Bis(chloromethylsilyl)amine and bis(chloromethylsilyl)methylamine; preparation, reactivity and spectroscopic studies of their stereoisomers and conformers

Article abstract of DOI:10.1039/a708220g

The compounds NH(SiHMeCl)₂ I and NMe(SiHMeCl)₂ 2 have been prepared by treating SiHMeCl₂ with CaCl₂·8NH₃ and NH₂Me respectively. Each was characterised by elemental analysis, mass spectrometry, NMR and IR spectroscopy. Dipole moments were also measured. The NMR spectra indicate that both compounds form 1:1 mixtures of the rac and meso diastereomers, their abundances corresponding to a statistically controlled synthetic pathway. The NMR and mass spectra also show that 1,5-dichloro-1,2,3,4,5-pentamethyltrisilazane, SiHMe[NMe(SiHMeCl)]₂ 3, which is formed as a side product in the synthesis of 2, also consists of two diastereomers. Variable-temperature ¹H NMR spectra of NH(SiHMeCl)₂ indicate participation of H(N) in hydrogen bonding. The compound is decomposed by heat and reacts with pyridine to form NH₄Cl, SiHMeCl₂ and polysilazanes, whereas NMe(SiHMeCl)₂ shows only slight decomposition up to 80°C and does not react with pyridine. Infrared spectra in the v(SiH) region are interpreted in terms of the results of ab initio calculations of frequency, intensity and conformer abundance. The two bands near 2200 cm^-1 in the spectrum of 2 have their origin in two effects: different orientations of the Si-H bonds relative to the Si-N-Si plane in several conformers and an unprecedented strong dipole-dipole coupling between the two Si-H bond stretching motions in situations where the bonds are roughly parallel. The absence of such an observed splitting for 1 is likely to be due in part to signal averaging during a free internal rotation. Significant couplings are also calculated to occur between Si-Cl bond-stretching motions, whose source must be different from that for the Si-H bond stretches.

Full text of DOI:10.1039/a708220g

View Article Online / Journal Homepage / Table of Contents for this issue
Bis(chloromethylsilyl)amine and bis(chloromethylsilyl)methylamine;
preparation, reactivity and spectroscopic studies of their
stereoisomers and conformers‡
Holger Fleischer,*,†,a Donald C. McKean,a Colin R. Pulhama and Michael Bühlb
a Department of Chemistry, University of Edinburgh, West Mains Road, Edinburgh, UK EH9 3JJ
b Institut für Organische Chemie der Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich,
Switzerland
The compounds NH(SiHMeCl)2 1 and NMe(SiHMeCl)2 2 have been prepared by treating SiHMeCl2 with
CaCl2ؒ8NH3 and NH2Me respectively. Each was characterised by elemental analysis, mass spectrometry, NMR
and IR spectroscopy. Dipole moments were also measured. The NMR spectra indicate that both compounds form
1:1 mixtures of the rac and meso diastereomers, their abundances corresponding to a statistically controlled
synthetic pathway. The NMR and mass spectra also show that 1,5-dichloro-1,2,3,4,5-pentamethyltrisilazane,
SiHMe[NMe(SiHMeCl)]2 3, which is formed as a side product in the synthesis of 2, also consists of two
diastereomers. Variable-temperature 1H NMR spectra of NH(SiHMeCl)2 indicate participation of H(N) in
hydrogen bonding. The compound is decomposed by heat and reacts with pyridine to form NH4Cl, SiHMeCl2 and
polysilazanes, whereas NMe(SiHMeCl)2 shows only slight decomposition up to 80 ЊC and does not react with
pyridine. Infrared spectra in the ν(SiH) region are interpreted in terms of the results of ab initio calculations of
frequency, intensity and conformer abundance. The two bands near 2200 cmϪ1 in the spectrum of 2 have their
origin in two effects: different orientations of the Si᎐H bonds relative to the Si᎐N᎐Si plane in several conformers
and an unprecedented strong dipole–dipole coupling between the two Si᎐H bond stretching motions in situations
where the bonds are roughly parallel. The absence of such an observed splitting for 1 is likely to be due in part to
signal averaging during a free internal rotation. Significant couplings are also calculated to occur between Si᎐Cl
bond-stretching motions, whose source must be different from that for the Si᎐H bond stretches.
Substituted disilylamines have been the subject of a number
of recent chemical and structural investigations.1–4 To date,
the species studied for structural purposes have involved
symmetrically substituted silyl groups, viz. NH(SiHMe2)2,1
R
R
R
Cl
Cl
Si 3
H
Cl
Me
Cl
Si
Me
Cl
Si 3
H
Me
N2
Me Me
N2
Cl
N2
Me
3
1
Si
1
Si
1
Si
3
NMe(SiHMe2)2,2 NMe(SiH2Me)2,2 NH(SiHCl2)2 and NMe-
H
H
H
H
(SiHCl2)2,3 where the interest has been mainly in the con-
formations adopted following the substitution of hydrogen by
chlorine or methyl groups.
1R, 3S
1R, 3R
1S, 3S
Me
N2
Me
N4
Me
N2
Me
N4
Asymmetrically substituted silylamines have the additional
interest as being potential precursors for the synthesis of optic-
ally active polysilylazanes, inorganic equivalents of peptides
and proteins, and for SiN-based ceramics bearing chiral and
functionalised silicon surface groups. The present investigation
of two such disilylamines, 1 and 2, was designed to discover
which of the possible stereoisomers of the compounds were
present and to study their chemical, spectroscopic and struc-
tural properties. As shown in Scheme 1, both compounds can
give rise to two diastereoisomeric forms, the optically inactive
isomer with the 1S,3R (= 1R,3S) configuration, denoted meso-1
and meso-2, and a racemic mixture of the two optically active
forms with 1S,3S and 1R,3R configurations, denoted rac-1 and
rac-2. While this work was in progress, the synthesis of 1 and 2
was reported, together with some infrared and NMR data.5
In the present paper we describe new low-temperature
methods of preparation of the two compounds, explore further
their chemical properties and present spectroscopic and dipole
moment evidence bearing on the nature of the conformers pres-
ent. In the accompanying paper,4 gas-phase electron diffraction
and ab initio studies are reported. These include geometry and
Cl
Cl
Me
Me
Cl
Me
Me
Cl
1
Si
1
Si
5Si
5Si
H
S3i
S3i
H
H
H
H
Me
H
Me
1R, 5S-3
1S, 5R-3
Scheme 1 Structural formulae of bis(chloromethylsilyl)amine 1 (R =
H), bis(chloromethylsilyl)methylamine 2 (R = Me) and 1,5-dichloro-
1,2,3,4,5-pentamethyltrisilazane 3
energy calculations at the MP2 level and SCF calculations of
force constants, vibration frequency and infrared intensity.
Results from the latter SCF calculations are used here in the
interpretation of the infrared spectra, which focuses mainly on
the bands appearing in the Si᎐H stretching region.
