Steady state and time-resolved spectroscopic studies of the photochemistry of 1-arylsilacyclobutanes and the chemistry of 1-arylsilenes

Article abstract of DOI:10.1139/v99-114

Direct photolysis of 1-phenylsilacyclobutane and 1-phenyl-, 1-(2-phenylethynyl)-, and 1-(4'-biphenylyl)-1-methylsilacyclobutane in hexane solution leads to the formation of ethylene and the corresponding 1-arylsilenes, which have been trapped by photolysis in the presence of methanol. Quantum yields for photolysis of the three methyl-substituted compounds have been determined to be 0.04, 0.26, and 0.29, respectively, using the photolysis of 1,1-diphenylsilacyclobutane (Φ(silene) = 0.21) as the actinometer. The corresponding silenes have been detected by laser flash photolysis; they have lifetimes of several microseconds, exhibit UV absorption maxima ranging from 315 to 330 nm, and react with methanol with rate constants on the order of (2-5) x 10⁹ M^-1 s^-1 in hexane. Absolute rate constants for reaction of 1-phenylsilene and 1-methyl-1-phenylsilene with water, methanol, tert-butanol, and acetic acid in acetonitrile solution have been determined, and are compared to those of 1,1-diphenylsilene under the same conditions. With the phenylethynyl- and biphenyl-substituted methylsilacyclobutanes, the triplet states can also be detected by laser flash photolysis, and are shown to not be involved in silene formation on the basis of triplet sensitization and (or) quenching experiments. Fluorescence emission spectra and singlet lifetimes have been determined for the three 1-aryl-1-methylsilacyclobutanes, 1,1-diphenylsilacyclobutane, and a series of acyclic arylmethylsilane model compounds. These data, along with the reaction quantum yields, allow estimates to be made of the rate constants for the excited singlet state reaction responsible for silene formation. 1-Methyl-1-phenylsilacyclobutane undergoes reaction from its lowest excited singlet state with a rate constant 10-80 times lower than those of the other three derivatives. The results are consistent with a stepwise mechanism for silene formation, involving a 1,4-biradicaloid intermediate that partitions between product and starting material.

Full text of DOI:10.1139/v99-114

1136
Steady state and time-resolved spectroscopic
studies of the photochemistry of
1-arylsilacyclobutanes and the chemistry of
1-arylsilenes
William J. Leigh, Rabah Boukherroub, Christine J. Bradaric,
Christine C. Cserti, and Jennifer M. Schmeisser
Abstract: Direct photolysis of 1-phenylsilacyclobutane and 1-phenyl-, 1-(2-phenylethynyl)-, and 1-(4′-biphenylyl)-1-
methylsilacyclobutane in hexane solution leads to the formation of ethylene and the corresponding 1-arylsilenes, which
have been trapped by photolysis in the presence of methanol. Quantum yields for photolysis of the three methyl-
substituted compounds have been determined to be 0.04, 0.26, and 0.29, respectively, using the photolysis of 1,1-
diphenylsilacyclobutane (Φsilene = 0.21) as the actinometer. The corresponding silenes have been detected by laser flash
photolysis; they have lifetimes of several microseconds, exhibit UV absorption maxima ranging from 315 to 330 nm,
and react with methanol with rate constants on the order of (2–5) × 109 M–1 s–1 in hexane. Absolute rate constants for
reaction of 1-phenylsilene and 1-methyl-1-phenylsilene with water, methanol, tert-butanol, and acetic acid in
acetonitrile solution have been determined, and are compared to those of 1,1-diphenylsilene under the same conditions.
With the phenylethynyl- and biphenyl-substituted methylsilacyclobutanes, the triplet states can also be detected by laser
flash photolysis, and are shown to not be involved in silene formation on the basis of triplet sensitization and (or)
quenching experiments. Fluorescence emission spectra and singlet lifetimes have been determined for the three 1-aryl-
1-methylsilacyclobutanes, 1,1-diphenylsilacyclobutane, and a series of acyclic arylmethylsilane model compounds.
These data, along with the reaction quantum yields, allow estimates to be made of the rate constants for the excited
singlet state reaction responsible for silene formation. 1-Methyl-1-phenylsilacyclobutane undergoes reaction from its
lowest excited singlet state with a rate constant 10–80 times lower than those of the other three derivatives. The results
are consistent with a stepwise mechanism for silene formation, involving a 1,4-biradicaloid intermediate that partitions
between product and starting material.
Key words: silene, silacyclobutane, photochemistry, biradical. 1147
Résumé : La photolyse directe du 1-phénylsilacyclobutane et des 1-phényl-, 1-(2-phényléthynyl)- et 1-(4′-biphénylyl)-1-
méthylsilacyclobutane, en solution dans l’hexane, conduit à la formation d’éthylène et des 1-arylsilènes correspondants
que l’on a piégés par photolyse en présence de méthanol. On a déterminé les rendements quantiques pour les
photolyses des trois composés portant un groupe méthyle; si l’on utilise la photolyse du 1,1-diphénylcyclobutane
(Φsilène = 0,21) comme actinomètre, les valeurs sont respectivement 0,04, 0,26 et 0,29. On a détecté les silènes
correspondants à l’aide de la photolyse éclair au laser; leurs temps de vie sont de plusieurs microsecondes, ils
présentent des maxima d’absorption UV à des valeurs allant de 315 à 330 nm et ils réagissent avec le méthanol avec
des constantes de vitesse de l’ordre de (2–5) × 109 M–1 s–1, dans l’hexane. On a déterminé les constantes de vitesse
absolue pour la réaction du 1-phénylsilène et du 1-méthyl-1-phénylsilène avec l’eau, le méthanol, l’alcool tert-butylique
et l’acide acétique, en solution dans l’acétonitrile, et on les a comparées à celles du 1,1-diphénylsilène dans les mêmes
conditions. Avec les méthylsilacyclobutanes substitués par des groupes phényléthynyle et biphényle, la photolyse éclair
au laser permet aussi de détecter les états triplets et, en se basant sur des expériences de sensibilisation de triplets et/ou
de désactivation, on a démontré qu’ils ne sont pas impliqués dans la formation du silène. On a déterminé les spectres
d’émission de fluorescence et les temps de vie des singulets des trois 1-aryl-1-méthylsilacyclobutanes,du 1,1-
diphénylsilacyclobutane et d’une série de composés arylméthylsilanes acycliques modèles. Ces données utilisées avec
les rendements quantiques des réactions ont permis d’évaluer les constantes de vitesse de réaction de l’état singulet
excité responsable de la formation du silène. La vitesse de réaction du 1-méthyl-1-phénylsilacyclobutane à partir de son
Received January 20, 1999.
This paper is dedicated to Jerry Kresge in recognition of his many achievements in chemistry.
W.J. Leigh,1 R. Boukherroub, C.J. Bradaric, C.C. Cserti, and J.M. Schmeisser.2 Department of Chemistry, McMaster
University, 1280 Main Street West, Hamilton, ON L8S 4M1, Canada.
1Author to whom correspondence may be addressed. Fax: (905) 522-2509. e-mail: leigh@mcmaster.ca
21998 Reactive Intermediate Student Exchange (RISE) Scholar.
Can. J. Chem. 77: 1136–1147 (1999)
© 1999 NRC Canada

