Generation and characterization of 1,2-Diaryl-1,1,2,2-tetramethyldisilane cation radicals

Article abstract of DOI:10.1021/jo100358y

Nanosecond laser flash photolysis methods were used to generate and spectrally characterize the cation radicals of 1,2-diaryl-1,1,2,2,- tetramethyldisilanes (Ar = p-X-Ph, X = H, CH₃, OCH₃) in hexafluoroisopropanol (HFIP) at room temperature. The disilane cation radicals rapidly reacted with methanol, with bimolecular rate constants ranging from 0.63 to 2.1 ×10⁸ M^-1 s^-1. The cation radicals were found to react with tert-butanol 4-5 times more slowly than methanol, consistent with a small steric effect for nucleophile-assisted fragmentation of the Si-Si bond. The standard potentials for oxidation of the disilanes in HFIP were determined by two different methods: first, by measuring equilibrium constants for electron exchange between the disilanes and the cation radical of hexaethylbenzene and, second, by combining electrochemical data from cyclic voltammetry with the lifetimes of the disilane cation radicals measured by laser flash photolysis in the same media. Agreement between the two methods was excellent (≤3 mV). The oxidation of 1,2-di-p-methoxyphenyl-1,1,2,2,- tetramethyldisilane by slow scan cyclic voltammetry in acetonitrile was not found to be reversible, in contrast to prior literature reports. Possible explanations for the prior results are proposed.

Full text of DOI:10.1021/jo100358y

pubs.acs.org/joc
Generation and Characterization of
,2-Diaryl-1,1,2,2-tetramethyldisilane Cation Radicals
1
†
Gonzalo Guirado, Olesya Haze, and Joseph P. Dinnocenzo*
Department of Chemistry and the Center for Photoinduced Charge Transfer, University of Rochester,
Rochester, New York 14627-0216. Present address: Universitat Aut oꢀ noma de Barcelona, Departament de
Quı´mica (Quimica-Fisica), Edifi C, 08193-Bellaterra (Barcelona), Spain (e-mail: gonzalo.guirado@uab.es).
†
jpd@chem.rochester.edu
Received February 25, 2010
Nanosecond laser flash photolysis methods were used to generate and spectrally characterize the
cation radicals of 1,2-diaryl-1,1,2,2,-tetramethyldisilanes (Ar = p-X-Ph, X = H, CH , OCH ) in
3
3
hexafluoroisopropanol (HFIP) at room temperature. The disilane cation radicals rapidly reacted
8
-1 -1
with methanol, with bimolecular rate constants ranging from 0.63 to 2.1 ꢀ 10 M
s . The cation
radicals were found to react with tert-butanol 4-5 times more slowly than methanol, consistent with
a small steric effect for nucleophile-assisted fragmentation of the Si-Si bond. The standard
potentials for oxidation of the disilanes in HFIP were determined by two different methods: first,
by measuring equilibrium constants for electron exchange between the disilanes and the cation
radical of hexaethylbenzene and, second, by combining electrochemical data from cyclic voltam-
metry with the lifetimes of the disilane cation radicals measured by laser flash photolysis in the same
media. Agreement between the two methods was excellent (e3 mV). The oxidation of 1,2-di-p-
methoxyphenyl-1,1,2,2,-tetramethyldisilane by slow scan cyclic voltammetry in acetonitrile was not
found to be reversible, in contrast to prior literature reports. Possible explanations for the prior
results are proposed.
Introduction
For example, electron impact ionizations of disilanes containing
a tethered alcohol do not show molecular ion peaks, consistent
with rapid intramolecular nucleophilic substitution on the
One-electron oxidation of disilanes under a variety of condi-
1
tions leads to fragmentation of the Si-Si bond. The fragmen-
tation mechanism of the disilane cation radical intermediates is
uncertain for the vast majority of these reactions;both uni-
molecular and nucleophile-assisted pathways have been pro-
posed. Most of the mechanistic evidence gathered to date that
supports a nucleophile-assisted mechanism has been indirect.
1f
parent ion. In addition, photooxidation of several unsymme-
trically substituted disilanes preferentially gave products that
could be rationalized from reaction of nucleophiles at the less
1g
sterically hindered silicon atom. Finally, the unusually low
oxidation potential found for 1,2-bis(2-(2-pyridyl)phenyl)-
1,1,2,2-tetramethyldisilane can be explained by intramolecular
1
j
coordination during one-electron oxidation. Only recently has
kinetic evidence been reported that the cation radical of the
highly sterically hindered 1,1,2,2-tetra-tert-butyl-1,2-diphenyl-
(
1) (a) Watanabe, H.; Kato, M.; Tabei, E.; Kuwabara, H.; Hirai, N.; Sato,
T.; Nagai, Y. J. Chem. Soc., Chem. Commun. 1986, 1662. (b) Mizuno, K.;
Nakanishi, K.; Chosa, J.; Nguyen, T.; Otsuji, Y. Tetrahedron Lett. 1989, 30,
1i
disilane reacts with methanol in a bimolecular reaction.
