n(H2O4) = N
,
H2O2
n(H2O3) + n(H2O4) = N
,
O2
n(H2O4)/n(H2O3) = a·I449/I500
,
Before the synthesis, the gas flow rates were set, and the discharge
was initiated. The reactor “finger” was filled with liquid nitrogen,
and the synthesis started. The gas mixture passed through the dis-
charge and dissociated before it flew onto the cold finger cap where
the peroxy radical condensate was formed in reactions of active
particles from the gas phase. The duration of the synthesis was 1.5–
5 h.
where n(H2O), n(H2O2), n(H2O3), and n(H2O4) are the quantities
of the condensate components, NH O, NH O , and NO are the quan-
tities of the decomposition products, I4249 and I500 are the peak
areas, and a is a calibration coefficient. The value a = 0.92 was
estimated on the basis of the theoretical Raman intensities of the
corresponding lines, calculated in this work. The relationships rep-
resent a nondegenerate system of linear equations whose solution
gives the composition of the condensate. The results of the compo-
sition determination are not highly precise. The main reasons are
that: (1) the weight of liquid decomposition products was deter-
mined with a relative error of 3–10%, (2) at an O2 content in the
initial gas mixture of more than 95%, ozone and occluded molecu-
lar oxygen in the condensate interfere with the determination of
2
2
2
When the synthesis was complete, the condensate was investigated
in situ by Raman spectroscopy. The spectra were recorded with a
Horiba Jobin Yvon LabRam HR 800 UV spectrometer (diffraction
grating 1800 or 300 lines/mm) equipped with an external sensing
unit (“superhead”). Green radiation (514.532 nm) from an ion ar-
gon laser was employed for excitation with the power at the sample
of 20 mW and the exposure time of about 100 s. All the spectra
Eur. J. Inorg. Chem. 2011, 5144–5150
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5149

A. V. Levanov et al.
FULL PAPER
1226–1228; b) A. R. Gromov, E. E. Antipenko, B. V. Strakhov,
Zh. Fiz. Khim. 1990, 64, 77–82; Russ. J. Phys. Chem. 1990, 64,
40–43.
T. V. Yagodovskaya, L. I. Nekrasov, Zh. Fiz. Khim. 1964, 38,
1750–1756; Russ. J. Phys. Chem. 1964, 38, 953–956; T. V. Yago-
dovskaya, L. I. Nekrasov, Zh. Fiz. Khim. 1966, 40, 1304–1312;
Russ. J. Phys. Chem. 1966, 40, 698–702.
P. A. Giguère, K. Herman, Can. J. Chem. 1970, 48, 3473–3482;
X. Deglise, P. A. Giguère, Can. J. Chem. 1971, 49, 2242–2247;
J. L. Arnau, P. A. Giguère, Can. J. Chem. 1973, 51, 1525–1529.
J. L. Arnau, P. A. Giguère, J. Chem. Phys. 1974, 60, 270–273.
A. Engdahl, B. Nelander, Science 2002, 295, 482–483.
T.-L. Tso, E. K. C. Lee, J. Phys. Chem. 1985, 89, 1618–1631.
J. T. Fermann, B. C. Huffman, G. S. Tschumper, H. F.
Schaefer III, J. Chem. Phys. 1997, 106, 5102–5108.
J. L. Arnau, P. A. Giguère, M. Abe, R. C. Taylor, Spectrochim.
Acta Part A 1974, 30, 777–796.
the quantity of molecular oxygen evolved on condensate decompo-
sition, and (3) the condensate composition is nonuniform along its
height.
[19]
[20]
Supporting Information (see footnote on the first page of this arti-
cle): Illustration of the deconvolution of the composite peak at ap-
proximately 880 cm–1 in the Raman spectra of peroxy radical con-
densates; description of the calculation methods; optimized struc-
tures for HOOH, HOOOH, HOOOOH calculated in this work.
[21]
[22]
[23]
[24]
Acknowledgments
This work was supported by the Russian Foundation for Basic Re-
search (grant # 09-03-00630-a).
[25]
[26]
[1] Polyoxides HOOOH and HOOOOH, as well as radical com-
plexes (OH), (HO2) and (HO2)2; see, for example: a) A. J. El-
liot, G. V. Buxton, J. Chem. Soc. Faraday Trans. 1992, 88,
2465–2470; b) M. A. Vincent, I. H. Hillier, J. Phys. Chem. 1995,
99, 3109–3113; c) X. Xu, R. P. Muller, W. A. Goddard III,
Proc. Natl. Acad. Sci. USA 2002, 99, 3376–3381; d) R. Zhu,
M. C. Lin, PhysChemComm 2001, 23, 1–6; e) R. Zhu, M. C.
Lin, Chem. Phys. Lett. 2002, 354, 217–226; f) M. C. Foti, S.
Sortino, K. U. Ingold, Chem. Eur. J. 2005, 11, 1942–1948; g)
J. M. Anglada, S. Olivella, A. Sole, J. Phys. Chem. A 2007, 111,
1695–1704.
[2] P. T. Nyffeler, N. A. Boyle, L. Eltepu, C. H. Wong, A. Esch-
enmoser, R. A. Lerner, P. Wentworth Jr., Angew. Chem. 2004,
116, 4756–4759; Angew. Chem. Int. Ed. 2004, 43, 4656–4659.