Previous infrared studies6–8 of the Si᎐H stretching regions
in silylamines and ethers have shown that isolated SiH stretch-
ing frequencies, νis(SiH), measured for preference in partially
deuteriated species, are sensitive to the orientation of the Si᎐H
bond with respect to the skeletal plane of the molecule and this
information may be used to deduce the conformation(s) present
in a sample.7 They also enable simple rules for the lowering of
νis(SiH) by methyl or its raising by halogen substitution to be
deduced, which are relevant when conformational analysis is
attempted.
† Present address: Institut für Anorganische Chemie und Analytische
Chemie, Johannes Gutenberg Universität, Mainz, Johann Joachim
Becher Weg 24, D-55099 Mainz, Germany.
‡ Non-SI units employed: Torr ≈ 133 Pa, D ≈ 3.33 × 15Ϫ30 C m.
Although spectra of partially deuteriated species were not
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592
585

View Article Online
Table 1 Proton, 13C and 29Si NMR spectral data for compounds 1, 2
and 3. Two entries denote values for different diastereomers
1
2
3
δ
H3CSi
H3CN
HN
0.60
—
1.90
5.12, 5.14
2.28, 2.34
—
0.61
2.67, 2.68
—
0.27,a 0.50,b 0.51b
2.51
—
HSi
5.12
4.51,a 5.02b
Ϫ2.76,a 1.06,b 1.12b
29.18, 29.32
—
H3CSi
H3CN
Si
1.00, 1.04
29.3, 29.6
Ϫ0.85
Ϫ5.33
1J(1H᎐29Si)/Hz
1J(1H᎐13CN)/Hz
1J(1H᎐13CSi)/Hz
3J(1HC᎐1HSi)/Hz
3J(1HN᎐1HSi)/Hz
253
—
122.5
2.44
1.60
254
137
123.0
2.63
—
227,a 256b
140
—
2.79,a 2.18,b 2.52b
—
a Signals originate from central (Si3) SiHMe group. b Signals originate
from terminal (Si1, Si5) SiHMe groups.
1
Table 2 Proton NMR shifts and J(1H᎐29Si) coupling constants for
some hydridosilanes and silylamines
Compound
1J(1H᎐29Si)/Hz
δ(HSi)
Ref.
SiHF3
SiHCl3
SiHBr3
NH(SiHCl2)2
NMe(SiHCl2)2
SiHI3
SiHMeCl2
SiHMe2Cl
SiHMe3
381
364
360
336
327
325
279
—
4.51
6.12
6.44
5.76
5.71
4.49
5.58
4.87
3.85
10
11
11
9
9
11
12
12
12
182
obtained in this work, a detailed study of the vibrational
properties of the molecules concerned has been made using the
aforementioned ab initio calculations, whereby valence force
constants for the individual bonds and their stretch–stretch
interactions have been obtained as well as frequencies for the
partially deuteriated species. These reveal unexpected couplings
between the bond stretching motions, a factor which must be
considered if conclusions about the conformations present are
to be drawn.
Results and Discussion
NMR studies
1
Fig. 1 Variable-temperature H NMR spectra of (a) compound 1 in
CDCl3, (b) 2 in [2H8]toluene
The data obtained are listed in Table 1. The resonances can be
satisfactorily assigned by comparison with the spectra of the
related molecules NH(SiHCl2)2,9 NMe(SiHCl2)2,9 SiHX3 (X =
F, Cl, Br, I or Me),10–12 SiHMe2Cl and SiHMeCl2,12 data for
NMR spectra (Fig. 1) showed no coalescence of peaks and no
variation in intensity ratios. [The most striking effect of chan-
ging temperature seen in Fig. 1 is that on δH(N) for compound
1. The high-field shift with rise in temperature indicates partici-
pation of the N᎐H bond in a hydrogen bridge.9] Furthermore,
internal rotation about the Si᎐N bonds is usually sufficiently
rapid to cause averaging of the NMR signals of different
conformers.
1
which are given in Table 2.§ Values of δH(Si) and J(1H᎐29Si)
are in agreement with the trends discerned for the compounds
in Table 2.
The origin of some of the signal splittings in the spectra of
compounds 1 and 2 does not lie in spin–spin couplings but in
distinct differences in chemical shift. This was shown by record-
ing spectra at different field strengths. Variable-temperature 1H
All these considerations point to the presence in solutions of
compounds 1 and 2 of chemically similar, but non-equivalent
species in equal abundance, i.e. equimolar or 1:1 mixtures of
rac and meso diastereomers. This equality of abundance implies
that the formation of these stereoisomers is statistically con-
trolled, i.e. the chirality of the first Si atom attached to the
nitrogen atom does not lead to a preferred chirality for the
second Si. We note also that the rac : meso ratio for 2 did not
depend on the solvent chosen for the synthesis, n-pentane
or tetrahydrofuran (thf) (see Experimental and Theoretical
Procedures).
§ The magnitude of 1J(1H᎐29Si) is a good indicator for the withdrawal of
electron density by the three other ligands in monohydridosilanes. Table
2 shows that high values correlate with very electronegative substitu-
ents. The chemical shift of the H atom directly bound to the Si atom,
δH(Si), does not reflect the electronegativity in every case, as can be
seen by comparison of SiHF3 and SiHI3. However, if only Cl, NR2 and
Me are considered as ligands, δH(Si) decreases with decreasing
electron-withdrawing force. The values of δH(Si) and 1J(1H᎐29Si) of 1–3
(see Table 1) fit in very well in the series, 3 clearly showing the different
δH(Si) values for SiHMe[NMe(SiHMeCl)]2 (4.51) and SiHMe-
[NMe(SiHMeCl)]2 (5.02).
586
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592

View Article Online
showed just two sets of non-equal intensities, we conclude that
3 consists of a non-equimolar mixture of the 1R,5S and 1S,5R
Table 3 Infrared bands (cmϪ1) observed for compounds 1 and 2
diastereomers only, which are the ones shown in Scheme 1.
1
2
Infrared spectra
Gas
This work
Liquid
Ref. 5
Gas
This work
Liquid
Ref. 5
Assignment
Table 3 lists the frequencies observed in the gas phase in this
work, and in the liquid from the earlier study.5 Bearing in mind
that only the stronger bands seen in the latter’s spectra were
reported, the agreement between the gas- and liquid-phase
spectra is good. Inspection of the liquid-phase spectra provided
by the authors showed several differences of detail. A down-
wards shift of 37 cmϪ1 in ν(NH) from the gas to the liquid
together with a marked increase in relative intensity indicate a
degree of hydrogen bonding in the latter phase which corrobor-
ates the evidence from the NMR results in solution (see above).
A similar shift of 33 cmϪ1 was found in the case of NH(Si-
HCl2)2.9 The doublet with 38 cmϪ1 spacing observed in the gas-
phase ν(SiH) region of compound 2 is replaced in the liquid by
a single broad, slightly asymmetric band with a half-width of
about 55 cmϪ1. (The corresponding band for 1 in the liquid is
also broad, but a peak absorption of about 100% prevented a
measurement of half-width.) In addition our spectrum of 1
contained a small amount of SiHMeCl2, identified from the
appearance in varying amounts of the tip of the very intense,
narrow Q branch at 2218 cmϪ1 of its ν(SiH) band, as seen in
Fig. 2. A minute trace of this impurity occurred also in 2, where
a very weak shoulder at 2218 cmϪ1 can just be discerned (Fig.
