The preparation and analysis of the phenyldimethylsilyllithium reagent and its reaction with silyl enol ethers

Article abstract of DOI:10.1039/a709111g

Phenyldimethylsilyllithium is formed from lithium and phenyldimethylsilyl chloride by slow cleavage of the Si-Si bond of 1,1,2,2-tetramethyl-1,2-diphenyldisilane after the rapid formation of the disilane. 1,1,2,2-Tetramethyl-1,2-diphenyldisiloxane, produced from the silyl chloride by reaction with oxides and hydroxides on the lithium metal surface, is cleaved by dimethyl(phenyl)silyllithium to give lithium dimethyl(phenyl)silanoxide. Dimethyl(phenyl)silyllithium reacts with 1,2-dibromoethane to give dimethyl(phenyl)silyl bromide, which is so rapidly consumed by excess silyllithium reagent that it does not interfere with the double titration used to measure its concentration. Dimethyl(phenyl)silane, produced by protonation of the silyllithium reagent, is also consumed by the silyllithium reagent to give 1,1,2,2-tetramethyl-1,2-diphenyldisilane, which regenerates the silyllithium reagent, as long as lithium is still present. By-products in the preparation of dimethyl(phenyl)silyllithium include 1,3-diphenyl-1,1,2,2,3,3-hexamethyltrisilane, dimethyldiphenylsilane and 1,4-bis[dimethyl(phenyl)-silyl]benzene. Dimethyl(phenyl)silyllithium displaces the silyl group from the tert-butyldimethylsilyl enol ether of cyclohexanone to give the lithium enolate under relatively mild conditions.

Full text of DOI:10.1039/a709111g

View Online / Journal Homepage / Table of Contents for this issue
The preparation and analysis of the phenyldimethylsilyllithium
reagent and its reaction with silyl enol ethers
Ian Fleming,*,a Richard S. Robertsa and Stephen C. Smithb
a Department of Chemistry, Lensfield Road, Cambridge, UK CB2 1EW
b Zeneca Agrochemicals, Jealott’s Hill, Bracknell, Berkshire, UK RG42 6ET
Phenyldimethylsilyllithium is formed from lithium and phenyldimethylsilyl chloride by slow cleavage
of the Si᎐Si bond of 1,1,2,2-tetramethyl-1,2-diphenyldisilane after the rapid formation of the disilane.
1,1,2,2-Tetramethyl-1,2-diphenyldisiloxane, produced from the silyl chloride by reaction with oxides
and hydroxides on the lithium metal surface, is cleaved by dimethyl(phenyl)silyllithium to give lithium
dimethyl(phenyl)silanoxide. Dimethyl(phenyl)silyllithium reacts with 1,2-dibromoethane to give
dimethyl(phenyl)silyl bromide, which is so rapidly consumed by excess silyllithium reagent that it does
not interfere with the double titration used to measure its concentration. Dimethyl(phenyl)silane,
produced by protonation of the silyllithium reagent, is also consumed by the silyllithium reagent to
give 1,1,2,2-tetramethyl-1,2-diphenyldisilane, which regenerates the silyllithium reagent, as long as
lithium is still present. By-products in the preparation of dimethyl(phenyl)silyllithium include 1,3-
diphenyl-1,1,2,2,3,3-hexamethyltrisilane, dimethyldiphenylsilane and 1,4-bis[dimethyl(phenyl)-
silyl]benzene. Dimethyl(phenyl)silyllithium displaces the silyl group from the tert-butyldimethylsilyl
enol ether of cyclohexanone to give the lithium enolate under relatively mild conditions.
H2O
Introduction
PhMe2SiLi
PhMe2SiH
+
+
LiOH
LiOH
For many years we, and countless others, have used the lithium
bis(phenyldimethylsilyl)cuprate reagent,1 and its precursor
phenyldimethylsilyllithium 2,2 as nucleophiles for introducing
a silyl group into a wide range of organic structures. For all
our experience with the phenyldimethylsilyllithium reagent,1,3
there remained some uncertain features in its preparation and
analysis. Gilman established that phenyldimethylsilyllithium 2
could be prepared by stirring phenyldimethylsilyl chloride 1
with lithium in THF (Scheme 1). He also showed that silyl
2
5
titrate against acid
H2O
PhMe2SiOLi
PhMe2SiOH
4
6
BrCH2CH2Br
+
+
PhMe2SiOLi
PhMe2SiLi
PhMe2SiOLi
PhMe2SiBr
7
4
2
4
titrate against acid
Scheme 2
Li
1
PhMe2SiCl
PhMe2SiLi
PhMe2SiSiMe2Ph
silyllithium reagent. Typically, we carry out a double titration
for organometallic compounds (Scheme 2), quenching one ali-
quot with water and one with 1,2-dibromomethane followed by
water, and titrating each against acid, with the difference giving
the concentration of the silyllithium reagent 2. Although the
reaction of an alkyllithium reagent with dibromoethane gives
an inoffensive alkyl bromide,7 the product in our case is pre-
sumably phenyldimethylsilyl bromide 7, which would hydrolyse
in the water to create an equivalent of hydrobromic acid. Any
acid created in this way would cause us to underestimate the
concentration of residual base and to overestimate the concen-
tration of silyllithium reagent. Furthermore, the formation of
acid after the addition of the dibromoethane should have
been obvious—there is always more silyllithium reagent than
residual base, and we should have obtained an acidic solution.
This never happened.
(iii) Our estimates of the concentration of silyllithium
reagent showed a rapid increase over the first 1–4 hours of
stirring with lithium, but the time taken to reach a maximum
concentration varied between 5 hours and several days. It was
not obvious what was happening during the long stirring,
following the first surge of silyllithium formation.
(iv) The silyllithium solutions appeared to maintain their titre
for silyllithium over many weeks and even months, although
using old solutions rarely gave good yields in whatever reaction
we were trying. The titre of old solutions merely revealed an
increase in residual base, which presumably interfered with our
reactions in some way.
Li
1
2
3
Scheme 1
chlorides in general react with lithium in THF to give initially a
silyllithium reagent, but these react with more of the silyl chlor-
ide to give disilanes. The disilane can then be cleaved with more
lithium to give the silyllithium reagent, but only when there is a
phenyl group attached to the silicon, as in the conversion of the
disilane 3 to the silyllithium reagent 2. We have shown that the
cleavage is remarkably sensitive to the presence of substituents
on the phenyl ring,4 with p-tolyldimethylsilyl chloride giving a
disilane, which is not cleaved by lithium.
