Electron paramagnetic resonance and computational studies of radicals derived from boron-substituted N-heterocyclic carbene boranes

Article abstract of DOI:10.1021/ja2038485

Fifteen second-generation NHC-ligated boranes with aryl and alkyl substituents on boron were prepared, and their radical chemistry was explored by electron paramagnetic resonance (EPR) spectroscopy and calculations. Hydrogen atom abstraction from NHC - BH

Full text of DOI:10.1021/ja2038485

ARTICLE
pubs.acs.org/JACS
Electron Paramagnetic Resonance and Computational
Studies of Radicals Derived from Boron-Substituted
N-Heterocyclic Carbene Boranes
John C. Walton,*,† Malika Makhlouf Brahmi,‡ Julien Monot,‡,§ Louis Fensterbank,‡ Max Malacria,‡
||
Dennis P. Curran,*,§ and Emmanuel Lac^ote*,‡,
†School of Chemistry, EaStChem, University of St. Andrews, St. Andrews, Fife KY16 9ST, United Kingdom
‡UPMC Univ Paris 06, Institut parisien de chimie molꢀeculaire (UMR CNRS 7201), C. 229, 4 place Jussieu, 75005 Paris, France
§Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States
S Supporting Information
b
ABSTRACT: Fifteen second-generation NHC-ligated boranes with aryl
and alkyl substituents on boron were prepared, and their radical chemistry
was explored by electron paramagnetic resonance (EPR) spectroscopy and
calculations. Hydrogen atom abstraction from NHCꢀBH2Ar groups
produced boryl radicals akin to diphenylmethyl with spin extensively
delocalized across the NHC, BH, and aryl units. All of the NHCꢀB HAr
3
radicals studied abstracted Br-atoms from alkyl bromides. Radicals with
bulky N,N0-dipp substituents underwent dimerization about 2 orders of magnitude more slowly than first-generation NHC-ligated
trihydroborates. The evidence favored head-to-head coupling yielding ligated diboranes. The first ligated diboranyl radical, with a
structure intermediate between that of ligated diboranes and diborenes, was spectroscopically characterized during photolysis of di-
t-butyl peroxide with N,N0-di-t-butyl-imidazol-2-ylidene phenylborane. The reactive site of B-alkyl-substituted NHCꢀboranes
switched from the boron center to the alkyl substituent for both linear and branched alkyl groups. The β-borylalkyl radicals obtained
from N,N0-dipp-substituted boranes underwent exothermic β-scissions with production of dipp-ImdꢀBH2 radicals and alkenes.
3
The reverse additions of NHCꢀboryl radicals to alkenes are probably endothermic for alkyl-substituted alkenes, but exothermic for
conjugated alkenes (addition of an NHCꢀboryl radical to 1,1-diphenylethene was observed). A cyclopropylboryl radical was
observed, but, unlike other R-cyclopropyl-substituted radicals, this showed no propensity for ring-opening.
’ INTRODUCTION
the NHC substituents. The resulting NHCꢀboryl radicals
(NHCꢀBH2 ) have π-delocalized structures, very different
3
Recent research has revealed that N-heterocyclic carbene
boranes (NHCꢀborane) possess novel features that make them
significantly different from boranes, borohydrides, and even other
boraneꢀLewis base complexes.1 NHCs are strong electron-pair
donors, so their complexes with boranes are remarkably stable as
compared to borane sulfides, ethers, and even amines. NHC-ligated
trihydroborates (NHCꢀBH3) are precursors for unusual boryl
radicals,2 borenium cations,3 boryl anions,4 and borylenes.5 NHCꢀ
BH3 complexes are robust and stable to ambient conditions and
to chromatography; hence they are also finding use as synthetic
reagents in radical,6 ionic,7 and organometallic reactions.8 They are
also promising co-initiators in radical photopolymerizations.9
The BꢀH bonds of NHCꢀboranes are dramatically weaker
than in uncomplexed boranes,10 having ΔHo(BꢀH) values of
80ꢀ88 kcal molꢀ1 according to ab initio computations2,11,12 and
experimental studies.13 Consequently, NHCꢀboryl radicals were
easily generated by H-atom abstraction from a first-generation of
complexes with the NHC ligated to BH3 (Figure 1). Conventional
radical initiators, including t-butoxyl,14 triplet benzophenone,9 and
C-centered radicals,15 abstract H-atoms from NHCꢀBH3 at rates
that depend strongly on the steric shielding at the boron center by
from the σ-electronic structures of amineꢀboryl and phosphi-
neꢀboryl radicals.11,16
Opportunities to alter properties by changing the NHC sub-
stituent are plentiful, so the NHC groups were extensively modified
during this prior work on first-generation NHCꢀboranes.9,14,15 Yet
the boron substituent was constant (BH3). Substituents on the
boron atoms are anticipated to both modify the steric environment
about B and perhaps change the strength of the BꢀH bonds. Novel
chemistry is therefore expected for a second generation of com-
plexes substituted at boron (NHCꢀBH2R).
Here, we prepare and study 15 NHCꢀboranes with a range of
aryl and alkyl substituents on boron. H-abstraction is rapid for BꢀAr
complexes, but the reactivity of the resultant B-aryl substituted
radicals is strikingly low in dimerizations. The chemistry of B-alkyl
NHCꢀboranes is influenced by a trade-off between substituents on
the NHC rings and the B-alkyl substituents, with radicals sometimes
being generated by CꢀH abstraction (from the alkyl substituent)
Received: April 26, 2011
Published: May 28, 2011
r
2011 American Chemical Society
10312
dx.doi.org/10.1021/ja2038485 J. Am. Chem. Soc. 2011, 133, 10312–10321
|

Journal of the American Chemical Society
ARTICLE
Figure 1. First- (ref 2) and second-generation (this work) NHCꢀ
borane radical precursors.
Figure 3. Structure and overlay of the experimental (black) and
computer-simulated (red) EPR spectra of NHCꢀB HC6H4OMe-4
3
radical (2b ) in PhH at 290 K.
3
available from boronic acids by a recently published route,17 and
details for their preparation and characterization are provided in
the Supporting Information.
Deaerated solutions of individual NHCꢀboranes and di-tert-
butyl peroxide (DTBP) in tert-butylbenzene or benzene were
irradiated with UV light in the resonant cavity of a 9 GHz EPR
spectrometer. We expected that the photochemically generated
t-BuO radicals would abstract hydrogen from the BH2 groups of
3
the substrates, thus generating the NHCꢀB HAr radicals. Satis-
3
factory spectra were obtained from each of 2aꢀc. Figure 3 shows
the structure and spectrum of 2b ; the spectra of the radicals
3
derived from 2a and 2c are in the Supporting Information.
The EPR parameters obtained from computer simulations of
the spectra are listed in Table 1. The 11B and HR hyperfine splitting
constants (hfs) of 2a ꢀ2c are smaller than for the standard radical
3
3
1 , indicating that spin density is delocalized away from the B-atom
3
into the phenyl rings. In line with this, hfs from phenyl ring H-atoms
were resolved for 2a and 2c and partly for 2b . The trend in the
3
3
3
magnitudes of a(11B) roughly corresponds with the mesomeric
effects of the para-substituents (Hammett σp constants of H, OMe,
and CF3 being 0.0, ꢀ0.16, and 0.53, respectively).18
Figure 2. Structures of N-heterocyclic carbene boranes for study of
BꢀAr substitution. Formal charges are shown on 1 for illustration. The
corresponding NHCꢀBH2 and NHCꢀB HAr radicals are designated
3
3
The structure of radical 2a was computed by DFT employing
3
as 1 , 2a , etc.
the UB3LYP functional with the 6-31Gþ(d) basis set,19 and this
is shown in Figure 4 along with the computed SOMO. The planes
of the 2,6-diisopropylphenyl rings are virtually perpendicular to the
plane of the imidazole ring and hence out of conjugation. This agrees
with the lack of hfs from the dipp-ring H-atoms in the EPR spectra.