It has been generally assumed that disilane cation radicals
are highly reactive intermediates with short lifetimes under
3
(
1
689. (c) Fukuzumi, S.; Kitano, T.; Mochida, K. Chem. Lett. 1989, 2177.
d) Fukuzumi, S.; Kitano, T.; Mochida, K. J. Chem. Soc., Chem. Commun.
990, 1236. (e) Nakadaira, Y.; Sekiguchi, A.; Funada, Y.; Sakurai, H. Chem.
Lett. 1991, 327. (f) Nakadaira, Y.; Otani, S.; Kyushin, S.; Ohashi, M.;
Sakurai, H.; Funada, Y.; Sakamoto, K.; Sekiguchi, A. Chem. Lett. 1991,
2
normal experimental conditions. It is surprising, therefore,
601. (g) Mizuno, K.; Nakanishi, K.; Chosa, J.; Otsuji, Y. J. Organomet.
Chem. 1994, 473, 35. (h) Kako, M.; Nakadaira, Y. Bull. Chem. Soc. Jpn. 2000,
7
3, 2403. (i) Al-Kaysi, R. O.; Goodman, J. L. J. Am. Chem. Soc. 2005, 127,
620. (j) Nokami, T.; Soma, R.; Yamamoto, Y.; Kamei, T.; Itami, K.;
(2) See, for example: Kako, M.; Hatakenaka, K.; Kakuma, S.; Ninomiya,
M.; Nakadaira, Y.; Yasui, M.; Iwasaki, F.; Wakasa, M.; Hayashi, H.
Tetrahedron Lett. 1999, 40, 1133.
1
Yoshida, J. Beilstein J. Org. Chem. 2007, 3.
3
326 J. Org. Chem. 2010, 75, 3326–3331
Published on Web 04/20/2010
DOI: 10.1021/jo100358y
r 2010 American Chemical Society

Guirado et al.
JOCArticle
that 1,2-di-p-anisyl-1,1,2,2-tetramethyldisilane (1) and sev-
eral related disilanes have been reported to be reversibly
oxidized electrochemically by slow scan (0.1-0.2 V/s) cyclic
3
voltammetry in CH CN at room temperature.
þ•
3
We describe herein the generation of 1 by nanosecond
laser flash photolysis and demonstrate that it has an extre-
mely short lifetime (<100 ns) in CH CN. Consistent with
3
this observation, electrochemical oxidation of 1 by cyclic
þ•
voltammetry is not reversible in CH CN. Generation of 1
3
is possible in the nonnucleophilic solvent (CF ) CHOH,
3
2
although only as a highly reactive, transient intermediate.
þ•
We provide direct kinetic evidence that 1 and related
disilane cation radicals undergo rapid Si-Si bond frag-
mentation by a nucleophile-assisted mechanism. In addi-
tion, we describe several methods to determine reliable
oxidation potentials for 1 and several structurally related
disilanes.
FIGURE 1. Transient spectrum produced from pulsed laser photo-
þ
lysis of a O
and 3. Inset: Plot of the pseudo-first-order decay rate constant at
2
-saturated HFIP solution containing NMQ , toluene,
Results
560 nm vs [TMB].
Generation and Reaction of Disilane Cation Radicals with
Nucleophiles. Our first experimental goal was to determine if
disilane cation radicals could be generated and spectrosco-
pically characterized in solution at room temperature. Dis-
ilanes 1-3 were chosen for initial study. One-electron
oxidation of the disilanes was accomplished by nanosecond
observed with all three disilanes. For example, in O -saturated
2
þ
HFIP, photolysis of NMQ /toluene solutions containing 3
(∼50mM) gave a newtransient withλmax ≈ 560nm (Figure 1).