We were aware of six uncertainties in the chemistry of
phenyldimethylsilyllithium.
(i) It had not been firmly established whether Gilman’s
preparation of the phenyldimethylsilyllithium reagent 2 was
successful only because the disilane 3 could be cleaved, as
seemed probable, or whether it was easy to make because the
rate of formation of the silyllithium reagent from the silyl chlor-
ide 1 was faster than its reaction with the silyl chloride to give
the disilane 3. It is known that tris(o-tolyl)silyl chloride2 and
mesityldimethylsilyl chloride5 give silyllithium reagents when
stirred with lithium, and it seemed possible that this success
with a substituted phenyl ring might be because the coupling of
the silyllithium reagent with the silyl chloride was slow.
(ii) Secondly, there was an anomaly in the analytical method6
most commonly used for measuring the concentration of the
J. Chem. Soc., Perkin Trans. 1, 1998
1209

View Online
Table 1 1H NMR chemical shifts δ (using C6D6 as internal lock for
colour typical of the reagent then develops more slowly, usually
spectra in THF) for the SiMe signals of the various silicon species
starting 30 minutes into the reaction, and approaching maxi-
mum intensity after 4–6 hours. During this period, we saw the
Compound
δH(CDCl3)
δH(THF)b
1
gradual development of the H NMR signal from the silyl-
lithium reagent and the disappearance of the signal from the
disilane.
1 PhMe2SiCl
2 PhMe2SiLi
3 (PhMe2Si)2
4 PhMe2SiOLi
5 PhMe2SiH
6 PhMe2SiOH
7 PhMe2SiBr
8 (PhMe2Si)O
9 (PhMe2Si)2Me2Si
10 Ph2SiMe2
11 PhMe2Si-p-C6H4-
SiMe2Ph
—
—
0.32
—
0.75
0.23
0.42
0.32
0.46a
0.38c
—
0.43
—
—
Lithium fresh from the bottle does not always cleave the
disilane, presumably because of the protective coat of oxides
and/or hydroxides. The fast initial reaction with the silyl
chloride, on the other hand, cleans the surface, causing
the appearance in the NMR spectrum of the signal from
diphenyltetramethyldisiloxane. We find that the reaction
between diphenyltetramethyldisilane and lithium fresh from a
bottle can be initiated by adding a little silyl chloride, either
phenyldimethylsilyl chloride or trimethylsilyl chloride, to clean
the surface of the lithium. We add our experience to that of
others11 in urging that this is a simple way to prepare active metals
for the organometallic chemistry of common main-group metals
like lithium, zinc and magnesium. Whereas traditional activ-
ation12 of these metals, by sonication, or by treatment with
alkyl halides or iodine, probably works because reaction with
the underlying metal mechanically dislodges the oxide film, a
silyl halide will combine with it chemically and remove it
completely.
The rapidity of the reaction between the silyllithium reagent
and the silyl chloride answered our second question too—as
dibromoethane is added to the silyllithium reagent 2, and silyl
bromide 7 is formed, we can now expect it to be quenched
rapidly by more of the silyllithium reagent. When we followed
the reaction by 1H NMR spectroscopy, no signal from the silyl
bromide 7 appeared on addition of dibromoethane, but
we did see the immediate appearance of the signal from the
disilane 3 (Scheme 4).
0.37a
0.42
0.92
0.33
0.30, 0.10
0.57
0.55
—
a Doublet, J 3.9 Hz. b Referenced to the THF signal as δ 1.85. c Doublet,
J 0.5 Hz.
(v) When we add aqueous ammonium chloride to quench our
reaction mixtures, which often use an excess of silyllithium
reagent (or a cuprate or zincate derived from it), we rarely
detect phenyldimethylsilane 5, which has a characteristic septet
(J 3.9 Hz) at δ 4.62 (in THF solution). Instead we obtain non-
polar material, which appears to be a mixture of diphenyl-
tetramethyldisilane and diphenyltetramethyldisiloxane.
(vi) The colour of phenyldimethylsilyllithium solutions is red
or reddish brown when derived from the silyl chloride and
greenish when derived by cleavage of the disilane. It is known
that the colour is not that of the silyllithium reagent itself, since
it can be quenched with a trace of oxygen without affecting the
1H and 7Li NMR signals of the silyllithium reagent.8 The
colour is associated with a quintet pattern in the EPR spectrum,
assigned to the radical anion of a 1,4-disilylated benzene.8,9 Is
this a correct assignment and where did such a by-product
come from?
BrCH2CH2Br
+
+
PhMe2SiOLi
PhMe2SiLi
PhMe2SiOLi
PhMe2SiBr
Having published some of our work in preliminary form,10
we report here our answers, as far as they go, in full, together
with the useful corollary that the silyllithium reagent generates
a lithium enolate from a silyl enol ether more rapidly than
methyllithium does.
2
4
7
4
fast
+
PhMe2SiBr
PhMe2SiLi
PhMe2SiSiMe2Ph
7
2
3
Scheme 4
Results and discussion
The slow continuous development of silyllithium reagent
over hours, is not so easily explained. One possibility is that
lithium phenyldimethylsilanoxide 4 reacts very slowly with the
silyllithium reagent to give silane 5 and lithium oxide, but we
were unable to demonstrate that this reaction is taking place.
We made up a mixture of these reagents, but saw the signals in
the 1H NMR spectra simply grow broader, making a definitive
statement risky. Thus it is likely that the variable slow rise in the
titre reflects only the different rates of stirring and the variable
quality of the lithium used over the years.