2). The presence of SiHMeCl2 is attributed to thermolysis in the
cell, rather than to hydrolysis, since no HCl was visible in the
spectrum. The contour of the 2218 cmϪ1 band limits inter-
ference to the immediate neighbourhood of its peak, leaving the
bands due to 1 and 2 unaffected.
3397w
2977mw
3360s
2970w
ν(NH)
2975w
2950w
2914w
2851w
2834w
2218(sh)
2208.1m
2180.5m
a
2970w
2950w
2900w
νasym[CH3(Si)]
νasym[CH3(N)]
νsym[CH3(Si,N)]
2 × δasym[CH3(Si)]
2 × δsym[CH3(N)]?
SiHMeCl2
ν(SiH)
2918w
2851w
2820w
2200s
2218w(sh)
2192.4s
2200s
1410m
1260s
1190s
ν(SiH)
a
1410vw
δasym[CH3(Si)],
δasym[CH3(N)]
δsym[CH3(Si)]
ρ[CH3(||)]
δ[NH(||)]
?
ν(CN)
ν(CN)
νasym(NSi2)
δ(SiH)
δ(SiH)
1266ms
1266m
1202mw
1260s
1200m
1188s
1102w(br)
1090m
1079m
946vs
885s
1090m
1070m
950vs
950vs
881s(sh)
858vs
805mb
754ms
633wb
566w
950s
840vs
750s
852s
850vs
750m
ρ(CH3)
ρ(CH3)
?
?
760m
517s
500vs
511s
500s
ν(SiCl)
a Region obscured by uncompensated atmospheric water lines and
window bands. b Corresponding features are seen in the liquid-phase
spectra.
The assignments given in Table 3 are based on the unscaled
frequencies from the Gaussian SCF/6-31G* calculations.4 Of
the modes which vary little with conformation, the ab initio
results indicate that there is little difference in the ν(CH) values
for the NMe and SiMe groups, but large ones in the δasym(CH3)
and δsym(CH3) modes, the latter markedly larger when the
methyl group is attached to the nitrogen atom. This is as
expected. The presence of overtones of both δasym(CH3) and
δsym(CH3) in the middle of the ν(CH3) region for the NMe
group therefore makes assessment of the very strong Fermi
resonances present here difficult.
Modes which are found to vary significantly with con-
formation include ν(NH), ν(SiH), δ(NH), δ(SiH), ρ(CH3),
νsym(NSi2), ν(SiCl) and skeletal bends, all of which are poten-
tially at least a source of information about the conformers
present. A minor variation in νasym(NSi2) due to coupling with
the highest δ(SiH) motion is likely to be an illusion arising from
the absence of the differing scale factors that will be needed if
the observed frequencies involving these two types of motion
are to be fitted exactly. In fact, a proper analysis of all except
one of these variations must await the calculation of scaled ab
initio force fields, for which purpose it will be necessary to trans-
fer scale factors from the parent molecules NH(SiH3)2 and
NMe(SiH3)2, currently under study. However information from
the ν(SiH) region needs no such assistance and we may proceed
directly to see what this reveals.
Fig. 2 The ν(Si᎐H) bands from the IR spectrum of compounds 1
(dotted) and 2
The 1H NMR spectrum of compound 3 also displays a
significant splitting of the resonances. Again this has to arise
from the presence of chemically similar but non-equivalent
species. However, in this case the resonances are of unequal
intensity, which requires closer scrutiny. There are four possible
stereoisomers for 3: 1R,5R-3, 1S,5S-3, 1R,5S-3 and 1S,5R-3.
The former two form a pair of enantiomers, thus 3 may consist
of three chemically different diastereomers. The last two
diastereomers are not identical, due to the presence of the
pseudo-asymmetrical atom Si3. The two NMe(SiHMeCl)
groups of 1R,5R-3 are diastereotopic to each other and should
give rise to two sets of signals. Each of these groups is enantio-
topic to one in 1S,5S-3, so that a racemic mixture of both
should result in two sets of signals having equal intensities. For
1R,5S-3, however, the two NMe(SiHMeCl) groups are enanti-
otopic to each other but diastereotopic to those of 1S,5R-3 and
to both sets of 1R,5R-3 and 1S,5S-3. As 1R,5S-3 and 1S,5R-3
are diastereomers, with differing sets of signals, the latter need
Analysis of the í(SiH) region. It is helpful if we place the data
for the current compounds 1 and 2 in the context of previous
work on disilylamines. Table 4 includes all such ν(SiH) data
so far available, while Fig. 2 shows the expanded ν(SiH) regions
of 1 and 2. Our eventual objective is to explain why all the
disilylamines bearing an N᎐H group show a single, broad, often
asymmetric band whereas all the disilylamines with an N᎐Me
group exhibit two peaks.
The starting point for our analysis of these spectra are the
1
not have equal intensities. Since the H NMR spectrum of 3
predictions from the ab initio calculations4 that for both com-
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592
587

View Article Online
pounds six conformers are present at room temperature, as
shown in Table 5. For each conformer we give the two dihedral
angles τ(HSiNX) (X = H for 1 and C for 2), the Si᎐H bond
lengths r(Si᎐H), the ν(SiH) frequencies ν1 and ν2 for the D0 (not
deuteriated) and D1 (one Si᎐D bond) species (the latter being
the νis values), the calculated infrared intensities for the D0
species bands [before (A) and after (Ax) weighting with the
abundance, x, calculated at the MP2/6-311ϩG* level for each
conformer] and finally the interaction force constant fЈ. (Dipole
derivatives with respect to bond stretching and related quan-
tities will be given in a later paper.15) The usual good correlation
between νis(SiH) and r(Si᎐H) obeys equation (1) with an R
factor of 0.9959.
νis(SiH)/cmϪ1 = 13 103(227) Ϫ 72.8(16) r(Si᎐H)/pm (1)
Some striking results emerge. Quite large splittings of the two
ν(SiH) frequencies occur in nearly all the conformers up to a
maximum of 38.5 cmϪ1. Those in compound 1, however, tend to
be about one-half of those in 2. The splittings arise from two
sources. The largest occur where the D0 and D1 frequencies are
identical, indicating zero coupling between the two Si᎐H
stretching motions, with a zero value of the interaction con-
stant fЈ. In each case the two Si᎐H bonds concerned have very
different orientations with respect to the skeletal plane, as indi-
cated by the torsion angles involved, e.g. 176.2 and Ϫ48.1Њ for
rac-2A [see Table 5(b)]. Similarly, their bond lengths are
different.
Table 4 Infrared bands (cmϪ1) in the ν(SiH) regions of disilylamines
in the gas phase
Molecule
Bands
Average Ref.