The computed hfs were in reasonable agreement with experiment
(Table 1) apart from a(11B).20 The structure (Figure 4) shows that
even with the i-Pr groups rotated as far away as possible, the steric
shielding of the B-atom is considerable. The planes of the Ph and
imidazole rings are disposed at an angle to each other of only about
15°. Thus, the SOMO that extends across these two rings and the
BH unit is practically a pure π-system. The computations imply these
3
3
and sometimes by BꢀH abstraction. The structures and onward
reactions (by themselves, with each other, and with alkyl bromides)
of all of these different radicals are described.
’ RESULTS AND DISCUSSION
NHCꢀBoranes with B-Aryl Substituents. Figure 2 shows the
structures of two first-generation standards (1 and 3) and nine
aryl-substituted NHCꢀboranes that were prepared to probe the
effects of B-aryl groups. In a previous report,14 the standard
imidazol-2-ylidene boryl radical 1 with large N-dipp substitu-
radicals are boron analogues of diphenylmethyl (Ph2CH ) and that
(like diphenylmethyl) resonance stabilization is very significant.
3
3
ents was much longer-lived and far less reactive than analogue 3
3
with small NꢀMe groups. The set of NHCꢀboranes 2aꢀc was
intended to reveal the effect of B-aryl substituents with electron-
releasing and electron-withdrawing properties on comparatively
crowded and long-lived intermediates. By way of contrast,
NHCꢀboranes 4aꢀc were chosen to yield analogously substi-
tuted but transient NHCꢀboryl radicals. Compounds 5ꢀ7 were
selected to probe the combined effect of B-phenyl and imidazo-
lylidene ring substituents. All of the BꢀAr compounds are readily
The additional resonance stabilization in radicals 2 favors
3
H-atom abstraction from the parent NHCꢀBH2Ar reagents.
However, the extra steric shielding of the B-atom as compared to
parent NHCꢀBH3 1 opposes the approach of the t-BuO radical.
3
To assess the effect of aryl substitution on the H-atom donor
ability of 2aꢀc, competitive H-abstractions were carried out with
propan-2-ol. UV irradiation of solutions containing the NHC-
borane and propan-2-ol, together with DTBP, gave rise to EPR
10313
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
Table 1. Experimental and Computed EPR Parameters for Aryl-Substituted NHCꢀBoryl Radicalsa
NHCꢀboryl
T/K or basis setb
a(1HR)
a(11B)
a(2N)
a(2Hγ)
a(2Ho)
a(2Hm)
a(Hp)
c
1
300
11.4(2H)
9.3
7.3
6.6
4.0
1.0
2.8
3
2a
300
3.9
1.1
ꢀ2.0
1.6
1.1
0.9
nre
0.8
2.8
3
2a
6-31Gþ(d)d
290
ꢀ9.7
9.8
11.6
5.9
3.6, 3.4
3.8
ꢀ1.4, ꢀ0.7
1.6
ꢀ2.4
3
2b
3
f
2c
295
9.6
6.9
4.2
1.8
1.8
3
a All g-factors 2.0028 ( 0.0005; hfs in gauss. Note that only the magnitudes and not the signs of hfs can be derived from EPR spectra. b DFT
computations with the UB3LYP functional; geometry optimized with the 6-31Gþ(d) basis set. c Data from ref 13. d Computations with the EPR-iii basis
set failed. e nr = not resolved. f Possibly a(3F) = 0.8 G: correlation coefficient virtually unchanged on including this.
Figure 4. DFT computed structure (left) and SOMO (right) of
NHCꢀboryl radical 2a .
3
spectra showing both the NHCꢀboryl radical and the 2-hydroxy-
Figure 5. Decay of the EPR signal of radical 2c as a function of time at
3
propan-2-yl radical. Stock solutions of known concentrations of
each of 2a, 2b, or 2c together with DTBP were prepared. To each
was added a known amount of propan-2-ol, and then the EPR
spectrum was recorded during photolysis. The ratio of the
concentrations of the two radicals was obtained from spectral
simulations. The spectra were relatively weak, so only rough
estimates of the ratios of the two radicals could be obtained (see
the Supporting Information for spectrum from 2b).
300 K. Blue: Experimental data points. Red: Second-order line for 2kt =
4.5 ꢁ 104 Mꢀ1 sꢀ1
.
toward alkyl radicals than 1, it is clear that 2a is not a promising
reagent for radical chain reductions involving alkyl radicals. Appar-
ently, for the slow reactions of alkyl radicals with 2a, the steric
shielding of the phenyl group outweighs its resonance stabilization.
The kinetics of the termination processes of the dipp-Imd
To analyze these data, it is easily shown that kH/ksH
[NHCꢀB HAr][i-PrOH]/[Me2C OH][NHCꢀBH2Ar],
=
ꢀB HAr radicals were investigated by following the decays of
3
3
3
their EPR signals as functions of time. Samples of each of the
dipp-Imdꢀboranes and DTBP were prepared in PhH and
degassed, and then the radicals were generated as before. The
EPR magnetic field was fixed in turn at the maximum intensity of
the signal of each radical. The decays of the signals were then
monitored as a function of time on shuttering the UV photolysis
beam. Several separate samples were examined, and a typical
where kH refers to H-atom abstraction from the NHC-borane by t-
BuO and ksH refers to the propan-2-ol standard. From a sample
3
containing excess propan-2-ol to 2a (molar ratio 47:1), the measured
ratio of the derived radicals was 2a :Me2C OH about 6:1; hence
3
3
kH/ksH ≈280. For 2b, kH/ksH ≈230. Spectra from 2c were too weak
for analysis. The rate constant for H-atom abstraction from propan-
2-ol by t-BuO is 1.8 ꢁ 106 Mꢀ1 sꢀ1 at 295 K;21 hence the kH values
3
at 295 K for 2a and 2b are about 5 ꢁ 108 and 4 ꢁ 108 Mꢀ1 sꢀ1
,
curve is shown in Figure 5 for the decay of 2c at 300 K.
3
The initial concentrations of the dipp-ImdꢀB HAr radicals
respectively. These are marginally larger than the kH of the unsub-
stituted parent 1 (kH = 1.4 ꢁ 108 at 295 K).13 Apparently, the
additional resonance stabilization conferred on the NHCꢀboryl
radicals by the aryl substitution marginally outweighs the increased
steric shielding in these H-abstractions with alkoxy radicals.
3
were determined from the average of the double integrals of the full
spectra before and after photolysis. The decays could not be satisfac-
torily fitted by first-order plots, but reasonably good fits were
obtained for second-order decays having the 2kt values listed in
Table 2. The fits were slightly imperfect (see Figure 5 and Supporting
Information), suggesting that there may be a first-order contribution
Alkyl radicals also abstract H-atoms from unsubstituted car-
bene boranes.13,15 These rate constants are much lower (104ꢀ
105 Mꢀ1 sꢀ1), but the reactions are still important in preparative
radical chain reductions. To assess the effect of a phenyl substituent
on such reactions, we used a published method15,22 to estimate
the rate constant of a primary-alkyl radical (nonyl) with 2a. At
<104 Mꢀ1 sꢀ1, the rate constant was below the lower limit of the
experimental method. In contrast, the rate constant for reaction of a
primary-alkyl radical with 1 is somewhat above the limit at (2ꢀ4) ꢁ
to the decay. The half-lives of the dipp-ImdꢀB HAr radicals are
3
about 10 s. Therefore, a combination of the dipp-ImdꢀB HAr
3
radicals with t-BuO radicals cannot be the process that is first order
3
in dipp-ImdꢀB HAr in the dark because all of the t-BuO radicals
3
3
must react within a few milliseconds of shuttering the UV beam.
