Raman-spectroscopy of oligomeric SiO species isolated in solid methane

Article abstract of DOI:10.1063/1.480123

From the IR-spectra of matrix isolated SiO species a D_2h-structure has been postulated for the dimer and a D_3h-structure for the trimer. High quality Raman-spectra - necessary for the complete characterization - were missing so far. Here we report the Raman-spectra especially of the totally symmetric vibrations for Si₂O₂ and Si₃O₃ and their ¹⁶O/¹⁸O isotopomers isolated in solid methane. We also detect the most intense A₁-vibration of Si₄O₄ and can assign it with its ¹⁶O/¹⁸O isotopic splitting. Ab initio calculations for all oligomers are presented in order to support the assignment of the spectra and to obtain geometric and energetic information about the oligomeric species which have been detected experimentally.

Full text of DOI:10.1063/1.480123

Raman-spectroscopy of oligomeric SiO species isolated in solid methane
Markus Friesen, Markus Junker, Andreas Zumbusch, and Hansgeorg Schnöckel
Citation: The Journal of Chemical Physics 111, 7881 (1999); doi: 10.1063/1.480123
View online: http://dx.doi.org/10.1063/1.480123
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/111/17?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Infrared spectra and theoretical studies of the C 2 O 3 − anion isolated in solid argon
J. Chem. Phys. 112, 7089 (2000); 10.1063/1.481304
Vibrational spectra of molecular ions isolated in solid neon. XV. Infrared spectroscopic evidence for NeHF + and
HFFH +
J. Chem. Phys. 110, 5037 (1999); 10.1063/1.478402
Infrared spectrum of matrix isolated ClSiO and ab initio calculations
J. Chem. Phys. 110, 3769 (1999); 10.1063/1.478267
Raman and IR spectroscopy of ( GeO ) n , with n=1,2,3,4, isolated in solid argon
J. Chem. Phys. 108, 8092 (1998); 10.1063/1.476248
Fourier transform infrared observation of the vibrational spectrum of the linear SiCCH radical in Ar at 10 K
J. Chem. Phys. 106, 6222 (1997); 10.1063/1.473616
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

JOURNAL OF CHEMICAL PHYSICS
VOLUME 111, NUMBER 17
1 NOVEMBER 1999
Raman-spectroscopy of oligomeric SiO species isolated in solid methane
Markus Friesen and Markus Junker
¨
¨
Institut fur Anorganische Chemie, Universitat Karlsruhe (TH), Engesserstr., Geb. 30.45,
D-76128 Karlsruhe, Germany
Andreas Zumbusch
¨
¨
¨
Institut fur Physikalische Chemie, Universitat Munchen, Butenandtstr. 5-13, Haus E,
¨
D-81377 Munchen, Germany
a)
¨
Hansgeorg Schnockel
¨
¨
Institut fur Anorganische Chemie, Universitat Karlsruhe (TH), Engesserstr., Geb. 30.45,
D-76128 Karlsruhe, Germany
͑Received 2 March 1999; accepted 10 August 1999͒
From the IR-spectra of matrix isolated SiO species a D2h-structure has been postulated for the dimer
and a D3h-structure for the trimer. High quality Raman-spectra—necessary for the complete
characterization—were missing so far. Here we report the Raman-spectra especially of the totally
symmetric vibrations for Si2O2 and Si3O3 and their 16O/18O isotopomers isolated in solid methane.
We also detect the most intense A1-vibration of Si4O4 and can assign it with its 16O/18O isotopic
splitting. Ab initio calculations for all oligomers are presented in order to support the assignment of
the spectra and to obtain geometric and energetic information about the oligomeric species which
have been detected experimentally. © 1999 American Institute of Physics.
͓S0021-9606͑99͒01641-4͔
I. INTRODUCTION
were of low quality we repeated these investigations in order
to obtain unequivocal experimental information about the
structures of dimeric, trimeric, and also of tetrameric SiO.
In contrast to the large number of IR-spectra, e.g., of SiO
͑Refs. 1–3͒, very few Raman-spectra of matrix isolated spe-
cies in general have been reported to date due to the intrinsic
weakness of the Raman-effect. Especially the recent devel-
opment of CCD techniques and the availability of laser
sources in the UV/VIS- and NIR-region has made Raman-
spectroscopy more attractive for investigations of matrix iso-
lated molecules. The usefulness of the technique is proven in
II. EXPERIMENTS
Molecular SiO is produced under high vacuum condi-
tions ͑about 10Ϫ6 mbar͒ by passing a stream of O2 over
heated silicon placed in a resistively heated furnace ͑about
1500 K͒ of Al2O3. The advantage of this procedure is that
the flow of SiO can be regulated reproducibly via the flow-
rate of O2. However, the O2 rate should not be allowed to
become larger than the vapor pressure of SiO, otherwise the
silicon pieces would be coated by a SiO/SiO2 film. In the gas
phase the SiO stream is mixed with a large excess of the
matrix gas ͑about 1:200͒ and subsequently the gas mixture is
condensed on a polished copper mirror. This mirror is cooled
with the help of a closed cycle cryostat ͑Leybold LB 510͒.