The lifetime of this transient was found to increase from ∼150
ns when no special precautions were used to remove adven-
titious water to >600 ns when solutions were dried over
molecular sieves. The transient species was found to react
with the good electron donors such as 1,2,4,5-tetramethoxy-
pulsed laser excitation (7 ns, 343 nm) of solutions containing
þ
∼
1 mM N-methylquinolinium hexafluorophosphate (NMQ )
as a photooxidant and toluene as a codonor. Although this
system for generating radical cations has been described in
6
benzene (TMB, Eox = 0.88 V vs SCE). A plot of the pseudo-
first-order rate constant for decay at 560 nm vs [TMB] was
linear (see Figure 1 insert) and the slope of the plot gave a
4
detail elsewhere, we outline here its essential features for
þ
9
-1 -1
clarity. Pulsed laser excitation of NMQ produces its singlet-
1
second-order rate constant of 1.3 ꢀ 10 M
s , which is near
þ
7
excited state ( NMQ *), which is rapidly intercepted by
toluene, which is present in high concentration (1 M), to
the diffusionlimit inHFIP. Concomitant withdisappearance
of the 560 nm transient, a relatively long-lived transient
appeared with λmax ≈ 460 nm, which was identical to inde-
þ•
efficiently produce toluene and the N-methylquinolinyl
•
þ•
radical (NMQ ) via photoinduced electron transfer. Reaction
of the disilanes, which are present in lower concentration
pendently generated TMB . These data are consistent with
þ•
the assignment of the 560 nm transient species to 3 . This
assignment is further supported by the spectral similarity
to 3 previously generated in a Freon matrix at 77K (λmax
þ•
(
20-50 mM) than toluene, with toluene by electron
þ•
transfer is expected to be highly exothermic and, therefore,
should rapidly generate the disilane cation radicals. In cases
8
≈ 580 nm). Transient species were also observed with
•
9
where the absorption of NMQ (λ
matic, it can be rapidly scavenged (<100 ns) in dioxygen-
≈ 540 nm) is proble-
disilanes 1 (λmax ≈ 800) and 2 (λmax ≈ 650). These transients
max
þ•
were also found to react with TMB to generate TMB ,
consistent with their assignment to 1 and 2 . The lifetimes
-
•
þ• þ•
saturated solution leading to O2 , which does not have
interfering absorptions in the visible region where cation
radicals typically absorb.
þ•
þ•
of 1 and 2 were ∼1200 and ∼900 ns, respectively, in dry
þ•
HFIP. Importantly, attempted generation of 1 ;the longest
lived disilane cation radical;in acetonitrile gave a weak
transient that completely decayed within 100 ns. We will
return to the significance of this observation when discussing
the electrochemical oxidation of the disilanes.
(5) For use of HFIP as a non-nucleophilic solvent, see: (a) Eberson, L.;
Hartshorn, M. P.; Persson, O. J. Chem. Soc., Perkin Trans. 2 1995, 1735.
(b) Kirmse, W.; Krzossa, B.; Steenken, S. Tetradedron Lett. 1996, 1197.
þ
In nanosecond, pulsed laser experiments with NMQ ,
toluene, and disilanes 1-3 in acetonitrile, no transient
absorptions between 350 and 800 nm (except for NMQ )
(c) Pienta, N. J.; Kessler, R. J. Am. Chem. Soc. 1993, 115, 8330. (d) Cozens,
F.; Li, J.; McClelland, R. A.; Steenken, S. Angew. Chem., Int. Ed. Engl. 1992,
31, 743. (e) Kirmse, W.; Kilian, J.; Steenken, S. J. Am. Chem. Soc. 1990,
•
1
12, 6399.
6) de Lijser, H. J. P.; Snelgrove, D. W.; Dinnocenzo, J. P. J. Am. Chem.
Soc. 2001, 123, 9698.
withlifetimes greater than∼100 nswereobserved. Incontrast,
(
when the experiments were conducted in the less nucleophilic
5
solvent hexafluoroisopropanol (HFIP), transients were
(7) The diffusion rate constant (k
Smoluchowski equation (k
d
) in HFIP can be estimated from the
s
-1 -1
9
-1 -1
d
≈ 8RT/(3000η) M
s
) to be ∼3 ꢀ 10 M
at 20 °C, with η = 1.94 cP: Murto, J.; Kivinen, A.; Lindell, E. Suom. Kemistil.
B 1970, 43, 28.
(8) Kumagai, J.; Yoshida, H.; Ichikawa, T. J. Phys. Chem. 1995, 99, 7965.
(
3) (a) Zhuikov, V. V. Russ. J. Gen. Chem. 1997, 67, 975. (b) Zhuikov,
V. V. Russ. J. Gen. Chem. 2000, 70, 879. (c) Zhuikov, V. V. Russ.
J. Electrochem. 2000, 36, 117.
þ•
þ•
þ•
(9) The λmax of 1 cannot be as well determined as that of 2 and 3 due
to the limited amount of analyzing light beyond 800 nm in our nanosecond
laser apparatus.
(
4) Dockery, K. P.; Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.; Gould,
I. R.; Todd, W. P. J. Am. Chem. Soc. 1997, 119, 1876.
J. Org. Chem. Vol. 75, No. 10, 2010 3327

Guirado et al.
