Matrix reactivity of Zn, Cd, or Hg atoms (M) in the presence of silane: Photogeneration and characterization of the insertion product HMSiH3 in a solid argon matrix

Article abstract of DOI:10.1021/jp0309769

Matrix-isolation experiments give evidence that broad-band UV-vis irradiation (200 ≤ λ ≤ 800 nm) of an Ar matrix doped with SiH₄ and a group 12 metal atom M (M = Zn, Cd, or Hg) induces metal insertion into an Si-H bond to give the silyl metal hydride molecule HMSiH₃ as the primary product. Si₂H₆ is a second product, irrespective of the identity of M, while the binary hydride MH₂ is also formed when M = Zn or Cd. The products have been identified by their IR spectra and experiments with SiD₄, together with the results of quantum chemical calculations, have provided the means of authentication. The properties of the HMSiH₃ molecules are compared with those of related species, and consideration is given to how the products come to be formed.

Full text of DOI:10.1021/jp0309769

J. Phys. Chem. A 2004, 108, 1393-1402
1393
Matrix Reactivity of Zn, Cd, or Hg Atoms (M) in the Presence of Silane: Photogeneration
and Characterization of the Insertion Product HMSiH3 in a Solid Argon Matrix
Victoria A. Macrae,† Tim M. Greene,*,‡ and Anthony J. Downs†
Inorganic Chemistry Laboratory, UniVersity of Oxford, South Parks Road, Oxford, OX1 3QR, U.K., and School
of Chemistry, UniVersity of Exeter, Stocker Road, Exeter, EX4 4QD, U.K.
ReceiVed: August 15, 2003; In Final Form: October 20, 2003
Matrix-isolation experiments give evidence that broad-band UV-vis irradiation (200 e λ e 800 nm) of an
Ar matrix doped with SiH4 and a group 12 metal atom M (M ) Zn, Cd, or Hg) induces metal insertion into
an Si-H bond to give the silyl metal hydride molecule HMSiH3 as the primary product. Si2H6 is a second
product, irrespective of the identity of M, while the binary hydride MH2 is also formed when M ) Zn or Cd.
The products have been identified by their IR spectra and experiments with SiD4, together with the results of
quantum chemical calculations, have provided the means of authentication. The properties of the HMSiH3
molecules are compared with those of related species, and consideration is given to how the products come
to be formed.
I. Introduction
The interaction of Zn, Cd, or Hg atoms with SiH4 in the gas
phase has also been investigated in some detail.12 Here,
excitation of the metal atom to its 3P or 1P state brings about a
Si-H bond activation in silane molecules urges attention
partly in its own right because of the practical importance of
the hydrosilation reaction and partly because of the obvious
parallels to be drawn with the still greater issue of C-H bond
activation.1 The reaction depends crucially on the initial
interaction between the silane and the substrate, and several
complexes with SiH4 or a derivative coordinated to a transition-
metal center have been characterized. By contrast, relatively
little is known about s- or p-block centers with regard either to
their interaction with a silane or to their capacity to activate
the Si-H bond.
•
reaction to produce not HMSiH3 but the fragments MH and
•
•SiH3. The vibrational and rotational properties of the MH
radical point to a mechanism in which the M (3P) atom inserts
into an Si-H bond with little or no activation barrier to form
a bent triplet intermediate [H-M-SiH3]*; this is too short lived
for statistical population of all degrees of freedom and may in
•
•
fact decompose to MH and SiH3 within one skeletal-bending
vibration. Ab initio and density-functional theory (DFT) calcula-
tions of the appropriate potential-energy surfaces support these
conclusions.13
Matrix isolation2,3 has proved a highly informative means of
studying weakly bound complexes involving molecules such
as SiH4, H2, and CO and of charting the reactions they undergo,
usually through photoactivation. For example, complexes of
SiH4 with NH3,4 HF,5 and HONO6 have been characterized in
this way. On the evidence of the IR, electron paramagnetic
resonance (EPR), and UV-vis spectra, the group 13 metal atoms
Al and Ga (M) have also been shown to form loosely bound
complexes M‚‚‚SiH4, in which the metal atom is η2 coordinated
by the silane.7-9 Irradiation at wavelengths near 410 or 254 nm
(corresponding to S r P or D r P excitation of the metal
atom) results in insertion of the metal atom into an Si-H bond
to form the M(II) derivative HMSiH3, a change that can be
reversed by exchanging UV for visible photolyzing radiation
(λ ) ca. 580 nm). Under broad-band UV-vis irradiation,
however, HMSiH3 favors an alternative reaction channel that
involves cleavage of the M-H bond and formation of the M(I)
derivative MSiH3. Under similar conditions, Hg atoms have also
been shown to react with SiH4 on photoactivation; the primary
reaction is again insertion giving HHgSiH3, but Si2H6 is a
significant secondary product.10 By contrast, Si atoms are
reported11 to react spontaneously with SiH4 to produce not only
HSiSiH3 but also its isomer H2SiSiH2.
Such differences in the outcome of gas and matrix reactions
are familiar enough.3 As part of our systematic studies of the
thermally and photolytically induced matrix reactions of the
group 12 metal atoms Zn, Cd, and Hg, we have now added
SiH4 to a list of simple substrate molecules that has hitherto
included H2,14 CH4,15 HCl,16 and H2O.17 Reactions have been
tracked and products identified by the IR spectra of Ar matrixes
containing the reagents; the conclusions are endorsed by the
observed effects of perdeuteration of the SiH4 and by the results
of DFT calculations. No sign of a weakly bound complex
M‚‚‚SiH4 (M ) Zn, Cd, or Hg) is detectable in the initial
deposits, but broad-band photolysis (200 e λ e 800 nm) gives
rise to the seemingly photostable insertion product HMSiH3,
together with Si2H6 as a secondary product. The behavior of
the group 12 metal atoms is compared with that of Al and Ga
in similar circumstances,7-9 and the properties of HMSiH3 are
analyzed in the light of what is known about related metal
hydrides.
2
2
2
2
II. Experimental and Theoretical Procedures
Zinc (Aldrich, purity 99.999%) and cadmium (Aldrich, purity
99.9998%) were each evaporated from a tantalum Knudsen cell
that was heated resistively to ca. 300 and 250 °C, respectively,
to generate the metal atoms in their ground electronic state.
Mercury atoms were generated simply by heating a reservoir
of doubly distilled mercury to 40-50 °C. An alternative source
of the metal atoms, including some in excited electronic states,
* Author to whom correspondence may be addressed. Tel: 0044-(0)-
1392-263452. Fax: 0044-(0)1392-263434. E-mail: T.M.Greene@
exeter.ac.uk.