The temperature of the copper block was controlled via a Si
diode. The lower limit of the temperature achievable with
our apparatus is 10 K. However, in the experiments de-
scribed here the block temperature is kept at 25 K in order to
get transparent matrices ͑cf. below͒. The deposition times are
long ͑150 min͒ in order to get a sufficient high amount of the
species under discussion for the detection of Raman-
scattering. The conditions for the SiO production are opti-
mized by the detection of O2 in the matrix. In all experiments
described here no O2 is present in the matrix which can be
monitored via the O2 vibration at 1552.6 cmϪ1 ͑in solid
CH4͒. Details of the experiment set up have been described
elsewhere.4
a
recently published work on matrix isolated GeO
oligomers,4 where the Raman-spectra served complete and in
a few cases to correct the information of the 30 year old
IR-spectroscopic information.5 Investigations about similar
SiO species are of interest because of its unknown structure
in the solid state. SiO finds use as an optical coating material.
In the gas phase over heated SiO only small amounts of
Si2O2 have been detected via mass-spectroscopy.6 During
deposition of SiO under matrix conditions monomeric,
dimeric and trimeric species have been obtained by means of
IR-spectroscopy.1–3 However, no hints for tetrameric SiO
have been published so far. The relative amounts of SiO,
Si2O2, and Si3O3 in the matrix depend strongly on the depo-
sition conditions, e.g., the Si:Ar ratio. The geometric struc-
tures derived from IR-spectroscopy are four and six-
membered rings with D2h- and D3h-symmetry. In contrast to
these findings which are supported by data for 16O/18O and in
some cases also by 28Si/29Si isotopic shifts,3 Raman-spectra
of matrix isolated Si2O2 have been published by Khanna
et al.,7 from which a chain like structure for Si2O2 with
C2-symmetry has been deduced. Since these Raman spectra7
In order to obtain high quality Raman-spectra allowing
the detection of weak bands an optimal signal to noise ratio
a͒
Author to whom correspondence should be addressed.
0021-9606/99/111(17)/7881/7/$15.00
7881
© 1999 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

7882
J. Chem. Phys., Vol. 111, No. 17, 1 November 1999
Friesen et al.
͑S/N͒ is necessary. As opaque matrices lead to an increased
background signal due to scattering, the production of trans-
parent matrices is of uttermost importance. Methane as ma-
trix material8 yields transparent samples if it is condensed at
higher temperatures between 20–25 K. The temperature
should not exceed 30 K, because at this temperature the va-
por pressure of CH4 is about 10Ϫ8 bar, that means the matrix
material would sublime. The Raman-band of solid CH4 ͑rela-
tive intensities in brackets ͓3017.8 cmϪ1 ͑8͒ T2 ; 2909.5 cmϪ1
͑100͒ A1 ; 1533.6 cmϪ1 ͑1͒ E͔ ͓The T2 mode at 1310 cmϪ1
exhibits a Raman intensity 104 times smaller than that of the
A1 vibration ͑calculated on a HF-SCF/SVP level͒.͔ are at
higher wave numbers than those expected for oligomeric SiO
species. Some experiments have been performed using the
typical matrix gases Ar and N2. In both cases the weak
Raman-bands are hidden under the strong scattering back-
ground.
Quantum chemical calculations are performed on an
IBM 390 workstation using the GAUSSIAN program package.9
Unless stated otherwise for MP2 calculations basis sets of
TZVP quality10 and for DFT calculations ͑with either B3-
LYP or BP-86 functionals͒ SVP basis sets11 are used.
III. RESULTS AND DISCUSSION
The Raman-spectrum of a typical experiment containing
16O species is presented in Fig. 1͑a͒. The most intense band
at 1216.3 cmϪ1, which can be detected after 100 min of ac-
quisition time has to be assigned to monomeric 28Si 16O.
Caused by the low scattering cross section of all SiO oligo-
meres further bands at 840.2 cmϪ1, 627.9 cmϪ1, 596.4 cmϪ1
,
556.2 cmϪ1, 530.4 cmϪ1, 460.4 cmϪ1 can only be observed
with long acquisition times. The band at 596.4 cmϪ1 is the
most intense. If the equivalent experiment is performed with
Si 18O the spectrum of Fig. 1͑b͒ is observed, in which the
following positions are measured: 1172.9 cmϪ1 (28Si 18O),
FIG. 1. Raman-spectra ͑␯Aϭ514 nm, Lϭ500 mW, ⌬␯ϭ2.2 cmϪ1͒ of ma-
trix isolated ͑a͒ oligomers of Si 16O in solid CH4; ͑b͒ oligomers of Si 18O in
solid CH4 ; ͑c͒ monomeric 28Si 16O ͑left͒ and 28Si 18O in solid CH4.
804.7, 616.8, 585.2, 548.9, 507.4, and 437.1 cmϪ1
.