Table 2 shows that the termination rate constant for the B-
phenyl-substituted radical 2a is more than 2 orders of magnitude
smaller than 2kt for the unsubstituted radical 1 dipp-ImdꢀBH2 .
3
104 Mꢀ1 s
.
ꢀ1 13,15 While it is not clear how much less reactive 2a is
3
3
10314
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
Table 2. Termination Rate Constants for dipp-ImdꢀBoryl
Table 3. Relative EPR Signal Intensities from Reactions of
NHCꢀBoryl Radicals with 1-Bromobutane and 2-Bromo-2-
methylpropanea
Radicalsa
radical B-substituent temp/K [NHCꢀBH2 ]i/Mb 2kt Mꢀ1 sꢀ1
3
3
4a
4b
4c
5
7
1
2a
2b
2c
5.1 ꢁ 10ꢀ7
3.6 ꢁ 10ꢀ6
2.1 ꢁ 10ꢀ6
2.9 ꢁ 10ꢀ6
9.0 ꢁ 106
5.0 ꢁ 104
3.8 ꢁ 104
4.5 ꢁ 104
3
3
3
3
3
3
3
3
3
3
1
H
300
292
292
300
3
n-Bu
t-Bu
3.4 1.2
1.0 0.8
0.4
0.5
0.6 1.8 1.6 0.6 0.06 0.06 0.00
0.3 1.0 0.9 2.2 0.5 0.3 0.4
2a
2b
2c
Ph
3
3
4-MeOC6H4
3
3
a Determined from the double integrals of the EPR spectra of the n-Bu
4-CF3C6H4
3
3
a In PhH solution except 1 , which was in PhBu-t. b Initial concentration
of dipp-Imdꢀboryl radical.
3
3
3
or t-Bu radicals relative to the t-Bu radical intensity obtained from 3 .
3
be obtained from any of these precursors (except 6; see below).
There are probably two reasons for this. First, the corresponding
NHCꢀboryl radicals may all be transient because the alkyl-
substituted imidazole N-atoms provide less steric shielding to termi-
Scheme 1. Possible Terminations of Phenyl-Substituted
Radical 2a
3
nation by dimerization than in 2aꢀc . So radical concentrations are
3
low. Second, because the interaction of the unpaired electron with
the H-atoms of N-alkyl groups is significant,14 the spectrum of
radical 4a (NCH3) could be divided into as many as 2160 lines.
3
Individual resonance lines from 4a (and the others in the series 4b,
3
c, 5ꢀ7) could therefore be below the noise level.
To learn whether boryl radicals were being generated by
H-abstraction from these NHC-boranes, we added 1-bromobu-
tane or 2-bromo-2-methylpropane to EPR samples prior to irradia-
tion. Abstraction of halogen atoms from alkyl halides by boryl
radicals has been observed before by EPR spectroscopy,14 by laser
flash photolysis,9 and in preparative experiments.25 Indeed, all of the
NHC-boryl radicals derived from 4aꢀc, 5ꢀ7 abstracted bromine
from alkyl bromides to some extent, and the relative intensities of
the n-Bu and t-Bu radical spectra observed are shown in Table 3.
3
3
The initial reactant concentrations and reaction conditions
were essentially the same for all of these experiments so the relative
intensities (that is, the yield of n-Bu or t-Bu ) are related to the
We suggest that radical 2 terminates by dimerization in a head-to-
3
head fashion to give 8.14 Dimer 8 is a ligated diborane analogous to
compounds NHCꢀBH2ꢀBH2ꢀNHC that have been recently
reported by Robinson.23 We suggested that first-generation radical
3
3
overall efficiency of the bromine abstraction. However, the relative
intensities will also depend on the rates of termination reactions.
1a terminates to make a diborane,13 and recently Braunschweig has
Table 3 shows that the yield of n-Bu was highest for the least
3
3
observed such a dimer.5 The Ph substituent in 2a imposes addi-
shielded radical 3 and lowest for most shielded radicals 2aꢀc , with
3
3
3
tional shielding on both partners in a head-to-head combination
reaction yielding dimer 8, which probably explains the large decrease
in rate of termination of 2a as compared to 1a .
the other radicals giving intermediate yields. Consitent with this, LFP
studies show that 3 is more reactive toward 1-iodopropane than 2 .
3
3
The picture is less clear for the t-Bu series. The spread of values is
3
3
3
Matsumoto and Gabbaï have reported that an electrochemically
comparatively less, and the sterically congested radicals 1 and
3
generated NHCꢀB (mesityl)2 radical (mesityl = 2,4,6-trimethyl-
2aꢀc gave appreciable yields of t-Bu . This can possibly be
3
3
3
phenyl) is highly persistent,3 so much so that they even attempted to
obtain an X-ray structure. So we considered that termination might
take place by coupling of the boron center of one radical with the
para-site of the phenyl ring in the second radical to give head-to-tail
dimer 9as shown in Scheme 1. This type of coupling is subject to less
steric hindrance than head-to-head coupling and predominates in the
terminations of triphenylmethyl and related radicals.
attributed to a slower cross-combination of the larger t-Bu with
3
the congested NHCꢀboryl radical.
In addition to this evidence, in parallel studies we have
observed radicals 4a and 4b by LFP experiments.10b We conclude
3
3
that boryl radicals can be generated to some extent by H-atom
abstraction from precursors 4aꢀc, 5, and 7, but that these radicals
must be relatively short-lived. In contrast to the restrained behavior
of these N-methyl, N-isopropyl, and N-cyclohexyl substrates, irra-
diation of N-t-butyl substrate 6 rewarded us with a big surprise.
A Unique Diboranyl Radical. Samples of di-t-BuꢀImdꢀ
borane 6 in PhH with DTBP were irradiated, and in time strong
spectra were obtained; however, these spectra (Figure 6) exhibited
four unusual features: (1) the strong signals took several minutes to
develop; (2) the hyperfine splitting pattern was substantially greater
in extent than that of the other NHCꢀboryls; (3) only extremely
weak n-Bu and t-Bu spectra were observed when alkyl bromides
To assess this possibility, we studied 4-methoxy- and 4-trifluoro-
methyl-substituted radicals 2b and 2c . Like Matsumoto and
3
3
Gabbaï’s radical, the para-positions in these radicals are blocked
to shut off head-to-tail dimerization. The 2kt values for 2b and
3
2c are very similar to 2kt for the phenyl derivative 2a (Table 2),
3
3
so we conclude that only head-to-head coupling to yield dimer 8
occurs. Accordingly, the extreme persistence of Matsumoto and
Gabbaï’s radical is not primarily due to the para-methyl groups
on the two mesityl rings. The reluctance of boron to form double
bonds to carbon is well documented,24 and this may explain why
head-to-tail dimers like 9 are not formed.