JOCArticle
TABLE 1. Substituent Effects on the Second-Order Rate Constants for
þ•
þ•
the Reaction of Disilane Cation Radicals 1 -3 with Methanol (kHOMe
and tert-Butanol (kHOBu
) in HFIP at 20 °C
rate constants, M
)
t
a
-1 -1
s
þ•
disilane
k
HOMe
k
HOBu
t
k
HOMe/kHOBu
t
þ•
þ•
þ•
7
8
8
7
7
7
1
2
3
a
6.3(2) ꢀ 10
1.3(1) ꢀ 10
2.1(2) ꢀ 10
1.3(2) ꢀ 10
2.9(2) ꢀ 10
4.9(2) ꢀ 10
4.8
4.5
4.3
Rate constants are an average of at least three independent determi-
nations; standard deviation in the last significant digit is given in
parentheses.
the disilanes and HEB was set up by using experimental
þ•
þ•
conditions similar to those used to generate 1 -3 . Pulsed
laser excitation of a O -saturated HFIP solution containing
þ
2
NMQ and toluene (1 M) generated the toluene cation
radical, as described above. The toluene cation radical
subsequently oxidized the disilane and HEB, which were
present in solution at lower concentration (∼1-20 mM). The
FIGURE 2. Plot of the pseudo-first-order rate constants for decay
of 1 -3 in HFIP as a function of the concentration of added
methanol.
þ•
þ•
þ•
þ•
þ•
spectra for the disilane cation radical and HEB were
Disilane cation radicals 1 -3 were found to rapidly
react with deliberately added nucleophiles, such as alcohols.
Shown in Figure 2 is a plot of the pseudo-first-order rate
constants for decay of the transients vs [HOMe]. The corres-
ponding second-order rate constants are given in Table 1.
monitored until an equilibrium was reached for the electron
transfer reaction (eq 1). The electron transfer equilibrium
þ•
constants (K ) were calculated from [disilane ][HEB]/
eq
þ•
[
trations of the radical cations (e10 M) are very small
disilane][HEB ]. At low pulse energies, where the concen-
-
5
t
Kinetic plots with added HOBu were also linear; the second-
order rate constants from these plots are listed in Table 1.
The rate constant ratio kHOMe/kHOBut was found to be small
compared to those of the neutral species, [disilane] and
[HEB] are indistinguishable from their initial concentra-
tions. The difference in oxidation potential between the
and nearly independent of the disilane cation radical struc-
ture (see Table 1).
disilane and HEB (ΔE ) was calculated from the Nernst
ox
equation (eq 2).
Oxidation Potentials of 1-3 by Redox Equilibrium. As
þ•
þ•
shown in Table 1, the rate constants for reaction of 1 -3
t
with HOMe or HOBu vary by less than a factor of 4.
Interestingly, this is substantially smaller than the rate con-
stant range of >65 observed for the reaction of HOMe with
benzyltrimethylsilane cation radical vs the p-methoxybenzyl
4
derivative. The large difference in reactivity of the benzylsi-
lane cation radicals has been correlated with their relative
ΔEox ¼ EoxðdisilaneÞ - EoxðHEBÞ ¼ - ðRT=FÞ ln Ket
ð2Þ
Shown in Figure 3 are typical redox equilibrium spectra
1
0
oxidation potentials, which differ by nearly ∼0.5 V. It was,
therefore, of interest to determine the oxidation potentials of
1
þ•
þ•
þ•
þ•
-3. Given the short lifetimes of 1 -3 described above,
for HEB and 3. Spectra of HEB and 3 are given by traces
a and b, respectively. Equilibrium mixtures of the two cation
radicals at three different concentrations of HEB and 3 are
also shown. An isosbestic point is observed at ∼530 nm, as
expected for the simple, electron transfer equilibrium in eq 1.
the one-electron oxidations of 1-3 with conventional elec-
trochemical methods are not expected to be reversible.
Indeed, as described in the next section, they are not.
The oxidation potentials of 1-3 in HFIP were determined
by two different methods. The first was by redox equilibra-
1
2
As described before, the equilibrium spectra can be fit
with linear combinations of the pure cation radical spectra.