In order to assign these bands to the different oligomeres
of SiO, we have performed experiments with different ratios
of SiO:CH4. Reactions of SiO molecules in the gas phase
can be neglected because of the low pressure. Since also
annealing of the matrix after condensation has nearly no in-
fluence on the band intensity of any oligomer, the process of
SiO oligomerization occurs via diffusion during the trapping
of SiO in the matrix materials, that means in the short period
when the species pass the fluid phases on the way to the solid
state. In order to influence the probability for the formation
of dimers or trimers, the diffusion time or diffusion path way
has to be changed which can be done either by variation of
the SiO concentration or by changing the deposition tem-
perature.
larization ratio could be detected, that means this band be-
longs to a totally symmetric vibration. All other bands were
of an too low intensity, so after the addition of a polarization
filter the bands could not be distinguished from the back-
ground. ͑Longer acquisition times should have a positive ef-
Further support for a correct assignment of bands to dif-
ferent oligomeres comes from experiments with mixtures of
28Si 16O and 28Si 18O species. In Fig. 2 the Raman-spectrum
of an experiment containing 60% 18O species is presented.
By inspection of the band shapes ͑splitting and intensity͒
conclusions on the numbers of equivalent O-atoms partici-
pating in a certain vibration can be drawn.
FIG. 2. Raman-spectrum of an experiment with 60% 18O enrichment. The
enlarged sector shows bands of the tetramer from 400 cmϪ1 to 500 cmϪ1
͑␯Aϭ514 nm, Lϭ500 mW, ⌬␯ϭ2.2 cmϪ1͒.
Also we tried to measure the depolarization ratios. How-
ever, only for the stronger band at 596.4 cmϪ1 a low depo-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

J. Chem. Phys., Vol. 111, No. 17, 1 November 1999
Spectroscopy of oligomeric SiO species
7883
TABLE I. Experimental ͑Raman-spectra in solid CH4 , ␯Aϭ514 nm, Lϭ500 mW, ⌬␯ϭ2.2 cmϪ1͒ and calcu-
lated vibrational frequencies (cmϪ1) of monomer 28Si16O and 28Si18O with the isotopic shifts ⌬16–18 (cmϪ1).
Gas a
Expt.
B3-LYP b
B3-LYP c
BP86 c
MP2 b
MP2 c
28Si16O
28Si18O
1241.6
1193.3
45.3
1216.3
1172.9
43.4
1245.4
1200.4
44.9
1240.5
1195.8
44.7
1179.9
1137.3
42.6
1176.0
1133.6
42.4
1181.8
1139.2
42.6
⌬
16–18O
aHarmonic frequencies of gaseous SiO ͑Ref. 12͒.
b
**
Basis set, 6-31G
.
cBasis set, SVP.
fect. However, instabilities within our optical arrangement
limited this possibility.͒
vibrations of a D2h-symmetric dimer ͑see Fig. 3͒.3 In the
Raman-spectrum of this four membered ring molecule two
A1g- and one B1g-vibration would be present. Their frequen-
cies have been calculated by a normal coordinate analysis
based on a simple force field to be ͓1͔: ␯(A1g): 766.6 cmϪ1
IV. ASSIGNMENT OF BANDS
Ab initio calculations are performed to determine the
lowest energy of the different isomers of the molecules. On
the basis of these results vibrational frequencies, IR- and
Raman-intensities and the corresponding isotopic shifts are
calculated. To include electron correlation effects, that are
known to be a source of error in the frequency calculations
on the Hartree–Fock level, we use density functional theory
and perturbation theory ͑MP2͒.
and 216.0 cmϪ1; ␯(B1g): 774.9 cmϪ1
.
These predictions are now corrected by our experiments.
Under 60% 18O substitution an unsymmetrical triplet at
840.2, 823.4, and 804.7 cmϪ1 is observed. From the intensity
pattern of this triplet it can be concluded that two identical
O-atoms are involved in this vibration, that means they be-
long indeed to a Si2O2 species. Furthermore from experi-
ments with different amounts of oligomeres it must be con-
cluded that the bands at 840.2 and 556.2 cmϪ1 belong to the
same species. 18O isotopic substitution shifts these bands to
804.7 and 548.9 cmϪ1, respectively. Though no depolariza-
tion measurements were possible ͑cf. above͒ based on the
results of ab initio calculations we assigned these two bands
to the totally symmetric representation A1g . The third band
to be expected (B1g) could not be observed which is plau-
sible since its intensity should only a fourth of that of the
most intense A1g-mode (840 cmϪ1). Further confirmation for
the correct assignment of both A1g bands of Si2O2 is the
agreement of calculated and measured isotopic shifts. The
high quality of the calculation with respect to the absolute
frequencies and with respect to their 16O/18O shifts can be
judged by inspection of Table II. These experimental find-
ings cannot be reconciled with the chainlike structure pro-
posed by Khanna et al. for Si2O2.7
A. SiO
The vibrational frequency of matrix isolated monomeric
28Si 16O is well known from IR-spectroscopy 1225.9 cmϪ1
͑Ar͒, 1223.9 cmϪ1 (N2). The gas phase value is
1229.7 cmϪ1. The Raman-spectrum of 28Si 16O in solid meth-
ane ͓Fig. 1͑c͔͒ shows an intense signal at 1216.3 cmϪ1 which
is shifted to 1172.9 cmϪ1 under 18O substitution. This shift of
43.4 cmϪ1 compares well to the known shifts in solid Ar
(43.8 cmϪ1) and in solid N2 (44.4 cmϪ1). The observation of
additional 29Si and 30Si shifts ͑29Si 16O: 1208.3 cmϪ1
30Si 16O: 1201.7 cmϪ1 29Si 18O: 1164.8 cmϪ1 30Si 18O:
1158.1 cmϪ1͒ confirms the assignment of the band.