However, splittings as large as 19.8 cmϪ1 occur in situations
such as in meso-2B, where the Si᎐H bonds are identical with
respect to their orientations, lengths and force constants. The
coupling constant responsible, fЈ, can be as large as 2.42 N mϪ1
(rac-2B). This constant, linking as it does bonds which are
gamma related, is of comparable size with the similar constant
between bonds which are alpha related, in SiH3 or SiH2
groups,16 but is much larger than those for beta-related bonds,
either trans or gauche, measured in disilanes.16–20
NH(SiH3)2
2164.4 (br)as
2144.5, ca. 2184
2128.0, 2165.3
2128.3
2164.4
2164.3
2146.7
2128.3
2133.3
2247
13
8
7
7
7
9
14
NMe(SiHD2)2
NMe(SiH2Me)2
NH(SiHMe2)2
NMe(SiHMe2)2
NH(SiHCl2)2
NMe(SiHCl2)2
2113.4, 2153.2
2247 (br)
2232.5, 2237 (sh),a
2253.4, 2260 (sh)b
2243.0
NH(SiHMeCl)2
2192.4 (br)as, 2218sp (sh)c 2192.4
This work
This work
NMe(SiHMeCl)2 2180.5, 2208, 2218 (sh)d
2194.3
a Possibly due to SiH2Cl2. b Possibly due to SiHCl3. c Variable intensity
The corresponding fЈ in CH᎐X᎐CH systems is quite
negligible, as may be seen by the close similarity of the ν(CH)
and due to SiHMeCl2 impurity. d Probably due to SiHMeCl2 impurity.
Table 5 The SCF/6-31G* data for Si᎐H bonds
Conformer
rac
meso
A
B
C
A
B
C
(a) Compound 1
τ1(HSi1NH)/Њ
τ2(HSi2NH)/Њ
r1(Si᎐H)/pm
r2(Si᎐H)/pm
ν1(D0)/cmϪ1
ν2(D0)/cmϪ1
(ν1 Ϫ ν2)/cmϪ1
νis,1/cmϪ1
150.5
161.7
Ϫ17.9
Ϫ17.9
146.68
146.68
2427.61
2423.95
3.7
2425.77
2425.77
204.0
198.9
0.233
47.5
155.2
167.8
Ϫ10.6
51.7
146.72
146.76
2423.92
2417.11
6.8
2423.34
2417.69
212.6
201.4
0.072
15.3
Ϫ52.8
146.51
146.77
2435.98
2417.11
18.9
2435.92
2417.15
166.7
209.4
0.102
17.0
161.7
146.43
146.43
2448.34
2437.07
11.3
2442.74
2442.74
259.6
40.5
0.165
42.8
6.7
1.33
1.58
27.7
146.48
146.69
2439.85
2424.71
15.1
2439.85
2424.71
154.5
199.5
0.348
53.8
69.4
Ϫ167.8
146.46
146.46
2447.76
2433.91
13.9
(2440.8)a
(2440.8)a
300.0
2.9
νis,2/cmϪ1
A1/km molϪ1
A2/km molϪ1
xb
0.081
24.3
0.2
A1x/km molϪ1
A2x/km molϪ1
fЈ(full)/N mϪ1
fЈ(HLM)/N mϪ1
21.4
0.20
0.35
46.3
0.50
0.51
14.5
0.50
0.53
0.04
c
1.94
(b) Compound 2
τ1(HSi1NC)/Њ
τ2(HSi2NC)/Њ
r1(Si᎐H)/pm
r2(Si᎐H)/pm
ν1(D0)/cmϪ1
ν2(D0)/cmϪ1
(ν1 Ϫ ν2)/cmϪ1
νis,1/cmϪ1
176.2
Ϫ48.1
146.31
146.82
2451.88
2413.34
38.5
2451.87
2413.34
160.5
189.8
0.196
31.5
169.8
173.1
146.35
146.36
2456.06
2436.96
19.1
2447.11
2446.09
279.5
19.4
0.258
72.1
5.0
2.42
2.67
Ϫ30.5
Ϫ28.1
146.63
146.70
2426.76
2421.83
4.9
2425.71
2422.88
135.6
222.5
0.046
6.2
166.2
37.0
146.34
146.71
2449.95
2421.17
28.8
2449.93
2421.17
160.0
179.1
0.300
48.0
53.7
174.8
Ϫ175.0
146.38
146.38
2454.61
2434.80
19.8
Ϫ22.6
54.2
146.63
146.83
2426.66
2412.35
14.3
2426.32
2412.70
184.9
186.5
0.069
12.8
(2444.7)a
(2444.7)a
295.3
7.7
νis,2/cmϪ1
A1/km molϪ1
A2/km molϪ1
xb
0.131
38.7
A1x/km molϪ1
A2x/km molϪ1
fЈ(full)/N mϪ1
fЈ(HLM)/N mϪ1
37.2
0.04
0.00
10.2
0.56
0.56
1.0
12.9
0.57
0.61
<0.01
<0.01
c
1.98
a Average of ν1 and ν2. b The molar fractions x are those from MP2 calculations, Table 3 in ref. 4, for slightly different torsional angles. c Cartesian
force constants not available.
588
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592

View Article Online
region of the IR spectra of species such as NMe3 and
NMe(CD3)2 etc., or by examining scaled ab initio force
fields.21,22
Table 6 The SCF/6-31G* torsional frequencies (cmϪ1) for compounds
1 and 2
There are several clues as to the origin of this large value
of fЈ in SiH᎐Y᎐SiH systems. (1) It is associated with Si᎐H
bonds which are almost parallel, but is absent where the bonds
are not so, for example in rac-1C, where the two bonds are
identically disposed but point in very different directions. The
widely differing infrared intensity distributions of ν1 and ν2 for
the various conformers reflect these differences in the mutual
orientation of the Si᎐H bonds. (2) In a similar calculation for
the C2v conformer of NH(SiHCl2)2,15 where both SCF and MP2
treatments were available, the value of fЈ diminishes as the
infrared intensity diminishes, from the SCF to the MP2 calcul-
ation. These facts suggest a transition dipole–transition dipole
mechanism for the interaction which reflects the much greater
infrared intensity associated with the stretching of Si᎐H bonds
as compared to C᎐H ones. This will be examined quantitatively
in a later publication.15
Conformer
rac
meso
Compound
A
B
C
A
B
C
1
19.8
38.8
16.6
35.9
24.8
27.9
21.0
28.1
29.0
38.5
21.2
24.4
25.3
32.2
23.4
29.1
12.4
39.3
20.1
36.1
25.5
38.6
17.9
39.1
2
For our present purpose in trying to understand the differ-
ence between the IR spectra of compounds 1 and 2, we now
compare the predicted infrared spectra as column diagrams for
the two compounds in Fig. 3. There is a very clear-cut separ-
ation of the absorption for compound 2 into two distinct
regions separated by about 33 cmϪ1. By contrast, for 1, while
most of the absorption occurs near 2425 cmϪ1, a high propor-
tion occurs near 2440 and 2448 cmϪ1, enough to make a doublet
appearance possible, but one with a much smaller spacing than
for the NMe compound. Any broadening effect could readily
merge the whole into one broad band.