Leigh et al.
1137
état singulet excité le plus bas est de 10 à 80 fois plus rapide que les vitesses des trois autres dérivés. Ces résultats
sont en accord avec un mécanisme par lequel la formation du silène se ferait par étapes impliquant un intermédiaire
1,4-biradicaloïde qui se répartirait entre le produit de la réaction et le produit de départ.
Mots clés : silène, silacyclobutane, photochimie, biradical.
[Traduit par la Rédaction]
two early papers on the photochemistry of 2-phenyl-
substituted silacyclobutanes (4, 25).
Introduction
Leigh et al.
Silacyclobutanes are well known to yield products consis-
tent with the formation of silene-reactive intermediates upon
photolysis in solution or the gas phase, or upon pyrolysis at
high temperatures (1–8). These compounds have proven to
be tremendously useful as photochemical precursors for
transient silenes, in experiments aimed at detecting the
silenes directly and studying the kinetics and mechanisms of
their reactions with alcohols and other “silenophiles” (9–17).
Photochemical silene formation proceeds with useful effi-
ciencies from silacyclobutanes that bear alkyl, vinyl, ethy-
nyl, silyl, and phenyl substituents at silicon, and possess UV
absorption maxima between 190 and 250 nm (2, 4, 12, 13,
17–20). In contrast, alkoxy substitution either directly at sili-
con or on the aromatic rings of phenylated derivatives ren-
ders the silacyclobutane moiety inert to photocycloreversion
(13, 17).
Several details of the mechanism for the photochemical
cycloreversion of silacyclobutanes are known, at least for
1,1-dialkyl-substituted derivatives that absorb below
~230 nm. The reaction is initiated from the lowest (σ,σ*) ex-
cited singlet state, by cleavage of one of the ring Si—C
bonds to form a biradicaloid intermediate that cleaves to
silene and alkene, recloses to starting material, and under-
goes intramolecular disproportionation if an alkyl substituent
is present at C2 (20). The presence of substituents at C2 of
the silacyclobutane ring directs the regiochemistry of the ini-
tial cleavage toward the Si—C2 bond, leading to an interme-
diate whose lifetime is long enough to allow bond rotations
that lead to partial scrambling of the stereochemistry at this
carbon (20). This mechanism contrasts that of the thermal
cleavage of silacyclobutanes, which has been shown to pro-
ceed by a stepwise mechanism involving initial C2—C3
bond cleavage (21–24).
The situation is less clear with 1-phenyl-substituted deriv-
atives, in which the phenyl substituent is the primary
chromophore. 1,1-Diphenylsilacyclobutane (1) affords 1,1-
diphenylsilene (2; eq. [1]) with reasonable efficiency
(Φ2 = 0.21), and time-resolved experiments have shown that
silene formation occurs on the nanosecond time scale and
cannot be quenched by saturation of the solution with oxy-
gen or other triplet quenchers (11). The reactive excited state
is thus most likely the lowest excited singlet state, which in
these cases is predominantly of benzenoid π,π* character.
For efficient cleavage to occur, however, the ring Si—C
bonds in the reactive excited state must obviously be weak-
ened compared to the situation in the ground state, so the
lowest benzenoid excited singlet state is presumably mixed
with σ,σ*, σ,π* or π,σ* configurations that activate the ring
toward cleavage. Whether the reaction occurs by a concerted
or stepwise mechanism has not been conclusively estab-
lished, although the question has in fact been addressed in
hν
Ph Si
2
Ph Si CH
Ph Si CH
+
[1]
C2H4
2
2
2
2
hexane
2
1
?
Φ = 0.21
In this paper, we report the photochemistry of 1-phenyl-
silacyclobutane (3) and the series of 1-methyl-1-arylsila-
cyclobutane derivatives 4–6, in which the aryl substituent is
varied between phenyl, phenylethynyl, and biphenyl. The
main goal of this work was to determine how the
photophysical properties of the chromophore affect the rates
and efficiencies of silene formation in arylsilacyclobutane
photolysis, and hence to better define the scope and limita-
tions of silacyclobutane photolysis as a source of reactive
silenes. It seemed likely to us that the phenylacetylene and
biphenyl chromophores in 5 and 6 would allow study of the
photochemistry and photophysics of arylsilacyclobutanes in
CH3
Si
H
Si
3
4
CH3
CH3
Si
C C Si
5
6
much greater detail than is convenient with simple phenyl
derivatives, since these chromophores generally fluoresce
strongly and promote efficient intersystem crossing to the
lowest triplet states, which are usually detectable by laser
flash photolysis. Thus, we have studied the photochemistry
of the three 1-aryl-1-methylsilacyclobutanes (4–6) in detail,
using steady state photolysis and time-resolved UV absorp-
tion and fluorescence emission spectroscopic techniques.
Photolysis of all four compounds in hydrocarbon solvents
yields the corresponding silenes (7–10), which have been
detected directly by laser flash photolysis methods and char-
acterized on the basis of their reactivity toward methanol
and tert-butanol. Absolute rate constants for the reactions of
7 and 8 with the two alcohols, water, and acetic acid have
been determined in acetonitrile solution, and are compared
to those for 1,1-diphenylsilene (2) under similar conditions.
Fluorescence spectra and lifetimes have also been deter-
mined for 1 and the acyclic arylsilanes 11–14, to provide a
more complete framework with which to understand the
© 1999 NRC Canada

1138
Can. J. Chem. Vol. 77, 1999
H3C
H3C
Si CH2
H3C
Si CH2
H
C
C
Si CH2
Si CH2
7
8
9
10
SiMe3
C C SiMe3
12
SiMe3
Ph2SiMe2
11
13
14
variations in photochemistry and photophysics of these
arylsilacyclobutanes as a function of substituent.
≈350 nm. The absorptions decay on the microsecond time
scale with mixed first- and second-order kinetics to within
10% of the prepulse level, and exhibit lifetimes in the 2–3 µs
range in scrupulously dried solvent. Time-resolved UV ab-
sorption spectra were recorded for air-saturated solutions of
the two compounds, 0.1–0.4 µs after the laser pulse. The
spectra recorded in hexane solution are shown in Fig. 1a, b
along with typical decay traces, recorded at a monitoring
wavelength of 315 nm. The lifetimes and intensities of the
signals are unaffected by saturation of the solutions with ei-
ther nitrogen or oxygen.
Results
Silacyclobutanes 5 and 6 were prepared by reaction of
1-chloro-1-methylsilacyclobutane with phenylethynyllithium
and 4-biphenylmagnesium bromide, respectively, using a
procedure similar to those reported previously for the syn-
thesis of 3 (26) and 4 (3). All four compounds were purified
to >99% (GC) purity by vacuum distillation (3–5) or re-
peated recrystallization (6).
Addition of water, methanol, tert-butanol, or acetic acid
((0.5–5) × 10–3 M) results in a reduction in the lifetimes of
both transients and a change in decay kinetics to pseudo first
order. Plots of the rate constant for decay (kdecay) versus con-
centration of added quencher according to eq. [3] (where k0
is the first-order rate constant for the decay of the transient
in the absence of added quencher and kROH is the second-
order quenching rate constant) were linear in all cases. Bi-
molecular rate constants for reaction of 3 and 4 with the four
reagents in acetonitrile solution are listed in Table 1, while
those for reaction with MeOH and t-BuOH in hexane are
listed in Table 2. Similar experiments were carried out using
MeOD, t-BuOD, and acetic acid-d in MeCN solution; the
kH/kD values calculated from these rate constants and those
for reaction of the corresponding protiated reagents are in-
cluded in Table 1. Previously reported rate constants for re-
action of 2 with these reagents in acetonitrile are included
for comparison.
Photolysis (254 nm) of the four compounds as deoxy-
genated 0.04 M solutions in hexane containing methanol
(0.05 M) led to the formation of ethylene and the meth-
oxysilane 15 (eq. [2]). The methoxysilanes 15a–15c were
identified by GC–MS and GC coinjection with independ-
ently prepared authentic samples, while 15d was isolated
1
and identified on the basis of its H NMR and mass spectra.
No other products could be detected in the crude photo-
lysates by GC analysis, up to ~20% conversion of the sila-
cyclobutane, within the detection limits of our GC method.
R
R
hν
hexane
MeOH
Ar Si
C2H4
+
[2]
Ar Si CH
OMe
3
3-6
15
[3]
kdecay = k0 + kROH[Q]
a. R = H; Ar = Ph
b. R = Me; Ar = Ph
Laser flash photolysis of deoxygenated hexane solutions
of the phenylethynyl derivative 5 (2 × 10–5 M) led to com-
plex behavior that could be shown to result from competing
formation of the silacyclobutane triplet state and silene 9,
whose absorption spectra overlap almost exactly but differ
considerably in intensity. Absorptions due to the triplet dom-
inate the transient behavior in deoxygenated hexane solu-
tion. For example, flash photolysis of a deoxygenated
solution of 5 gave rise to a long-lived (τ ~ 5 µs), strongly ab-
sorbing, transient with λmax = 315 nm. The lifetime of the
transient was reduced substantially upon addition of oxygen
or 1,3-octadiene, but was apparently unaffected upon addi-
tion of 0.005 M methanol. Monitoring at 260 nm, within the
long-wavelength absorption band of the silacyclobutane,
gave rise to a bleaching signal that recovered almost to the
prepulse level with the same lifetime as that of the transient
≡
c. R = Me; Ar = C C-Ph
d. R = Me; Ar = 4-C6H4-C6H5
Photolysis (350 nm) of a deoxygenated 0.05 M solution of
benzophenone in hexane containing 0.01 M 6 and 0.1 M
MeOH led to the slow formation of benzhydrol (identified
by MS), but no destruction of the silacyclobutane.
Methoxysilane 15d could not be detected in the mixture,
within the limits of our GC method.
Nanosecond laser flash photolysis (λex = 248 nm) of con-
tinuously flowing, air-saturated solutions of 3 and 4 (~0.014
M) in hexane or acetonitrile leads to weak, but readily de-
tectable, transient absorptions in the spectral region below
© 1999 NRC Canada