;
;
;
In order to determine the accuracy of ab initio methods,
the measured frequency of the SiO IR-absorption is com-
pared to the values of different theoretical calculations
͑Table I͒. However the frequencies from these calculations
cannot be compared directly with those observed in matrix
isolation Raman-experiments since only the harmonic poten-
tial of gaseous molecules are calculated with ab initio meth-
ods. To account for anharmonicity and matrix interactions
we calculated for each matrix gas a factor by dividing the gas
phase value ␻ of 1241.6 cmϪ1 (28Si 16O) ͑Ref. 12͒ by the
experimentally determined anharmonic frequency in the ma-
trix. The resulting factors are 1.0208 (CH4); 1.0128 ͑Ar͒;
1.0145 (N2). The frequencies of all oligomeric SiO species,
which can only be detected under matrix conditions, are mul-
tiplied with this factor. Subsequently the corrected harmonic
frequencies of the oligomeres from the experiment can be
used as a confidential measure for the quality of the quantum
chemical calculations.
B. Si2O2
IR-bands of Ar-matrix isolated Si2O2 at 803.1 and
FIG. 3. In-plane vibrations of Si2O2 with D2h-symmetry with calculated
͑MP2/SVP͒ frequencies ͑oxygen, black; silicon, white͒.
767.5 cmϪ1 have been assigned to the B3u- and B2u
-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

7884
J. Chem. Phys., Vol. 111, No. 17, 1 November 1999
Friesen et al.
TABLE II. Experimental, corrected ͑due to anharmonicity and matrix shift͒, and calculated vibrational frequen-
cies (cmϪ1) of 28Si216O2 and isotopic shifts ⌬16–18 (cmϪ1).
B3-LYP a
⌬
MP2 b
⌬
Raman c
16–18
16–18
16–18
16–18
Expt.
⌬
Corr.
⌬
dSi–O /pm
ЄOSiO /°
¯
¯
¯
¯
¯
¯
¯
¯
171.46
87.1
¯
¯
172.3
87.1
¯
¯
¯
¯
B1u
B1g
Ag
B2u
B3u
Ag
¯
¯
¯
8.7
28.5
28.5
35.5
¯
¯
¯
8.9
28.9
28.9
36.2
207.4
530.0
549.2
773.5
799.0
840.0
7.5
19.8
8.6
29.7
28.8
35.2
213.6
535.6
552.4
782.2
819.2
846.9
7.7
20.0
9.5
28.2
29.6
34.6
¯
¯
¯
25
556.2 d
767.5 e
803.1 e
840.2 d
568.8
778.6
814.7
857.8
36
¯
¯
100
a
**
Basis set, 6-31G
.
bBasis set, SVP.
cRaman-intensities resulting from a HF-SCF calculation ͑SVP basis͒.
dRaman-spectra in solid CH4.
eIR-spectra in solid N2 ͑Ref. 1͒.
C. Si3O3
The assignment of the second Raman-band to the A1Ј-
representation is also confirmed by the observed 16O/18O
splitting pattern in an experiment with 60% 18O enrichment.
Though due to its low intensity the fourth band in the ex-
pected quartet ͑28Si3 18O3 could not be detected͒
(530.3/522.3/515.1 cmϪ1), the measured isotopic shift in dif-
ferent experiments with pure 16O and 18O species of
22.9 cmϪ1 is in line with the results of DFT calculations and
confirms the assignment ͑cf. Table III͒.
The three absorptions in the IR-spectrum of matrix iso-
lated Si3O3 in solid N2 at 972.6, 631.5, and 311.5 cmϪ1 were
assigned to the E -modes of the six membered
Ј
D3h-symmetric Si3O3 molecule.1 The 12 normal modes be-
long to the following representations. The in-plane vibrations
are visualized in Fig. 4,
2AЈϩ1AЈϩ1AЉϩ3E ϩ1E .
Ј
Љ
1
2
2
For the partially substituted species splitting of vibra-
tions is expected as a result of removal of the degeneracy.