The answer to our fourth question is that any proton source
reacts with the silyllithium reagent to give phenyldimethylsilane
5, which then reacts with the silyllithium reagent to give the
disilane 3 and lithium hydride (Scheme 5). The disilane in turn
The phenyldimethylsilyllithium reagent
We monitored the reaction in THF by 1H NMR spectro-
scopy, following the growth and diminution of the SiMe
signals from the various species present (Table 1). The signal
from phenyldimethylsilyl chloride 1 disappeared over a few
minutes at 0 ЊC, and the signal from diphenyltetramethyl-
disilane 3 grew at an equal rate. Clearly the slow step in the
formation of the silyllithium reagent 2 is the cleavage of the
disilane, as expected. We also see the appearance of the signal
from the disiloxane 8, some of it derived from the more or less
unavoidable hydrolysis of the silyl chloride before we made up
the reaction mixture, and some of it derived, as soon as we
mixed the reagents, from the coat of oxides and/or hydroxides
on the lithium. We then see its disappearance, starting to take
place as soon as the silyl chloride has been consumed, to be
replaced by the signal from lithium phenyldimethylsilanoxide 4
(Scheme 3). This reaction takes place as fast as the silyllithium
+
+
LiH
PhMe2SiLi
PhMe2SiH
PhMe2SiSiMe2Ph
3
2
5
Scheme 5
+
+
PhMe2SiOLi
PhMe2SiLi
PhMe2SiOSiMe2Ph
PhMe2SiSiMe2Ph
is cleaved by lithium (Scheme 1) to regenerate the silyllithium
reagent, and all that happens to the titre, as long as there is an
excess of lithium present, is an increase in residual base. We
confirmed the reasonableness of this pathway by adding the
silyllithium reagent to phenyldimethylsilane, and observing
the formation of disilane within 15 minutes at room temper-
ature.
2
8
4
3
Scheme 3
is being formed, and is typically complete within 30 minutes.
During these two phases, wisps of red colour flow off the
lithium, rapidly to be quenched as disilane forms. The full red
However, water is not the only potential contaminant—
1210
J. Chem. Soc., Perkin Trans. 1, 1998

View Online
oxygen can be expected to give successively silyl peroxide anion
and silanoxide, which will remove silyllithium and contribute to
residual base. We do not know why this is not a conspicuous
problem, but there is some evidence that the reaction with
oxygen is not straightforward. We quenched the silyllithium
reagent 2 with oxygen, and observed the formation of the
silanol 6 in 26% yield, the silane 5 in 3% yield and disilane 3 in
38% yield, similar to Gilman’s results with triphenylsilyllithium
and oxygen.13 Since disilane is formed, it will regenerate the
silyllithium reagent in the presence of lithium, and the incur-
sion of oxygen is not as serious as one might at first suppose.
The observation that silane 5 is easily cleaved by the silyl-
lithium reagent also provides the answer to our fifth question.
The phenyldimethylsilane that is presumably formed in an
aqueous quench is itself unstable, both towards any residual
silyllithium reagent not yet quenched, and towards reaction
with hydroxide ion in the alkaline solution, giving silanoxide
and hence the disiloxane to a greater or lesser extent, depending
upon the precise conditions of the workup. The reaction with
hydroxide ion is known to be too fast in aqueous media to
follow the kinetics.14
We checked that the development in intensity of the signals
assigned to the silyllithium reagent 2 and to the silanoxide 4
corresponded to the concentration of the lithium reagent and
of residual base that we were measuring by titration. We carried
out three runs with, successively, a good sample of the silyl
chloride 1, a 2:1 mixture of the silyl chloride 1 with the
disiloxane 8, and a 1:2 mixture. The agreement between the two
methods of estimating concentration was moderately good,
with the concentration of the silyllithium reagent measured by
NMR spectroscopy consistently about 10–15% lower than the
value from titration, but the concentration of lithium phenyl-
dimethylsilanoxide measured by NMR spectroscopy and of
residual base measured by titration agreed rather better ( 5%).
disilane to give octamethyltrisilane and regenerate methyl-
lithium, which will attack at silicon in the trisilane to displace
pentamethyldisilanyllithium.
We also detected by GC–MS a significant quantity of 1,4-
bis[dimethyl(phenyl)silyl]benzene 11, identified by comparison
with an authentic sample prepared from the silyllithium reagent
and 1,4-dibromobenzene. Although we have no evidence for
how this substance is formed, it adds support to the assign-
ment8,9 of its radical anion as the source of the colour in sol-
utions of the phenyldimethylsilyllithium reagent. Our finding
dimethyl(diphenyl)silane 10 in the reaction mixture also pro-
vides for the first time a substrate for para-coupling18 with the
silyllithium reagent that could lead to this compound.
Cleavage of silyl enol ethers using the phenyldimethylsilyllithium
reagent
The speed with which the disiloxane 8 was cleaved by the silyl-
lithium reagent 2 suggested that other ethers might also be
cleaved easily, just as triphenylsilyl ethers are cleaved by tri-
phenylsilyllithium.19 We find that the trimethylsilyl ether of
benzyl alcohol is cleaved in THF solution even at Ϫ78 ЊC, with
a 30% yield of pentamethyl(phenyl)disilane formed after four
hours. More significantly, the silyl enol ether 12 was cleaved
within 10 minutes at Ϫ78 ЊC in THF to give the lithium enolate
13, identified by methylation giving the ketone 14 (Scheme 7).
OSiMe2Ph
Ph
OLi
O
i
ii
Ph
Ph
Ph
Ph
Ph
12
13
14
77%
OSiMe2But
OLi
O
ii
i
By-products in the formation of the phenyldimethylsilyllithium
reagent
We also saw weak SiMe signals not assignable to any of the
species mentioned so far. Accordingly, we kept a mixture of the
silyllithium reagent 2 and the disilane 3 at 4 ЊC for 7 days, and
examined the mixture of products by GC–MS. One of the
products was 1,3-diphenyl-1,1,2,2,3,3-hexamethyltrisilane 9, a
plausible product if the silyllithium reagent attacks diphenyl-
tetramethyldisilane at silicon,15 and displaces phenyllithium
(Scheme 6). Phenyllithium will react in turn with disilane 3 to
15
16
17 74%
Scheme 7
Stork and Hudrlik, in their pioneering work on the use of silyl
enol ethers as precursors to lithium enolates,20 typically cleaved
comparable trimethylsilyl enol ethers with methyllithium at
0 ЊC in 1 hour in diethyl ether and in a few minutes in glyme.