In comparing these with the observed spectra several factors
must be borne in mind. First, the frequencies will be overestim-
ated by the SCF/6-31G* calculation by about 9%. Our experi-
ence suggests that the primary effect of this on force constants
expressed in internal coordinates is on the diagonal constants,
not the interaction ones.17–19 Secondly, the ab initio approach
may not reproduce quantitatively the effects of the torsion
angle on the Si᎐H force constants. The effect of chlorine
substitution, for instance, is overestimated at this level of
treatment.17–19 Thirdly, if the coupling between the stretching
motions arises from a dipole–dipole mechanism, this will have
been overestimated in so far as ab initio intensities generally
exceed experimental ones for this type of motion by at least a
factor of two.23 In such a case, therefore, ‘observed’ coupling
constants are likely to be smaller than those predicted here.¶ This
last consideration leaves little doubt that the observed separ-
ation of about 30 cmϪ1 in the two ν(SiH) peaks of 2 arises
primarily from the conformational dependence of νis(SiH) in
rac-2A and meso-2A which together constitute 50% of the
sample, according to the MP2 estimates of abundance. If the
above is the correct interpretation, then the infrared spectrum
in the ν(SiH) region of the partially deuteriated isotopomer
of compound 2, NMe(SiHMeCl)(SiDMeCl), should be very
similar to that of the undeuteriated one, as is the case for the
corresponding species of NMe(SiH3)2.13
Fig. 3 Infrared spectra in the ν(Si᎐H) regions of compounds 1 and 2,
as predicted ab initio. Frequencies and intensities are from SCF/6-31G*
calculations; conformer abundancies were determined at the MP2/
6-311ϩG* level
unchanged, that is the chemical effect of N-methyl substitution
on the Si᎐H bond is negligible: average values are: 2, 2433.9; 1,
2435.9 cmϪ1; when weighted by the MP2 abundancies these
become 2437.4 and 2434.2 cmϪ1 respectively. A more reliable
indication comes from calculations for the structures of
NH(SiHCl2)2 and NMe(SiHCl2)2 where both Si᎐H bonds
are exactly in the skeletal plane and roughly parallel to each
other.3 The difference in the average νis(SiH) is only 1.7 cmϪ1
,
NMe(SiHCl2)2 > NH(SiHCl2)2.23
Since the frequency of the single band for ν(SiH) observed in
the IR spectrum of compound 1 lies close to the average of the
two seen in the spectrum of 2, it is likely that it represents an
average of frequencies which are continuously varying. The
interconversion of conformers would then be taking place on a
time-scale short compared with about 2.2 × 10Ϫ11 s. This, how-
ever, raises the question as to why the same effect does not
operate in 2. Table 6 compares the two lowest vibration fre-
quencies for the six conformers of 1 and 2, which may be taken
to be essentially torsional motions of the SiHMeCl groups
against the Si᎐N᎐Si skeleton. The values are all real, in contrast
to what is found for NH(SiHCl2)2,3 while the frequencies for 1
are slightly larger than those for 2 except in the case of the
meso-B conformers. Taking into account also that many
molecules will be in torsionally excited states well above the
likely barrier to internal rotation of the SiHMeCl group, we
conclude that the time-scales for interconversion are likely to be
essentially the same in the two compounds. The ν(SiH) split-
tings must therefore be either just above or just below the
critical value at which signal averaging takes place.
The ν(SiH) region in the IR spectrum of compound 1 is
harder to understand. The two highest component bands calcul-
ated to lie near 2448 cmϪ1 are both intense and if the coupling
that has produced them has been overestimated their absorp-
tion would be expected at lower energy, coinciding with the
2440 cmϪ1 band. An overall doublet spacing of about 15 cmϪ1
would then result. It is interesting that the ab initio calculations
of frequency indicate that the average frequency is almost
¶ Calculations at both SCF and MP2 levels for NH(SiH3)2 and
NMe(SiH3)2 show that the MP2-based interaction constants are slightly
less than the SCF ones and indistinguishable from those observed
experimentally. By contrast, variations in νis(SiH) and r(Si᎐H) of the
kind found in this work are slightly larger at the MP2 level, the former
again in excellent agreement with experiment.13
If free rotation is occurring on the infrared time-scale for
compound 1, it is natural to extend the explanation to the other
disilylamines bearing an NH group (see Table 5), namely
NH(SiH3)2, NH(SiHMe2)2 and NH(SiHCl2)2, as has been sug-
gested for the first two of these.7 However this involves the
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592
589

View Article Online
Table 7 The SCF/6-31G* data for Si-Cl bonds in compounds 1 and 2
Conformer
rac
meso
A
B
C
A
B
C
Compound 1
τ3(ClSi1NH)/Њ
τ4(ClSi2NH)/Њ
r1(Si1᎐Cl)/pm
r2(Si2᎐Cl)/pm
f(Si1᎐Cl)/N mϪ1
f(Si2᎐Cl)/N mϪ1
fЈ/N mϪ1
Ϫ93.8
61.4
208.22
208.06
291.74
294.11
4.14
Ϫ82.9
Ϫ82.9
208.52
208.52
288.39
288.39
2.32
96.7
96.7
208.58
208.58
287.52
287.52
1.76
Ϫ89.5
Ϫ86.7
208.55
208.56
287.96
287.93
2.10
Ϫ76.0
76.0
208.06
208.06
*
*
*
104.9
Ϫ62.4
208.14
208.05
292.83
294.20
4.10
Compound 2
τ3(ClSi1NH)/Њ
τ4(ClSi2NH)/Њ
r1(Si1᎐Cl)/pm
r2(Si2᎐Cl)/pm
f(Si1᎐Cl)/N mϪ1
f(Si2᎐Cl)/N mϪ1
fЈ/N mϪ1
Ϫ67.6
65.6
208.42
208.50
289.37
288.47
4.19
Ϫ74.4
Ϫ71.3
208.69
208.92
286.18
283.57
2.61
83.5
85.5
209.05
208.78
281.81
285.09
2.16
Ϫ78.2
Ϫ76.8
208.67
208.98
286.41
282.80
2.42
Ϫ68.8
68.6
208.33
208.33
*
*
*
92.3
Ϫ59.4
208.57
208.44
287.47
289.37
4.16
* Cartesian force constants not available.