The substitution of one 16O-atom by an 18O-atom in Si3O3
will change the symmetry from D3h to C2v and in conse-
quence E and E vibrations will split; E →A ϩB and
Vibrations of AЈ and E symmetry are only Raman-active
while those of AЉ2-symmetry are only IR-active. Vibrations
Љ
1
of E -symmetry should be observed in the Raman-, as well
Ј
as in the IR-spectrum. Experiments in which different
amounts of oligomeres are deposited show, that the ratio of
the intensity of the bands at 627.9, 596.4, and 530.3 cmϪ1
remains unchanged alluding to the fact that these three
Raman-bands belong to the same species. After 18O isotopic
substitution these bands are shifted to 616.8, 585.2, and
Ј
Љ
Ј
1
2
E →A ϩB . In Fig. 5 the calculated splitting for the E
Љ
Ј
2
1
vibration is visualized for species generated under 60% 18O
enrichment. The intensity distribution of the bands corre-
sponds to that expected for the 18O enrichment. Thus a pat-
tern of six bands should be observed if the single vibrations
could be resolved. However, in our experiments ͑60% 18O
enrichment with a resolution of only 2 cmϪ1͒ the band at
627.9 cmϪ1 is split into only two bands which are broadened
by the absorptions of the other isotopic isomers shown in
Fig. 1͑c͒.
507.4 cmϪ1
.
The most intense Raman-signal at 596.4 cmϪ1 splits
under 60% 18O substitution into a quartet (596.4/590.8/
587.8/585.2 cmϪ1). The intensity pattern is typical for a vi-
bration in which three identical oxygen-atoms are involved.
Since for this band the depolarization ratio could be deter-
mined to be 0.2, the assignment to the AЈ1-motion of the
Si3O3 species seems to be sure.
Thus the assignment of the bands at 627.9 cmϪ1 (E )
Ј
and 530.3 (A1Ј) is confirmed by quantum chemical calcula-
tions and with the determination of the 16O/18O isotopic shift.
In contrast there is no corresponding fit for the 596.4 band,
since the measured shift is about twice as large as a calcu-
lated one ͓11.2 cmϪ1 ͑expt.͒, 5.8 cmϪ1 ͑DFT͒, and 6.8 cmϪ1
͑MP2͔͒. The reason for this discrepancy is a Fermi-resonance
with the first overtone of the 311.5 cmϪ1 (E ) vibration,
Ј
which causes a larger interaction with the Si3 18O3 isoto-
pomer which leads to larger than expected red shift of this
E -mode, which is visualized in Fig. 6.
Ј
D. Si4O4
In contrast to the heavier homologs GeO,4 SnO,13 and
PbO ͑Ref. 14͒ there exists no experimental evidence for a
tetrameric SiO species. In our SiO experiments we observed
FIG. 4. In plane vibrations of Si3O3 with D3h-symmetry with calculated
͑MP2/SVP͒ frequencies ͑oxygen, black; silicon, white͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

J. Chem. Phys., Vol. 111, No. 17, 1 November 1999
Spectroscopy of oligomeric SiO species
7885
TABLE III. Experimental, corrected ͑due to anharmonicity and matrix shift͒, and calculated vibrational fre-
quencies (cmϪ1) of 28Si316O3 and isotopic shifts ⌬16–18 (cmϪ1).
B3-LYP a
⌬
MP2 b
⌬
Raman c
16–18
16–18
16–18
16–18
Expt.
⌬
Corr.
⌬
dSi–O /pm
ЄOSiO /°
ЄSiOSi /°
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
168.9
101.7
138.7
¯
¯
¯
169.3
101.5
138.5
¯
¯
¯
¯
¯
¯
E
A2Љ
E
A1Ј
A1Ј
¯
¯
¯
¯
¯
¯
¯
¯
104.4
222.4
298.8
524.6
582.5
617.5
963.5
4.2
8.0
10.4
24.3
6.8
100.7
212.5
308.2
537.5
600.1
635.4
992.3
4.1
7.7
10.7
25.1
6.8
0
¯
5
11
100
51
1
Љ
¯
¯
311.5 d
530.3 e
596.4 e
631.5 d
972.6 d
316.0
541.3
608.8
635.9
985.9
Ј
22.9
28.5
11.1
40.7
23.4
11.4
11.3
41.3
E
E
9.7
42.6
9.9
43.1
Ј
Ј
a
**
Basis set, 6-31G
.
bBasis set, SVP.
cRaman-intensities resulting from a HF-SCF calculation ͑SVP basis͒.
dIR-spectra in solid N2 ͑Ref. 1͒.
eRaman-spectra in solid CH4.
a band at 460.2 cmϪ1 with a shift under 18O substitution to
437.1 cmϪ1. In an experiment with partial substitution ͑60%
18O͒ a triplet (443.4/449.6/454.2 cmϪ1) with nearly similar
intensities is observed. The ‘‘pure’’ isotopic isomers Si4 16O4
and Si4 18O4 are not detected. This observed pattern could be
reconciled with a vibration in which four identical O-atoms
are involved. However this assignment is not unequivocal.