Fitting of the equilibrium spectra in Figure 3 gave an average
value for K of 3.0(4). The average value of K from several
1
1,12
tion. This method, which is described in detail elsewhere,
involves measuring the equilibrium constant for electron
transfer between a compound of unknown oxidation poten-
tial and the cation radical of a compound whose reduction
potential is known. It turned out that hexaethylbenzene
HEB) served as a suitable reference compound for disilanes
-3. The oxidation of HEB is fully reversible in HFIP
Eox = 1.114(3) V vs SCE). The redox equilibrium between
et
et
independent experiments was 2.8(4). This value of K corres-
et
ponds to E (3) - E (HEB) = -0.026(4) V. The error in
ox
ox
(
ΔEox is small (4 mV), but typical of the high precision that
can be obtained by the redox equilibrium method. Combin-
1
(
1
3
eq
ing ΔEox with E (HEB) gives E (3) = 1.088(5) V vs SCE in
ox
ox
HFIP. Analogous redox equilibrium experiments were car-
ried out with 1 and 2, which form long-lived disilane cation
radicals, with equally good results. The oxidation potentials
from redox equilibrium measurements with 1/HEB and
(
10) Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.; Gould, I. R.; Todd,
W. P.; Mattes, S. L. J. Am. Chem. Soc. 1989, 111, 8973.
11) For reviews, see: (a) Steenken, S. Landolt-B o€ rnstein 1985, 13e, 147.
b) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637. (c) Stanbury, D. M.
(
(
In General Aspects of the Chemistry of Radicals; Alfassi, Z. B., Ed.; Wiley:
New York, 1999; Chapter 11.
2
/HEB are given in Table 2.
(
12) (a) Guirado, G.; Fleming, C. N.; Lingenfelter, T. G.; Williams, M. L.;
Zuilfhof, H.; Dinnocenzo, J. P. J. Am. Chem. Soc. 2004, 126, 14086.
b) Merkel, P. B.; Luo, P.; Dinnocenzo, J. P.; Farid, S. J. Org. Chem. 2009,
Oxidation Potentials of Disilanes 1-3 from Cyclic Voltam-
þ•
þ•
metry in HFIP. Given the short lifetimes of 1 -3 in HFIP,
the oxidations of 1-3 are not expected to be reversible
by conventional cyclic voltammetry. Indeed, they are not.
(
7
4, 5163.
(13) Standard deviation of the last significant digit is given in parentheses.
3
328 J. Org. Chem. Vol. 75, No. 10, 2010

Guirado et al.
JOCArticle
FIGURE 4. Anodic peak potentials (Epa) for disilanes 1-3 as a
þ•
FIGURE 3. Spectra of hexaethylbenzene cation radical (a), 3 (b),
and the equilibrium, electron transfer spectra (gray lines) in HFIP at
2
least-squares fit to it (thin black line).
-1
function of the logarithm of the scan rate (ν, V s ) in HFIP. Formal
0
oxidation potentials (E° ) calculated from eq 3 for 1-3 as a function
of the logarithm of the scan rate.
0 °C. Inset: One of the equilibrium spectra (gray line) and the best
pseudo-first-order) reaction of the disilane cation radicals
and, finally, a second electron transfer from the products of
the chemical reaction, i.e., an ECE mechanism. Taking into
account the results from the above laser flash photolysis
experiments, the oxidations of 1-3 can be explained by one-
electron oxidation of the disilanes followed by rapid nucleo-
philic substitution by the solvent. The leaving groups from
the substitution;the aryldimethylsilyl radicals;presumably
undergo rapid oxidation at the electrode, thus accounting for
TABLE 2. Differences in Oxidation Potentials for Disilanes 1-3 and
HEB (ΔEox) Determined from Redox Equilibrium Measurements in HFIP
eq
at 20 °C, Oxidation Potentials for 1-3 (Eox) Based on ΔEox, and Formal
0
Oxidation Potentials (E° ) from Cyclic Voltammetry-Laser Flash
Photolysis
a
eq b
)
0
c
disilane
(ΔEox
)
(Eox
(E° )
1
2
3
a
-0.045(3)
-0.032(4)
-0.026(4)
1.069(4)
1.082(5)
1.088(5)
1.067(4)
1.083(5)
1.091(3)
1
5
the two-electron wave.
Average Eox(disilane) - Eox(HEB), V; standard deviation in the last
b
eq
The preliminary electrochemical experiments demonstrate
that 1-3 undergo rapid oxidation at the electrode surface.
On the basis of this observation, the anodic potentials (Epa)
for the electrochemical oxidations are expected to follow
significant digit is given in parentheses. Eox = (HEB) þ ΔEox; V vs SCE.
c
Formal oxidation potential from cyclic voltammetry-laser flash
photolysis in HFIP at 20 °C, V vs SCE (see text).