† University of Oxford.
‡ University of Exeter.
10.1021/jp0309769 CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/03/2004

1394 J. Phys. Chem. A, Vol. 108, No. 8, 2004
Macrae et al.
involved the use of a microwave-powered reactive resonance
lamp, as described previously.14,15 SiH4 was synthesized by the
action of LiAlH4 on SiCl4 in Et2O solution and purified by
fractional condensation in vacuo18 or used as supplied by Aldrich
(99.9%). SiD4 was prepared from LiAlD4 and SiCl4 and purified
in a similar manner, or used as supplied by CDN Isotopes (98
atom % D). Argon, nitrogen, and krypton were used as supplied
by BOC (research grade).
Figure 1, to the appearance of new IR absorptions; details are
given in Table 1. The features are concentrated in three main
regions: (i) 1820-1880 cm-1; (ii) 930-950 cm-1; and (iii)
825-880 cm-1. The first of these regions is characteristic of
the ν(Zn-H) vibrations of HZnX species in which the sub-
stituent X is relatively electropositive (cf. ZnH2 1870.2,14
HZnCH3 1866.1,15 HZnCl 1952.3,16 and HZnOH 1955.0
cm-1).17,25 However, recent studies of the reactions of laser-
ablated Si atoms with H226 suggest that it may also accommodate
ν(Si-H) vibrations of anionic silicon hydrides, with wavenum-
The metal vapor was co-deposited with an excess of argon
doped with either SiH4 or SiD4 on a CsI window cooled
normally to ca. 12 K by means of a Displex closed-cycle
refrigerator (Air Products model CS202). Details of the ap-
paratus are given elsewhere.19 The matrix gas doped with SiH4
or SiD4 to levels varying between 0.1 and 10% was deposited
at a rate of 1.0-1.5 mmol h-1, typically over a period of 1.5-2
h. Following deposition, the IR spectrum of the resulting matrix
was recorded. The sample was then exposed to the radiation
from a Spectral Energy Hg-Xe arc lamp operating at 800 W.
A water filter was used to absorb IR radiation and so reduce
any heating of the matrix, while the effects of selective
photolysis were investigated with the aid of a range of filters:
Pyrex (λ > 290 nm), soda glass (λ > 310 nm), “vis block”
(λ ) 200-400 nm), UV block (λ > 400 nm), and Oriel
interference filters for λ ) 313 nm (full width at half height
(fwhh) 16 nm) and 254 nm (fwhh 10 nm).
-
bers of 1856/1837 and 1823 cm-1 being attributed to SiH3
and SiH2-, respectively. Both the second and third regions are
most obviously linked with δ(SiHn) modes of one sort or
another26,27 (cf. the following observed wavenumbers (in cm-1):
SiH3 923.9,26 Si2H6 938, 834,10,26 HAlSiH3 846.2,9 HGaSiH3
-
845.0;9 and the calculated values: SiH3- 956.6, 868.4,26 SiH2
963.1.26 In addition, a very weak absorption was observed to
grow at 630.2 cm-1; in that it correlated with a weak absorption
at 1870.8 cm-1, this appeared to confirm the presence of ZnH2
as a minor photoproduct.14 Extending the period of photolysis
to ca. 2 h led to the continuing buildup of the main bands at ca.
1840, 940, 860, and 835 cm-1; there was a simultaneous growth
of the satellites appearing on the low-wavenumber flanks of
the 1840 and 860 cm-1 bands but at a distinctly slower rate.
Several experiments were attempted in which the photolyzing
radiation was confined to a narrower band of wavelengths.
Hence it was shown that radiation with λ ) 290-400 nm was
effective in promoting the changes described above, although
the use of a filter necessarily restricted the flux of active
radiation and so reduced the yield of products in a given time.
Irradiation with visible light (λ > 400 nm) had no effect on the
matrix, either before or after UV photolysis. Experiments with
low matrix concentrations of metal and SiH4 were unable to
exploit fully the superior definition of the IR bands because
the products were then liable to be formed at concentrations
too low for ready detection.
IR spectra were recorded in the range 4000-400 cm-1
,
typically at a resolution of 0.5 cm-1 and with a wavenumber
accuracy of (0.1 cm-1, with a Nicolet Magna-IR FTIR
spectrometer operating with a liquid N2 cooled MCTB detector.
UV-vis spectra were recorded in the range 300-900 nm using
a Perkin-Elmer-Hitachi Model 330 spectrophotometer.
DFT calculations were carried out using the Gaussian 98
program suite.20 The method applied throughout was B3LYP
with a CEP-31G basis set,16 a combination that has been shown
to give satisfactory results for small molecules of the type
described here.
III. Results
Similar experiments were carried out with SiD4 in place of
SiH4. Hence it was found that all the product features in the IR
spectra reported above were strongly redshifted (see Table 1),
the three main regions of absorption now (i) 1310-1330, (ii)
680-690, and (iii) 610-640 cm-1, giving H/D ratios of ca.
1.389:1, 1.373:1, and 1.36:1, respectively. The group of absorp-
tions at highest wavenumber is thus more likely to be due to
ν(Zn-H) (for which H/D is typically 1.38-1.39:1)14-17,26 than
to ν(Si-H) vibrations (for which H/D is typically ca. 1.37:1).26,27
To complement the studies with thermally generated Zn
atoms, some experiments were also carried out with a microwave-
powered resonance lamp as a source of the metal atoms in their
electronic ground as well as excited states. The IR spectrum of
the matrix recorded immediately after deposition of these atoms
with an SiH4/Ar sample was notable for including bands at
1870.8 and 630.2 cm-1 associated with ZnH2 and also at 935-
950 and ca. 835 cm-1. At this stage however, there were only
traces of the absorptions at 1820-1850 or 850-880 cm-1 that
were the most prominent signs of UV photolysis of a similar
matrix including thermally evaporated atoms. Only on broad-
band photolysis (200 e λ e 800 nm) did these particular features
develop to produce something approaching the pattern observed
on photolysis of matrixes containing thermally generated Zn
atoms.
The IR spectra associated with the products of the reactions
of the metal atoms with SiH4 and SiD4 will be reported in turn
for Zn, Cd, and Hg. Bands have been assigned on the basis of
the following criteria: (i) their growth or decay characteristics
under different conditions; (ii) comparisons with the spectra
registered in control experiments and with the spectra of related
species; (iii) the observed effects of exchanging SiH4 for SiD4;
and (iv) consideration of how well the measured spectra are
reproduced by the results of DFT calculations for a particular
product.