Therefore quantum chemical calculations have been per-
formed in order to find the favored Si4O4 isomer which fits in
with the observed frequency and its isotopic shifts. Geometry
optimization have been performed on the DFT levels using
the following starting geometries: ͑a͒ an eight-membered
ring of C4v-symmetry ͑like S8 molecules͒; ͑b͒ a cagelike
structure of D2d-symmetry ͑like S4N4͒; ͑c͒ a heterocubane-
like structure ͑like Ge4O4͒ ͑Ref. 4͒; and ͑d͒ a planar eight-
membered ring of D4h-symmetry ͑like Si4O4Cl8͒.15 Subse-
quently frequency calculations have shown whether the ob-
tained structure is a minimum or only a saddle point on the
hypersurface. The energetic global minimum of all isomers
is the puckered eight-membered ring of D2d-symmetry. Only
a few kJ molϪ1 higher in energy is the planar D4h isomer
(7 kJ molϪ1). The energetically most unfavorable species
(ϩ132 kJ molϪ1) is the heterocubanelike Td-isomer. How-
ever, only the heterocubane (Td-symmetry͒ and the
D
2d-symmetric eight-membered ring represent minima on
the potential hypersurface. Therefore only these two isomers
will be discussed.
Bonding in the SiO heterocubane should be weaker
caused by differences in coordination numbers. The three-
fold coordination in this isomer causes an increase of the SiO
bond length of 17.6 pm in comparison to Si3O3 which is
reflected in the decrease of the shared electron number
͑SEN͒ of the SiO bond—a measure for the covalent bond
strength16—by about 0.4 (1.26→0.89). Therefore it can be
expected that the different bonding in both isomers ͑geom-
etry and force constants͒ will also be seen in quite different
vibrational spectra including the 16O/18O shifts. This expec-
tation is confirmed by DFT calculations shown in Table IV.
From these calculated frequencies it is evident, that the
Raman-active A1 breathing mode of both isomers, in which
mainly the O-atoms are involved are quite different;
420.1 cmϪ1 for the eight-membered Si4O4 ring and
674.4 cmϪ1 for the Td-symmetric heterocubane.
Comparing these results with the observed Raman-
spectra (460.4 cmϪ1) it seems reasonable to assign this band
to a Si4O4 isomer with D2d-symmetry. The calculated struc-
ture of this isomer is presented in Fig. 7. Another argument
FIG. 6. Illustration of the Fermi resonance in 28Si3 18O3 with calculated
frequencies ͑MP2/SVP͒. The black bands show the situation without reso-
nance, the marked ͑ࡗ͒ bands belong to the unshifted 28Si3 16O3.
FIG. 5. Calculated spectrum of the E -vibration of Si O (D3h) with 60%
Ј
3
3
18O enrichment.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

7886
J. Chem. Phys., Vol. 111, No. 17, 1 November 1999
**
Friesen et al.
TABLE IV. Calculated (B3-LYP/6-31G ) vibrational frequencies ␯ (cmϪ1), isotopic shifts ⌬16–18 (cmϪ1
)
and geometric data of 28Si416O4. Intensities ͑SCF calculation͒ are listed in brackets ͓IR-intensity in ͑km/mol͒
and relative Raman-intensity͔.
16–18
16–18
␯
⌬
IR/Raman
␯
⌬
IR/Raman
D
2d-symmetry
dSi–O /pm: 167.7; ЄOSiO /°: 102.8, ЄSiOSi /°: 159.1
B1
A1
A2
E
B2
B2
E
43.3
65.3
76.1
121.4
122.4
277.8
344.0
2.3
0.0
4.2
4.9
1.4
0/3.6
0/51
0/0
0.14/1.2
0.08/4.6
20/17
100/0.60
A1
E
A1
B1
A2
E
420.0
555.2
556.0
584.7
782.0
23.2
1.1
1.2
0/65
1.6/27
0/45
0/100
0/0
1551/0.23
255/0.17
2.4
31.7
44.2
47.5
10.0
17.0
1003
1029
B2
Td-symmetry
dSi–O /pm: 186.4; dSi–Si /pm: 278.9; ЄSiSiO /°: 85.0. ЄSiOSi /°: 96.8
E
T1
E
278.1
307.0
396.0
398.2
0.0
12.1
22.7
6.0
0/43
0/0
0/27
37/18
T2
A1
T2
A1
573.5
579.3
648.4
674.4
20.6
0.0
25.1
38.7
125/2.5
0/100
511/10
0/10
T2
for this structure is the comparison between the measured
and calculated 16O/18O shifts. The measured shift of about
23.1 cmϪ1 compares better to the calculated value the ring
molecule Si4O4 (23.2 cmϪ1) than to that of a heterocubane
(38.7 cmϪ1). For the second A1-vibration in a hypothetical
heterocubanelike Si4O4 at 579.3 cmϪ1 the highest Raman-
intensity has been calculated, but no 16O/18O shift should be
observed, since only the Si-atoms are involved in this mode.
However the equivalent vibration for the heterocubanelike
Ge4O4—it has been observed at 340 cmϪ1 ͓From the relation
of the G-matrix elements it has been concluded that the
A1-vibrations at 579 cmϪ1 ͑calc. for Si4O4͒ and at 340 cmϪ1
͑obs. for Ge4O4͒ are equivalent (340/579Ϸ(mSi /
cages.17,18 For the hydrosilsesquioxane H8Si8O12 it has been
shown that a typical ring vibration within the Si4O4 rings of
the Si8O12 cages with Oh-symmetry ͑a cube of 8 Si-atoms
connected by 12 O-atoms on each of the edges͒ is observed
in the Raman-spectrum at 456 cmϪ1
.