3
3
were added; and (4) decay of the radical species was much slower
than anticipated. From measurements of the initial radical con-
centration and decay curve fitting, a second-order rate constant of
2kt = 6 ꢁ 105 Mꢀ1 sꢀ1 was obtained at 280 K. The decay was also
Radical generation from aryl-substituted NHCꢀboranes 4aꢀc
and 5ꢀ7 was also examined, but no well-defined EPR signals could
10315
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
spectrum. The SOMO is a π-type orbital (Figure 7) spanning the
NHCꢀB (Ph)BH moiety. The calculated BꢀB bond length of
3
1.756 Å is nicely intermediate between the observed (X-ray) values
of 1.561 and 1.828 Å for the NHC-ligated diborene NHCꢀBHd
BHꢀNHC and diborane NHCꢀBH2ꢀBH2ꢀNHC, respec-
tively.23 The BꢀHβ bond length and BBHβ bond angle of 11m
are 1.233 Å and 105.0°, respectively. The computations show that
the BβꢀHβ bond made an angle of 23° with the plane of the SOMO
at Bβ. This, together with the extensive π-spin delocalization,
explains why the a(Hβ) is somewhat smaller than the maximum
permitted value (about 20 G). The computed Hβ hfs (and the hfs of
Figure 6. Overlay of the experimental (black) and simulated (red) EPR
spectra obtained from photolysis of DTBP and 6 in PhH at 300 K.
monitored at 300 K, and the data were more scattered, but the same
rate constant was derived. These factors suggested a different kind of
radical was produced from 6 and that this radical had more steric
the H’s in the PhB ring) are very sensitive to the dihedral angles
3
shielding around the B-atom than the expected radical 6 .
3
subtended with the SOMO. Considering the comparative ease of
rotation about the single bonds involved, the agreement of calcula-
tion with experiment (Table 4) is impressive.
A satisfactory computer simulation of the spectrum was
eventually obtained with the parameters shown in Table 4. The
EPR parameters show couplings to two N-atoms, two H-atoms, and
one B-atom with magnitudes suggesting the presence of an
NHCꢀBoranes with B-Alkyl Substituents. We previously
showed that H-abstraction from thexyl-borane 14 by t-BuO
3
NHCꢀB unit. In addition, the three hfs of 2.8 G (ortho and para
3
H-atoms) and two of 0.6 G (meta H-atoms) show the unpaired
electron is also delocalized into a phenyl ring. There remain a
coupling to a second B-atom and one comparatively large coupling
to an H-atom. These data suggest that the observed species is the
ligated diboranyl radical 11 (Scheme 2 and Figure 7) with one Hβ.
We suggest that radical 11 is formed by H-atom abstraction
from dimer 10. In turn, this is produced during photolysis by
combination of the initial radicals 6 , so the delay in the
3
development of the signal is sensible. Similarly, the greater extent
of steric shielding in 11 as compared to 6 explains the slow
3
decay and reluctance to abstract Br-atoms.
NHC-ligated diboranes are unusual species that have recently
been identified by Robinson and co-workers from the reduction
of NHCꢀBBr3 with potassium/graphite in THF.23 Rare dibor-
enes of type NHCꢀBHdBHꢀNHC were also produced from the
same reactions. These were thermally stable, although air- and
moisture-sensitive. We now find that neutral diboranyl radical 11,
which represents an intermediate stage between a diborene and a
diborane, is also a viable species with a half-life of ∼1 s in solution.
Unlike all of the other NHCꢀboryl radicals of this study, 11 contains
a labile β-H-atom; hence termination could take place by dispro-
portionation to re-form ligated diborane 10 together with an equal
amount of the novel ligated diborene 12. Dimerization to yield
ligated tetraborane 13 also is possible (Scheme 2). Radical 11 can be
viewed as the adduct from H-atom addition to diborene 12 (or
Figure 7. Structures of radicals 11 and 11m (left), and DFT computed
structure and overlaying SOMO of model diboranyl radical 11m (H-
atoms replace t-Bu groups on the imidazole rings of 11).
Scheme 2. Formation and Decay of Ligated Diboranyl
Radicals
indeed as the adduct from Ph radical addition to a diborene of struc-
3
ture NHCꢀBPhdBHꢀNHC). The structures of several neutral
and anionic diboranes with aromatic and aryloxyl substituents have
been reported,26 but the chemistry of ligated diborenes is in its
infancy.27 Our results suggest analogies with alkenes are appropriate.
The computed structure of 11m shows the NHCꢀB B unit as
3
virtually planar with the PhꢀB ring twisted somewhat out of this
3
plane (dihedral 44°). The imidazole and Ph rings attached to the BH
center are orthogonal, in agreement with their lack of hfs in the EPR
Table 4. EPR Parameters of Ligated Diboranyl Radical 11a
radical
T/K or DFT basis set
a(11BR)
a(Hβ)
a(11Bβ)
a(2N)
a(2Hi)
a(3Ho,p
)
a(2Hm)
11
300
11.7
14.4
5.5
11.2
15.1
14.9
3.4
ꢀ5.3
ꢀ4.8
3.4
3.3
2.3
2.4
2.8
0.6
0.6
0.5
model 11mb,c
model 11mb,c
6-31Gþ(d)
6-311þ(d,p)
ꢀ0.8, ꢀ1.2
ꢀ0.8, ꢀ1.1
ꢀ1.0, ꢀ1.2, ꢀ1.5
ꢀ1.0, ꢀ0.9, ꢀ1.3
a g-factor 2.0030 ( 0.0005; hfs in gauss. Only the magnitudes and not the signs of hfs can be derived from EPR spectra. b Model 11m as 11 but H replaces
t-Bu on N-atoms of the imidazole rings. c DFT computations with the UB3LYP functional; geometry optimized with the 6-31Gþ(d) basis set. Hfs for all
other Hs were computed to be e(0.1 G (except the NHs). Computations with the EPR-iii basis set failed.
10316
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
about 240 K by a new spectrum. This sharpened at 290 K to the
spectrum shown in the bottom of Figure 9. This is the spectrum
of previously reported NHCꢀboryl radical 1 .
3
We attempted to record spectra at lower temperatures in
cyclopropane, but these were ill-defined and weak, probably because
of poor solubility of 17. Computer simulation of the 220 K spectrum
provided the EPR parameters shown in Table 5. Comparison of
these data with those reported for radical 20b14 suggested that the
220 K spectrum belonged to C-centered radical 20a. This conclusion
was supported by the results of a DFT computation with model
radical 19 bearing N-methyl substituents (Figure 10 and Table 5). It
follows that the t-BuO radical selectively abstracts the tertiary
3
H-atom from the isobutyl group in preference to one of the BH2
hydrogens. At higher temperatures, radical 20a undergoes β-scission,
thus releasing NHCꢀboryl 1 and isobutene (see Scheme 3).
3
TheEPRspectraweretooweakforradical concentration measure-
ments, but we estimated by visual inspection and simulation that
radical 20a and the released radical 1 were of equal concentration at
3
about Tmid = 210 K. The activation energies (Ed) of many uni-
molecular reactions correlate with EPR -derived Tmid values accord-
ing to eq 1:28
Figure 8. The NHCꢀboranes with B-alkyl substituents studied.
Ed=kcal molꢀ1 ¼ 0:044Tmid þ 0:22
ð1Þ
Accordingly, Ed = 9.5 kcal molꢀ1 for the β-scission of 20a.
Strictly, this linear correlation is only expected to hold for processes
in which the rearranged and unrearranged radicals terminate at the
diffusion-controlled limit. The ejected radical 1 undergoes dimer-
3
ization more slowly than this, but termination may be dominated by
the much faster cross-combination with 20a. So the value obtained
from eq 1 is probably reasonably reliable. Assuming a normal pre-
exponential factor of 1013 sꢀ1, the rate constant kd for the β-scission
of 20a at 300 K is about 106 sꢀ1
.
The estimated Ed value compares well with the Ed = 8.1 kcal molꢀ1
previously obtained for dissociation of 20b.14 The higher activa-
tion energy for β-scission of 20a is as expected because the
isobutene released is less stable than the 2,3-dimethylbut-2-ene
released from 20b. The process is fast for both radicals, which
accords well with the idea that NHC-ligation weakens BꢀC
bonds as well as BꢀH bonds.11
Figure 9. Top: Overlay of computer simulation (red) and experimental
(black; after digital removal of broad central feature) EPR spectra
obtained from 17 in t-BuPh at 220 K. Bottom: EPR spectrum from 17
at 290 K (identified as NHCꢀboryl 1 ).