In principle, it is nonetheless possible to determine the
1
6
0
þ•
þ•
eq 3, where E° is the formal oxidation potential, k is the
rate constant for the chemical reaction coupled to the
electron transfer, ν is the scan rate, and n is the number of
electrons transferred in the initial oxidation (R, T, and F have
their usual meaning). The rate constants for decay of
disilane oxidation potentials if the lifetimes of 1 -3 are
known under the reaction conditions by using well-known
electrochemical relationships between the peak potential
and the scan rate. Determination of the cation radical
lifetimes by laser flash photolysis on electrochemical
þ•
þ•
1
4
1 -3 were measured by laser flash photolysis from the
-
solutions provided a rare opportunity to obtain inde-
pendent measurements of the disilane oxidation poten-
tials. To accomplish this objective, it was first necessary to
better characterize the electrooxidation process.
þ
same solutions (including electrolyte, 0.1 M n-Bu N BF )
previously used to determine the variation of E as a
4
4
pa
0
function of log(ν). The values of E° were calculated at each
scan rate from the E values and the cation radical decay
rate constant (k) by using eq 3. Gratifyingly, the calculated
pa
The electrochemical oxidations of 1-3 by cyclic voltam-
metry using a glassy carbon electrode as a working electrode
were not reversible in HFIP at scan rates varying from 0.1
0
E° values for all three disilanes were found to be independent
-
1
of scan rate (Figure 4, filled symbols). As shown in Table 2,
the formal potentials derived from the combined electro-
chemistry-laser flash photolysis experiment are in excellent
agreement with those determined by the redox equilibrium
method.
to 50 V s . The anodic peak potential currents (i ) in-
pa
creased linearly with the concentrations of the disilanes (C =
1
/2
-15 mM) and with the square root of the scan rate (ν ).
1
1/2
The ratio of i /Cν remained constant over the entire scan
pa
rate range and a comparison of the ratio to that measured for
ferrocene under identical conditions provided the number
of electrons transferred (1.8-1.9). At a scan rate of 0.1 V/s,
the peak widths (Epa - Epa/2) for the oxidation waves were
ꢀ
ꢁ
RT RT
-
k RT
v nF
0
Epa ¼ E° þ 0:780
ln
ð3Þ
nF 2nF
5
6 ( 2 mV and were independent of disilane concentration.
Cyclic Voltammetry of 1 in CH CN. On the basis of the
3
Plots of the anodic peak potentials (E ) vs log(ν) were linear
pa
þ•
extremely short lifetime of 1 in CH CN estimated from
3
and had slopes of 32 ( 2 V/log(ν) (Figure 4, unfilled symbols).
These data are consistent with fast heterogeneous electron
transfer at the electrode followed by a rapid first-order (or
(15) The electrochemical experiment was performed under an argon
atmosphere, otherwise the silyl radicals might have partially reacted with
dioxgen present in solution (cf.: Mizuno, K.; Tamai, T.; Hashida, I.; Otsuji,
Y. J. Org. Chem. 1995, 60, 2935).
(14) For an elegant use of this approach, see: Gould, I. R.; Wosinska,
Z. M.; Farid, S. Photochem. Photobiol. 2006, 82, 104.
(16) Nicholson, R. S.; Shain, I. Anal. Chem. 1964, 36, 706.
J. Org. Chem. Vol. 75, No. 10, 2010 3329

Guirado et al.
JOCArticle
This reactivity difference is similar to that found for nucleo-
philic substitution at silicon on the similarly reactive cation
radical of benzyltrimethylsilane. On the basis of this simi-
larity and the known propensity for Si-Si bond fragmenta-
1
tion in disilane oxidations, we ascribe the reactions of
þ•
þ•
1
-3 with alcohols to nucleophile-assisted fragmentation
of the Si-Si bond.
þ•
þ•
As noted above, the reactivities of 1 -3 are remarkably
similar. This contrasts with the >65-fold difference in reac-
tivity for the cation radicals of benzylsilanes 4 and 5 (see
4
below) with HOMe in HFIP. The differences can now be
readily explained by the relative oxidation potentials, which
are a good predictor of relative cation radical reactivity. For
example, the oxidation potential of 4 is 0.5 V greater than 5,
10
þ•
which explains the much higher reactivity of 4 . In contrast,
the oxidation potentials of 1-3 differ by only ∼0.02 V, which
explains the similar reactivities of their cation radicals.