This band of H8Si8O12 is assigned to a ring opening vi-
bration within the model compound for some zeoliths. Be-
sides the frequencies the displacement of the Si- and
O-atoms within this normal mode are quite similar to the
breathing mode of ‘‘our’’ Si4O4 molecule since for both vi-
brations the 16O/18O shift has been calculated to be at
22 cmϪ1. ͓Calculations on a B3-LYP/6-31G -level have
**
mGe)
1/2).4͔—exhibits nearly no 16O/18O shift. In the spec-
shown, that the SiO-bond distance in the Si8O12 ͑169.9 pm͒
cages is similar to that of Si4O4 ͑164.3 pm͒. Furthermore—
as shown by SCF calculations—for both molecules the
breathing modes are very similar with respect to frequency
(475Ϯ5 cmϪ1) and 16O/18O shift (22Ϯ2 cmϪ1).͔ That is, be-
sides geometry the force constants seem to be similar within
both kinds of Si4O4 systems, though the oxidation states of
the Si-atoms are different. Obviously this difference in oxi-
dation state plays a minor role in larger molecules than in
smaller ones, e.g., in the following Si2O2 four-membered
ring molecules different force constants are observed:
OvSiO2SivO,19 Si2O2.3 Furthermore it seems to be signifi-
cant, that the calculated Raman-intensities ͑SCF level͒ are of
minor reliability. Obviously, like for Ge4O4, not the pre-
dicted two strong breathing modes are observed but only that
with the lower energy.
trum of Si4O4 just this band is missing, which is a further
hint against the formation of a heterocubanelike isomer of
Si4O4 in the matrix. That is the absolute values of the fre-
quencies, their 16O/18O shifts, the energetic argument from
ab initio calculations, and the different Raman-spectra in
comparison to Td-like Ge4O4 favor the eight-membered
Si4O4 ring against the heterocubanelike isomer.
A further confirmation for our assignment of the Raman-
band at 460 cmϪ1 to an eight-membered Si4O4 molecule
comes from a quite different area. Raman-spectroscopic in-
vestigations on different zeolith like SiO cages have shown,
that there is a correlation between the most intense Raman-
band and the size of the SiO ring system within the
In contrast to Ge4O4 with its heterocubanelike structure
we find evidence for an eight-membered ring structure for
Si4O4. This tendency seems to be plausible, since the GeO
bond should have more ionic character and therefore a het-
erocubanelike structure as the smallest unit of the obviously
polar NaCl lattice should be favored.
FIG. 7. Molecular structure of Si4O4 with D2d-symmetry due to DFT cal-
**
culations ͑B3-LYP; 6-31G basis͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29

J. Chem. Phys., Vol. 111, No. 17, 1 November 1999
Spectroscopy of oligomeric SiO species
7887
E. Energetic and kinetic relations within the
oligomerization process
more ionic bonding in the heterocubanelike Ge4O4 seems
plausible. However, the final aim of our investigations to
gain knowledge about the so far unknown structure of solid
SiO ͑Ref. 22͒ cannot be reached with matrix isolation spec-
troscopy, because it is impossible to make definite conclu-
sions on the basis of the structure of the obtainable small SiO
oligomeres. Furthermore there is another serious problem:
Though the D2d-symmetric Si4O4 molecule shows a certain
similarity to violet phosphorus—like in all types of phos-
phorus only threefold coordinated P-atoms are present—it
can be assumed that in solid SiO the Si-atoms are fourfold
and the O-atoms are twofold coordinated. This means, that
Si–Si bonding will be necessary and therefore no analogies
to simple AB lattice structures can be expected.
The oligomerization energies for the different SiO spe-
cies have been calculated with DFT methods. Since the con-
centration of monomeric SiO is high in the beginning of the
condensation process the energetic situation for the follow-
ing reactions shall be considered:
SiOϩSiO→Si2O2 ⌬EϭϪ201 kJ/mol,
SiOϩSi2O2→Si3O3 ⌬EϭϪ241 kJ/mol,
SiOϩSi3O3→Si4O4 ⌬EϭϪ166 kJ/mol.
From these values it is reasonable to assume that the concen-
tration of Si3O3 should be high, while that of Si2O2 should be
much lower and that of Si4O4 should represent a minimum.
However, it is important to mention that the observed con-
centrations do not represent the situation of a thermodynami-
cal equilibrium. As under the conditions of matrix isolation
the thermodynamic equilibrium is not reached, the formation
of the dimer should represent the final step, because the sub-
sequent formation of a rigid matrix will prevent a further
diffusion process of SiO. However, during deposition a fluid
like state exists in which SiO can diffuse and oligomerize to
form Si2O2 and eventually Si3O3, since the matrix becomes
smooth for a short time around the area of the high exother-
mic oligomerization process. After the reaction to Si3O3 the
pathway for the continuously condensed high temperature
molecule SiO is long and therefore the probability for the
formation of Si4O4 is drastically reduced. Its concentration
could only be increased by a higher deposition temperature.