Leigh et al.
1139
Fig. 1. Time-resolved UV absorption spectra of silenes 7 (a), 8 (b), 9 (c), and 10 (d) in hexane solution at 23°C. The spectra of 9 and 10
were recorded in the presence of ~0.008 M 1,3-octadiene to shorten the lifetime of the corresponding silacyclobutane triplet to τT <20 ns.
∆-OD
0.03
∆-OD
(b)
315-nm
(a)
315-nm
0.01
0.02
0.01
0
0
5
10
µs
0
5
10
15
µs
0.005
0
250
300
350
400
450
500
250
300
350
400
450
500
Wavelength (nm)
Wavelength (nm)
∆-OD
0.03
0.02
0.01
0
∆-OD
0.008
0.006
0.004
0.002
0
(c)
320-nm
330-nm
(d)
0
5
10
15
µs
0
0.5
1
µs
250
300
350
400
450
500
250
300
350
400
450
500
Wavelength (nm)
Wavelength (nm)
Table 1. Bimolecular rate constants for reaction of substituted 1-phenylsilenes 7, 8, and 2 with H2O, MeOH(D), t-BuOH(D), and
AcOH(D) in air-saturated acetonitrile solution at 23°C.a
Ph
Si CH2
H2O
kq / 109 M–1 s–1 kq / 109 M–1 s–1 kH/kD
MeOH
t-BuOH
kq / 109 M–1 s–1
HOAc
kq / 109 M–1 s–1
R
R
kH/kD
kH/kD
H (7)
Me (8)
Ph (2)b
0.72 ± 0.01
0.56 ± 0.01
0.76 ± 0.09
1.63 ± 0.08
1.16 ± 0.02
1.2 ± 0.1
1.3 ± 0.1
1.6 ± 0.1
1.5 ± 0.2
0.67 ± 0.04
0.20 ± 0.09
0.22 ± 0.02
1.5 ± 0.1
1.4 ± 0.1
1.6 ± 0.1
1.85 ± 0.08
1.55 ± 0.04
1.23 ± 0.07
1.1 ± 0.1
1.0 ± 0.1
1.1 ± 0.1
aErrors are listed as twice the standard deviation from least-squares analysis of kdecay vs. concentration data according to eq. [3].
bData from ref. 11.
Table 2. Time-resolved UV absorption spectra and bimolecular rate constants for reaction of substituted 1-phenylsilenes 7–10 and 2
with methanol and tert-butanol in hexane solution at 23°C.a
R
Si CH2
λmax (nm)
kMeOH / 109 M–1 s–1
kt-BuOH / 109 M–1 s–1
Ar
R
Ar
7
8
Ph
Ph
Ph
PhCϵC
4-PhC6H4
H
315
315
325
320
330
3.1 ± 0.1
3.2 ± 0.2
1.9 ± 0.2
5 ± 2
1.2 ± 0.1
1.2 ± 0.1
0.40 ± 0.07
—
Me
Ph
Me
Me
2b
9
10
2.9 ± 0.3
—
aMeasured by laser flash photolysis. Errors in the rate constants are listed as twice the standard deviation from least-squares analysis of kdecay vs.
concentration data according to eq. [3].
bData from ref. 11.
© 1999 NRC Canada

1140
Can. J. Chem. Vol. 77, 1999
Fig. 2. Triplet–triplet absorption spectra for silacyclobutanes 5 (a) and 6 (b) in deoxygenated hexane solution containing 0.03 M
methanol at 23°C.
∆-OD
∆-OD
(a)
(b)
0.2
0.04
0.02
0
0.08
0.08
0.06
0.04
0.02
0
0.15
0.1
0.05
0
0.06
0.04
0.02
0
0
10 20 30
µs
0
20
40 60
µs
250
300
350
400
450
500
300 350 400 450 500 550 600
Wavelength (nm)
Wavelength (nm)
absorption at 315 nm. The lifetimes of the absorption and
bleaching signals were also identical (τ ~ 500 ns) in the
presence of air, but differed in oxygen-saturated solution,
where the 315 nm transient decay exhibited an average life-
time of τ ~ 240 ns and the bleaching signal at 260 nm recov-
ered with a clean pseudo-first-order lifetime τ ~ 100 ns.
Different behavior was noted upon addition of 0.0027 M
1,3-octadiene, where the bleaching signal was unresolvable
from the laser pulse and the 315 nm transient decayed with a
lifetime of τ ~ 600 ns, was much weaker in intensity, and
was superimposed on a strong fluorescence signal.
The kinetic behavior of the bleaching signal is consistent
with the formation of a transient whose decay mainly regen-
erates the precursor, as would be expected for a triplet state.
We thus assign the transient absorption spectrum recorded
under deoxygenated conditions as consisting largely of the
triplet–triplet absorption spectrum of silacyclobutane 5, su-
perimposed on a second, more weakly absorbing, transient
(assigned to the silene 9, vide infra) that is resolvable only
under conditions where the triplet is strongly quenched. The
spectrum of the triplet, recorded in deoxygenated hexane so-
lution containing 0.03 M methanol to remove contributions
from the silene (Fig. 2a), exhibits λmax = 315 nm, in the
range expected for a phenylacetylene triplet (27, 28). The
rate constant for quenching of the triplet by 1,3-octadiene in
hexane was found to be (9.9 ± 0.6) × 109 M–1 s–1, which is
also consistent with its assignment. A rate constant for trip-
let quenching by oxygen was estimated as kq ~ 6 × 108 M–1 s–1
from the rates of recovery of the precursor bleaching signal
in air- and oxygen-saturated solution.
diene (e.g., 0.003 M) led to no improvement in accuracy be-
cause the silene signal was weaker due to screening of the
excitation light by the diene. The fact that the silene is ob-
servable under conditions where the triplet lifetime is re-
duced to <25 ns indicates that the silene is formed from the
lowest excited singlet state of the silacyclobutane, and not
from the triplet.
Similar results were obtained in laser flash photolysis ex-
periments with the 4′-biphenyl-substituted derivative 6. Ac-
cordingly, flash photolysis of deoxygenated 6 × 10–5
M
solutions of 6 in hexane containing 0.03 M MeOH gave rise
to a long-lived (τ = 12.5 µs) transient that decayed with
pseudo-first-order kinetics at very low laser intensities, and
exhibited λmax = 375 nm (Fig. 2b). This transient is assigned
to the triplet state of 6 on the basis of the similarity of its
spectrum to that of the 4-trimethylsilyl-4′-biphenyl (13) trip-
let (λmax = 375 nm under similar conditions) and the fact that
it is quenched by 1,3-octadiene with a rate constant kq = (4.9
± 0.5) × 109 M–1 s–1. Alternatively, flash photolysis of air-
saturated solutions of 6 in hexane containing 0.008 M 1,3-
octadiene gave rise to a transient that decayed with clean
pseudo-first-order kinetics (τ = 2.1 µs) and exhibited λmax
=
330 nm. The species proved to be sensitive to the presence
of methanol; a plot of kdecay vs. [MeOH] according to eq. [2]
was linear, and afforded a slope kMeOH = (2.9 ± 0.3) × 109
M–1 s–1. On the basis of these data, this transient is assigned
to silene 10.
The time-resolved UV absorption spectra of silenes 9 and
10 in air-saturated hexane solution containing 0.008 M 1,3-
octadiene are shown in Figs. 2c,d. The bimolecular rate con-
stants for their reaction with MeOH under the same condi-
tions are summarized in Table 2, along with the
corresponding data for 7, 8 and 1,1-diphenylsilene (2) (11).
Quenching by t-BuOH was not investigated in these two
cases.
Quantum yields for disappearance of 4–6 were determined
by 254 nm merry-go-round photolysis of deoxygenated
0.03–0.04 M solutions in hexane containing methanol (0.05
M), using 1,1-diphenylsilacyclobutane (1) as actinometer.
Photolysis of 1 in methanolic hexane solution affords ethyl-
ene and diphenylmethoxymethylsilane (16) as the only de-
tectable products (2, 4, 11) — the latter due to trapping of
In hexane solution containing 0.001 M 1,3-octadiene or
saturated with oxygen, the lifetime of the residual transient
absorption at 315 nm is reduced upon addition of methanol.
The spectrum of the transient observed under these condi-
tions exhibits λmax = 320 ± 2 nm (Fig. 1c), and can thus be
most reasonably assigned to 1-methyl-1-phenylethynylsilene
(9). The transient is quenched efficiently by MeOH, though
unfortunately we were able to obtain only a qualitative esti-
mate of the rate constant (kMeOH = (5 ± 2) × 109 M–1 s–1),
because of interference from the triplet absorptions and sam-
ple fluorescence at higher alcohol concentrations. Experi-
ments with solutions containing higher concentrations of
© 1999 NRC Canada