Although our reagent works faster, there is little difficulty
cleaving trimethylsilyl enol ethers, so we turned to the tert-
butyldimethylsilyl enol ether 15. In a comparable case, Stork
and Hudrlik found that methyllithium took 20 h at room tem-
perature, and the reaction had to be carried out in glyme. We
find that phenyldimethylsilyllithium cleaves this silyl enol ether
in THF within two hours at 20 ЊC to give the lithium enolate 16,
identified in the 1H NMR spectrum and by methylation giving
2-methylcyclohexanone 17 in 74% yield together with 1-tert-
butyl-1,1,2,2-tetramethyl-2-phenyldisilane, in 99% yield. The
silyllithium reagent even cleaves the silyl enol ether 15 with a
half-life of about eight hours in a 1:4 mixture of THF and
diethyl ether. Clearly phenyldimethylsilyllithium is a powerful
reagent for preparing lithium enolates from silyl enol ethers,
avoiding long reaction times and special solvents.
+
+
PhLi
PhMe2SiLi
PhMe2SiSiMe2Ph
PhMe2SiSiMe2SiMe2Ph
2
3
9
PhLi + PhMe2SiSiMe2Ph
Ph2Me2Si + PhMe2SiLi
3
10
2
PhMe2Si
SiMe2Ph
11
Scheme 6
give dimethyldiphenylsilane 10, and we do indeed observe the
presence of this compound both in the 1H NMR spectrum and
by GC–MS. An alternative route to the silane 10 is by attack of
the phenyllithium on the trisilane 9, this time displacing 1-lithio-
2-phenyl-1,1,2,2-tetramethyldisilane. However, the signals from
the trisilane and the silane 10 were always present in constant
proportions, suggesting that the latter pathway is not import-
ant. Pathways like these are almost certainly responsible for the
unexpected formation of pentamethyldisilanyllithium when
hexamethyldisilane is treated with methyllithium on a large
scale.16 Methyllithium will first displace trimethylsilyllithium,17
but on a large scale this reagent will have time to attack the
Experimental
Unless otherwise stated, light petroleum refers to the fraction
bp 40–60 ЊC. Ether refers to diethyl ether.
Dimethyl(phenyl)silyllithium 2
Typically, lithium shot (0.5 g in mineral oil, 70 mmol) was
stirred rapidly for 15 min in dry hexane (20 cm3) under argon.
J. Chem. Soc., Perkin Trans. 1, 1998
1211

View Online
The hexane was removed and the lithium suspended in dry
THF (20 cm3). The mixture was stirred rapidly with chloro-
dimethyl(phenyl)silane (5.0 cm3, 29.8 mmol) at 0 ЊC for 6 h to
give a deep red solution; δH(270 MHz; THF–[2H6]benzene,
85:15) 7.45 (2 H, d, J 7, o-H), 7.06 (2 H, t, J 7, m-H), 6.87 (1 H,
t, J 7, p-H) and 0.23 (6 H, s, SiMe2); δC(THF–[2H6]benzene,
85:15) 165.3, 133.0, 125.7, 121.9 and 6.9.
silyllithium (1.0 mol dmϪ3 in THF, 10 cm3, 10.0 mmol) in tolu-
ene (20 cm3) was added and the mixture refluxed for 2.5 h. The
mixture was cooled to room temperature, quenched with dilute
hydrochloric acid (0.3 mol dmϪ3, 50 cm3), extracted with ether
(3 × 100 cm3), the organic layers washed with brine (100 cm3),
dried (MgSO4) and evaporated under reduced pressure. Light
petroleum (60 cm3) was added precipitating a solid, which was
filtered off, and the solution evaporated under reduced pressure.
Chromatography (SiO2, light petroleum) gave the bis-silane 11
(0.59 g, 34%) as needles, mp 56–60 ЊC (from hexane at low tem-
perature) (lit.,24 59 ЊC); Rf (light petroleum) 0.18; νmax(Nujol)/
cmϪ1 1587 (Ph), 1250 (SiMe2), 1134 (SiPh) and 1115 (SiPh);
δH(250 MHz; CDCl3) 7.6 (4 H, m, o-H), 7.53 (4 H, s, SiC6H4Si),
7.3 (6 H, m- and p-H) and 0.55 (12 H, s, SiMe2); δC(CDCl3)
139.1, 138.1, 134.2, 133.5, 129.1, 127.8 and Ϫ2.5; m/z (EI) 346
(25%, Mϩ), 331 (100), 193 (25) and 143 (30) (Found: Mϩ,
346.1583. C22H26Si2 requires M, 346.1573).
1,1,2,2-Tetramethyl-1,2-diphenyldisilane 3
Chlorodimethyl(phenyl)silane 1 (20 cm3, 117 mmol) was stirred
with a suspension of hexane-washed lithium shot (0.90 g in
mineral oil, 128 mmol) in dry THF (80 cm3) at 0 ЊC under argon
for 4 h. 1,2-Dibromoethane (1 cm3) and water (100 cm3) were
added, and the aqueous layer was extracted with ether–hexane
(1:1, 2 × 50 cm3). the combined organic layers were washed
with brine (50 cm3), dried (MgSO4) and the solvent evaporated.
Distillation (Kugelrohr, oven temperature 120 ЊC at 1 mmHg)
gave the disilane (9.4 g, 59%) as needles, mp 31–35 ЊC (from
hexane at low temperature) (lit.,21 mp 35 ЊC, bp 130–133 ЊC at 2
mmHg); Rf (hexane) 0.40; νmax(Nujol)/cmϪ1 1586 (Ph), 1245
(SiMe2) and 830 (SiMe2); δH(250 MHz; CDCl3) 7.4–7.2 (10 H,
m, Ph) and 0.32 (12 H, s, SiMe2); δC(CDCl3) 139.1, 134.0, 128.5,
Titration of silyllithiums
Following the method of Gilman,6 an aliquot of silyllithium
solution (typically 1 cm3) was quenched in water (5 cm3) and the
basic solution titrated against standardised hydrochloric acid
(0.1 mol dmϪ3) using phenolphthalein. A second aliquot was
quenched in 1,2-dibromoethane (5 cm3), shaken vigorously
with water (5 cm3) and the mixture titrated against hydrochloric
acid (0.1 mol dmϪ3). The silyllithium reagent was typically 1
mol dmϪ3.
ϩ
ϩ
127.8 and Ϫ3.8; m/z (CI, NH4 ) 288 (20%, M ϩ NH4 ); (EI)
270 (15%, Mϩ), 197 (20) and 135 (100).