3
radicals takes place selectively at the tertiary H-atom of the thexyl
group.14 The C-centered radical generated in this way undergoes
β-scission to give 2,3-dimethylbut-2-ene and the dipp-
ImdꢀBH2 radical 1 . To more generally assess the regioselec-
EvansꢀPolanyi equations relating the activation energies of β-
scission reactions to the overall reaction enthalpies (ΔHo) have
been obtained.29 A basic relationship derived for hydrocarbon
radicals and ignoring polar effects is:
3
3
tivity of H-atom abstraction from B-alkylꢀNHCꢀboranes, new
compounds 15ꢀ18 were investigated (Figure 8; see Supporting
Information for preparation and characterization details).
Surprisingly, when a solution containing n-butyl precursor 15
and DTBP in t-BuPh was photolyzed in the EPR resonant cavity,
the only detectable species in the temperature range 220ꢀ290 K
was parent radical 1 . t-BuO radicals are normally unselective in
Ed ðkcal molꢀ1Þ ¼ 12:4 þ 0:79ΔHo
ð2Þ
The estimated reaction enthalpies derived from use of eq 2
with the Ed data for 20a,b indicate both dissociations are exothermic
by 4ꢀ5 kcal molꢀ1! This was unexpected because NHCꢀboryl
radicals add rapidly (by experiment) and irreversibly (by calcula-
tion) to methyl acrylate.9 The rapid addition is key to the success of
NHCꢀboranes as co-initiators in radical photopolymerization.
NHCꢀboryl radicals are nucleophilic,9,25 so polar effects operating
in the transition state favor addition to methyl acrylate and not
isobutene. Yet could the additions of NHCꢀboryl radicals to simple
alkenes really be endothermic? We turned to calculations to
independently address the reaction enthalpies.
3
3
abstracting from hydrocarbon chains. Thus, we expected that an
assortment of the C-centered radicals from abstraction of all types of
secondary H-atoms of the butyl chain might be formed along with
the NHCꢀB HBu-n radical. That only 1 was detected suggests
3
3
that CH3CH2CH CH2BH2ꢀNHC was selectively formed and
3
subsequently dissociated to give 1 and 1-butene. We tried to
observe the C-centered radical at low temperature, but the spectra
3
were broad and poorly resolved.
A DFT computational study of the closely related model radical
19 was undertaken with the UB3LYP functional and a 6-31Gþ(d)
basis set.14 The reaction coordinate for the β-scission of 19 to
When the isobutyl precursor 17 was examined, the complex
EPR spectrum shown in the top of Figure 9 was obtained at low
temperatures (<220 K) in t-BuPh solvent. On raising the
temperature, this spectrum weakened and was replaced above
NHCꢀboryl 3 and isobutene was followed by incrementation of
3
the BꢀC bond while allowing all other geometrical parameters to
10317
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
Table 5. EPR, Kinetic, and Thermodynamic Data for NHCꢀBora-alkyl Radicals
radical T/K or DFT basis set a(6Hβ)/G a(2Hβ)/G a(11B)/G a(2Hγ)/G Tmid/K Ed/kcal/mol kd/sꢀ1300 K ΔHokcal/mol
ref
20aa
19
220
21.1
19.6
20.8
13.1
11.6
19.3
20.2
16.0
1.5
210
180
9.5
7.8
8.1
106
ꢀ3.8
ꢀ4.0
ꢀ5.4
this work
this work
14
6-31Gþ(d)
155
ꢀ1.6
20b
107
a In PhBu-t solution; g = 2.0027.
Figure 11. Overlay of simulated (red) and experimental (black) EPR
spectra of 21 obtained from addition of 3 to 1,1-diphenylethene in PhH
3
at 290 K.
ready formation in this reaction (and others).11 The energies of the
reactant, products, and TS were corrected for their zero point
energies and for thermal effects to 298 K (Table 5). The DFT
results supported the finding of exothermic β-scission, and the
magnitudes of the calculated activation energy and reaction enthalpy
were in pleasing agreement with the experimental values obtained for
radical 20a (Table 5 and Figure 10).
These β-scissions are, of course, the reverse of NHCꢀBH2
3
radical additions to alkenes and have the same TSs. The BꢀCꢀC
angle of approach of 3 to isobutene was computed to be 104.3°. For
3
Figure 10. DFT reaction coordinate and transition state for β-scission
of model radical 19.
additions of C-centered radicals to various types of alkenes, the
computed CꢀCꢀC angles of approach lie in the alkyl carbon range
106ꢀ112°, and the lengths of the forming CꢀC bonds are between
2.17 and 2.31 Å.30,31 These values are not very different from the
BꢀCꢀC angle and BꢀC bond length that we calculated for disso-
Scheme 3. β-Scission of IsobutylꢀNHCꢀBorane 20a
ciation of 19. However, addition of 3 to isobutene was computed
3
to be endothermic by 4 kcal molꢀ1, and the addition activation
energy was 11.8 kcal molꢀ1 (close to the experimental value of
13.3 kcal molꢀ1). C-centered radicals provide a stark contrast. For
example, the experimental activation energy for addition of methyl
radical to isobutene is only 6.2 kcal molꢀ1, and the reaction is
exothermic by 19.5 kcal molꢀ1 30,31
.
The resonance-stabilized benzyl radical has been suggested as
a model for NHCꢀboryls.11,13 Experimental data for addition
of benzyl radical to isobutene are not available, but data for its
addition to other alkenes show the activation energy to be
about 9 kcal molꢀ1 for an exothermic reaction.30 We conclude
that NHCꢀboryl radical additions to unactivated alkenes are
inherently slower than are C-centered radical additions and
that atom transfer addition reactions to such alkenes (that is,
radical hydroborations) may be difficult to accomplish. On the
other hand, addition/elimination reactions32 of assorted ra-
dicals to B-allyl or B-alkenyl substituted NHCꢀboranes could
well be feasible.
To further ascertain whether boryl radicals are inclined to add to
alkenes, we examined the EPR spectra from photolyses of solutions
containing the most reactive NHCꢀborane 3, DTBP, and several
different alkenes including cyclohexene and acrylonitrile. However,
only weak and/or ill-defined signals were obtained. Reasoning that
addition might occur if the energy of the product radical were
lowered by significant resonance stabilization, we investigated the
relax. The BꢀC bond lengthened from its initial value of 1.683 to
2.230 Å in the transition state (TS) (Figure 10).
The computed HOMO of the TS showed that there was extensive
delocalization of the unpaired electron into the π-system of the
NHCꢀboryl unit. This supports the idea that resonance stabilization
of NHCꢀboryl radicals is an important contributing factor to their
reaction of 3 with 1,1-diphenylethene. This produced the EPR
3
spectrum shown in Figure 11.
10318
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
Scheme 4. EPR hfs (/G) of Adduct Radical 21 with Model R-
Ethyldiphenylmethyl Radical
Analyzing the EPR spectrum, the hfs to boron (4.5 G in 18a )
3
is significantly smaller than observed for other NHCꢀboryl
radicals (see, for example, Table 1). The HOMO of cyclopro-
pane is a degenerate pair of orbitals constructed from p-orbitals
in the plane of the ring (Walsh orbitals).35 A p-orbital at an
adjacent planar sp2 atom is well aligned to interact with either
member of the Walsh pair. This interaction leads to a small
activation of methylene groups adjacent to cyclopropane rings
toward H-atom abstraction36 and is responsible for the small but
significant stabilization energy of the cyclopropylmethyl radical
(2.4 kcal molꢀ1).37 In radical 18a the NHCꢀBH unit is planar
3
and its π-system can also align to interact with the cyclopropane
Walsh orbitals.