Leigh et al.
1141
Table 3. Reaction quantum yields (Φsil),a fluorescence emission maxima, and singlet lifetimes for 1-aryl-1-methylsilacyclobutanes 4–6,
1,1-diphenylsilacyclobutane (1), and the corresponding model compounds 11–14, determined in deoxygenated isooctane solution at
23°C. Estimates of excited state reaction rate constants and state efficiencies for formation and reaction of biradicaloid intermediates
are also included.
Compound
Φsil
λF (nm)b
τS (ns)b,c
ksil/107 s–1 d
kBIR/107 s–1 e
φ
φ
sil
e
f
BIR
4
0.04 ± 0.02
284
284
300
300
310
310
285
285
30 ± 2
38 ± 3
0.13 ± 0.07
—
3.9 ± 0.9
—
4.1 ± 0.5
—
8.1 ± 0.9
—
0.7 ± 0.4
—
10.2 ± 0.8
—
6.7 ± 0.5
—
9.0 ± 0.9
—
0.21
0.67
0.47
0.23
0.19
0.38
0.61
0.90
11 (PhSiMe3)
5
12 (PhCCSiMe3)
6
13 (4′-BiphSiMe3)
1
14 (Ph2SiMe2)
—
0.26 ± 0.04
—
0.29 ± 0.03
—
6.6 ± 0.2
20.0 ± 0.1
7.0 ± 0.3
13.2 ± 0.4
2.6 ± 0.1
3.4 ± 0.2
0.21 ± 0.03
—
aFor disappearance of substrate, from photolysis of deoxygenated 0.03–0.04 M solutions of silacyclobutane in hexane containing 0.05 M MeOH, using 1
as actinometer. Errors are listed as the standard deviation from the mean of duplicate determinations.
bRecorded in deoxygenated isooctane solution, at concentrations of 10–5–10–4 M.
cDetermined by time-correlated single photon counting. Errors are listed as ±2σ.
SCB
dCalculated from Φ and τS
(see eq. [5]).
sil
eEstimated from the fluorescence lifetimes of silacyclobutane and corresponding model compounds according to eqs. [8] and [9].
fCalculated from Φ and φBIR according to eq. [7].
sil
the eight compounds matched the static absorption spectra
Fig. 3. Concentration vs. time plots for merry-go-round
photolysis of deoxygenated 0.03–0.04 M solutions of
silacyclobutanes 4 (ٗ), 5 (᭜), 6 (᭝), and 1 (᭹) in hexane
containing methanol (0.05 M) at 25°C.
(in the region above ~240 nm) in each case. Fluorescence
lifetimes were determined for all eight compounds at the
emission maxima by time-correlated single photon counting,
under the same conditions as those employed for the deter-
mination of the steady state fluorescence spectra. In all eight
cases, the fluorescence decay profiles analysed either to
single exponential kinetics or to two exponentials, which
consisted of a prominent (>98%) short-lived decay superim-
posed on a very weak (<2%) long-lived decay whose origin
was not investigated. Identical fluorescence lifetimes were
obtained from solutions of ca. 2-fold higher concentration in
substrate, within experimental error, indicating that, at these
concentrations, self-quenching of fluorescence (due to
intermolecular excimer formation, for example) is insignifi-
cant. The fluorescence emission maxima and singlet life-
times obtained from these experiments are listed in Table 1,
along with various kinetic parameters calculated from the re-
action quantum yields and singlet lifetimes (vide infra).
0.04
0.03
0.02
0.01
0.00
0
10
20
time (min)
1,1-diphenylsilene (2) — and is characterized by a quantum
yield of Φ2 = 0.21 ± 0.03 (11). Quantum yields were calcu-
lated from the relative slopes of plots of concentration vs.
photolysis time, which are shown in Fig. 3. The results of
these experiments are collected in Table 3. Qualitative ob-
servations indicated the quantum yield for photolysis of 3 to
be similar to that of 4 under similar conditions.
Steady state fluorescence emission spectra of 4–6 and 1
were recorded in deoxygenated isooctane solutions at
concentrations in the 10–5–10–4 M range. The fluorescence
spectra of 4–6 were slightly weaker but otherwise indistin-
guishable from those of the corresponding model com-
pounds 11–13 in each case. The fluorescence spectrum of
dimethyldiphenylsilane (14) consisted of a prominent band
centered at 285 nm along with a weaker shoulder centered at
~335 nm that is tentatively attributed to intramolecular
excimer formation. A similar spectrum was observed for 1,
except the long-wavelength shoulder was much weaker than
in the spectrum of 14. The fluorescence excitation spectra of
Discussion
The steady state photochemistry and time-resolved spec-
troscopic behavior of all four of the silacyclobutanes 3–6 are
consistent with the formation of ethylene and the corre-
sponding arylsilenes 7–10 as the exclusive products of
photolysis, within the limits of our (GC and 1H NMR) detec-
tion limits. The four silenes exhibit the expected trends in
lifetime (τ = 2–5 µs), λmax (315–330 nm), and reactivity to-
ward alcohols (kMeOH = (2–5) × 109 M–1 s–1 in hexane at
23°C), considering our previously reported data for 1,1-
diphenylsilene (2) (9, 11). The steady state results reported
here for photolysis of 3 in the presence of methanol are anal-
ogous to those reported previously for the photolysis of this
compound in the presence of other alcohols (26).
As we reported previously for 2 (11), the reactivities of
silenes 7 and 8 are significantly lower in acetonitrile solu-
tion than in hexane, probably due to weak complexation by
the nitrile solvent, and in all cases the rate constants follow
© 1999 NRC Canada