Given the relatively large differences in both the oxidation
þ•
þ•
þ•
þ•
potentials and relative reactivities of 4 -5 vs 1 -3 , it is
interesting to note that the cation radicals exhibit dramati-
cally opposite trends in their visible absorptions. The ab-
þ•
þ•
sorption spectra of 4 (530 nm) and 5 (500 nm) differ by
only 30 nm, with the p-methoxy derivative absorbing at
FIGURE 5. Cyclic voltammograms (single scan) of disilane 1(5.0mM,
top), 4,4 -dimethoxybiphenyl (DBP, 12 mM, middle), and a solution
0
4
shorter wavelength. In contrast, the p-methoxy-substituted
containing 2.5 mM 1 and 5.8 mM DBP (bottom) in acetontrile contain-
ing 0.1 M tetra-n-butylammonium tetrafluoroborate recorded with a
þ•
disilane cation radical, 1 , absorbs at longer wavelength
þ•
-
1
.
scan rate of 0.50 V s
than 3 , and by ∼240 nm! A compelling rationalization of
these markedly different results will require an assignment of
the electronic transitions for the benzylsilane and disilane
cation radicals.
laser flash photolysis (τ < 100 ns), it is not surprising that
the oxidation of 1 by cyclic voltammetry in CH CN was not
found to be reversible. This result is in disagreement, how-
3
3
ever, with earlier reports that 1 and several related disilanes
are reversibly oxidized in CH CN at room temperature at
3
low scan rates. These reports claim that 1 exhibits a reversible
CV wave at ∼1.3 V vs SCE and a second, irreversible wave at
∼
1.7 V. We find that oxidation of 1 in CH CN shows
3
irreversible oxidation waves at ∼1.2 and ∼1.9 V vs SCE
The ability to determine the oxidation potentials of 1-3 by
the redox equilibrium method is noteworthy, especially given
the high reactivities of the disilane cation radicals. This pro-
(
Figure 5, top panel). The reversible wave previously re-
0
ported at ∼1.3 V is consistent with that observed for 4,4 -
1
7
dimethoxybiphenyl (DBP), a plausible contaminant from
1
9
1
8
vides yet another example of the power of the methodology.
Finally, we note that, due to the high reactivities of disilane
cation radicals toward nucleophiles, the oxidations of disilanes
the preparation of 1. Indeed, when we deliberately add
DBP to a solution containing 1 (Figure 5, bottom panel), we
can closely reproduce the cyclic voltammogram previously
3
b
1-3 are not reversible by conventional cyclic voltammetry.
published. It is worth noting that a reversible CV peak at
1.2-1.3 V can also be generated by multiple oxidation-
Nor do we expect the electrochemical oxidations of similarly
substituted disilanes to be reversible, although it may be
possible to obtain reversible behavior with ultrafast potentio-
∼
reduction cycles on pure 1. Thus, we suggest that prior
literature results on the electrooxidation of 1 can be ex-
plained by either the presence of DBP as an impurity or the
generation of DBP from electrooxidation of 1.
20
stats and microelectrodes. Nevertheless, as described here,
slow scan CV combined with rate data from laser flash
photolysis (LFP) can provide redox potentials with accuracy
and precision comparable to those established by the redox
equilibrium method. The CV-LFP approach should find great-
er use in determining thermodynamically meaningful redox
potentials for compounds that form highly reactive ion radicals
upon oxidation or reduction.
Discussion
The cation radicals of 1,2-diaryl-1,1,2,2-tetramethyldisi-
lanes 1-3 were successfully generated in hexafluoroisopro-
panol (HFIP) at room temperature. 1 -3 were found to
rapidly react with both HOMe and HOBu . We attribute
þ•
þ•
t
t
the ∼4- to 5-fold lower reactivity of HOBu to a steric effect.
(19) For application of this method to the determination of oxidation
potentials for primary amines, see: Bourdelande, J. L.; Gallardo, I.;
Guirado, G. J. Am. Chem. Soc. 2007, 129, 2817.
(
17) (a) Ronl ꢁa n, A.; Coleman, J.; Hammerich, O.; Parker, V. D. J. Am.
Chem. Soc. 1974, 96, 845. (b) Park, S. M.; Bard, A. J. Chem. Phys. Lett. 1976,
8, 257.
18) 4,4 -Dimethoxybiphenyl is a plausible byproduct of the 4-methoxy-
phenyl magnesium bromide used to prepare 1 in ref 3. Our samples of 1-3
were shown to contain <0.03% of the corresponding biphenyls (see the
Experimental Section).
(20) (a) Wightman, R. M.; Wipf, D. O. Electroanal. Chem. 1989, 15, 267.
(b) Andrieux, C. P.; Hapiot, P.; Sav ꢁe ant, J.-M. Chem. Rev. 1990, 90, 723.
(c) Heinze, J. Angew. Chem., Int. Ed. Engl. 1993, 32, 1268. (d) Amatore, C.;
Bouret, Y.; Maisonhaute, E.; Abruna, H. D.; Goldsmith, J. I. C. R. Chim.
2003, 6, 99. (e) Amatore, C.; Maisonhaute, E.; Simmoneau, G. Electrochem.