Zinc. Various experiments were carried out with thermally
generated Zn atoms being co-deposited with an excess of SiH4-
doped argon. The IR spectrum of the deposit consisted only of
absorptions attributable to isolated SiH4 molecules,21 as well
as weak features associated with trace impurities (H2O,22
[H2O]n,22 CO2,23 CO,24 and Zn‚OH2)17,25 that could be kept to
a minimum but never wholly eliminated. By contrast with
similar experiments involving Al or Ga atoms,8,9 no additional
bands could be discerned as a result of the inclusion of Zn atoms
in the matrix. Even matrixes containing only 0.1% SiH4, which
displayed sharp, well-resolved bands, gave no hint of shoulders
or satellites attributable to Zn‚‚‚SiH4 contact pairs. In this
respect, the behavior parallels that of Zn atoms with CH415 but
not with HCl16 and H2O,17,25 which give clear signs of
perturbation in the immediate presence of the metal atoms.
Exposure of the matrix to broad-band UV-vis irradiation
(200 e λ e 800 nm) for as little as 1 min led, as illustrated in
Cadmium. Figure 2 shows the IR spectra recorded in an
experiment in which thermally evaporated Cd atoms were co-
deposited with SiH4-doped Ar (Ar/SiH4 ) 100:1); details of

HMSiH3 in a Solid Argon Matrix
J. Phys. Chem. A, Vol. 108, No. 8, 2004 1395
Figure 1. (a) IR spectrum of the deposit formed by cocondensing Zn atoms with a 1% mixture of SiH4 in Ar at 12 K, and (b) IR spectrum of the
same deposit after broad-band UV-vis photolysis.
TABLE 1: Wavenumbers (cm-1) of IR Absorptions
Observed Following UV Photolysis of an Ar Matrix
Containing Zn Atoms and SiH4 or SiD4
from their counterparts in the Zn experiments, are most likely
to be due to ν(Cd-H) modes (cf. CdH2 1753.5,14 HCdCH3
1760.5,15 HCdCl 1835.8,16 HCdOH 1837.0 cm-1).17 Indeed, a
feature at 1753.8 cm-1, together with a weak doublet at 604.4/
601.7 cm-1, is identifiable with the presence of the CdH2
molecule.14 The regions 930-950 and 830-870 cm-1 match
closely those in the Zn experiments, and the relevant signals
originate most probably in δ(SiH3) vibrations. The close parallel
between the behaviors of Zn and Cd makes it unlikely that
silicon hydride anions (expected to absorb at 1920-1860
cm-1 26
) contribute to either set of spectra. The parallel extended
to the behavior of the Cd/SiH4 matrix on continued photolysis,
which led to the growth of the main bands at ca. 1740, 940,
860, and 835 cm-1 but to a slower buildup of the satellites to
low wavenumbers of 1740 and 860 cm-1
.
The same results could be achieved rather slower on selective
photolysis with light having λ ) ca. 313 nm. Moreover, the
products showed no signs of photolability, either on continued
UV irradiation or on exposure to visible light (λ ) 400-800
nm).
Substitution of SiD4 for SiH4 resulted in marked redshifts of
all the major IR bands arising from photoproducts, with those
at 1700-1760 moving to 1220-1270 cm-1 (H/D ) 1.39-
1.40:1), those at 930-950 to ca. 690 cm-1 (H/D ) 1.38:1),
and those at 830-870 to 610-640 cm-1 (H/D ) 1.345:1). The
results of deuteration are thus wholly consistent with the
attribution of the original bands to either ν(Cd-H) or δ(SiH3)
fundamentals.
a Feature common to Zn and Cd experiments.
wavenumbers are given in Table 2. After deposition, the matrix
displayed no IR bands that could not be ascribed to free SiH4
or trace impurities. After only 1 min of broad-band UV-vis
photolysis (200 e λ e 800 nm), however, several new bands
appeared and grew on continued irradiation. As in the Zn
experiments, these were confined mainly to three regions, viz.,
(i) 1700-1760 cm-1, (ii) 930-950 cm-1, and (iii) 830-870
cm-1. The bands at highest wavenumber, appreciably redshifted
Introduction of the Cd vapor not from a Knudsen cell but
from a resonance lamp gave matrixes which showed evidence

1396 J. Phys. Chem. A, Vol. 108, No. 8, 2004
Macrae et al.
Figure 2. (a) IR spectrum of the deposit formed by cocondensing Cd atoms with a 1% mixture of SiH4 in Ar at 12 K, and (b) IR spectrum of the
same deposit after broad-band UV-vis photolysis.
TABLE 2: Wavenumbers (cm-1) of IR Absorptions
Observed Following UV Photolysis of an Ar Matrix
Containing Cd Atoms and SiH4 or SiD4
In the case of the Zn/SiH4 experiments, no useful information
could be gleaned from measurements of the UV-vis spectrum
of the Ar matrix that were effectively limited by a short-
wavelength cutoff of ca. 300 nm. Similar measurements on a
matrix containing thermally generated Cd atoms revealed a UV
3
1
absorption near 310 nm attributable to the P r S transition
of atomic Cd.14,15 Significantly, the presence of SiH4 appeared
to produce a broadening of the band but no change in its
2
2
position. By contrast, the S r P transition of atomic Al in
similar circumstances is redshifted by some 45 nm.8,9 Any
interaction between SiH4 and the group 12 metal atom in its
ground electronic state must therefore be very weak. As noted
3
1
above, selective irradiation into the P r S absorption band
of Cd with light having λ ) ca. 313 nm induced the growth of
the IR bands at 1700-1760, 930-950, and 830-870 cm-1
.
Mercury. Photoexcitation of Hg atoms to the 3P1 electronic
state in the presence of SiH4 in a solid inert matrix at low
temperatures has been shown previously10 to result in insertion
into an Si-H bond. In addition to HHgSiH3, the primary product
thus formed, Si2H6 emerges as a significant secondary product.
We have carried out experiments along the lines described in
the preceding sections by substituting thermally generated Hg
for Zn or Cd vapor. The IR spectrum of the Ar matrix initially
formed gave no hint of a spontaneous reaction between SiH4
and Hg atoms in their ground electronic state. On the other hand,
broad-band UV-vis irradiation of the matrix gave rise to the
IR results illustrated in Figure 3 and detailed in Table 3. New
bands appeared and grew with continued photolysis near 1887,
940, 870, and 835 cm-1. Similar findings were reported earlier10
on the basis of photoexcitation with the output from either a
a Feature common to Zn and Cd experiments.
of reaction with SiH4 on deposition. Thus, the IR spectrum of
a matrix prepared in this way exhibited bands characteristic of
CdH2.14 In addition, the bands centered at ca. 938 and 835 cm-1
were relatively prominent compared with those at 1700-1750
and 840-870 cm-1. However, subsequent broad-band UV-
vis photolysis led to the progressive buildup of the latter features
and the ultimate development of a spectrum analogous to that
observed with thermally generated metal atoms.