1142
Can. J. Chem. Vol. 77, 1999
the trend kHOAc ~ kMeOH > kH O > kt-BuOH. The reactivities of
oxygen or diene). In fact, the presence of diene is necessary
in order to detect the corresponding silenes (9 and
10) cleanly by laser flash photolysis, because their spectra
overlap severely with those of the triplet states of the precur-
sors. Similarly, benzophenone-sensitization experiments
with 6, at a concentration corresponding roughly to 90% ef-
ficient triplet sensitization, failed to produce detectable
amounts of the methanol trapping product of silene 10.
These experiments conclusively demonstrate that the reac-
tive excited state of arylsilacyclobutanes, with respect to [2
+ 2] photocycloreversion to the corresponding silenes, is the
lowest excited singlet state.
2
the three phenylsilene derivatives with each reagent are re-
markably similar to one another in this solvent, probably be-
cause the small differences in rate that are expected, based
on the electronic effects of the substituents (17), are offset
by opposing steric and solvation effects. The three com-
pounds show a slightly greater spread in reactivity toward
methanol in hexane solution than in acetonitrile, which
might reflect differences in solvation throughout the series
by the nitrile solvent. As we found previously for 2 (11), the
rate constants for reaction with water and the alcohols are
subject to small, but clearly primary, deuterium kinetic iso-
tope effects. This is consistent with a reaction mechanism
involving initial, reversible complexation followed by rate-
controlling proton transfer from oxygen to carbon (eq. [4])
(10, 11, 13, 17, 29–33).
Comparison of the singlet lifetimes of the silacyclobu-
tanes 1 and 4–6 with those of the corresponding model com-
pounds (11–14), in which the silacyclobutane ring carbons
are replaced by two methyl groups, provides additional evi-
dence for the direct involvement of the lowest excited sin-
glet state in silene formation. Compared to the corresponding
silacyclobutanes, the latter four compounds are all essen-
tially non-photoreactive, allowing the conclusion that their
singlet lifetimes are controlled mainly by nonproductive de-
cay processes such as fluorescence, intersystem crossing,
and internal conversion. In fact, the singlet lifetimes of the
three trimethylsilyl-substituted compounds 11–13 are similar
to those of the corresponding methylarenes under similar
conditions (toluene, 34 ns (35); 1-phenylpropyne, 15 ns;3 4′-
methylbiphenyl, 15.2 ns (35)), indicating that trialkylsilyl
substitution has little intrinsic effect on the photophysics of
the arene chromophore in each of these compounds. In con-
trast, the singlet lifetime of dimethyldiphenylsilane (14; τs =
3.4 ns) is substantially shorter than that of diphenylmethane
(τS = 25.3 ns (35)). We tentatively attribute this to intra-
molecular excimer formation on the basis of the long-wave-
length shoulder in the static fluorescence emission spectrum
of the diarylsilane. This weak emission band occurs at a po-
sition characteristic of inter- and intramolecular excimer
emissions in substituted benzene derivatives (36). A similar
band appears in the emission spectrum of 1, though it is
weaker than in that of 14. Diphenylmethane does not form
an intramolecular excimer (35, 36), and consequently its sin-
glet lifetime is similar to that of toluene. In any event, the
singlet lifetimes of the silacyclobutane derivatives are in
each case significantly shorter than those of the correspond-
ing model compounds, which is consistent with the presence
of an additional mode of excited singlet state decay due to
reaction involving the four-membered ring. The alternate
possibility that the shorter singlet lifetimes of the cyclics are
due to an enhancement in one or more of the nonproductive
decay modes cannot, however, be rigorously ruled out. Fluo-
rescence and intersystem crossing could be quantitatively
evaluated through quantum yield measurements but, even
then, possible differences in the rates of nonproductive inter-
nal conversion cannot be easily evaluated.
δ+
R''OH
k1
[4]
R'RSi CH + R''OH
2
R'RSi CH
δ-
k-1
2
k2
R''O
R'RSiCH
3
Laser flash photolysis studies of 1,1-diphenylsilacyclobu-
tane (1) in nitrogen-saturated hexane solution have shown
that 1,1-diphenylsilene (2) is formed within the duration of
the excitation pulse from a KrF excimer laser (~15 ns) (11).
This establishes an upper limit (of the same value) for the
lifetime of the excited state responsible for the formation of
the reactive intermediate. In this case, the identification of
the reactive excited state (i.e., singlet or triplet) by conven-
tional triplet sensitization or quenching experiments is not a
trivial task, and we can rule out the triplet state as the pre-
cursor only providing that its lifetime is in excess of about
10 ns. We have been unsuccessful in our attempts to detect
the triplet state of 1 directly by laser flash photolysis, but
this is not unexpected for a benzene derivative bearing
nonconjugative substituents (27).
The results reported here for 5 and 6 provide strong evi-
dence that the lowest triplet states of 1-arylsilacyclobutanes
are not involved in silene formation. In these two cases, the
triplet states are easily detectable by laser flash photolysis
because intersystem crossing is more efficient in phenyl-
alkynes and biphenyls than in simple benzene derivatives,
and the triplets absorb considerably more strongly (27, 28,
34). The triplet lifetimes of these two molecules were deter-
mined to be ~14 and 12 µs, respectively, in deoxygenated
hexane solution at room temperature, which should be con-
sidered to be lower limits. Quenching of the triplets by oxy-
gen or 1,3-octadiene proceeds rapidly in both cases, but has
no discernible effect on the yield of silene (as measured by
the initial ∆OD of the silene absorption in the presence of
We were surprised to discover that the quantum yield for
silene formation from 1-methyl-1-phenylsilacyclobutane
(4) is ~5 times lower than that for 1,1-diphenylsilacyclo-
butane (1), and up to ~8 times lower than those for 5 and 6,
which have lower excited singlet state energies. This must
be due to a substituent effect on the rate constants for reac-
3 In deoxygenated isooctane at 25°C. We thank T.L. Morkin for this measurement.
© 1999 NRC Canada

Leigh et al.
1143
Scheme 1.
Scheme 2.
*
*
R'
Si
R'
Si
R'
R'
Si
R'
R'
Si
knp
R
hν
knp
R
Si
S0
1/τSSCB = kBIR+knp
R
R
hν
R
Si
R
S1
S0
S0
S0
S1
kcouple
ksil
kBIR
R'
R'
R'
kcleave
1/τSSCB = ksil + knp
Si CH2
C2H4
R
Si
CH2
Si CH2
C2H4
R
R
+
+
BIR
SCB
Φ
sil = kBIRφsilτS
(φsil = kcleave/(kcleave+kcouple))
tive excited state decay and (or) (if silene formation is non-
concerted) the partitioning of the biradicaloid intermediate
between silene and starting material. Scheme 1 depicts the
simplest possible mechanism to account for excited state de-
cay in these compounds, where silene formation occurs di-
rectly from the excited singlet state by concerted
cycloreversion with rate constant ksil, and knp is the sum of
the rate constants for all nonproductive pathways for excited
state decay. The reaction rate constants can be calculated
from the quantum yields and singlet lifetimes according to
eq. [5], and are listed in Table 3. It should be noted that
these are true elementary rate constants only if silene forma-
tion occurs directly from the lowest excited singlet state by a
concerted mechanism.
SCB
mod
[8]
[9]
kBIR 1/
≈
τS
– 1/
τS
SCB
mod
φBIR=kBIR τS
≈ 1 – (τSSCB/τS
)
The ksil and kBIR values summarized in Table 1 both indi-
cate that the excited singlet state of 1-methyl-1-
phenylsilacyclobutane (4) is significantly less reactive to-
ward bond cleavage than those of 1, 5, and 6. The most rea-
sonable explanation for this is that the mixing of benzenoid
π and (or) π* MOs with Si—C σ and (or) σ* MOs in the
lowest excited singlet state, which is required in order to ac-
tivate the silacyclobutane ring toward bond cleavage, is
much less efficient in 4 than in the other three derivatives.
Presumably, the more extended π-systems in 1, 5, and 6 as-
sist in this mixing by perturbing the symmetries of the rele-
vant π-type molecular orbitals.
The calculated φBIR and φ values for the four silacyclo-
butanes also show some intseilresting trends. The data for 1
and 4 suggest that the difference in overall reaction quantum
yield for these two compounds is primarily due to differ-
ences in the behavior of the biradicaloid intermediates, not
to differences in the efficiencies of excited state Si—C bond
SCB
[5]
Φsil = ksil τS
If silene formation occurs by a stepwise mechanism, then
the ksil values calculated above are a more complex function
of the primary rate constant for excited state bond cleavage
to form the biradicaloid intermediate (kBIR) and an efficiency
factor (φ ) that describes its partitioning between silene and
sil
the ground state of the silacyclobutane (Scheme 2 and
eq. [6]). In other words, the overall quantum yield for reac-
tion is given by the product of the state efficiencies for
biradicaloid formation (φBIR = kBIR τSSCB) and for silene for-
cleavage. The approximately equal φ values for the two
BIR
mation from the biradicaloid (φ = kcleave τBIR), as shown in
compounds result from the fact that substituent-induced per-
turbations that affect kBIR also tend to affect the rates of the
nonproductive decay processes: in spite of the fact that 4 is
~10 times less reactive than 1, the intrinsic lifetime of its ex-
cited singlet state (as measured by τSmod) is also ~10 times
longer. In fact, the data indicate that φBIR varies relatively
little throughout the entire series of compounds. On the
sil
eq. [7]. The rigorous determination of these parameters re-
quires additional information, which could be obtained if the
intermediate could be detected either directly or indirectly
via chemical trapping. Unfortunately, no additional transient
absorptions that might be assigned to such species have been
observed on the nanosecond time scale, in these or any of
the other arylsilacyclobutanes that we have studied.
other hand, φ varies smoothly throughout the series, sug-
sil
gesting that the biradicaloid from 1 mainly cleaves (kcleave
9kcouple) while that from 4 mainly couples to regenerate start-
ing material (kcouple ~ 4kcleave).
~
[6]
[7]
ksil=kBIR·φ
sil
Φsil=φBIR·φ
sil
Such a large variation in the relative rate constants for
cleavage and coupling of the putative 1,4-biradicaloid inter-
mediates is surprising. The two main factors that should af-
fect the partitioning of such species between cleavage and
coupling are the relative thermodynamic stabilities of the
coupling and cleavage products and the conformational pref-
erences of the 1,4-biradical as they are reflected in the rela-
tive transition state energies for cleavage and coupling (37–
41). In the present cases, thermodynamics would clearly be
expected to strongly favor coupling over cleavage in all
The best we can do at present is to approximate kBIR val-
ues from the singlet lifetimes of the silacyclobutanes and
their model compounds, making the admittedly crude as-
sumption that the sum of the rates of the nonproductive de-
cay processes in each of the silacyclobutanes is equal to the
inverse singlet lifetime of its model compound. To the extent
that this assumption is reasonable, kBIR and φ can be esti-
BIR
mated using eqs. [8] and [9], and φ then follows from
sil
eq. [7]. These parameters are collected in Table 1.
© 1999 NRC Canada