Lithium dimethyl(phenyl)silanoxide 4
Dimethyl(phenyl)silanol 6 (8 mg, 0.05 mmol), (2,2-dimethyl-
propyl)benzene (0.1 cm3, 0.06 mmol), [2H6]benzene (0.05 cm3),
n-butyllithium (2.5 mol dmϪ3 in hexane, 0.02 cm3, 0.05 mmol)
and dry THF (0.6 cm3) were mixed in a dry NMR tube under
argon. The doublet of silanol 6 at δH 0.38 (1 H, d, J 1.5, SiMe2)
was replaced by the singlet of silanoxide 4; δH(270 MHz; THF–
[2H6]benzene 85:15) (partial) 0.32 (6 H, s, SiMe2).
Reaction of silyl chloride 1 with lithium observed by 1H NMR
spectroscopy
Hexane-washed lithium shot (30 mg, 5 mmol), silyl chloride 1
(0.1 cm3, 0.6 mmol), (2,2-dimethylpropyl)benzene (0.1 cm3, 0.6
mmol), dry THF (1 cm3) and [2H6]benzene (0.05 cm3) were
sonicated in an NMR tube under argon, maintaining the son-
icator bath below 10 ЊC, and recording the 1H NMR spectra at
intervals (Table 1).
Dimethyl(phenyl)silanol 6
Sodium hydroxide (0.8 g, 20 mmol), methanol (3 cm3), propan-
2-ol (7 cm3) and 1,1,2,2-tetramethyl-1,2-diphenyldisiloxane 8
(2.85 g, 10 mmol) were heated, initially at 50 ЊC and then at
120 ЊC to evaporate the solvents. Methanol (15 cm3) was added
and the solvents were distilled (bath temperature 120 ЊC at 15
mmHg). This was repeated three times, the final distillation giv-
ing a crude solid on cooling. The solid was dissolved in boiling
light petroleum (bp 80–100 ЊC), hot filtered and evaporated
under reduced pressure. Ether (30 cm3) was added and the mix-
ture poured into glacial acetic acid (1.8 cm3), water (30 cm3) and
ether (15 cm3). The organic layer was washed with water, dried
(MgSO4) and evaporated under reduced pressure. Distillation
(Kugelrohr, oven temperature 115 ЊC at 15 mmHg) (bp lit.,22
101–101.5 ЊC at 14 mmHg) gave the silanol (1.30 g, 43%); Rf
(EtOAc–hexane, 20:80) 0.26; νmax(film)/cmϪ1 3286 (OH), 1253
(SiMe2), 1119 (SiPh), 866 (SiO) and 829 (SiMe2); δH(270 MHz;
CDCl3) 7.6 (2 H, m, o-H), 7.4 (3 H, m, m- and p-H), 1.9 (1 H, br
s, OH) and 0.42 (6 H, s, SiMe2).
Comparison of molarities of silyllithium 2 measured by titration
and 1H NMR spectroscopy
Hexane-washed lithium shot (three molar equivalents based on
silicon content), silyl chloride 1, disiloxane 8, (2,2-dimethyl-
propyl)benzene (1.51 g, 10.2 mmol) and dry THF (10 cm3) were
stirred rapidly for 6 h at 0 ЊC under argon. Aliquots (1 cm3) of
the silyllithium solution were titrated against hydrochloric acid
1
(0.1 mol dmϪ3). The H NMR spectra were acquired from
another aliquot (0.6 cm3) added to [2H6]benzene (0.05 cm3).
Two samples were kept for 90 d at Ϫ20 ЊC. The molarities of the
silyllithium reagent and residual base were calculated from the
integrals of the peaks for 2 at δ 0.26 (6 H, s, SiMe2), for 4 at
δ 0.38 (6 H, s, SiMe2) and for (2,2-dimethylpropyl)benzene at
δ 1.00 (Table 2).
Reaction of silyllithium 2 with 1,2-dibromoethane
Method A. Hexane-washed lithium shot (0.5 g, 83 mmol),
silyl chloride 1 (1.5 cm3, 8.86 mmol), disiloxane 8 (2.0 cm3, 6.66
mmol), (2,2-dimethylpropyl)benzene 1.2 cm3, 9.70 mmol) and
dry THF (14 cm3) were stirred for 6 h at 0 ЊC under argon. The
1H NMR spectra of aliquots (0.6 cm3) in [2H6]benzene (0.05
cm3) were acquired after 0.5 and 6 h. 1,2-Dibromoethane (0.04
cm3, 0.46 mmol) was added to the NMR tube and the spectrum
reacquired. The signal from the silyl chloride 1 had disappeared
after 0.5 h, the signal from the disilane 3 disappeared after 6 h,
the signal from the silanoxide 4 remained constant throughout,
the signal from the silyllithium reagent 2 grew between 0.5 and
6 h, and disappeared after the treatment with dibromoethane,
and signals for the disilane 3 and ethane (δ 5.10) appeared.
Method B. 1,2-Dibromoethane (1 cm3) was added to silyl-
lithium 2 (0.75 mol dmϪ3 solution in THF, 7.5 cm3, 5.6 mmol)
under argon. Ether (20 cm3) was added at 0 ЊC, and the mixture
filtered through Florisil. The solvent was evaporated under
1,1,2,2-Tetramethyl-1,2-diphenyldisiloxane 8
Chlorodimethyl(phenyl)silane 1 (10 cm3, 59 mmol) was stirred
in water (0.5 cm3), acetone (25 cm3) and saturated sodium
hydrogen carbonate solution (40 cm3) for 1 h at 0 ЊC. An aque-
ous work-up and distillation (Kugelrohr, oven temperature
170 ЊC at 5 mmHg) (bp lit.,23 135 ЊC at 2 mmHg) gave the
disiloxane (7.5 g, 89%); Rf (hexane) 0.40; νmax(film)/cmϪ1 1590
(Ph), 1254 (SiMe2), 1119 (SiPh) and 831 (SiMe2); δH(270 MHz;
CDCl3) 7.46 (4 H, m, o-H), 7.3 (6 H, m, m- and p-H) and 0.33
(12 H, s, SiMe2); δC(CDCl3) 139.9, 133.1, 129.3, 127.8 and 0.9;
m/z (EI) 286 (35%, Mϩ), 271 (85), 193 (100) and 89 (25).