HMSiH3 in a Solid Argon Matrix
J. Phys. Chem. A, Vol. 108, No. 8, 2004 1397
Figure 3. (a) IR spectrum of the deposit formed by cocondensing Hg atoms with a 1% mixture of SiH4 in Ar at 12 K, and (b) IR spectrum of the
same deposit after broad-band UV-vis photolysis.
TABLE 3: Wavenumbers (cm-1) of IR Absorptions
Observed Following UV Photolysis of an Ar Matrix
Containing Hg Atoms and SiH4 or SiD4
gave way to SiD4 with H/D ratios varying from 1.393:1 to
1.344:1 (see Table 3) and thereby helping to identify the nature
of the vibration as well as the carrier of the absorption (q.v.).
As in the studies with Zn and Cd, introduction of the Hg
vapor from a resonance lamp caused some reaction with SiH4
to occur during deposition of the Ar matrix. The IR spectrum
of the resulting deposit then included the absorptions near 940
and 835 cm-1, but of the absorptions at ca. 1880 and 870 cm-1
,
there was no sign. Only on UV photolysis of the matrix were
these features observed to develop.
IV. Discussion
The reactions of group 12 metal atoms with both methane
and silane have attracted interest not only through experimental
studies3,10,12,15 compassing the gaseous as well as the matrix
phases but also through quantum chemical analysis.12,13 As
described above, new IR features are seen to develop as a result
(a) of photolytically induced reactions between SiH4 and Zn,
Cd, or Hg atoms (M) isolated together in a solid Ar matrix or
(b) of cocondensation of the same metal atoms from a reactive
resonance lamp with an excess of SiH4-doped Ar. The main
products will be shown to be the silyl metal hydride, HMSiH3,
formed by insertion of the electronically excited M atom into
an Si-H bond, and disilane, Si2H6, in proportions that vary with
the conditions of formation. Analogous studies of the systems
M/HCl,16 M/H2O,17,25 and M′/SiH48,9 (M ) Zn, Cd, or Hg; M′
) Al or Ga) have provided clear spectroscopic evidence that
the metal atom in its ground electronic state adds to the substrate
to form a loosely bound preinsertion complex. Unlike the group
13 metal atoms Al and Ga,8,9 the group 12 metal atoms show
deuterium lamp or a KrF excimer laser. In that case, the features
near 1887 and 870 cm-1 were associated with the insertion
product HHgSiH3 and those near 940 and 835 cm-1 with Si2H6.
The results of our experiments with Hg differed somewhat from
those of the corresponding experiments with Zn and Cd in that
28
there was no sign of the binary mercury hydride HgH2 and
the IR bands near 940 and 835 cm-1 appeared to be significantly
weaker compared with those at high wavenumber (1700-1900
cm-1) and near 870 cm-1. In keeping with the previous report,10
the main bands near 1887, 940, 870, and 835 cm-1 built up on
continued photolysis, but the weaker features near 1875 and
860 cm-1 grew more slowly.
Selective photolysis of the initial Hg/SiH4-doped matrix with
radiation having λ ) ca. 254 nm produced similar results, albeit
at a reduced rate of product growth. By contrast, visible radiation
had no effect either before or after exposure to UV photolysis.
All the product bands showed a pronounced redshift when SiH4

1398 J. Phys. Chem. A, Vol. 108, No. 8, 2004
Macrae et al.
a
TABLE 4: Reaction Energies, ∆E in kJ mol-1, Calculated for Selected Changes Involving M (Zn, Cd, or Hg) and SiH4
change
M (1S) f M (3P)
M ) Zn
M ) Cd
M ) Hg
+410a/391b
+385a/374b
+495a/500b
M (3P) + SiH4 f HMSiH3 (3A)
M (3P) + SiH4 f HMSiH3 (1A1)
M (3P) + SiH4 f MH• + •SiH3
M (3P) + SiH4 f H• + •MSiH3
M (3P) + SiH4 f HMH + SiH2 (3A1)
-138
-96
-170
-397
-335
-415
-127a/96b
-78
-87a/64b
-46
-165a/151b
-132
-91
-21
-103
a Basis set CEP-31G. b Experimental value.13a
TABLE 5: Calculated Bond Lengths (Å) and Angles (deg)
of Equilibrium Geometries of HMSiH3 Moleculesa
no sign of forming such a complex with SiH4. Accordingly,
we consider first the potential for interaction with Zn, Cd, or
Hg atoms before going on to analyze the properties of the
HMSiH3 molecules and the likely mechanisms controlling the
formation of these and Si2H6.
fragment
M ) Zn
M ) Cd
M ) Hg
H-M
1.5545
2.3836
1.5159
106.18
112.58
1.7127
2.5495
1.5158
106.18
112.59
1.6962
2.5311
1.5134
106.97
111.87
M-Si
M‚‚‚SiH4. Our DFT calculations using the B3LYP method
confirm expectations and the results of earlier calculations
involving both CH429 and SiH413 that in its 1S ground electronic
state a Zn, Cd, or Hg atom is incapable of activating the SiH4
molecule. With unrestrained geometry, only very shallow
potential-energy minima can be found for the M (1S)‚‚‚SiH4
assembly with depths calculated not to exceed 1 kJ mol-1. The
only significant interaction, occurring between the valence ns
orbital of M and the bonding t2 orbitals of the SiH4, can lead
only to full occupation of bonding and antibonding MOs and
no net bonding. The situation is not unlike that between a noble
gas and SiH4, with neither component able to offer occupied or
vacant orbitals of suitable energy to provide a useful basis for
a specific donor-acceptor interaction. This contrasts with the
behaviors of M (1S) atoms in the presence of HCl16 and H2O17
in similar circumstances; vibrational perturbation of these
substrates attests to the formation of loosely bound M‚‚‚HCl
and M‚‚‚H2O adducts. SiH4 is capable of forming such adducts
with ground-state (2P) Al and Ga atoms.8,9 Coordination of the
metal via an edge of the SiH4 tetrahedron is then calculated to
result in a binding energy of ca. 10 kJ mol-1, with the
perturbation leading to a slight distortion and measurable shifts
of the vibrational fundamentals of the SiH4 molecule under
appropriate conditions of matrix isolation. In keeping with the
reduced basicity of the group 12 metals and theoretical forecasts,
however, our experiments with Zn, Cd, or Hg under the same
conditions were unable to detect any comparable IR features.