Therefore the chances for the formation of larger oligomeres
than Si4O4 are slim. For these larger species matrix isolation
technique seems not to be a suitable preparation method.
ACKNOWLEDGMENTS
This work was supported by the Deutsche Forschungs-
gemeinschaft and in part by the Fonds der Chemischen In-
dustrie. We thank R. Ahlrichs for helpful discussions.
1 J. S. Anderson and J. S. Ogden, J. Chem. Phys. 51, 4189 ͑1969͒.
2 J. W. Hastie, R. H. Hauge, and J. L. Margrave, Inorg. Chim. Acta 3, 601
͑1969͒.
3
¨
H. Schnockel, T. Mehner, H. S. Plitt, and S. Schunk, J. Am. Chem. Soc.
111, 4578 ͑1989͒.
4
¨
A. Zumbusch and H. Schnockel, J. Chem. Phys. 108, 8092 ͑1998͒.
5 J. S. Ogden and M. J. Ricks, J. Chem. Phys. 52, 353 ͑1970͒.
6 K. F. Zmbov, L. L. Ames, and J. L. Margrave, High. Temp. Sci. 5, 235
͑1973͒; C. H. Wu, H. R. Ihle, and K. J. Zmbov, J. Chem. Soc., Faraday
Trans. 2 76, 447 ͑1980͒.
7 R. K. Khanna and D. D. Stranz, J. Chem. Phys. 74, 2108 ͑1981͒.
8 M. Hawkins, M. J. Almond, and A. J. Downs, J. Phys. Chem. 89, 3326
͑1985͒.
9
GAUSSIAN 94, Revision B.3, M. J. Frisch, G. W. Trucks, H. B. Schlegel, P.
M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A.
Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G.
Zakrzewski, J. V. Ortiz, J. B. Foresman, C. Y. Peng, P. Y. Ayala, W.
Chen, M. W. Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L.
Martin, D. J. Fox, J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M.
Head-Gordon, C. Gonzalez, and J. A. Pople, Gaussian, Inc., Pittsburgh,
Pennsylvania 1995.
V. CONCLUSION
During the isolation of the high temperature molecule
SiO in solid noble gases ͑or N2, CH4͒ oligomeric species
(SiO)n are formed. With help of the hitherto unknown
Raman-spectra it is possible to complete the force field and
to decide experimentally which of the possible isomers is the
most stable one. For Si2O2 and Si3O3 planar D2h- and
D3h-structures, postulated on the basis of former IR-
investigations, are confirmed and completed. However for
Si2O2 allusion to a different isomer based on Raman-
spectroscopic investigations could not be confirmed.7,20 Fur-
thermore no evidence is found for the previously discussed
Si–Si bond within Si2O2.3,4,20,21 Obviously the Si–Si inter-
action is stronger than, e.g., in OvSiO2SivO,21,19 but by far
not as strong as a typical Si–Si single bond.
10
¨
A. Schafer and R. Ahlrichs, J. Phys. Chem. 100, 589 ͑1993͒.
11
¨
A. Schafer, H. Horn, and R. Ahlrichs, J. Chem. Phys. 97, 2571 ͑1992͒.
12 K. P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure
IV. Constants of Diatomic Molecules ͑Van Nostrand, New York, 1979͒.
13 J. S. Ogden and M. J. Ricks, J. Chem. Phys. 53, 896 ͑1970͒.
14 J. S. Ogden and M. J. Ricks, J. Chem. Phys. 56, 1658 ͑1972͒.
15 M. Binnewies, M. Jerzembeck, and A. Kornick, Angew. Chem. 103, 762
͑1991͒.
16 C. Ehrhardt and R. Ahlrichs, Theor. Chim. Acta 68, 231 ͑1985͒; R. Ahl-
richs and C. Ehrhardt, Chem. Unserer Zeit 19, 120 ͑1985͒.
17 P. Bornhauser and G. Calzaferri, J. Phys. Chem. 100, 2035 ͑1996͒.
18
´
¨
C. Marcolli, P. Laine, R. Buhler, and G. Calzaferri, J. Phys. Chem. B 101,
1171 ͑1997͒.
19
¨
¨
T. Mehner, H. J. Gocke, S. Schunk, and H. Schnockel, Z. Anorg. Allg.
Chem. 580, 121 ͑1990͒.
20 N. Goldberg, M. Iraqi, W. Koch, and H. Schwarz, Chem. Phys. Lett. 225,
404 ͑1994͒.
Experimental results attributable to Si4O4 with a
21
¨
H. Schnockel, in Tailor-Made Silicon-Oxygen Compounds, edited by R.
D
2d-ring structure are presented and supported by ab initio
Corriu and P. Jutzi ͑Vieweg and Sohn, Braunschweig/Wiesbaden, 1996͒.
calculations. The different structure in comparison to the
22 B. Friede and M. Jansen, Z. Anorg. Allg. Chem. 624, 1159 ͑1998͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.138.73.68 On: Sun, 21 Dec 2014 18:48:29