difficulty that replacement of a methyl group by a chlorine
atom raises ν(SiH) by 64 cmϪ1, on passing from NH(SiHMe2)2
to NH(SiHMeCl)2, while the further replacement between
NH(SiHMeCl)2 and NH(SiHCl2)2 causes an additional rise of
55 cmϪ1. These shifts are markedly larger than those from
SiH2Me2 to SiH2MeCl (43.8 cmϪ1) and from SiHMe2Cl to
SiHMeCl2 (43.4 cmϪ1).7 By contrast, the low frequency shift
from NH(SiH3)2 to NH(SiHMe2)2 (36 cmϪ1) is normal for a pair
of methyl groups.7
Two explanations of the enhanced shifts arising from
chlorine substitution suggest themselves. (1) The internal
rotation is not completely free but limited to a range of torsion
angles favouring the preferred staggered arrangement of Me or
Cl groups as seen along the Si ؒ ؒ ؒ Si axis, a common feature of
the structure of compounds of this type.4 However, without a
more detailed study of the dependence of ν(SiH) on the torsion
angle, it is not obvious how this will affect the distribution of
ν(SiH) values to be averaged. This is an area we are currently
exploring.13 (2) There may be a chemical influence on the Si᎐H
bond not only from α substituents on the same silicon atom
but also from those on the remote silicon. Evidence from SCF
calculations on NH(SiHCl2)2 and NMe(SiHCl2)2 supports this
latter view.15
A final question is: why are the splittings in the ν(SiH) fre-
quencies in the conformers of compound 1 much smaller than
those in 2? The change in the dipole–dipole coupling prominent
in the rac-B conformers can be wholly accounted for by the
change in the H᎐Si᎐N᎐H(C) torsion angles from 1 to 2 and the
resultant effects on the H ؒ ؒ ؒ H distance and mutual orientation
of the two Si᎐H bonds. The dipole derivatives in the two com-
pounds are found to be identical.15 About half of the difference
in the splittings in νis(SiH) can also be attributed to this cause.
Thus a calculation of ν1 and ν2 for a non-equilibrium rac-1
structure in which the H᎐Si᎐N᎐H angles are constrained to
176.2 and Ϫ48.1Њ (values for rac-2A), yields the values 2446.4
and 2418.9 cmϪ1, a separation of 27.5 cmϪ1, somewhat larger
than the 18.9 cmϪ1 found for the equilibrium structure rac-1A,
but still smaller than the 38.5 cmϪ1 for rac-2A. Further explor-
ation of these and related compounds will be needed to eluci-
date the sources of these small structural changes.
stretching frequencies for a compound is necessarily an indic-
ation of the presence of non-equivalent Si᎐H bonds.
Ab initio results for Si᎐Cl bonds. While there are no group
frequencies to be expected from Si᎐Cl stretching motions, such
as occur for Si᎐H ones, we may use the ab initio force fields in
internal coordinates to examine Si᎐Cl stretching force con-
stants for signs of variation with respect to orientation in the
molecule. The appropriate data are shown in Table 7. The excel-
lent correlation between bond length and force constant obeys
equation (2) with an R factor of Ϫ0.9994.
f(Si᎐Cl)/N mϪ1 = 2831(23) Ϫ 112.19(11)R(Si᎐Cl)/pm(2)
In most cases, the stronger Si᎐Cl bonds are associated with
smaller values of |τ3| or |τ4|. There is a general tendency for the
Si᎐Cl bonds to be slightly stronger in compound 1 than in 2,
but whether this arises from small differences in the torsion
angles or from a direct chemical effect of methyl substitution at
the nitrogen atom is not clear. Perhaps more surprising is the
prevalence of sizeable stretch–stretch interaction constants,
which are almost identical for corresponding conformers of 1
and 2, despite differences in the torsion angles involved. (By
contrast, there is no interaction between the two SiC stretches
in any conformation.) The origin of these interactions would
appear to be quite different from the dipole–dipole one for the
ν(SiH) motions postulated above. A plausible source is the
n(N)–σ*(Si᎐Cl) interaction which is thought to influence the
orientations found in the stable conformers.4 If such is the case,
the coupling term fЈ arises from an effect transmitted through a
delocalised orbital centred on the nitrogen atom, rather than
being a ‘through-space’ interaction as exemplified by the
ν(SiH)–ν(SiH) coupling.
Thermolysis
Kept at 80 ЊC for 24 h, compounds 1 and 2 show a rather
different behaviour: 1 decomposes to NH4Cl, HSiMeCl2 and
1
oligosilazanes as could be shown by H NMR and mass spec-
trometry, whereas 2 stayed nearly unchanged and only small
amounts of oligosilazanes could be detected by mass
spectrometry.
One important new conclusion regarding the interpretation
of the infrared spectra of this type of compound can now be
drawn. The theoretical prediction of long-range couplings
between Si᎐H stretching motions means that it can no longer be
assumed, as in earlier work,7 that observation of several ν(SiH)
Reaction with pyridine
The reaction of compounds 1 and 2 with pyridine should reveal
590
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592

View Article Online
the Brønsted acidity of 1 and the Lewis acidity of both com-
pounds. This type of reaction has been used to study the Lewis
and Brønsted acidities of NH(SiCl3)2 24 and NH(SiHCl2)2.9
No change of temperature was observed, either for com-
pound 1 or 2, when pyridine was added to a solution of the
disilylamine in chloroform. Any reaction has thus to be slow
and/or only slightly exothermic. In the case of 1, all its charac-
NH3) was added over a period of 4 h to a stirred solution of
SiHMeCl2 (40 cm3, 386 mmol) in n-pentane (100 cm3) main-
tained at Ϫ45 ЊC. The resulting suspension was subsequently
stirred for 40 h at room temperature. The compounds CaCl2ؒ
4NH3 and NH4Cl were filtered off and washed twice with n-
pentane (20 cm3). After removing unchanged SiHMeCl2 and n-
pentane in vacuo at Ϫ45 ЊC, the residue was warmed to 0 ЊC and
fractionally distilled. The compound NH(SiHMeCl)2 was con-
densed as a colourless liquid at Ϫ45 ЊC (7.0 g, 83%). On cooling
to Ϫ196 ЊC it initially forms a glass which subsequently changes
audibly into polycrystalline material. Differential thermal
analysis did not reveal any melting process between Ϫ150 ЊC
and room temperature, but instead showed conversion of a
crystalline phase into a glass at Ϫ130 ЊC (Found: C, 12.7; H, 4.4;
Cl, 39.6; N, 5.8; Si, 29.3. Calc. for C2H9Cl2NSi2: C, 13.8; H, 5.2;
Cl, 40.7; N, 8.0; Si, 32.3%.).** EI mass spectrum: m/z 173
(95, M), 158 (100, M Ϫ Me), 143 (5, M Ϫ 2Me), 138 (50,
M Ϫ Cl), 123 (9, M Ϫ Cl Ϫ Me), 102 (8, M Ϫ Cl Ϫ2Me)
and 79 (28%, SiHMeCl). µ (0.001–0.01 mol dmϪ3 solutions in
1
teristic peaks in the H NMR spectrum disappeared, the only
clearly identifiable product being SiHMeCl2. Peaks in the H(Si)
chemical shift range suggested that a variety of compounds
containing SiHMeCl groups were formed. A very broad
signal at δ 17.8 indicated the formation of either an N ؒ ؒ ؒ H᎐N
bridge or else of NH4ϩ. As the peaks due to pyridine resembled
strongly those of neat pyridine in CHCl3 the presence of NH4Cl
was inferred. Furthermore, we conclude that neither 1 nor any
of its decomposition products forms a complex with pyridine,
the latter acting as a decomposition catalyst towards 1. In
contrast, the 1H NMR spectrum of a mixture of 2 and pyridine
was essentially the sum of the spectra of the pure compounds,
indicating the absence of any reaction.
n-hexane) 2.27(6)†† D. d20 1.10 g cmϪ3
.