1,4-Bis[dimethyl(phenyl)silyl]benzene 11
1,4-Dibromobenzene (1.2 g, 5 mmol), tetrakis(triphenyl-
phosphine)palladium (0.5 g, 0.4 mmol) and dry toluene (40
cm3) were refluxed for 30 min under argon. Dimethyl(phenyl)-
1212
J. Chem. Soc., Perkin Trans. 1, 1998

View Online
Table 2 Comparison of the concentration of silyllithium reagent by double titration and by integration in the 1H NMR spectra
Conc. silyllithium (mol dmϪ3 Conc. residual base (mol dmϪ3 Conc. after 90 daysa
)
)
Reagents
By titration
0.76
By 1H NMR
By titration
0.01
By 1H NMR
0.02
Silyllithium
0.99
Res. base
Silyl chloride 1
0.71
0.59
(2.50 cm3, 14.7 mmol)
and no disiloxane
Silyl chloride 1
0.77
0.65
0.62
0.41
0.58
0.42
0.62
0.81
—
1.66
(1.25 cm3, 7.3 mmol)
and disiloxane 8
(1.6 g, 3.5 mmol)
Silyl chloride 1
0.50
—
(1.25 cm3, 7.3 mmol)
and disiloxane 8
(4.27 g, 14.9 mmol)
a Measured by titration in mol dmϪ3
.
reduced pressure, and the residue chromatographed (SiO2,
hexane) to give a mixture of disilane 3 (45%, 1H NMR),
identical (TLC, 1H NMR, GC) to an authentic sample, and 1,3-
diphenyl-1,1,2,2,3,3-hexamethyltrisilane25 9 (8%, 1H NMR); Rf
(hexane) 0.32; δH(250 MHz; CDCl3) 7.6–7.0 (10 H, m, Ph), 0.30
(12 H, s, SiMe2Ph) and 0.10 (6 H, s, SiSiMe2Si); m/z (EI) 328
(5%, Mϩ), 313 (10), 193 (30), 135 (100), 116 (60) and 73 (20).
showed the presence of [yields based on the (2,2-dimethyl-
propyl)benzene]: dimethyl(phenyl)silane 5 (2.5%), dimethyl-
(phenyl)silanol 6 (26%), dimethyl(diphenyl)silane 10 (6%,
unstandardised), 1,1,2,2-tetramethyl-1,2-diphenyldisilane
3
(38%) and 1,1,2,2-tetramethyl-1,2-diphenyldisiloxane 8 (0.3%).
1,2-Diphenyl-1-[dimethyl(phenyl)silyloxy]ethene 12
1,2-Diphenylethanone (0.42 g, 2.2 mmol) in dry THF (4 cm3)
was refluxed with hexane-washed sodium hydride (60% disper-
sion in oil, 0.20 g, 5 mmol) in dry THF (6 cm3) under argon for
3 h. Dry triethylamine (0.46 cm3, 3.2 mmol) and chlorodi-
methyl(phenyl)silane 1 (0.55 cm3, 3.2 mmol) were added at
room temperature, and the mixture was stirred for 15 min,
diluted with ether (30 cm3), washed with cold saturated sodium
hydrogen carbonate solution (3 × 20 cm3), brine (20 cm3), dried
(MgSO4) and evaporated under reduced pressure. Rapid chro-
matography (SiO2, EtOAc–light petroleum, 2:98) gave the silyl
enol ether 12 (0.44 g, 60%) (Z-12:E-12, >99:1) as an oil; Rf
(EtOAc–light petroleum, 2:98) 0.33; νmax(film)/cmϪ1 1630
The reaction of silane 5 and silyllithium 2
Dimethyl(phenyl)silyllithium (1.06 mol dmϪ3 in THF, 0.31 cm3,
0.33 mmol) was added to dimethyl(phenyl)silane (0.05 cm3, 0.33
mmol), (2,2-dimethylpropyl)benzene (0.05 cm3) and [2H6]-
benzene (0.05 cm3) in dry THF (0.5 cm3) under argon in a dry
NMR tube. The 1H NMR spectrum was acquired after 15 min
at room temperature and after standing for 16 h, when a white
precipitate (probably lithium hydride) had precipitated. Quanti-
tative conversion to disilane 3 was observed [1H NMR, by com-
parison with the integral of the peak at δ 1.00 from (2,2-
dimethylpropyl)benzene] after 16 h, with most of the reaction
(ca. 80%) complete within 15 min. Carbon dioxide was bubbled
through the mixture for 1 min, the mixture washed with sodium
hydroxide (5%, 1 cm3), the organic layer was dried (MgSO4) and
evaporated under reduced pressure to give the disilane 3 (65 mg,
(C᎐C), 1600 (Ph) and 1592 (Ph); δ (250 MHz; CDCl ) 7.7–7.2
᎐
H
3
(15 H, m, Ph and SiPh), 6.16 (1 H, s, ᎐CH) and 0.34 (6 H, s,
᎐
SiMe2); δC(CDCl3) 151.0, 139.6, 136.9, 136.6, 133.6, 129.9,
128.9, 128.2, 128.2, 128.1, 127.8, 126.5, 126.2, 111.0 and Ϫ0.7;
m/z (EI) 330 (90%, Mϩ), 135 (100) (Found: Mϩ, 330.1438.
C22H22OSi requires M, 330.1440).
1
74%) as an oil, identical (IR, H NMR, GC) to an authentic
sample.
The reaction of silyllithium 2 with disilane 3
Reaction of silyllithium 2 with silyl enol ether 12
Dimethyl(phenyl)silyllithium (0.95 mol dmϪ3 in THF, 3.2 cm3,
3.0 mmol) and disilane 3 (0.27 g, 1.0 mmol) were kept for 8 d at
4 ЊC. Carbon dioxide was bubbled through the mixture for 1
min, ether (15 cm3) was added, the solution washed with
sodium hydroxide (5% in H2O, 3 × 10 cm3) and the organic
layer dried (MgSO4) and evaporated under reduced pressure. 1H
NMR and GC analysis of the residue showed the presence of
the following compounds (yields in mmol by 1H NMR spectro-
scopy, based on total silicon content): 1,2-diphenyltetramethyl-
disilane 3 (2.2 mmol, 55%), identical (1H NMR, GC, MS) to an
authentic sample, 1,3-diphenyl-1,1,2,2,3,3-hexamethyltrisilane
9 (0.4 mmol, 10%), identical (1H NMR, GC) with another
sample, dimethyl(diphenyl)silane26 10 (0.4 mmol, 10%); Rf
(hexane) 0.32; δH(250 MHz; CDCl3) 7.6–7.0 (10 H, m, Ph), 0.57
(6 H, s, SiMe2); m/z (EI) 212 (15%, Mϩ), 197 (100), 135 (15), 105
(20) and 77 (10), and 1,4-bis[dimethyl(phenyl)silyl]benzene 11
(1 mmol, 25%), identical (δH 0.55, δC Ϫ2.5, and MS) to an
authentic sample, and no trace of benzoic acid.