None of the possible reactions 1-4 between SiH4 and a metal
atom M ) Zn, Cd, or Hg is predicted by either present or
Si-H
M-Si-H
H-Si-H
a Basis set CEP-31G, symmetry C3V.
favorable option involves relaxation to the singlet (1A1) ground
state, which is characterized by a linear H-M-Si skeleton
(thereby completing reaction 1). On the other hand, calculations
indicate that little or no barrier opposes the dissociation of
•
•
HMSiH3 (3A′) into MH and SiH3, which are the products
observed in experimental studies of the interaction between M
(3P) and SiH4 in the gas phase (reaction 2).3,10,12,13 Of the
reaction channels 3 and 4, no direct experimental evidence has
yet come to light.
HMSiH3 (1A1). DFT calculations using the B3LYP functional
find an equilibrium geometry with C3V symmetry for each of
the molecules HMSiH3 in its singlet ground state. The dimen-
sions and wavenumbers and intensities in IR absorption of the
fundamental vibrational transitions calculated for both HMSiH3
and DMSiD3 are given in Tables 5 and 6; the results for HMSiH3
are close to those reported previously by Alikhani.13 The
most intense IR absorptions are predicted to correspond to the
ν(M-H) (ν2) and symmetric δ(SiH3) (ν3) modes with wave-
numbers of 1700-1900 and ca. 830 cm-1, respectively. Next
in order of intensity are the ν(Si-H) modes (ν1 and ν5) at 2100-
2200 cm-1, δ(SiHM) (ν8) at 300-400 cm-1, and antisymmetric
δ(SiH3) (ν6) near 920 cm-1. However, the earlier experience
of Legay-Sommaire and Legay10 with matrix-isolated HHgSiH3
is that the proximity of ν1 and ν5 to the ν3 (t2) absorption of the
much more abundant SiH4 precursor is liable in practice to
prevent them from being observed. The ν8 absorption is
predicted to be less than one-fifth the intensity of the strongest
absorption (due to ν2) as well as occurring in a region of reduced
detector sensitivity. The remaining fundamentals ν6, ν7, and ν4
are expected to have IR intensities that decrease in that order,
being at least 1 order of magnitude weaker than the strongest
absorption.
1
earlier13 studies to be exothermic when M is in its S ground
electronic state. Clearly, therefore, energy has to be injected if
any such change is to occur. There is now ample theoretical
evidence that excitation of M to its lowest-lying excited state,
M + SiH4 f HMSiH3
(1)
(2)
M + SiH4 f •MH + •SiH3
M + SiH4 f H• + •MSiH3
M + SiH4 f HMH + SiH2
(3)
(4)
On this basis, the two regions of strongest IR absorption that
were observed to develop at 1700-1900 and 850-880 cm-1
on UV irradiation of an Ar matrix containing SiH4 and Zn, Cd,
or Hg atoms are most plausibly identified with ν2 and ν3 of the
appropriate insertion product HMSiH3. Such an inference is
supported not only by the effects of deuteration but also by the
clear parallels that can be drawn with the earlier study of the
Hg/SiH4 system.10 Despite every endeavor, no sign of the
ν(Si-H) or ν(Si-D) fundamentals ν1 and ν5 could be elicited
for any of the molecules, the relevant features being obscured
presumably by the much stronger absorption due to ν3 (t2) of
the parent silane molecule.21 No similar problem beset the
detection of the ν8 mode, and the failure to descry any feature
3P, gives rise to an exciplex M*‚SiH4, now with a significant
binding energy and also with the thermodynamic potential to
undergo many of the reactions 1-4 (see Table 4). Moreover,
there appears to be little or no barrier to the breaking of the
Si-H bond. Depending on M and the conditions, this may lead
initially to an HMSiH3 intermediate in a 3A′ excited state, having
a highly angular H-M-Si skeleton and an M-H bond length
very similar to that of the free •MH molecule. Of the channels
open to this excited molecule, the thermodynamically most

HMSiH3 in a Solid Argon Matrix
J. Phys. Chem. A, Vol. 108, No. 8, 2004 1399
TABLE 6: Comparison of Calculated and Observed Wavenumbers (cm-1) for HMSiH3 Molecules: Calculated Intensities (km
mol-1) in Parentheses
SiH4
SiD4
metal
Zn
mode
obs
calca
obs
calca
ν1
ν2
ν3
ν4
ν5
ν6
ν7
ν8
ν1
ν2
ν3
ν4
ν5
ν6
ν7
ν8
ν1
ν2
ν3
ν4
ν5
ν6
ν7
ν8
2090.1 (73)
1887.5 (312)
831.6 (225)
325.9(4)
2116.1 (156)
921.0 (33)
486.7 (14)
336.6 (42)
2088.3 (91)
1737.9 (347)
823.1 (276)
287.3 (4)
2116.8 (156)
921.8 (36)
470.4 (18)
337.7 (41)
2098.3 (77)
1857.4 (478)
827.7 (307)
294.7 (3)
1488.9 (40)
1345.1 (162)
615.2 (112)
311.8 (7)
1527.1 (83)
658.5 (16)
364.4 (8)
240.6 (22)
1487.5 (50)
1234.8 (177)
608.8 (139)
275.8 (6)
1527.8 (82)
659.3 (17)
349.0 (9)
241.0 (21)
1494.2 (43)
1317.0 (241)
612.4 (156)
281.6 (5)
1536.2 (78)
658.7 (18)
393.6 (1)
1846.6-1821.6
874.2-855.2
1329.8-1312.6
639.9-637.5
Cd
Hg
1745.9-1704.2
860.2-849.7
1252.4-1224.1
639.5-633.0
1886.7-1855.9
871.2-860.6
1354.4-1332.0
647.7-643.3
2127.6 (143)
920.7 (36)
534.8 (2)
364.5 (20)
260.6 (10)
a Basis set CEP-31G, symmetry C3V.
attributable to this source must be put down to a combination
of the low yield, inherent low intensity, and reduced sensitivity
of detection. It is possible that ν6 contributes to the group of
bands observed to develop at 930-950 cm-1 irrespective of
whether the metal was Zn, Cd, or Hg. However, the intensity
of the bands and their preferential growth in experiments using
a resonance lamp as the source of the metal vapor leaves little
doubt that they arise primarily from some other product (q.v.).
The multiplet patterns characterizing the two regions of
absorption we associate with ν2 and ν3 of the HMSiH3 molecules
cannot be ascribed wholly to simple matrix-site effects.2 Apart
from the primary features that we associate with well-isolated
HMSiH3 molecules, the satellites appearance typically at lower
energy suggests the presence of loosely bound contact pairs
HMSiH3‚‚‚X, where X is most likely to be SiH4, M, or even
H2O impurity. Somewhat surprisingly, the relative intensities
of the components in each multiplet did not appear to change
appreciably when the concentration of SiH4 was varied, or on
annealing the matrix. Accordingly, it was impossible to establish
with any certainty the precise origins of the individual bands.