1144
Can. J. Chem. Vol. 77, 1999
cases. Cleavage requires a colinear arrangement of the cen-
tral C—C bond with the p-orbitals on the trivalent centers,
but has little dependence on conformation about the central
C—C bond. Coupling, on the other hand, can only occur
from gauche conformers. While steric effects would clearly
be predicted to lead to an increase in cleavage over coupling
in the biradicaloid from 1 compared to that from 4, it is dif-
to the effects of substituents on the degree of mixing of the
benzenoid π-type molecular orbitals of the chromophore
with σ-type molecular orbitals localized in the silacyclo-
butane ring. Modest variations in excited singlet state energy
have no discernible effect on the rate constant, within this
limited series of compounds. The relative rates of cleavage
and coupling in the putative 1,4-biradicaloid intermediate
are very sensitive to substitution at the silicon center; cou-
pling predominates in the biradical from 1-methyl-1-
phenylsilacyclobutane, while cleavage predominates in that
from 1,1-diphenylsilacyclobutane.
Future work in this area will be directed at developing a
better understanding of the various factors that contribute to
the efficient formation of reactive silenes from photolysis of
arylsilacyclobutanes, and exploiting this reaction for poten-
tially useful applications.
ficult to rationalize the variation in φ between 4, 5, and 6
sil
on the basis of these simple arguments.
We stress again that our analysis of the data in terms of a
stepwise reaction mechanism is based on the crude assump-
tion that the singlet lifetimes of the model compounds 11–14
provide a reasonable estimate of the nonproductive contribu-
tions to the singlet lifetimes of the corresponding silacyclo-
butanes (i.e., kF + kIC + kISC). Further work will be required
in order to address the validity of this assumption. To the ex-
tent that it is valid (and equally so in all cases), then the
intermediacy of a 1,4-biradical en route to the corresponding
silene is only really demanded for the monoaryl-substituted
derivatives 4–6, for which the ksil and kBIR values calculated
from the lifetime data are significantly different. In the case
of 1, our data cannot rule out the possibility that silene for-
mation occurs by a concerted σ2s+σ2s cycloreversion mecha-
nism.
Experimental
Ultraviolet absorption spectra were recorded on a Perkin
Elmer Lambda 9 spectrometer interfaced to a Pentium mi-
crocomputer. Fluorescence emission spectra and lifetimes
were determined using a Photon Technologies Inc. LS-100
spectrofluorimeter, which also enables lifetime determina-
tions by the time-correlated single-photon counting tech-
nique. Gas chromatographic (GC) analyses were carried out
using a Hewlett–Packard 5890II + gas chromatograph
equipped with a conventional heated injector, a flame ioniza-
tion detector, a Hewlett–Packard 3396A integrator, and a
DB1 megabore capillary column (15 m × 0.53 mm; Chro-
matographic Specialties, Inc.). Mass spectra and GC–MS
analyses were recorded on a Hewlett–Packard 5890II gas
chromatograph equipped with an HP-5971A mass selective
detector and a DB5 fused silica capillary column (30 m ×
0.25 mm; Chromatographic Specialties, Inc.). Semiprepara-
tive gas chromatographic separations were carried out using
a Varian 3300 gas chromatograph equipped with a thermal
conductivity detector and a 3.8% UCW982 on 80/100
Supelcoport (24 ft × 0.25 in.; Supelco, Inc.) stainless steel
column. Elemental analyses were performed by Guelph
Chemical Laboratories, Inc.
2,2,4-Trimethylpentane (isooctane) and hexane were BDH
Omnisolv grade and were used as received from the sup-
plier. Acetonitrile (BDH Omnisolv) was distilled from cal-
cium hydride and then further dried by several passages
through activated alumina. Methanol, methanol-O-d, tert-
butanol, tert-butanol-O-d, glacial acetic acid, acetic acid-d,
trimethylphenylsilane (11), and phenyl(trimethylsilyl)acety-
lene (12) were of the highest purity available, and were used
as received from Aldrich Chemical Co. 4′-Trimethylsilyl-
biphenyl (13; mp 49–50°C) (42) and dimethyldiphenylsilane
(14; bp 99°C (0.75 Torr; 1 Torr = 133.3 Pa) (43)) were pre-
pared by the reported procedures, and purified by
recrystallization or vacuum distillation to purities in excess
of 99%, as estimated by capillary GC.
Conclusions
The formation of 1-arylsilenes by formal σ2s+σ2s photo-
cycloreversion of 1-arylsilacyclobutanes is a reaction of the
lowest excited singlet state, which is shown by the behavior
of derivatives for which the arylsilacyclobutane triplet state
can be detected directly by laser flash photolysis.
Photocycloreversion proceeds with a quantum yield of only
0.04 in 1-methyl-1-phenylsilacyclobutane, which is 5 times
lower than in 1,1-diphenylsilacyclobutane. The quantum yield
is even higher in derivatives in which the phenyl group is re-
placed with 2-phenylethynyl or 4′-biphenyl substituents. The
resulting silenes, 1-phenyl-, 1-methyl-1-phenyl-, 1-methyl-1-
(2-phenylethynyl)-, and 1-(4′-biphenylyl)-1-methylsilene, re-
act with aliphatic alcohols at rates on the order of (2–5) ×
109 M–1 s–1 in hydrocarbon or acetonitrile solution at ambi-
ent temperature. Deuterium isotope and solvent effects on
the rate constants are consistent with a mechanism involving
reversible complexation of the silene with the alcohol, fol-
lowed by rate-determining proton transfer from silicon to
carbon.
Comparison of the singlet lifetimes of the three 1-aryl-1-
methylsilacyclobutanes and the corresponding aryltrimethyl-
silane suggests that cycloreversion proceeds by cleavage of
one of the ring Si—C bonds to yield a 1,4-biradicaloid that
partitions between silene and starting material. A similar
analysis for 1,1-diphenylsilacyclobutane is consistent with
either concerted or stepwise mechanisms for silene forma-
tion. The lifetime data allow calculation of rate constants for
the primary excited state ring cleavage process and state ef-
ficiencies for formation of the intermediate and its collapse
to silene, assuming that the singlet lifetimes of the model
compounds approximate the rates of the nonproductive con-
tributions to excited state decay in the corresponding
silacyclobutane. Variations in the rate constant for excited
state cleavage throughout the series are tentatively attributed
1-Phenylsilacyclobutane (3) was prepared by lithium alu-
minum hydride reduction of 1-phenyl-1-chloro-1-silacyclo-
butane as reported by Bertrand et al. (26). The compound
was purified by vacuum distillation (bp 97–100°C/20 Torr
(26)), and exhibited the following spectroscopic properties.
IR: 2963.0 (m), 2924.0 (s), 2854.2 (s), 2122.6 (m), 1731.4
© 1999 NRC Canada