Dimethyl(phenyl)silyllithium (0.90 mol dmϪ3 in THF, 3.55 cm3,
3.2 mmol) was stirred with the silyl enol ether 12 (96 mg, 0.29
mmol) in dry THF (1 cm3) under argon at Ϫ78 ЊC for 10 min.
Iodomethane (0.20 cm3, 3.2 mmol) was added, the mixture
warmed to room temperature, methanol (3 cm3) and dilute
hydrochloric acid (3 mol dmϪ3, 2 drops) were added and the
mixture kept overnight. An aqueous work-up and chrom-
atography (SiO2, EtOAc–light petroleum, 8:92) gave the ketone
14 (47 mg, 77%) as a solid, mp 52–54 ЊC (from MeOH–H2O)
(lit.,27 53–54 ЊC); Rf (Et2O–light petroleum, 12:88) 0.33;
νmax(film)/cmϪ1 1682 (C᎐O), 1597 (Ph) and 1582 (Ph); δ (250
᎐
H
MHz; CDCl3) 7.97 (2 H, dd, J 7.5 and 1, PhCO o-H), 7.5–7.1
(8 H, m, other Ph), 4.70 (1 H, q, J 7.5, CHMe) and 1.55 (3 H, d,
J 7.5, Me).
1-tert-Butyl-1,1,2,2-tetramethyl-2-phenyldisilane
Dimethyl(phenyl)silyllithium 2 (1.1 mol dmϪ3 in THF, 3.0 cm3,
3.3 mmol) and tert-butylchloro(dimethyl)silane (0.5 g, 3.3
mmol) were kept in dry THF (5 cm3) under argon at 0 ЊC for
4 h. 1,2-Dibromoethane (0.5 cm3) was added and the solvents
evaporated under reduced pressure. The residue was taken up in
hexane (20 cm3), filtered and the solution evaporated under
reduced pressure. Distillation (Kugelrohr, oven temperature
135 ЊC at 21 mmHg) gave the silane (485 mg, 59%); Rf (light
petroleum) 0.32; νmax(film)/cmϪ1 1245 (SiMe2), 1106 (SiPh) and
831 (SiMe2); δH(250 MHz; CDCl3) 7.52 (2 H, m, o-H), 7.36 (3
The reaction of silyllithium 2 with oxygen
Oxygen was bubbled through a solution of silyllithium 2 (1.04
mol dmϪ3 in THF, 0.50 cm3, 0.52 mmol) and (2,2-dimethyl-
propyl)benzene (0.05 cm3, 0.29 mmol) at 0 ЊC for 1 min, the red
colour rapidly disappearing. Water (5 cm3) was added, the mix-
ture extracted with ethyl acetate (5 cm3), the organic layer dried
(MgSO4) and evaporated under reduced pressure. GC analysis
J. Chem. Soc., Perkin Trans. 1, 1998
1213

View Online
3 I. Fleming and U. Ghosh, J. Chem. Soc., Perkin Trans. 1, 1994, 257;
I. Fleming, M. Solay and F. Stolwijk, J. Organomet. Chem., 1996,
521, 121; S. Saito, K. Shimada, H. Yamamoto, E. Martínez de
Marigorta and I. Fleming, Chem. Commun., 1997, 1299; I. Fleming,
R. S. Roberts and S. C. Smith, J. Chem. Soc., Perkin Trans. 1, 1998,
1215; I. Fleming, J. Frackenpohl and H. Ila, J. Chem. Soc., Perkin
Trans. 1, 1998, 1229.
H, m, m- and p-H), 0.90 (9 H, s, CMe3), 0.45 (6 H, s, SiMe2Ph)
and 0.05 (6 H, s, SiMe2CMe3); δC(CDCl3) 140.3, 133.8, 128.2,
127.7, 27.6, 17.9, Ϫ2.4 and Ϫ6.1; m/z (EI) 250 (85%, Mϩ), 193
(100), 135 (90) and 73 (45) (Found: Mϩ, 250.1573. C14H26Si2
requires M, 250.1573).
4 N. Abd. Rahman, I. Fleming and A. B. Zwicky, J. Chem. Res. (S),
1992, 292; J. Chem. Res. (M), 1992, 2041.
5 H. Müller, U. Weinzierl and W. Seidel, Z. Anorg. Allg. Chem., 1991,
603, 15.
6 H. Gilman, F. K. Cartledge and S.-Y. Sim, J. Organomet. Chem.,
1963, 1, 8.
7 H. Gilman and D. Aoki, J. Org. Chem., 1959, 24, 426; H. Gilman
and H. J. S. Winkler, in Organometallic Chemistry, ed. H. Weiss,
Reinhold, New York, 1960, pp. 270–345.
8 J. B. Lambert and M. Urdaneta-Pérez, J. Am. Chem. Soc., 1978, 100,
157.
9 Y.-P. Wan, D. H. O’Brien and F. J. Smentowski, J. Am. Chem. Soc.,
1972, 94, 7680 and references cited therein.
Reaction of silyllithium 2 with benzyloxy(trimethyl)silane
Dimethyl(phenyl)silyllithium 2 (1.28 mol dmϪ3 in THF, 0.13
cm3, 1.67 mmol), benzyloxy(trimethyl)silane28 (0.03 cm3, 0.15
mmol) and 2,2-dimethyl(phenyl)propane (0.026 cm3, 0.15
mmol) in dry THF (0.5 cm3) were kept under argon at Ϫ78 ЊC
for 4 h with occasional shaking. 1,2-Dibromoethane (0.2 cm3,
excess) was added, the mixture shaken with water (0.3 cm3),
diluted with ether (1 cm3) and the organic layer dried (MgSO4).