Compound 2 was prepared according to equation (4).
Experimental and Theoretical Procedures
3 NH2Me ϩ 2 SiHMeCl2 → NMe(SiHMeCl)2 ϩ
Materials and instrumentation
2 NH3MeCl (4)
All manipulations were performed either in a dry-box under an
atmosphere of argon or nitrogen (dried with activated 4 Å
molecular sieves and Sicapent) or using a vacuum line. Sol-
vents were distilled from CaH2 and stored over activated
molecular sieves; gaseous NH3 and NH2Me were dried by pass-
ing the vapours over 3 Å molecular sieves; SiHMeCl2 (Aldrich,
97%) was used without further purification and CaCl2ؒ8NH3
was prepared according to the method of Hüttig.25 Elemental
analyses (C,H,N) were performed with a LECO-1000 combus-
tion analyser; Cl was determined potentiometrically with
AgNO3. Differential thermal analysis was performed using a
Mettler TA 3000 instrument. Dipole moments of the liquids
were measured using a WTW Dipolemeter DM 01. Gas-phase
IR spectra were recorded in the range 4000–400 cmϪ1 on a
Perkin-Elmer Paragon 1000 spectrometer using a 10 cm gas cell
at a resolution of 1 cmϪ1, filled with about 5 Torr (saturated)
vapour pressure. The 1H, 13C and 29Si NMR spectra were
recorded on Bruker AMX 200 and AMX 400 instruments.
Methylamine (30 cm3, 0.70 mol) was condensed over a
period of 8 h into a stirred solution of SiHMeCl2 (365 g, 3.17
mol) in n-pentane (250 cm3) maintained at Ϫ30 ЊC, and the
mixture was stirred for 12 h at room temperature. The
compound NH3MeCl was filtered off and washed twice with
n-pentane (50 cm3); SiHMeCl2 and n-pentane were distilled
from the combined liquid phases at normal pressure and the
residue was fractionally distilled in vacuo. Compound 2 was
obtained as the main product (40 ЊC, 8 Torr; 17.0 g, 39%).
Alternatively, a 2.0 mol dmϪ3 solution of NH2Me in thf (50
cm3, 0.10 mol) was added dropwise over a period of 3 h to a
stirred solution of SiHMeCl2 (25 cm3, 0.24 mol) in thf main-
tained at Ϫ30 ЊC and the mixture stirred for 12 h. The com-
pound NH3MeCl was filtered off and washed twice with thf (50
cm3); SiHMeCl2 and thf were removed in vacuo (0 ЊC) and the
residue, kept at room temperature, fractionally distilled in
vacuo. Compound 2 condensed as a colourless liquid at Ϫ30 ЊC
(2.67 g, 43%). On cooling to Ϫ150 ЊC a glass is formed which
changed to a polycrystalline material at Ϫ105 ЊC which in turn
melted sharply at Ϫ56 ЊC (Found: C, 17.8; H, 5.29; Cl, 36.4; N,
6.1; Si, 29.4. Calc. for C3H11Cl2NSi2: C, 19.1; H, 5.90; Cl, 37.7;
N, 7.4; Si, 29.8%).** EI mass spectrum: m/z 187 (89, M), 172
(100, M Ϫ Me), 157 (13, M Ϫ 2Me), 152 (61, M Ϫ Cl), 142 (17,
M Ϫ 3Me), 137 (13, M Ϫ Cl Ϫ Me), 122 (31, M Ϫ Cl Ϫ 2Me),
117 (10, M Ϫ 2Cl), 108 (68, M Ϫ Cl Ϫ 3Me) and 79 (65%,
SiHMeCl). µ (0.001–0.01 mol dmϪ3 solutions in n-hexane)
Preparations
As both compounds 1 and 2 were reported to be thermally
unstable even at ambient temperature,5 we opted for a low-
temperature synthesis which avoided the several hours of reflux
at 55 ЊC associated with the previous method.5 Our method also
replaces the previously used starting materials of NH(SiMe3)2
and NMe(SiMe3)2 with the cheaper CaCl2ؒ8NH3 and NH2Me.||
(An unexpected bonus was the isolation of compound 3 from
the preparation of 2.) Compound 1 was therefore synthesized
according to equation (3), a variant of which has recently been
2.59(6)†† D. d20 1.13 g cmϪ3
.
Compound 3 was isolated as a side product in the synthesis
of 2 by the first route above. The product was obtained at 65 ЊC,
0.1 Torr (3.9 g, 15%). EI mass spectrum: m/z 260 (4.5, M) and
225 (100%, M Ϫ Cl).
All three compounds are very sensitive towards moisture, a
characteristic feature of those with a Si᎐Cl moiety. The thermal
decomposition of 1 and 2 was tested as follows: small amounts
of 1 and 2 were condensed into Pyrex tubes and sealed before
heating to 80 ЊC for 24 h. The tubes were opened under an inert
atmosphere and samples submitted for mass and NMR spectral
examination. Compound 1: EI mass spectrum: m/z 411 {66,
[᎐NSiHMe(SiHMeCl)᎐]3}, 274 {35, [᎐NSiHMe(SiHMeCl)᎐]2},
173 (81, 1) and 114 (38%, SiHMeCl2); 1H NMR (CDCl3, 25 ЊC,
3 CaCl2ؒ8NH3 ϩ 8 SiHMeCl2 → 4 NH(SiHMeCl)2 ϩ
8 NH4Cl ϩ 3 CaCl2ؒ4NH3 (3)
used to prepare NH(SiHCl2)2.9 It was established earlier that
only four of the eight NH3 molecules in CaCl2ؒ8NH3 are
reactive towards SiHCl3 and elemental analysis of the solid
residue confirmed that the same was true for the reaction with
SiHMeCl2.
The compound CaCl2ؒ8NH3 (8.95 g, 41.4 mmol, 290 mmol
|| A further advantage of our approach is that the excess of SiHMeCl2 in
the reaction products from the preparation of compound 2 may be
readily recovered and used for a further reaction. When NMe(SiMe3)2 is
the starting material the reaction yields an inseparable mixture of
SiHMeCl2 and SiMe3Cl.
** The less than perfect agreement between experimental and calculated
values reflects the extreme moisture sensitivity of the compounds and
the concomitant difficulties in manipulating them.
†† Value in parentheses denotes the estimated standard deviation.