Commun. 2000, 2, 81. (f) Amatore, C.; Maisonhaute, E.; Simmoneau, G.
J. Electroanal. Chem. 2000, 486, 141.
3
0
(
3
330 J. Org. Chem. Vol. 75, No. 10, 2010

Guirado et al.
JOCArticle
Experimental Section
measurements were made at least 100 ns after the laser pulse. At
the end of each equilibrium experiment the spectrum of the
sample containing the single equilibrium component was re-
recorded to ensure that the laser power had not changed during
Materials. Acetonitrile was purified by passing the solvent
over a bed of activated alumina. Hexafluoroisopropanol was
2
1
successively distilled from sodium bicarbonate and activated
˚
the course of the experiment. Equilibrium spectra were fit as
previously described.
molecular sieves (3 A) under N
2
. Toluene was distilled under N
2
.
12a
The oxidation potential differences
Spectral-grade methanol and tert-butanol were obtained from
commercial sources and used as received. 1,2,4,5-Tetramethoxy-
benzene was generously provided by Kodak. N-Methylquinoli4-
nium hexafluorophosphate was prepared by a literature method.
Hexaethylbenzene (Kodak) was purified by recrystallization from
ethanol. Tetra-n-butylammonium tetrafluoroborate (Fluka, elec-
trochem. grade) was used as received. Disilanes 1-3 were prepared
by a literature method and were purified by repeated recrystalli-
zation from ethanol/ethylacetate, methanol/ethyl acetate, and
methanol, respectively. Gas chromatographic analysis (HP Ultra
given in Table 2 are an average of at least three independent
determinations. In experiments where the influence of adventi-
tious water was to be minimized, the solutions were dried by
˚
storage over 3 A activated molecular sieves for >2 h.
Cyclic Voltammetry. Oxidation potentials were measured with
a BAS-100B/W Electrochemical Workstation or a computer-
controlled VSP-potentiostat (Versatile Modular Potentiostat).
Scan rates were in the range 0.1-50 V/s and positive feed-back
iR compensation was used throughout. Experiments were con-
ducted with a standard three-electrode setup in a conical electro-
chemical cell enclosed in a jacket that allowed the temperature to
be maintained at 20 °C by means of a thermostated circulating
bath. The working electrode was a glassy carbon disk (1.0 mm
dia.) that was polished with a 1 μm diamond paste. The counter
electrode was a glassy carbon disk (0.3 cm diameter). All of the
potentialsare reported vs an SCE electrode that was isolated from
the working electrode compartment by a salt bridge with a
ceramic frit that allowed ionic conduction between the two
solutions while avoiding appreciable contamination. Solutions
were prepared with HFIP or acetonitrile as a solvent and they
were purged with argon before the measurements were made. The
22
2
, 9.6 m ꢀ 0.2 mm ꢀ 0.33 μm) of 1-3 showed that the disilanes
contained <0.03% of the corresponding biphenyls as impurities.
Transient Absorption Experiments. Detailed descriptions of
the nanosecond and picosecond transient absorption apparatus
1
2
are given elsewhere. N-Methyl quinolinium hexafluoropho-
sphate solutions (OD ≈ 0.5-1.0 at 343 nm) containing 1 M
toluene were prepared in HFIP or acetonitrile. Experiments
were conducted in quartz cuvettes equipped with high vacuum
stopcocks that carried serum caps. Solvent-saturated dioxygen
was bubbled through the solutions for 10-15 min before
equilibrium measurements were made. For redox equilibration
experiments, separate solutions were initially prepared with
each of the components in order to record the spectra of their
corresponding cation radicals. Then, increasing concentrations
of the second component were added to one of the cuvettes from
a dioxygen-saturated stock solution via a calibrated gas-tight
syringe equipped with a 6 in. 22 gauge fixed needle. All equili-
brium measurements were made at 20 °C. In general, equilibrium
-3
concentration of the electroactive substances was ∼10 M. The
supporting electrolyte was 0.1 M tetra-n-butylammonium tetra-
fluoroborate.
Acknowledgment. Research was supported by a grant
from the National Science Foundation (CHE-0749919).
G.G. is grateful to the Spanish Ministerio de Ciencia e Innova-
cion for financial support (CTQ2009-07469). We are grateful
to Eastman Kodak for generously providing a sample of
1,2,4,5-tetramethoxybenzene.
(
21) (a) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518. (b) Alaimo, P. J.; Peters,
D. W.; Arnold, J.; Bergman, R. G. J. Chem. Educ. 2001, 78, 64.
(22) Kelling, H. Z. Chem. (Stuttgart, Ger.) 1967, 7, 237.
J. Org. Chem. Vol. 75, No. 10, 2010 3331