Leigh et al.
1145
(m), 1625.5 (s) 1569.4 (s), 1510.1 (m), 1428.9 (m), 1119.5
(s), 699.4 (s), 499.2 (m); H NMR, δ (ppm): 1.25 (m, 4H),
2.20 (m, 2H), 5.33–5.35 (m, 1H), 7.41 (m, 3H), 7.66 (m,
2H); 13C NMR, δ (ppm): 12.7, 18.6, 19.7, 128.0, 128.2,
129.9, 134.4, 135.2, 135.8; MS, m/z (I): 148 (30), 120 (83),
105 (100), 79 (15), 67 (8), 53 (25), 43 (8).
was stirred vigorously for 10 min and then placed in a sepa-
ratory funnel to separate the aqueous and organic layers. The
yellow ether solution was washed with water (3 × 25 mL)
and dried over anhydrous magnesium sulfate. Evaporation of
the solvent on the rotary evaporator afforded an oily yellow
1
1
solid (5.1 g), whose H NMR spectrum was consistent with
the desired product 6. The solid was recrystallized twice
from 95% ethanol to yield the product as slightly yellow
plates (2.5 g, 0.0105 mol, 42%). A portion of the material
(0.2 g) was further purified by column chromatography on
silica gel using hexane as the eluant. Three more
recrystallizations from 95% ethanol afforded the product as
colourless plates (mp 41–41.5°C), which were identified as
1-(4-biphenylyl)-1-methylsilacyclobutane (6) on the basis of
the following data: IR (neat): 3068.7 (w), 3021.8 (w),
2960.8 (m), 2926.0 (m), 1382.8 (m), 1249.7 (m), 1116.1 (s),
1-Methyl-1-phenylsilacyclobutane (4) was prepared by
the method of Auner and Grobe (3), and purified by column
chromatography on silica gel using hexane as eluant. The
compound was characterized on the basis of the following
spectroscopic data. IR: 2963.8 (s), 2856.5 (m), 1428.2 (s),
1396.3 (s), 1249.9 (s), 1115.5 (s), 867.2 (s), 772.1 (s), 732.5
1
(s), 697.6 (s), 427.7 (s); H NMR, δ (ppm): 0.55 (s, 3H),
1.26 (m, 4H), 2.19 (p, 2H), 7.40 (m, 3H), 7.64 (m, 2H);
13C NMR, δ (ppm): –1.8, 14.3, 18.2, 127.9, 129.4, 133.5;
MS, m/z (I): 162 (10), 134 (100), 119 (45), 105 (20), 93 (7),
79 (5), 53 (10), 43 (13).
863.2 (s), 757.4 (s), 723.2 (s), 695.9 (s), 653.5 (s); UV, λmax
:
257 nm, ε248 nm: 24 000 ± 2000 M–1 cm–1; H NMR, δ: 0.58
(s, 3H), 1.22 (m, 2H), 1.29 (m, 2H), 2.17 (m, 2H), 7.53 (m,
3H), 7.65 (m, 6H); 13C NMR, δ: –1.81, 14.37, 18.25, 92.4,
126.58, 127.08, 127.37, 128.71, 133.94, 137.24, 140.93,
142.08; MS, m/z: 238 (17), 211 (22), 210 (100), 195 (49),
167 (18), 165 (22), 152 (10), 53 (25), 43 (67). Anal. calcd.
for C16H18Si: C 80.63, H 7.62; found: C 80.60, H 7.80.
Methoxymethylphenylsilane (15a) and methoxydimethyl-
phenylsilane (15b) were prepared by reaction of methyl-
phenylchlorosilane and dimethylphenylchlorosilane, respectively,
with methanol in ether in the presence of trimethylamine
(43, 44).
1
1-Methyl-1-phenylethynylsilacyclobutane (5)
To
a cooled (–78°C) solution of phenylacetylene
(0.55 mL, 5.0 mmol) in anhydrous ether (20 mL) was added
a 1.6 M hexane solution of n-butyllithium (3.1 mL, 5.0 mmol).
The solution was allowed to warm to room temperature and
stirred for an additional 2 h. The solution was recooled to
–78°C, and a solution of 1-chloro-1-methylsilacyclobutane
(0.48 g, 4 mmol) in dry diethyl ether (10 mL) was added
dropwise over ~10 min with stirring. The solution was al-
lowed to warm to room temperature over 12 h, and then hy-
drolyzed with water (20 mL). The aqueous layer was
extracted with ether (3 × 10 mL). The ether extracts were
combined and dried over anhydrous magnesium sulfate, fil-
tered, and stripped of solvent to yield a colorless liquid that
was purified by silica gel column chromatography using
hexane as eluant. The product (0.63 g, 3.4 mmol, 85%) was
obtained as a colorless liquid and identified as 1-methyl-1-
phenylethynylsilacyclobutane (3) on the basis of the follow-
ing data: IR (neat): 3096 (w), 2970 (m), 2928 (m), 2156 (s),
Methoxydimethyl(phenylethynyl)silane (15c) was pre-
pared by the method of Corriu et al. (45), and exhibited ana-
lytical and spectroscopic data similar to those reported. 4-
Methoxydimethylsilyl-4′-biphenyl (15d) was identified after
isolation from a semipreparative scale photolysis of a deoxy-
genated solution of 6 (0.24 g, 1 mmol) and methanol
(0.4 mL, 10 mmol) in hexane (10 mL), in a quartz photo-
lysis tube with 8 RPR-2537 low-pressure mercury vapor
lamps. The photolysis was carried out to ca. 70% conversion
of 6 (ca. 5 h). The resulting bright yellow photolysate was
stripped of solvent to yield a yellow oil, which was chroma-
tographed on a Florisil column using hexane–dichloro-
methane mixtures. The product, which was isolated in
modest purity (ca. 95% by GC) as a light yellow oil (0.05 g,
1488 (s), 1251 (m), 1220 (m), 1121 (s), 879 (s); UV, λmax
:
247-nm, ε248 nm: 78 990 ± 1200 M–1 cm–1; H NMR, δ: 0.5
(s, 3H), 1.12 (m, 2H), 1.28 (m, 2H), 2.15 (m, 2H), 7.32 (m,
3H), 7.48 (m, 2H); 13C NMR, δ: –0.1, 15.4, 18.4, 92.4,
107.1, 122.8, 128.3, 128.8, 132.0; MS, m/z: 186 (20), 158
(67), 143 (100), 129 (20), 117 (11), 103 (10), 77 (8), 53
(15), 43 (25); HRMS, calcd. for C12H14Si: 186.08647;
found: 186.08633.
1
1
0.2 mmol, 20%), exhibited H NMR and mass spectral data
1
consistent with its proposed structure: H NMR, δ: 0.06 (s,
6H), 3.48 (s, 3H), 7.34–7.53 (m, 9H); MS, m/z: 243 (5), 242
(27), 228 (21), 227 (100), 197 (31), 181 (8), 167 (9), 165
(11), 152 (10), 59 (21).
1-(4′-Biphenylyl)-1-methylsilacyclobutane (6)
Magnesium turnings (6.0 g, 0.25 g-at.) were placed in a
250 mL 2-necked round-bottom flask fitted with a con-
denser, addition funnel, magnetic stirrer, and gas inlet, and
covered with anhydrous ether. A solution of 4-bromobi-
phenyl (6.38 g, 0.027 mol) in anhydrous ether (100 mL) was
then added dropwise, at a rate sufficient to sustain reflux of
the reaction mixture (~1 mL/min). To the resulting dark gray
reaction mixture was added, dropwise with stirring over
15 min, a solution of 1-chloro-1-methylsilacyclobutane
(3.0 g, 0.025 mol) in anhydrous ether (20 mL). The mixture
was stirred for a further 16 h at room temperature, the stirrer
was removed, and the resulting grey-brown solution was de-
canted from the suspended magnesium salts into a 500 mL
Erlenmeyer flask. Water (200 mL) was added; the mixture
Analytical-scale photolyses were carried out using a
Rayonet reactor fitted with 4–8 RPR-2537 lamps and a
merry-go-round apparatus. Aliquots of hexane or cyclohex-
ane-d12 solutions containing 3–6 (0.03–0.04 M) and metha-
nol (0.05 M) were placed in Suprasil quartz cuvettes (3 mm
× 10 mm), sealed with rubber septums, and saturated with
dry argon. Hexadecane (0.01 M) was included as an internal
GC standard for the runs carried out in hexane. The only
products detected in these photolyses were ethylene, the cor-
responding methoxysilane (15a–15d), and small amounts of
products identified as the corresponding hydroxysilanes on
the basis of GC–MS analysis, presumably due to the pres-
ence of small amounts of water. The alkoxysilanes were
© 1999 NRC Canada