The DFT computed SOMO of model radical 18b (Figure 12
3
and Scheme 5) shows how the π-arrangement of the NHCꢀboryl
unit extends into the plane of the three-membered ring giving a
coherent π-system extended over the whole molecule. The calcu-
lated BꢀC3ring bond length of 1.556 Å is significantly less than the
analogous BꢀC bond length of the alkyl substituted 20a (1.668 Å)
The EPR parameters derived from the simulation are dis-
played on structure 21 in Scheme 4. Comparison of these hfs
with those of R-ethyldiphenylmethyl 22,33 which serves as a
model, supports the conclusion that addition takes place to afford
β-borylalkyl radical 21. We conclude that although NHCꢀboryls
do not readily add to alkyl-substituted alkenes, addition can take
place at ambient temperature to alkenes with substituents that
provide significant resonance stabilization in the adduct radical.
Compound 16 containing a B-thexyl group combined with
N-i-Pr substituents provided a pivotal comparison with com-
pound 14 containing B-thexyl and N-dipp substituents. To our
and even slightly less than that of the aryl-substituted 4a (1.560 Å).
3
The small a(11B) observed for 18a is a consequence of the signi-
3
ficant electron delocalization into the three-membered ring. The very
small magnitude of the hfs from Hβ in radical 18a is also unusual.
3
However, this is also explained by 18a adopting the delocalized
3
structure shown in Figure 12. This places Hβ in the nodal plane of the
SOMO so that its interaction with the unpaired electron is negligible.
To learn if the cyclopropyl group activates the adjacent BH2
bonds, we carried out competitive H-abstraction experiments
between 18a and propan-2-ol as described above (see the
surprise, reaction of 16 with t-BuO radicals gave the spectrum of
3
the corresponding B-centered radical 16 (see the Supporting
3
Information and Table 6). In 14, the dipp substituents shield the
boron center to such an extent that H-abstraction at the tertiary
H-atom of the thexyl group is preferred. Similar behavior is
observed for 15 and 17, reinforcing the conclusion that this is due
to the N-substituent, not the B-substituent (which is bulky in 14
but not in 15 or 17). On the other hand, in 16, steric shielding by
the i-Pr substituents is much less important and H-abstraction at
the weaker BꢀH bonds proceeds. This is an outstanding example
of the steric regio-switching that can be accomplished by means
of NHC scaffolds (Scheme 3).
Figure 12. DFT computed structure and SOMO for cyclopropylꢀ
boryl radical 18b .
Ring-opening is the defining reaction of cyclopropylcarbinyl
and related radicals (cyclopropyloxy, cyclopropylaminyl),34 so
we wondered whether cyclopropylboryl radicals would open.
Hydrogen abstraction from the cyclopropyl-substituted precur-
3
Scheme 5. Lack of Ring-Opening of the Ligated Cyclopro-
pylboranyl Radical 18a
sor 18a by t-BuO radicals gave a good quality EPR spectrum
3
3
(see the Supporting Information) that was analyzed as shown in
Table 6. This spectrum was essentially unchanged in the range of
170ꢀ350 K. Upon addition of 1-bromobutane or 2-methyl-2-
bromobutane, the spectra of the corresponding alkyl radicals
were observed. These results are good evidence that radical 18a
has formed.
3
Table 6. EPR Parametersa for NHCꢀB -alkyl Radicals
3
radical
T/K or DFT basis set
a(11B)
a(1HR)
a(2N)
a(2Hi)
a(1Hβ)
a(2Hγ)
a(2Hγ)
16
7.8
4.5
10.8
9.0
4.1
4.2
3.8
2.4
nr
3
18a
18b
18b
290
∼1.2
ꢀ1.0
ꢀ1.6
nr
∼1.2
1.1
nr
3
6-31Gþ(d)
EPR-iii
10.3
3.2
ꢀ8.9
ꢀ8.3
0.2
0.1
ꢀ0.7
0.7
3
1.2
3
a g-factors = 2.0030 ( 0.0005. nr = not resolved.
10319
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
Supporting Information). A rate constant ratio kH(18a)/
kH(i-PrOH) = 77.0 was determined at 295 K. Use of the
known21 rate constant for H-atom abstraction from propan-2-ol
open even up to 350 K. This is a striking demonstration of the
reluctance of BdC double bonds to form.
by t-BuO leads to kH(18a) = 1.4 ꢁ 108 Mꢀ1 sꢀ1 at 295 K. This
3
’ EXPERIMENTAL SECTION
rate constant is the same as that for H-abstraction from the
standard dipp-ImdꢀBH3 1.14 If there is indeed a small activation
of the BH2 bonds in 18a by the cyclopropyl group, then this is
counter-balanced by increased steric shielding of the boron.
Cyclopropylcarbinyl, cyclopropylaminyl, and cyclopropyloxy
radicals ring-open extremely rapidly to relieve ring strain.37
Preparation and Characterization of the NHCꢀBoranes.
See the Supporting Information.
EPR Spectroscopy. EPR spectra were obtained with a Bruker EMX
10/12 spectrometer fitted with a rectangular ER4122 SP resonant cavity
and operating at 9.5 GHz with 100 kHz modulation. Stock solutions of
each NHCꢀborane (2ꢀ15 mg) and DTBP (ca. 0.1 mL) in tert-
butylbenzene or benzene (1.0 mL) were prepared and sonicated if
necessary. An aliquot (0.2 mL), to which any additional reactant had
been added, was placed in a 4 mm o.d. quartz tube, deaerated by
bubbling nitrogen for 15 min, and photolyzed in the resonant cavity by
unfiltered light from a 500 W super pressure mercury arc lamp. Solutions
in cyclopropane were prepared on a vacuum line by distilling in the
cyclopropane, degassing with three freezeꢀpumpꢀthaw cycles, and
finally flame sealing the tubes. Most of the EPR spectra were recorded
with 2.0 mW power, 0.8 Gpp modulation intensity, and gain of ca. 106. In
all cases where spectra were obtained, hfs were assigned with the aid
of computer simulations using the Bruker SimFonia and NIEHS
Winsim2002 software packages. For kinetic measurements, precursor
samples were used mainly in “single shot” experiments; that is, new
samples were prepared for each temperature and each concentration to
minimize sample depletion effects. In a few cases, second shot data were
included. EPR signals were double integrated using the Bruker WinEPR
software, and radical concentrations were calculated by reference to the
double integral of the signal from a known concentration of the stable
radical 2,2-diphenyl-1-picrylhydrazyl (DPPH) [1 ꢁ 10ꢀ3 M in PhMe],
run under identical conditions, as described previously.39
However, the observation of the spectrum of 18a demonstrates
3
that ring-opening of this ligated cyclopropylboryl radical to
produce radical 23 does not occur even up to 350 K.38 By using
eq 1, we estimated that the activation energy for this ring-open-
ing is >15.6 kcal molꢀ1 and that the rate constant at 300 K is
<40 sꢀ1. The reluctance of 18a to undergo β-scission is again a
3
consequence of the instability of BdC double bonds. In agree-
ment with this, a DFT computation [UB3LYP/6-31Gþ(d)]
with the model N,N0-dimethylimidazol-2-ylidene analog 18b
showed a strongly endothermic β-scission with only the shal-
lowest of minima for the ring-opened radical, the latter having a
distorted geometry (see the Supporting Information).