A further complication arose from the proximity of the 850-
The insensitivity to the nature of the metal clearly points to
a metal-free silicon hydride. The various possibilities include
the neutral molecules SiH2, SiH3, Si2H2, Si2H4, and Si2H6.
•
Comparison with the results of earlier matrix studies10,26 leads
us to conclude that only one of these has the right IR credentials
to be the carrier of the bands at ca. 940 and 835 cm-1, namely,
Si2H6. Thus, we find no sign of the low-wavenumber ν(Si-H)
absorptions characterizing SiH2, of the symmetric δ(SiH3)
absorption at 720-730 cm-1, which is the strongest feature on
the spectrum of •SiH3, or of the distinctive absorptions attributed
to either Si2H2 or Si2H4.26 That Si2H6 should be formed in our
experiments need not be a matter for surprise in that it was
also recognized as a significant photoproduct of earlier experi-
ments involving matrixes doped with Hg and SiH4.10
Matrix-isolated metal atoms are also susceptible to photo-
ionization, especially in the presence of a suitable electron
trap.3,30 Accordingly, we cannot altogether discount the pos-
sibility that short-wavelength UV photolysis results in electron
transfer to give products of the type M+SiHn-. Recent matrix
studies of the reactions of laser-ablated Si atoms with hydrogen
•
-
have led to the assignment of IR bands to the anions SiH2
- 26
880 cm-1 absorptions to a second set centered near 835 cm-1
.
and SiH3
.
Although some of our product bands appear at
The latter normally grew in at the same rate as the first, but
being much more prominent in the spectra of the deposits
formed in resonance lamp experiments, it must originate in a
silicon hydride product other than HMSiH3.
similar wavenumbers, they are unlikely to arise from these or
other anionic species for several reasons. First, they build up
on continued UV photolysis, conditions more likely to cause
photodecay of such products.26 Second, the bands that might
be so mistaken shift markedly when the metal is changed.
Finally, the H/D ratios do not support such an assignment.
HMH. Experiments with Zn or Cd but not with Hg revealed
the growth of weak bands signaling the formation of the
dihydride molecules as secondary products of UV photolysis.
The wavenumbers of the relevant bands are in close agreement
with those reported previously for these molecules trapped in
Ar matrixes.14 On the other hand, none of the experiments gave
any hint of the monohydride that would be expected to absorb
Si2H6. Hence two absorptions attributable to a major photo-
product have still to be positively assigned. These occur at ca.
940 and 835 cm-1 with relative intensities clearly implying that
they originate in the same molecule, and it is also noteworthy
that the wavenumbers and profiles of the absorptions did not
vary appreciably with the nature of the metal. The wavenumbers
and H/D ratios afford strong circumstantial evidence that the
bands represent deformation modes of an SiHn unit. No
additional features traceable to the same source could be
observed in the ν(Si-H) region of the spectrum, and so it must
be supposed that the relevant transitions were again masked by
the intense absorption hereabouts due to SiH4.
in the region ca. 1200-1500 cm-1 14,15,28
nor was there anything
,
to suggest a role for dimetal species such as HMMSiH3, HMM,
or HMMH.14 It may be noted that the dihydride is also a

1400 J. Phys. Chem. A, Vol. 108, No. 8, 2004
Macrae et al.
SCHEME 1
TABLE 7: ν(M-H) Wavenumbers (in cm-1) for HMX
Molecules Isolated in Ar Matrixes
secondary product that is observed to be formed when excited
Zn, Cd, or Hg atoms are trapped with CH4 in Ar matrixes.15
The only other changes in the IR spectra brought about by
UV photolysis could be traced to photoinduced reactions of Zn
and Cd (M) atoms with H2O impurity, giving the insertion
products HMOH.17,25 These could be recognized by the ap-
pearance of very weak absorptions at 1955.0 (M ) Zn) and
1837.0 cm-1 (M ) Cd).
molecule
M ) Zn
M ) Cd
M ) Hg
ref
HMH
HMCl
HMOH
HMCH3
HMSiH3
1870.2a
1952.3
1955.0
1866.1
1846.6
1753.5a
1835.8
1837.0
1760.5
1745.9
1943.7a,b
2092.0
2116.5
1955.3
1886.7
14, 28
16
17, 25
15
this work
a Antisymmetric ν(M-H) mode. b N2 matrix.
Reaction Mechanisms. The conditions of photolysis used
in our experiments were such as to excite the group 12 metal
1
atoms, M, trapped with SiH4 in the Ar matrixes from their S
excited state with energy sufficient to overcome the barrier to
insertion. No such conflict exists for the reactions of M (3P)
with SiH4, although this does not mean of course that the
possibility of two-photon processes is to be excluded.
3
ground electronic state to the P excited state. The chances of
1
populating the more energetic P state were relatively slight,
except when a resonance lamp was used to deliver the metal
vapor to the matrix; in that case, the discharge populated both
In the gas phase, the initial insertion product [HMSiH3]* (3A′)
undergoes fragmentation with the formation of the radicals
3
1
the P and P states and generated the resonance radiation to
excite other atoms. Earlier experiments and theoretical stud-
ies3,10,12,13 leave little doubt, it seems, that M (3P) atoms are
capable of rupturing an Si-H bond of the SiH4 molecule and
that this occurs in the sequence set out in Scheme 1. Although
neither the complex [M (3P)‚‚‚SiH4]* nor the first insertion
product [HMSiH3]* (3A′) has been detected directly by experi-
•
•MH and SiH3 (reaction 2). It may be that fragmentation also
occurs when the reagents are held in a solid Ar matrix but that
the failure of either fragment to escape from the matrix cage
results in recombination to produce HMSiH3 in its 1A1 ground
electronic state; thermodynamically, this is the best favored of
all the possible outcomes short of returning to M (1S) + SiH4.
Unfortunately, our experiments give no clue to whether this or
physical relaxation of the [HMSiH3]* (3A′) molecule, made
possible by spin-orbit coupling and/or matrix mediation, is the
principal agency leading to the observed insertion product. The
IR spectra of the photolyzed matrixes certainly gave no signs
suggesting the presence of either •MH14,15,28 or •SiH326 radicals.
Traces of the dihydrides HMH14 were observed in our experi-
ments with M ) Zn or Cd but not with Hg. These may be
formed from traces of H2 impurity in the matrix, or they may
be the outcome of secondary changes. For example, reaction 3
would give rise to H• atoms, which are known to be quite mobile
in Ar matrixes30 and could compete with •SiH3 to trap a fraction
of the •MH radicals. Alternatively, reaction 4 would deliver the
dihydride directly, but of the expected coproduct SiH2,26 we
could find no trace.