1146
Can. J. Chem. Vol. 77, 1999
10. C.J. Bradaric and W.J. Leigh. J. Am. Chem. Soc. 118, 8971
(1996).
11. W.J. Leigh, C.J. Bradaric, C. Kerst, and J.H. Banisch.
Organometallics, 15, 2246 (1996).
12. R.K. Vatsa, A. Kumar, P.D. Naik, H.P. Upadhyaya, U.B.
Pavanaja, R.D. Saini, J.P. Mittal, and J. Pola. Chem. Phys.
Lett. 255, 129 (1996).
13. C.J. Bradaric and W.J. Leigh. Can. J. Chem. 75, 1393 (1997).
14. C. Kerst, R. Boukherroub, and W.J. Leigh. J. Photochem.
Photobiol. A: 110, 243 (1997).
15. C.J. Bradaric and W.J. Leigh. Organometallics, 17, 645 (1998).
16. W.J. Leigh, C. Kerst, R. Boukherroub, T.L. Morkin, S.I.
Jenkins, K. Sung, and T.T. Tidwell. J. Am. Chem. Soc. 121,
4744 (1999).
17. W.J. Leigh, R. Boukherroub, and C. Kerst. J. Am. Chem. Soc.
120, 9504 (1998).
18. G. Bertrand, G. Manuel, P. Mazerolles, and G. Trinquier. Tet-
rahedron, 37, 2875 (1981).
19. T. Brix, N.L. Arthur, and P. Potzinger. J. Phys. Chem. 93, 8193
(1989).
20. M.G. Steinmetz and H. Bai. Organometallics, 8, 1112 (1989).
21. T.J. Barton, G. Marquardt, and J.A. Kilgour. J. Organomet.
Chem. 85, 317 (1975).
22. C.M. Golino, R.D. Bush, P. On, and L.H. Sommer. J. Am.
Chem. Soc. 97, 1957 (1975).
23. R.T. Conlin, M. Namavari, J.S. Chickos, and R. Walsh.
Organometallics, 8, 168 (1989).
24. M.S. Gordon, T.J. Barton, and H. Nakano. J. Am. Chem. Soc.
119, 11 966 (1997).
25. P.B. Valkovich, T.I. Ito, and W.P. Weber. J. Org. Chem. 39,
3543 (1974).
26. G. Bertrand, J. Dubac, P. Mazerolles, and J. Ancelle. Nouv. J.
Chim. 6, 381 (1982).
27. I. Carmichael and G.L. Hug. J. Phys. Chem. Ref. Data, 15, 1
(1986).
28. C. Kerst, R. Ruffolo, and W.J. Leigh. Organometallics, 75,
5804 (1997).
identified by GC–MS and GC coinjection with the authentic
samples. Quantum yield experiments were carried out in
similar fashion, simultaneously irradiating solutions of 1 and
4–6 in hexane containing 0.05 M methanol.
Fluorescence emission spectra and lifetimes were re-
corded on argon-saturated isooctane solutions, whose con-
centrations were adjusted to yield an absorbance of 0.1–0.2
at the excitation wavelength (245–260 nm). In each case, the
fluorescence excitation spectrum matched the UV absorption
spectrum in the region above ~240 nm.
Nanosecond laser flash photolysis experiments employed
the pulses (248 nm; ca.15–25 ns; 70–150 mJ) from Lumonics
510 or Lambda Physik Compex 120 excimer lasers filled
with F2–Kr mixtures in helium or neon, and a microcom-
puter-controlled detection system (46). The intensity of the
beam was reduced to <10 mJ at the cell using a series of
stainless steel wire meshes as neutral density filters. Solu-
tions were prepared at concentrations such that the absorb-
ance at the excitation wavelength was ca. 0.7 (~8 × 10–4 M),
and were flowed continuously through a 3 × 7 mm Suprasil
flow cell connected to a calibrated 100 mL reservoir. Nitro-
gen or oxygen was bubbled continuously through the reser-
voir throughout the experiments. Solution temperatures were
controlled to within 0.1°C with a VWR 1166 constant tem-
perature circulating bath plumbed to a brass sample holder,
and measured with a Teflon-coated copper–constantan ther-
mocouple which was inserted directly into the flow cell.
Quenchers were added directly to the reservoir by microlitre
syringe as aliquots of standard solutions. Rate constants
were calculated by linear least-squares analysis of decay rate
vs. concentration data (six or more points) that spanned at
least one order of magnitude in the transient decay rate. Er-
rors are quoted as twice the standard deviation obtained
from the least-squares analysis in each case.
29. S. Nagase and T. Kudo. J. Chem. Soc. Chem. Commun. 363
(1983).
Acknowledgments
30. N. Wiberg. J. Organomet. Chem. 273, 141 (1984).
31. M. Kira, T. Maruyama, and H. Sakurai. J. Am. Chem. Soc.
113, 3986 (1991).
32. G.W. Sluggett and W.J. Leigh. J. Am. Chem. Soc. 114, 1195
(1992).
We wish to thank the McMaster Regional Centre for Mass
Spectrometry for high-resolution mass spectra and exact
mass determinations, Tracy Morkin and Nicholas Toltl for
technical assistance, and the Natural Sciences and Engi-
neering Research Council of Canada for financial support.
33. W.J. Leigh and G.W. Sluggett. J. Am. Chem. Soc. 116, 10 468
(1994).
34. I. Carmichael and G.L. Hug. In CRC handbook of organic
photochemistry. Vol. I. Edited by J.C. Scaiano. CRC Press,
Boca Raton, Fla. 1989. p. 369.
35. I.B. Berlman. Handbook of fluorescence spectra of aromatic
molecules. Academic Press, London. 1965.
36. J.B. Birks. Organic molecular photophysics. Wiley–Interscience,
New York. 1971.
37. F.D. Lewis and T.A. Hilliard. J. Am. Chem. Soc. 92, 6672
(1970).
38. P.J. Wagner, P.A. Kelso, A.E. Kemppainen, J.M. McGrath,
H.N. Schott, and R.G. Zepp. J. Am. Chem. Soc. 94, 7506
(1972).
39. J.C. Scaiano, E.A. Lissi, and M.V. Encina. Rev. Chem.
Intermed. 2, 139 (1978).
40. S. Ariel, S.V. Evans, M.G. Garibay, B.R. Harkness, N.
Omkaram, J.R. Scheffer, and J. Trotter. J. Am. Chem. Soc.
110, 5591 (1988).
References
1. L.E. Gusel’nikov and M.C. Flowers. J. Chem. Soc. Chem.
Commun. 864 (1967).
2. P. Boudjouk and L.H. Sommer. J. Chem. Soc. Chem.
Commun. 54 (1973).
3. N. Auner and J. Grobe. J. Organomet. Chem. 197, 147 (1980).
4. P. Jutzi and P. Langer. J. Organomet. Chem. 202, 401 (1980).
5. L.E. Gusel’nikov, N.S. Nametkin, and V.M. Vdovin. Acc.
Chem. Res. 8, 18 (1975).
6. G. Raabe and J. Michl. Chem. Rev. 85, 419 (1985).
7. A.G. Brook. In The chemistry of organic silicon compounds.
Edited by S. Patai and Z. Rappoport. John Wiley & Sons, New
York. 1989. p.965.
8. A.G. Brook and M.A. Brook. Adv. Organomet. Chem. 39, 71
(1996).
9. W.J. Leigh, C.J. Bradaric, and G.W. Sluggett. J. Am. Chem.
Soc. 115, 5332 (1993).
41. L.J. Johnston and J.C. Scaiano. Chem. Rev. 89, 521 (1989).
© 1999 NRC Canada

Leigh et al.
1147
42. H.A. Clark, A.F. Gordon, C.W. Young, and M.J. Hunter. J.
Am. Chem. Soc. 73, 3798 (1951).
45. R.J.P. Corriu, J.J.E. Moreau, and H. Praet. Organometallics, 9,
2086 (1990).
43. R.M. Harrison, D.R.M. Walton, and T.M.J. Walton. Dictionary
of organometallic compounds. 2nd ed. Chapman and Hall,
London. 1995.
46. W.J. Leigh, M.S. Workentin, and D. Andrew. J. Photochem.
Photobiol. A: 57, 97 (1991).
44. P.E. Rakita, J.P. Srebro, and L.S. Worsham. J. Organomet.
Chem. 104, 27 (1976).
© 1999 NRC Canada