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592
591

View Article Online
3
3
SiMe4) δ 5.58 [q, J(H᎐H) 2.2, HSiMeCl2], 5.13 [q, J(H᎐H)
2.4, NH(SiHMeCl)2], 4.99 [m, H(Si), oligomer], 1.89 [br s,
NH(SiHMeCl)2], 0.87 [d, 3J(H᎐H) 2.2, HSiMeCl2], 0.60 [d,
3J(H᎐H) 2.4 Hz, NH(SiHMeCl)2] and 0.45 [m, Me(Si), oligo-
mer]. Compound 2: EI mass spectrum: m/z 333 (44, tetra-
silazane), 260 (50, 3), 219 [56, (᎐SiHMeNMe᎐)3], 187 (59, 2),
146 [89, (᎐SiHMeNMe᎐)2] and 114 (24%, SiHMeCl2); in the 1H
NMR spectrum, no peaks other than those of 2 were observed.
For the reaction with pyridine the disilylamine (ca. 5 mmol)
was dissolved in chloroform (25 cm3). The solution was stirred
at room temperature while pyridine (2.0 cm3, 24.8 mmol) was
added in portions of 0.2 cm3 and the temperature was moni-
tored during each addition. For an NMR study of the reaction,
a few drops of the disilylamine were dissolved in CDCl3 (1.0
cm3), pyridine (0.1 cm3) was added and the sample tube sealed
off.
and J. Dunoguès of the University of Bordeaux 1 for providing
us with their liquid-phase infrared spectra of compounds 1
and 2.
References
1 G. Gundersen and D. W. H. Rankin, Acta Chem. Scand., Ser. A,
1984, 38, 647.
2 G. Gundersen, D. W. H. Rankin and H. E. Robertson, J. Chem. Soc.,
Dalton Trans., 1985, 191.
3 H. Fleischer, D. Hnyk, D. W. H. Rankin, H. E. Robertson, M. Bühl
and W. Thiel, Chem. Ber., 1995, 128, 807.
4 H. Fleischer, P. T. Brain, D. W. H. Rankin, H. E. Robertson,
M. Bühl and W. Thiel, following paper and refs. therein.
5 E. Bacqué, J.-P. Pillot, M. Birot, J. Dunoguès, G. Bourgeois and
M. Pétraud, J. Organomet. Chem., 1994, 481, 167.
6 D. C. McKean, J. Mol. Struct., 1984, 113, 251.
7 D. C. McKean and I. Torto, Spectrochim. Acta, Part A, 1993, 49,
1095.
8 D. C. McKean, I. Torto, J. E. Boggs and K. Fan, J. Mol. Struct.,
1992, 260, 27.
9 H. Fleischer, K. Hensen, D. Burgdorf, R. Flindt, U. Wannagat,
H. Bürger and G. Pawelke, Z. Anorg. Allg. Chem., 1995, 621, 239.
10 H. J. Campbell-Ferguson, E. A. V. Ebsworth, A. G. MacDiarmid
and T. Yoshioka, J. Phys. Chem., 1967, 71, 723.
11 E. Hengge and F. Höfler, Z. Naturforsch., Teil A, 1971, 26, 768.
12 D. E. Webster, J. Chem. Soc., 1960, 5132.
13 H. Fleischer, D. C. McKean and I. Torto, unpublished work.
14 H. Fleischer, K. Hensen and D. C. McKean, unpublished work.
15 H. Fleischer, D. C. McKean, M. Bühl and W. Thiel, unpublished
work.
Force-field calculations
For each of the conformers calculated to be present4 the ab
initio Cartesian force field and atomic Cartesian coordinates
output by the GAUSSIAN program26 were input into the pro-
gram ASYM 40 (an update of ASYM 2027) for transformation
into an internal coordinate system, with output of valence force
constants and frequencies for partially deuteriated isotopomers.
The potential energy associated with the stretching of the two
Si᎐H or Si᎐Cl bonds in each molecule is given by equation (5)
16 D. C. McKean, Spectrochim. Acta, Part A, 1992, 48, 1335.
17 D. C. McKean, M. H. Palmer and M. F. Guest, J. Mol. Struct.,
1996, 376, 289.
18 D. C. McKean, A. L. McPhail, H. G. M. Edwards, I. R. Lewis,
V. S. Mastryukov and J. E. Boggs, Spectrochim. Acta, Part A, 1993,
49, 1079.
19 D. C. McKean, A. L. McPhail, H. G. M. Edwards, I. R. Lewis,
W. F. Murphy, V. S. Mastryukov and J. E. Boggs, Spectrochim. Acta,
Part A, 1995, 51, 215.
20 D. C. McKean, H. G. M. Edwards, I. R. Lewis, W. F. Murphy,
V. S. Mastryukov and J. E. Boggs, Spectrochim. Acta, Part A, 1995,
51, 2237.
21 W. F. Murphy, F. Zerbetto, J. L. Duncan and D. C. McKean, J. Phys.
Chem., 1993, 97, 581.
22 D. C. McKean, G. P. McQuillan, W. F. Murphy and F. Zerbetto,
J. Phys. Chem., 1990, 94, 4820.
23 A. M. Coats, D. C. McKean and D. Steele, J. Mol. Struct., 1994,
320, 269.
24 H. Moretto, P. Schmidt and U. Wannagat, Z. Anorg. Allg. Chem.,
1972, 394, 125.
25 G. F. Hüttig, Z. Anorg. Allg. Chem., 1922, 123, 31.
26 GAUSSIAN 94, Revision C.2, M. J. Frisch, G. W. Trucks, H. B.
Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheesman,
T. A. Keith, G. A. Petersson, J. A. Montgomery, K. Raghavachari,
M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman,
J. Cioslowski, B. B. Stefanov, A. Nanayakkara, M. Challacombe,
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, PA, 1995.
27 L. Hedberg and I. M. Mills, J. Mol. Spectrosc., 1993, 160, 117.
2V = f1r12 ϩ f2r22 ϩ 2fЈ r1r2
(5)
where r1 and r2 are the displacements in each bond, f1 and f2 the
valence force constants and fЈ the stretch–stretch interaction.
The program ASYM 40 yielded directly the diagonal (f1,f2)
and the off-diagonal (fЈ) constants, but for the Si᎐H bonds they
may also be obtained very simply through a harmonic local
mode (HLM) calculation from the ab initio-derived ν(SiH)
frequencies of the D0 and D1 species.16,17 In this type of cal-
culation, sometimes called an energy-factored force field, all
motions other than Si᎐H stretching are ignored. In general
the HLM value of f Ј is closely similar to the value from
the full normal coordinate treatment embodied in the
ab initio calculation, reflecting the very small amount of
motions other than Si᎐H stretching in the ν(SiH) normal
coordinates. The Si᎐Cl bonds cannot of course be similarly
treated, due to the extent to which observed ‘ν(SiCl)’ frequen-
cies are associated with modes in which there is coupling to
other motions in the molecule.
Acknowledgements
We thank the Royal Society (C. R. P.) and the Leverhulme Trust
(H. F.) for financial support. Part of the experimental work
described was carried out at the University of Frankfurt. We
thank Professor K. Hensen for laboratory facilities, the
University NMR service for the NMR spectra and Hoechst
for the elemental analyses. We are indebted to Drs. J. P. Pillot
Received 14th November 1997; Paper 7/08220G
592
J. Chem. Soc., Dalton Trans., 1998, Pages 585–592