With only two vibrational fundamentals that can be character-
ized in practice, the HMSiH3 molecules do not lend themselves
to detailed vibrational analysis. We note, however, that the
ν(M-H) modes occur at wavenumbers wholly in keeping with
those of related molecules (see Table 7). The two regions of
absorption associated with HMSiH3 displayed relatively rich
multiplet patterns that must be attributed to the occupation of
different matrix sites, including some which incorporate species
other than Ar and so give rise to loosely bound contact pairs of
the form HMSiH3‚‚‚X, where X ) SiH4, M, or trace impurities
such as H2O. All of the relevant features grow on continued
photolysis. That some do this at a faster rate than others (q.v.)
•
ment, the vibrational and rotational energies of MH, one of
the ultimate gaseous products, reflect the character of the
transient insertion product.12
Evidently the initial contact pair M (1S)‚‚‚SiH4 involves
minimal mutual perturbation of the metal atom in its ground
electronic state and the SiH4 molecule. In this respect, SiH4 does
not differ from its less polarizable homologue CH4.14 Only with
appreciably more acidic partners, such as HCl16 and H2O,17,25
is it possible to identify experimentally the formation of a
distinct 1:1 adduct with the group 12 metal atom. Following
excitation to its 3P excited state, the Zn, Cd, or Hg atom interacts
much more strongly with SiH4 on the evidence of theoretical
calculations;12,13 still more significantly, insertion of the metal
atom into an Si-H bond then occurs more or less spontaneously
in an exothermic reaction yielding [HMSiH3]* in its excited
3A′ state. This contrasts with the behavior of CH4 for which
the corresponding reaction is calculated to be met by an
appreciable activation barrier (ca. 75 kJ mol-1 in the case of
Zn), possibly as a result of the increased steric hindrance at the
more compact and congested carbon center.12,13 When the
reagents are isolated in a solid Ar matrix, however, the outcome
is the same for CH4 and SiH4, with HMEH3 (E ) C or Si; M
1
) Zn, Cd, or Hg) having a linear H-M-E skeleton in its A1
ground state being identified as the primary product. The conflict
between experiment and theory in the cases where M ) Zn or
Cd and E ) C cannot be easily resolved; it may be that the M
(3P)‚‚‚CH4 complex absorbs a second photon to reach a higher

HMSiH3 in a Solid Argon Matrix
J. Phys. Chem. A, Vol. 108, No. 8, 2004 1401
may be a sign of secondary reactions, but equally well it may
reflect merely the known phenomenon31 of varying photosen-
sitivity from one matrix site to another. The multiplets, which
were also observed in the earlier study of the Hg/SiH4 system,10
are consistent with a mechanism involving the breakup of the
initial insertion product [HMSiH3]* (3A′) followed by recom-
When the group 12 metal atoms are introduced from a
resonance lamp, reaction with SiH4 occurs during co-deposition
but with Si2H6 now as the dominant product. It is difficult to
know in these circumstances the parts played by metal atoms
in the more highly excited 1P state and by the greater mobility
enjoyed by reagents and products alike during condensation.
The conditions come closer to those prevailing in the gas phase
where •SiH3 formation becomes a major reaction channel in the
metal-sensitized decomposition of SiH4; the energy content of
the reaction channel may then be such as to dissociate the
weakly bound •MH radical.10 Dimerization of the •SiH3 radicals
during deposition would then account for the formation of Si2H6;
alternatively, reaction 5 may become important at the deposition
stage.
Our experiments give no reason to believe that well-isolated
HMSiH3 molecules differ from the HMCH3 molecules,14 which
are photostable under the conditions of broad-band photolysis
used to generate them. There is no hint of decomposition to
form either •MH or •MSiH3, matching the behavior of HAlSiH3
and HGaSiH3 in similar conditions, nor is there any definite
sign that insertion can be reversed, although this would actually
be difficult to establish when photoconversion is relatively
inefficient and in the absence of any distinctive mark of the
precursor, i.e., the M (1S)‚‚‚SiH4 contact pair.
•
bination of the HM• and SiH3 radicals within the confines of
the matrix cage. The various components then reflect changes
of orientation of the product molecule as well as differences in
the character of the matrix cage.
One of the most puzzling features of this and the earlier10
study is the provenance of Si2H6, which is the only other
photoproduct to appear in all experiments in concentrations
comparable with that of HMSiH3. Si2H6 is formed when matrix-
isolated SiH4 dimers are irradiated with far-UV light (λ ) 193
nm),10 but this two-photon process is unlikely to contribute
significantly to the yield of Si2H6 observed in the experiments
with group 12 metal atoms. Instead, Legay-Sommaire and Legay
have suggested10 that the excited metal atoms act on adjacent
SiH4 dimers to promote reaction 5; the dihydride HMH may
then be a secondary product as excited metal atoms compete
for the H2 thus formed.14,28 Alternatively HMSiH3 may be
M* + [SiH4]2 f Si2H6 + M + H2
(5)
Acknowledgment. We thank the EPSRC for the award of
an Advanced Fellowship to T.M.G., the funding of a studentship
for V.A.M., and financial support of the Oxford group.
photoinduced to react with a neighboring SiH4 molecule in
accordance with eq 6. Support for the second mechanism was
provided in the earlier Hg/SiH4 experiments10 by the finding
that prolonged irradiation caused a decrease in the intensity of
References and Notes
HMSiH3‚‚‚SiH4 f Si2H6 + M + H2
(6)
(1) Crabtree, R. H. Chem. ReV. 1985, 85, 245. Schubert, U. AdV.
Organomet. Chem. 1990, 30, 151. Crabtree, R. H. Angew. Chem., Int. Ed.
1993, 32, 789. Masaev, D. G.; Morokuma, K. J. Am. Chem. Soc. 1995,
117, 799. Corey, J. Y.; Braddock-Wilking, J. Chem. ReV. 1999, 99, 175.
Lin, Z. Chem. Soc. ReV. 2002, 31, 239.
(2) Almond, M. J.; Downs, A. J. AdV. Spectrosc. 1989, 17, 1. Dunkin,
I. R. Matrix-Isolation Techniques: A Practical Approach; Oxford University
Press: Oxford, U.K., 1998.
the IR absorptions attributed to HHgSiH3‚‚‚SiH4 contact pairs.