3
’ CONCLUSIONS
Hydrogen atom abstraction by t-BuO radicals readily took
3
place from the BH2 groups of all of the second-generation B-aryl
substituted NHCꢀboranes. Spin was extensively delocalized in the
resulting radicals, which had SOMOs extending across the NHC,
BH, and aryl units, thus making them bora-analogues of diphenyl-
methyl radicals. These NHCꢀB HAr radicals were substantially
3
less reactive than NHCꢀBH2 in abstracting Br-atoms from alkyl
3
’ ASSOCIATED CONTENT
bromides, probably because of increased steric shielding of their
radical centers.
S
Supporting Information. Experimental details of the
b
Radicals 2aꢀc , with large dipp substituents pendent from the
3
preparations of the boranes; NMR spectra of all new com-
pounds; sample EPR spectra of NHCꢀboryl and related radicals;
experimental and computed hfs of NHCꢀboryl radicals; kinetic
data for NHCꢀboryl radical decays; details of photochemical
reactions; and DFT-optimized geometries and energies for the
NHCꢀboryl radicals. This material is available free of charge via
the Internet at http://pubs.acs.org.
NHC groups, underwent dimerizations about 2 orders of magnitude
slower than analogues without the B-aryl substituents, a conse-
quence of the additional steric shielding. The evidence favored head-
to-head coupling yielding ligated diborane derivatives rather than
head-to-tail coupling through the aromatic ring of the substituent. In
the case of the di-t-butyl-substituted complex 6, a new type of ligated
diboranyl radical, with a structure intermediate between that of a
ligated diborane and a diborene, was generated from the head-to-
head dimer.
’ AUTHOR INFORMATION
The behavior of B-alkyl-substituted NHCꢀboranes depended
strongly on an interplay between the NHC-substituents and the
B-substituents. For ligated boranes with large pendant dipp substit-
uents, H-abstraction was directed away from boron to the alkyl
groups. The resulting β-bora-alkyl radicals underwent exothermic
β-scissions with production of dipp-ImdꢀBH2 radical (1 ) and an
Corresponding Author
jcw@st-andrews.ac.uk; curran@pitt.edu; emmanuel.lacote@
icsn.cnrs-gif.fr
Present Addresses
ICSN CNRS, B^atiment 27, Av. de la Terrasse, 91198
Gif-sur-Yvette Cedex, France.
3
3
alkene. The reverse reactions, that is, additions of NHCꢀBH2 to
3
simple alkenes, are projected to be unfavorable and endothermic,
but addition of 3 to 1,1-diphenylethene does occur. In a remarkable
3
’ ACKNOWLEDGMENT
example of the steric control achievable with NHC scaffolds, for
the ligated borane with N,N0-diisopropyl substituents, the site of
H-abstraction reverted to boron, and the alkyl-substituted-boryl
radical was obtained.
The B-cyclopropylꢀNHCꢀboryl radical generated from 18a
displayed evidence of electron delocalization into the Walsh
orbitals in the plane of the three-membered ring. Unlike other c-
This work was supported by grants from the U.S. National
Science Foundation (CHE-0645998 to D.P.C.), from Agence
Nationale de la Recherche (ANR, BLAN0309 Radicaux Verts,
and 08-CEXC-011-01, Borane), UPMC, CNRS, and IUF (M.M.,
L.F.). Technical assistance (MS, elemental analyses) was gener-
ously offered by FR 2769. J.C.W. thanks EaStChem for financial
support. We thank Jacques Lalevꢀee for helpful discussions.
C3H5X radicals (X = CR2, NR, O), this radical did not ring-
3
10320
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321

Journal of the American Chemical Society
ARTICLE
’ REFERENCES
(20) Previous work has shown computed a(11B) values depend
strongly on the basis set employed.
(21) (a) Paul, H.; Small, R. D.; Scaiano, J. C. J. Am. Chem. Soc. 1978,
100, 4520–4527. (b) Malatesta, V.; Scaiano, J. C. J. Org. Chem. 1982,
47, 1455–1459.
(1) (a) Kuhn, N.; Henkel, G.; Kratz, T.; Kreutzberg, J.; Boese, R.;
Maulitz, A. H. Chem. Ber. 1993, 126, 2041–2045. (b) Ramnial, T.; Jong,
H.; McKenzie, I. D.; Jennings, M.; Clyburne, J. A. C. Chem. Commun.
2003, 1722–1723. (c) Yamaguchi, Y.; Kashiwabara, T.; Ogata, K.; Miura,
Y.; Nakamura, Y.; Kobayashi, K.; Ito, T. Chem. Commun. 2004,
2160–2161.
(2) Ueng, S.-H.; Makhlouf Brahmi, M.; Derat, E.; Fensterbank, L.;
Lac^ote, E.; Malacria, M.; Curran, D. P. J. Am. Chem. Soc. 2008, 130,
10082–10083.
(3) (a) Weber, L.; Dobbert, E.; Stammler, H.-G.; Neumann, B.;
Boese, R.; Bl€aser, D. Chem. Ber. 1997, 130, 705–710. (b) Matsumoto, T.;
Gabbaï, F. P. Organometallics 2009, 28, 4252–4253.
(4) (a) Braunschweig, H.; Chiu, C.-W.; Radacki, K.; Kupfer, T.
Angew. Chem., Int. Ed. 2010, 49, 2041–2044. (b) Monot, J.; Solovyev, A.;
(22) We thank Mr. Andrey Solovyev for conducting this experiment.
(23) (a) Wang, Y.; Quillian, B.; Wei, P.; Wannere, C. S.; Xie, Y.;
King, R. B.; Schaefer, H. F., III; Schleyer, P. v. R.; Robinson, G. H. J. Am.
Chem. Soc. 2007, 129, 12412–12413. (b) Wang, Y.; Quillian, B.; Wei, P.;
Xie, Y.; Wannere, C. S.; King, R. B.; Schaefer, H. F., III; Schleyer, P. von
R.; Robinson, G. H. J. Am. Chem. Soc. 2008, 130, 3298–3299.
(24) See, for example: (a) Eisch, J. J. Adv. Organomet. Chem. 1996,
39, 355. (b) Berndt, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 985–988.
(25) Ueng, S.-H.; Fensterbank, L.; Lac^ote, E.; Malacria, M.; Curran,
D. P. Org. Biomol. Chem. 2011, 9, 3415–3420.
(26) (a) Moezzi, A.; Bartlett, R. A.; Power, P. P. Angew. Chem., Int.
Ed. Engl. 1992, 31, 1082–1083. (b) Moezzi, A.; Olmstead, M. M.; Power,
P. P. J. Am. Chem. Soc. 1992, 114, 2715–2717. (c) Grigsby, W. J.; Power,
P. P. Chem. Commun. 1996, 2235–2236. (d) Grigsby, W. J.; Power, P. P.
Chem.-Eur. J. 1997, 3, 368–375.
(27) Scheschkewitz, D. Angew. Chem., Int. Ed. 2008, 47, 1995–1997.
(28) Walton, J. C. J. Chem. Soc., Perkin Trans. 2 1998, 603–605.
(29) (a) Sabbe, M. K.; Reyniers, M.-F.; Van Speybroeck, V.;
Waroquier, M.; Marin, G. B. ChemPhysChem 2008, 9, 124–140.
(b) Sirjean, B.; Glaude, P. A.; Ruiz-Lopez, M. F.; Fournet, R. J. Phys.
Chem. A 2008, 112, 11598–11610.
(30) (a) Wong, M. W.; Pross, A.; Radom, L. J. Am. Chem. Soc. 1993,
115, 11050–11051. (b) Wong, M. W.; Pross, A.; Radom, L. J. Am. Chem.