GC analysis showed the presence of 1-phenyl-1,1,2,2,2-penta-
methyldisilane (30%, calibrated against 2,2-dimethyl(phenyl)-
propane).
10 I. Fleming, R. S. Roberts and S. C. Smith, Tetrahedron Lett., 1996,
38, 9395.
Reaction of silyllithium 2 with silyl enol ether 15
Dimethyl(phenyl)silyllithium (0.90 mol dmϪ3 in THF, 0.33 cm3,
0.30 mmol) was added to silyl enol ether 1529 (41 mg, 0.19
mmol), [2H6]benzene (0.1 cm3) and dry [2H8]THF (0.75 cm3)
under argon in a dry NMR tube at room temperature. The 1H
NMR spectra were recorded at intervals, measuring the relative
11 P. Knochel, M. C. P. Yeh, S. C. Berk and J. Talbert, J. Org. Chem.,
1988, 53, 2390; K. Takai, T. Kakiuchi and K. Utimoto, J. Org.
Chem., 1994, 59, 2671; A. Fürstner and A. Hupperts, J. Am. Chem.
Soc., 1995, 117, 4468.
12 E. Erdik, Tetrahedron, 1987, 43, 2203.
13 M. V. George and H. Gilman, J. Am. Chem. Soc., 1959, 81, 3288.
14 F. P. Price, J. Am. Chem. Soc., 1947, 69, 2600; H. Gilman,
B. Hofferth, H. W. Melvin and G. E. Dunn, J. Am. Chem. Soc., 1950,
72, 5767; H. Gilman and G. E. Dunn, J. Am. Chem. Soc., 1951, 73,
3404; G. Schott and C. Harzdorf, Z. Anorg. Allg. Chem., 1960, 307,
105.
integrals of the peaks at δ 4.95 (1 H, m, 15 ᎐CH) and δ 4.37
᎐
(1 H, m, 16 ᎐CH), until none of the starting material was visible
᎐
(approx. 2.5 h with a half-life of ca. 20 min). Iodomethane (0.2
cm3, excess) was added, the sample diluted with ether (20 cm3),
the organic layer washed with sodium metabisulfite solution
(10%, 10 cm3) and dried (MgSO4). GC (a known amount of
benzyl alcohol standard) showed 2-methylcyclohexanone (74%)
and 2-tert-butyl-1-phenyl-1,1,2,2-tetramethyldisilane (99%).
Enolate 16; δH(250 MHz; [2H8]THF–[2H6]benzene, 85:15)
15 H. Gilman and B. J. Gaj, J. Org. Chem., 1961, 26, 2471.
16 K. Krohn and K. Khanbabaee, Angew. Chem., Int. Ed. Engl., 1994,
33, 99. See also G. Wickham, H. A. Olszowy and W. Kitching,
J. Org. Chem., 1982, 47, 3788 and P. F. Hudrlik, M. A. Waugh and
A. M. Hudrlik, J. Organomet. Chem., 1984, 271, 69.
17 W. C. Still, J. Org. Chem., 1976, 41, 3063.
18 R. Balasubramanian and J. P. Oliver, J. Organomet. Chem., 1980,
197, C7.
19 D. Wittenberg, T. C. Wu and H. Gilman, J. Org. Chem., 1959, 24,
1349.
4.37 (1 H, m, ᎐CH) and 2.2–1.5 (8 H, m, cyclohexyl CH ). The
᎐
2
disilane was identical (GC, 1H NMR spectroscopy) to an
authentic sample.
20 G. Stork and P. F. Hudrlik, J. Am. Chem. Soc., 1968, 90, 4464.
21 R. Calas, A. Marchand, E. Frainnet and P. Gerval, Bull. Soc. Chim.
Fr., 1968, 2478.
Acknowledgements
22 J. F. Hyde, O. K. Johannson, W. H. Daudt, R. F. Fleming, H. B.
Laudenslager and M. P. Roche, J. Am. Chem. Soc., 1953, 75, 5615.
23 J.-C. Bonnet and E. Marechal, Bull. Soc. Chim. Fr., 1972, 3561.
24 T. I. Ponomareva, T. A. Krasovskaya and M. V. Sobolevskii, Plast.
Massy., 1963, 22; Chem. Abstr., 1963, 59, 14 017b.
25 K. E. Ruehl and K. Matyjaszewski, J. Organomet. Chem., 1991, 410,
1.
26 S. B. Nagelberg, C. E. Reinhold, B. R. Willeford, M. P. Bigwood,
K. C. Molloy and J. J. Zuckerman, Organometallics, 1982, 1, 851.
27 Dictionary of Organic Compounds, 5th edn., Chapman and Hall,
N. Y., 1982, D-08081.
We thank the EPSRC and Zeneca Agrochemicals for a CASE
award (R. S. R.).
References
1 D. J. Ager, I. Fleming and S. K. Patel, J. Chem. Soc., Perkin Trans. 1,
1981, 2520; I. Fleming and D. Marchi, Synthesis, 1981, 560;
I. Fleming, T. W. Newton and F. Roessler, J. Chem. Soc., Perkin
Trans. 1, 1981, 2527; I. Fleming, M. Rowley, P. Cuadrado, A. M.
González-Nogal and F. J. Pulido, Tetrahedron, 1989, 45, 413;
I. Fleming, D. Higgins, N. J. Lawrence and A. P. Thomas, J. Chem.
Soc., Perkin Trans. 1, 1992, 3331; I. Fleming, in Organocopper
Reagents: A Practical Approach, ed. R. J. K. Taylor, OUP, Oxford,
1995, ch. 12, pp. 257–292.
28 A. Holt, A. W. P. Jarvie and J. J. Mallabar, J. Organomet. Chem.,
1973, 59, 141.
29 I. Kopka and M. W. Rathke, J. Org. Chem., 1981, 46, 3771.
2 D. Wittenberg and H. Gilman, Quart. Rev., 1959, 13, 116; M. V.
George, D. J. Peterson and H. Gilman, J. Am. Chem. Soc., 1960, 82,
403; H. Gilman, R. A. Klein and H. J. S. Winkler, J. Org. Chem.,
1961, 26, 2474.
Paper 7/09111G
Received 22nd December 1997
Accepted 5th February 1998
1214
J. Chem. Soc., Perkin Trans. 1, 1998