No such decrease was observed in our experiments, but the
intensities of the main low-wavenumber satellites, most plausibly
identified with HMSiH3‚‚‚SiH4 pairs, did grow distinctly more
slowly than the features associated with HMSiH3 surrounded
exclusively by Ar atoms. A further possibility, not considered
previously, is that trapping of the HM• and •SiH3 radicals formed
(3) Himmel, H.-J.; Downs, A. J.; Greene, T. M. Chem. ReV. 2002, 102,
4191.
(4) Meininger, D. R.; Ault, B. S. J. Phys. Chem. A 2000, 104, 3481.
(5) Davis, S. R.; Andrews, L. J. Phys. Chem. 1989, 93, 1273.
(6) Mielke, Z.; Tokhadze, K. G. Chem. Phys. Lett. 2000, 316, 108.
(7) Lefcourt, M. A.; Ozin, G. A. J. Am. Chem. Soc. 1988, 110, 6888.
(b) Lefcourt, M. A.; Ozin, G. A. J. Phys. Chem. 1991, 95, 2616.
(8) Gaertner, B.; Himmel, H.-J. Angew. Chem., Int. Ed. 2002, 41, 1538.
(9) Gaertner, B.; Himmel, H.-J.; Macrae, V. A.; Downs, A. J.; Greene,
T. M. Chem.-Eur. J. Manuscript submitted to be considered for publication.
(10) Legay-Sommaire, N.; Legay, F. J. Phys. Chem. A 1998, 102, 8759.
(11) Maier, G.; Reisenauer, H. P.; Glatthaar, J. Chem.-Eur. J. 2002, 8,
4383.
(12) Breckenridge, W. H. J. Phys. Chem. 1996, 100, 14840.
(13) (a) Luna-Garc´ıa, H.; Castillo, S.; Ram´ırez-Sol´ıs, A. J. Chem. Phys.
1997, 107, 6627. (b) Luna-Garc´ıa, H.; Castillo, S.; Ram´ırez-Sol´ıs, A. J.
Chem. Phys. 1999, 110, 11315. (c) Alikhani, M. E. Chem. Phys. Lett. 1999,
313, 608. (d) Siegbahn, P. E. M.; Svensson, M.; Crabtree, R. H. J. Am.
Chem. Soc. 1995, 117, 6758.
(14) Greene, T. M.; Brown, W.; Andrews, L.; Downs, A. J.; Chertihin,
G. V.; Runeberg, N.; Pykko¨, P. J. Phys. Chem. 1995, 99, 7925.
(15) Greene, T. M.; Andrews, L.; Downs, A. J. J. Am. Chem. Soc. 1995,
117, 8180.
(16) Macrae, V. A.; Greene, T. M.; Green, J. C.; Downs, A. J. J. Phys.
Chem. A Manuscript submitted to be considered for publication.
(17) Macrae, V. A.; Greene, T. M.; Downs, A. J. Unpublished results.
(18) Norman, A. D.; Webster, J. R.; Jolly, W. L. Inorg. Synth. 1968,
11, 170.
3
by fragmentation of the A′ insertion product does not lead
exclusively to HMSiH3 (1A1), but that reaction of the fragments
with SiH4 contained in some of the matrix cages leads to the
reactions 7 and 8
HM• + SiH4 f HMSiH3 + H•
•SiH3 + SiH4 f Si2H6 + H•
(7)
(8)
A surprising feature of our results is that the relative yields
of HMSiH3 and Si2H6 did not vary appreciably with the
concentration of SiH4 in the matrix, although the degree of
photoconversion was invariably small. It is difficult therefore
to reconcile this result with reactions that require only one SiH4
molecule (to form HMSiH3 by reaction 1) and two SiH4
molecules (to form Si2H6, whether by reaction 5 or 6). It seems
altogether more likely that reactions 7 and 8 are actually the
primary sources of the two main products observed in our
experiments with thermally generated metal atoms. The need
for the intervention of two further SiH4 molecules to form either
HMSiH3 or Si2H6 would then account not only for the near
invariance of the relative yields with concentration but also for
the paucity of these yields. On this basis, it must be supposed
that, without an SiH4 molecule in the matrix cage, the triplet
insertion product reverts to M (1S)‚‚‚SiH4.
(19) Himmel, H.-J.; Downs, A. J.; Greene, T. M.; Andrews, L.
Organometallics 2000, 19, 1060.
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,

1402 J. Phys. Chem. A, Vol. 108, No. 8, 2004
Macrae et al.
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.3; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(21) Milligan, D. E.; Jacox, M. E. J. Chem. Phys. 1970, 52, 2594. Wilde,
R. E.; Srinivasan, T. K. K.; Harral, R. W.; Sankar, S. G. J. Chem. Phys.
1971, 55, 5681.
(22) Ayers, G. P.; Pullin, A. D. E. Specrochim. Acta 1976, 32A, 1629.
Fredin, L.; Nelander, B.; Ribbegård, G. J. Chem. Phys. 1977, 66, 4065.
(23) Fredin, L.; Nelander, B.; Ribbegård, G. J. Mol. Spectrosc. 1974,
53, 410. Guasti, R.; Schettino, V.; Brigot, N. Chem. Phys. 1978, 34, 391.
(24) Dubost, H. Chem. Phys. 1976, 12, 139.
(25) Kauffman, J. W.; Hauge, R. H.; Margrave, J. L. J. Phys. Chem.
1985, 89, 3541.
(26) Andrews, L.; Wang, X. J. Phys. Chem. A 2002, 106, 7696. (b)
Andrews, L.; Wang, X. J. Am. Chem Soc. 2003, 125, 6581.
(27) Bu¨rger, H.; Rahner, A. Vib. Spectra Struct. 1990, 18, 217.
(28) Legay-Sommarire, N.; Legay, F. Chem Phys. Lett. 1993, 207, 123.
(b) Legay-Sommarire, N.; Legay, F. J. Phys. Chem. 1995, 99, 16945.
(29) Castillo, S.; Ram´ırez-Sol´ıs, A.; Diaz, D.; Poulain, E.; Novaro, O.
Mol. Phys. 1994, 81, 825. (b) Castillo, S.; Ram´ırez-Sol´ıs, A.; Poulain, E.
Int. J. Quantum Chem., Quantum Chem. Symp. 1993, 27, 587.
(30) Bondybey, V. E.; Ra¨sa¨nen, M.; Lammers, A. Annu. Rep. Prog.
Chem., Sect. C 1999, 95, 331.
(31) See, for example: (a) Poliakoff, M.; Turner, J. J. J. Chem. Soc.,
Dalton Trans. 1973, 1351. (b) Fanfarillo, M.; Cribb, H. E.; Downs, A. J.;
Greene, T. M.; Almond, A. J. Inorg. Chem. 1992, 31, 2962.