Soc. 1994, 116, 6284–6292. (c) Wong, M. W.; Pross, A.; Radom, L. J. Am.
Chem. Soc. 1994, 116, 11938–11943.
ꢀ
ꢀ
Bonin-Dubarle, H.; Derat, E.; Curran, D. P.; Robert, M.; Fensterbank,
L.; Malacria, M.; Lac^ote, E. Angew. Chem., Int. Ed. 2010, 49, 9166–9169.
(5) Bissinger, P.; Braunschweig, H.; Kraft, K.; Kupfer, T. Angew.
Chem., Int. Ed. 2011, 50, 4704–4707.
(6) Ueng, S.-H.; Fensterbank, L.; Lac^ote, E.; Malacria, M.; Curran,
D. P. Org. Lett. 2010, 12, 3002–3005.
(7) (a) Chu, Q.; Makhlouf Brahmi, M.; Solovyev, A.; Ueng, S.-H.;
Curran, D. P.; Malacria, M.; Fensterbank, L.; Lac^ote, E. Chem.-Eur. J.
2009, 15, 12937–12940. (b) Lindsay, D. M.; McArthur, D. Chem.
Commun. 2010, 46, 2474–2476. (c) McArthur, D.; Butts, C. P.; Lindsay,
D. M. Chem. Commun. 2011, 47, 6650ꢀ6652.
(8) Monot, J.; Makhlouf Brahmi, M.; Ueng, S.-H.; Robert, C.;
Desage-El Murr, M.; Curran, D. P.; Malacria, M.; Fensterbank, L.;
Lac^ote, E. Org. Lett. 2009, 11, 4914–4917.
(9) (a) Tehfe, M.-A.; Monot, J.; Makhlouf Brahmi, M.; Bonin-
Dubarle, H.; Curran, D. P.; Malacria, M.; Fensterbank, L.; Lac^ote, E.;
Lalevꢀee, J.; Fouassier, J.-P. Polym. Chem. 2011, 2, 625–631. (b) Tehfe,
M.-A.; Makhlouf Brahmi, M.; Fouassier, J.-P.; Curran, D. P.; Malacria,
M.; Fensterbank, L.; Lac^ote, E.; Lalevꢀee, J. Macromolecules 2010,
43, 2261–2267.
(10) (a) Rablen, P. R. J. Am. Chem. Soc. 1997, 119, 8350–8360.
(b) Rablen, P. R.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 4648–4653.
(11) Walton, J. C. Angew. Chem., Int. Ed. 2009, 48, 1726–1728.
(12) Hioe, J.; Karton, A.; Martin, J. M. L.; Zipse, H. Chem.-Eur. J.
2010, 16, 6861–6865.
(31) Fischer, H.; Radom, L. Angew. Chem., Int. Ed. 2001, 40, 1340–1371.
(32) Rosenstein, I. In Radicals in Organic Synthesis, 1st ed.; Renaud,
P., Sibi, M., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 1, pp 50ꢀ71.
(33) Tiꢁno, J.; Borsig, E.; Pilaꢀr, J. Collect. Czech. Chem. Commun.
1969, 34, 66.
(34) (a) Newcomb, M. In Radicals in Organic Synthesis, 1st ed.;
Renaud, P., Sibi, M., Eds.; Wiley-VCH: Weinheim, 2001; Vol 1,
pp 317ꢀ336. (b) Walton, J. C. In Carbocyclic Three- and Four-Membered
Ring Compounds; de Meijere, A., Ed.; Georg Thieme Verlag: Stuttgart, 1997;
Vol. E 17c of Houben Weyl, Methoden der Organischen Chemie,
pp 2438ꢀ2525. (c) Nonhehel, D. C. Chem. Soc. Rev. 1993, 22, 347–359.
(35) (a) Walsh, A. D. Trans. Faraday Soc. 1949, 45, 179. (b) Hoffmann,
R. J. Am. Chem. Soc. 1968, 90, 1475–1485. (c) Hemmersbach, P.; Klessinger,
M. Tetrahedron 1980, 36, 1337–1343.
(13) Ueng, S.-H.; Solovyev, A.; Yuan, X.; Geib, S. J.; Fensterbank, L.;
Lac^ote, E.; Malacria, M.; Newcomb, M.; Walton, J. C.; Curran, D. P.
J. Am. Chem. Soc. 2009, 131, 11256–11262.
(36) Roberts, C.; Walton, J. C. J. Chem. Soc., Chem. Commun. 1984,
1109–1111. (b) Ingold, K. U.; Walton, J. C. Acc. Chem. Res. 1986, 19, 72–77.
(37) Walton, J. C. Magn. Reson. Chem. 1987, 25, 998–1000.
(38) Even more dramatic changes in equilibrium constants between
a closed shell bicyclic ring and a diradical have been observed by nuclear
substitutions of borons for carbons in a bis-bora-analog of cyclobutane-
1,3-diyl, see: Scheschkewitz, D.; Amii, H.; Gornitzka, H.; Schoeller, W.
W; Bourissou, D.; Bertrand, G. Science 2002, 295, 1880–1881.
(39) Maillard, B.; Walton, J. C. J. Chem. Soc., Perkin Trans. 2 1985,
443–450.
(14) Walton, J. C.; Makhlouf Brahmi, M.; Fensterbank, L.; Lac^ote,
E.; Malacria, M.; Chu, Q.; Ueng, S.-H.; Solovyev, A.; Curran, D. P. J. Am.
Chem. Soc. 2010, 132, 2350–2358.
(15) Solovyev, A.; Ueng, S.-H.; Monot, J.; Fensterbank, L.; Malacria,
M.; Lac^ote, E.; Curran, D. P. Org. Lett. 2010, 12, 2998–3001.
(16) (a) Baban, J. A.; Marti, V. P. J.; Roberts, B. P. J. Chem. Soc.,
Perkin Trans. 2 1985, 1723–1733. (b) Paul, V.; Roberts, B. P. J. Chem.
Soc., Perkin Trans. 2 1988, 1183–1193. (c) Baban, J. A.; Roberts, B. P.
J. Chem. Soc., Perkin Trans. 2 1988, 1195–1200. (d) Kirwan, N.; Roberts,
B. P. J. Chem. Soc., Perkin Trans. 2 1989, 539–550. (e) Dang, H. S.;
Roberts, B. P. Tetrahedron Lett. 1992, 33, 6169–6172. (f) Dang, H. S.;
Diart, V.; Roberts, B. P.; Tocher, D. A. J. Chem. Soc., Perkin Trans. 2
1994, 1039–1045. (g) Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin
Trans. 2 1994, 2411–2422. (h) Lucarini, M.; Pedulli, G. F.; Valgimigli, L.
J. Org. Chem. 1996, 61, 1161–1164. (i) Barton, D. H. R.; Jacob, M.
Tetrahedron Lett. 1998, 39, 1331–1334. (j) Sheeler, B.; Ingold, K. U.
J. Chem. Soc., Perkin Trans. 2 2001, 480–486.
(17) Makhlouf Brahmi, M.; Monot, J.; Desage-El Murr, M.; Curran,
D. P.; Fensterbank, L.; Lac^ote, E.; Malacria, M. J. Org. Chem. 2010,
75, 6983–6985.
(18) Exner, O. In Conformational Analysis in Chemistry; Chapman,
N. B., Shorter, J., Eds.; Plenum: New York, 1978; pp 439ꢀ540.
(19) Frisch, M. J.; et al. Gaussian 03, revision A.1; Gaussian, Inc.:
Pittsburgh, PA, 2003.
10321
dx.doi.org/10.1021/ja2038485 |J. Am. Chem. Soc. 2011, 133, 10312–10321