[3] B. Plesnicˇar, Acta Chim. Slov. 2005, 52, 1–12, and references
cited therein; A. Bergant, J. Cerkovnik, B. Plesnicˇar, T. Tuttle,
J. Am. Chem. Soc. 2008, 130, 14086–14087.
I. I. Skorokhodov, V. B. Golubev, L. I. Nekrasov, V. B. Evdoki-
mov, N. I. Kobozev, Zh. Fiz. Khim. 1962, 36, 93–97; Russ. J.
Phys. Chem. 1962, 36, 47–49; M. Y. Antipin, E. E. Antipenko,
B. V. Strakhov, E. V. Lunina, L. I. Nekrasov, Zh. Fiz. Khim.
1977, 51, 3064–3067.
[27] In principle, H2OO (isomeric hydrogen peroxide, oxywater) and
·
(HO2)2 (cyclic dimer of the HO2 radical) should be taken into
account as possible components of peroxy radical condensates.
However, the calculated vibrational spectra available from the
literature allow us to exclude these compounds from consider-
ation. Indeed, the vibrational spectrum of H2OO should have
only one band due to oxygen framework oscillations (harmonic
frequency 665 cm–1) (H. H. Huang, Y. Xie, H. F. Schaefer III,
J. Phys. Chem. 1996, 100, 6076–6080). The spectrum of
(HO2)2 should have three Raman-active oscillations (harmonic
frequencies 1233, 230, 182 cm–1) whose frequencies do not es-
sentially change on substitution of deuterium for hydrogen
(J. T. Fermann, B. C. Huffman, G. S. Tschumper, H. F.
Schaefer III, J. Chem. Phys. 1997, 106, 5102–5108). Lines with
these features are not present in our spectra.
[4] R. A. Lerner, A. Eschenmoser, Proc. Natl. Acad. Sci. USA
2003, 100, 3013–3015, and references cited therein; X. Zhu, P.
Wentworth Jr., A. D. Wentworth, A. Eschenmoser, R. A. Ler-
ner, I. A. Wilson, Proc. Natl. Acad. Sci. USA 2004, 101, 2247–
2252.
[28] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T.
Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R.
Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morok-
uma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzew-
ski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian 03 (Re-
vision E.01), Gaussian, Inc., Wallingford, CT, 2004.
[29] P. A. Denis, F. R. Ornellas, J. Phys. Chem. A 2009, 113, 499–
506.
[5] D. J. Mckay, J. S. Wright, J. Am. Chem. Soc. 1998, 120, 1003–
1013.
[6] K. Suma, Y. Sumiyoshi, Y. Endo, J. Am. Chem. Soc. 2005, 127,
14998–14999.
[7] M. Berthelot, Compt. Rend. 1880, 90, 656.
[8] D. I. Mendeleev, Osnovy Khimii, 6th ed., V. Demakov Printing
House, St. Petersburg, 1895, p. 151.
[9] M. Venugopalan, R. A. Jones, Chemistry of Dissociated Water
Vapor and Related Systems, Interscience Publishers, New
York – London – Sydney, 1968.
[10] E. Boehm, K. F. Bonhoeffer, Z. Phys. Chem. 1926, 119, 385–
399.
[11] G. I. Lavin, F. B. Stewart, Proc. Natl. Acad. Sci. USA 1929, 15,
829–832.
[12] A. K. Brewer, J. W. Westhaver, J. Phys. Chem. 1930, 34, 2343–
2355.
[13] K. H. Geib, P. Harteck, Ber. Deutsch. Chem. Ges. 1932, B65,
1551–1555.
[14] E. Ohara, J. Chem. Soc. Jpn. 1940, 61, 569–582.
[15] N. I. Kobozev, L. I. Nekrasov, E. N. Eremin, Zh. Fiz. Khim.
1956, 30, 2580–2581; N. I. Kobozev, I. I. Skorokhodov, L. I.
Nekrasov, E. I. Makarova, Zh. Fiz. Khim. 1957, 31, 1843–1850.
[16] a) A. B. Tsentsiper, M. S. Danilova, A. S. Kanishcheva, A. I.
Gorbanev, Zh. Neorg. Khim. 1959, 4, 1952–1957; Russ. J. Inorg.
Chem. 1959, 4, 886–889; b) J. A. Wojtowicz, F. Martinez, J. A.
Zaslowksy, J. Phys. Chem. 1963, 67, 849–852.
[30] P. A. Giguère, Chemistry 1975, 48, 20–22.
[31] F. C. Fehsenfeld, K. M. Evenson, H. P. Broida, Rev. Sci. In-
strum. 1965, 36, 294–298.
[32] PeakFit, Software for Peak Separation and Analysis, version
4.12, SeaSolve Software Inc., SYSTAT Software Inc., San Jose,
CA, 2003.
[17] I. I. Skorokhodov, L. I. Nekrasov, N. I. Kobozev, V. B. Evdoki-
mov, Zh. Fiz. Khim. 1962, 36, 274–281; Russ. J. Phys. Chem.
1962, 36, 136–140.
[18] a) O. Yu. Berezin, E. E. Antipenko, B. V. Strakhov, Zh. Fiz.
Khim. 1980, 54, 2145–2147; Russ. J. Phys. Chem. 1980, 54,
[33] See for example: A. I. Vogel (Ed.), Vogel’s Textbook of Quanti-
tative Chemical Analysis, 5th ed., Longman Group UK Lim-
ited, Harlow (Essex, England), 1989, pp. 372–373.
Received: July 23, 2011
Published Online: October 18, 2011
5150
www.eurjic.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2011, 5144–5150