2
J.H. Docherty et al. / Tetrahedron xxx (xxxx) xxx
Scheme 1. Utility of silane reagents in modern synthetic chemistry. a Metal-catalysed hydrosilylation of alkene or alkynes. b Reductive functionalisation of alkenes by hydrogen
atom transfer. c Ligand-redistribution reaction of phenylsilane to diphenylsilane, silane and triphenylsilane. d Use of ligand-redistribution as masked hazardous silanes.
2. Results and discussion
catalysed hydrosilylation suggested that the presence of an elec-
trophile, or electrophilic metal complex, may hinder, or retard, the
formation of these hazardous species. We therefore investigated
whether common electrophiles could inhibit the ligand-
redistribution reaction (Fig. 2). By adding the electrophile after
the nucleophile, the initiator of hydrosilane redistribution, any
subsequent reactivity could be monitored to assess whether ligand-
redistribution had stalled. In separate reactions, the ligand-
redistribution of phenylsilane initiated by NaOtBu was monitored
by 1H NMR to which an electrophile was added after 10 min. The
standard reaction, without any added electrophile, displayed ca.
40% conversion after 10 min and continued to a plateau at ca. 84%
conversion after 7 h. However, when benzaldehyde was added,
10 min after initiation, the ligand-redistribution reaction was
completely retarded and the conversion of phenylsilane at each
subsequent time point was equivalent to that observed at 10 min
(Fig. 2, ).
The addition of ethyl benzoate produced the same observation,
with complete cessation of ligand-redistribution (Fig. 2, ). The
addition of either a cobalt(II) dichloride complex (EtBIPCoCl2) or
iron(II) dichloride complex (EtBIPFeCl2) proved similarly effective at
inhibiting the ligand-redistribution reaction (Fig. 2, , ) [24]. Addi-
tion of acetonitrile showed some retardation activity, with a
significantly lower observed rate of reaction when compared to the
control reaction with no added acceptor (Fig. 2, ).
To gain insight into these processes we opted to assess the effect
of the nucleophile on the rate of the reaction. Specifically we
questioned if a variation of the nucleophile loading showed any rate
change in the ligand-redistribution reaction of phenylsilane. Using
sodium tert-butoxide as the nucleophile we monitored the reaction
progress using different loadings (2.5, 3.75 and 5 mol%, Fig. 3, a).
The kinetic data obtained showed similar patterns of reactivity to
that obtained when using other nucleophiles (Fig. 1). On changing
the loading, or concentration, of sodium tert-butoxide there was a
slight but noticeable change in the rate of reactions whereby
increased loading of sodium tert-butoxide led to increased rates of
ligand-redistribution. Taking these results we could use the
These precedents prompted us to consider the factors important
to hydrosilane ligand-redistribution reactions. One important
question was whether any generic nucleophile could be used to
initiate ligand-redistribution. In answering this possibility we
selected phenylsilane as a test substrate in combination with
various nucleophiles (Fig. 1). Using 1H NMR, the reaction progress
of the ligand-redistribution reaction could be directly monitored.
Lithium tert-butoxide proved to be a slow initiator for the reaction,
only achieving 10% conversion after 13 h. Lithium aluminium hy-
dride displayed significantly better reactivity, and steadily con-
verted phenylsilane to a mixture of diphenylsilane, silane and
triphenylsilane over the course of ca. 17 h to 80% total conversion.
Hexamethyldisilazide (HMDS) anions of both sodium and potas-
sium reacted faster than both previous reagents, but achieved less
overall conversion, plateauing at ca. 70%. Sodium- and potassium
tert-butoxide demonstrated the best reactivity towards the ligand-
redistribution reaction, both showing the fastest rate of reaction
and highest conversion. Reaction profiling of all components of the
reaction mixture containing phenylsilane and sodium tert-butoxide
(5 mol%) showed concomitant consumption of phenylsilane and
formation of silane and diphenylsilane. The latter is subsequently
consumed as triphenylsilane is formed (Fig. 1, bottom).
Additionally we tested a range of alkoxy-hydrosilane reagents
in combination with sodium tert-butoxide to both assess the rate
of ligand-redistribution and identify the products of these
reactions (Scheme 2). Methyldiethoxysilane underwent ligand-
redistribution to quickly generate methylsilane and methyl-
triethoxysilane. The analogous dimethylethoxysilane reacted to
form dimethylsilane and dimethyldiethoxysilane. Triethoxysilane
rapidly underwent redistribution to form silane and tetraethox-
ysilane. Tetramethyldisiloxane (TMDS) reacted significantly slower
to give a complex mixture of products, as observed by 1H and 29Si
NMR, with dimethylsilane being one component of the mixture.
1,1,1,3,5,5,5-Heptamethyltrisiloxane (MD’M) slowly decomposed
to an unidentifiable mixture of silicon containing products.
The prevailing reactivity across all alkoxy-hydrosilanes examined
is suggestive of a general pathway towards the entropically
favoured gaseous alkyl-hydrosilanes (or silane). The general reac-
tivity trend observed was: (EtO)3SiH > Me(EtO)2SiH > Me2(EtO)
SiH > PhSiH3 > Me2SiHOSiHMe2 > (Me3SiO)2MeSiH.
graphical variable time normalisation analysis method (VTNA) re-
t
ꢀ
ported by Bures to determine the reaction order in [NaO Bu] [25].
The normalized time scale directly uses reaction profile kinetics
adjusted for t[NaOtBu]n on the x-axis, where t represents time and n
is the order in sodium tert-butoxide. When n ¼ 0.5, overlay be-
tween all reaction profiles was found suggesting that the order with
The absence of a build-up of alkyl-hydrosilane gases in metal-
Please cite this article as: J.H. Docherty et al., Nucleophile induced ligand rearrangement reactions of alkoxy- and arylsilanes, Tetrahedron,