N-heterocyclic carbene boryl radicals: A new class of boron-centered radical

Article abstract of DOI:10.1021/ja904103x

Reduction of xanthates by N-heterocyclic carbene boranes (NHC-boranes) has been suggested to occur by a radical chain mechanism involving heretofore unknown NHC-boryl radicals. In support of this suggestion, both the expected borane dithiocarbonate product and an unexpected borane xanthate product have now been isolated. These are the first NHC-boranes with boron-sulfur bonds, and their structures have been secured by spectroscopic and crystallographic means. The first rate constants for H-atom transfer from an NHC borane complex were determined by using the ring opening of a substituted cyclobutylcarbinyl radical as a clock reaction. The rate constant for reaction of the NHC-borane with a secondary alkyl radical at ambient temperature is 4 × 10⁴ M^-1 s^-1, and the Arrhenius function displayed an entropic term (log A term) that was typical for a bimolecular reaction. The B-H bond dissociation energy of an NHC-borane complex has been estimated at 88 kcal/mol. The putative NHC-boryl radical in these transformations has been detected by EPR spectroscopy. Spectral analysis suggests that it is a π-radical, analogous to the benzyl radical.

Full text of DOI:10.1021/ja904103x

Published on Web 07/16/2009
N-Heterocyclic Carbene Boryl Radicals: A New Class of
Boron-Centered Radical
Shau-Hua Ueng,†,‡ Andrey Solovyev,† Xinting Yuan,§ Steven J. Geib,†
Louis Fensterbank,‡ Emmanuel Lacoˆte,‡ Max Malacria,*,‡ Martin Newcomb,*,§
John C. Walton,*,| and Dennis P. Curran*,†
Department of Chemistry, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15260, UPMC
UniV Paris 06, Institut Parisien de Chimie Mole´culaire (UMR CNRS 7201), C. 229, 4 place
Jussieu, 75005 Paris, France, Department of Chemistry, UniVersity of Illinois at Chicago,
845 West Taylor Street, Chicago, Illinois 60607, and School of Chemistry, EaStChem, UniVersity
of St. Andrews, St. Andrews, Fife KY16 9ST, U.K.
Received May 28, 2009; E-mail: max.malacria@upmc.fr; men@uic.edu; jcw@st-andrews.ac.uk;
curran@pitt.edu
Abstract: Reduction of xanthates by N-heterocyclic carbene boranes (NHC-boranes) has been suggested
to occur by a radical chain mechanism involving heretofore unknown NHC-boryl radicals. In support of this
suggestion, both the expected borane dithiocarbonate product and an unexpected borane xanthate product
have now been isolated. These are the first NHC-boranes with boron-sulfur bonds, and their structures
have been secured by spectroscopic and crystallographic means. The first rate constants for H-atom transfer
from an NHC borane complex were determined by using the ring opening of a substituted cyclobutylcarbinyl
radical as a clock reaction. The rate constant for reaction of the NHC-borane with a secondary alkyl radical
at ambient temperature is 4 × 104 M-1 s-1, and the Arrhenius function displayed an entropic term (log A
term) that was typical for a bimolecular reaction. The B-H bond dissociation energy of an NHC-borane
complex has been estimated at 88 kcal/mol. The putative NHC-boryl radical in these transformations has
been detected by EPR spectroscopy. Spectral analysis suggests that it is a π-radical, analogous to the
benzyl radical.
Introduction
bonds in the H-atom donor. Inherently strong hydrogen-element
bonds can sometimes be considerably weakened by Lewis
Carbon-radical chain reaction methodology has become an
essential component of organic syntheses.1 In the simplest
version of a chain reaction, a carbon radical is generated from
a precursor and reduced by H-atom transfer from a donor. In
turn, the radical formed from the H-atom donor reacts with
another precursor to produce a second carbon radical, and the
cycle repeats. Group 14 metal hydrides, especially tin hydrides
such as tributyltin hydride, are well suited for reductive chain
reactions based on the kinetics of their reactions with carbon
radicals,2 but concerns about toxicity of tin and an emphasis
on green chemistry have led to an increasing interest in
alternative hydrogen atom donors for carbon radical reductions.3
acid-base complexation, and recent work has shown that the
O-H bonds of trialkylborane-water (alcohol) and titanium-
(III)-water complexes can serve as H-atom donors to carbon
radicals.4
Free boranes have strong boron-hydrogen bonds so they do
not behave as hydrogen atom donors.5 In contrast, complexes
of boranes such as amine boranes (R3N-BH3)6,7 and phosphine
boranes (R3P-BH3)8 can serve as radical hydrogen atom donors
in some settings. However, the B-H bonds of such complexes
(4) (a) Spiegel, D. A.; Wiberg, K. B.; Schacherer, L. N.; Medeiros, M. R.;
Wood, J. L. J. Am. Chem. Soc. 2005, 127, 12513–12515. (b) Pozzi,
D.; Scanlan, E. M.; Renaud, P. J. Am. Chem. Soc. 2005, 127, 14204–
14205. (c) Cuerva, J. M.; Campan˜a, A. G.; Justicia, J.; Rosales, A.;
Oller-Lo´pez, J. L.; Robles, R.; Ca´rdenas, D. J.; Bun˜uel, E.; Oltra, J. E.
Angew. Chem., Int. Ed. 2006, 45, 5522–5526. (d) Medeiros, M. R.;
Schacherer, L. N.; Spiegel, D. A.; Wood, J. L. Org. Lett. 2007, 9,
4427–4429. (e) Allais, F.; Boivin, J.; Nguyen, V. T. Beilstein J. Org.
Chem. 2007, 3, No. 46. (f) Jin, J.; Newcomb, M. J. Org. Chem. 2008,
73, 7901–7905.
Useful alternatives to metal hydrides must react rapidly
enough with carbon radicals to maintain chain reactions, and
fast H-atom transfer reactions require weak hydrogen-element
† University of Pittsburgh.
‡ UPMC.
§ University of Illinois at Chicago.
| University of St. Andrews.
(5) Rablen, P. R.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 4648–
4653.
(1) (a) Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-
VCH: Weinheim, 2001. (b) Radicals In Synthesis; Gansäuer, A., Ed.;
Springer-Verlag: Berlin, 2006; Vols. 263 and 264.
(2) Chatgilialoglu, C.; Newcomb, M. AdV. Organomet. Chem. 1999, 44,
67–112.
(6) Reviews of amine and phosphine boranes: (a) Carboni, B.; Monnier,
L. Tetrahedron 1999, 55, 1197–1248. (b) Carboni, B.; Carreaux, F.
In Science of Synthesis, Organometallics: Boron Compounds; Kauf-
mann, D. E., Matteson, D. S., Eds.; Verlag: Stuttgart, 2004; Vol. 6,
pp 455-484. (c) Gaumont, A. C.; Carboni, B. In Science of Synthesis,
Organometallics: Boron Compounds; Kaufmann, D. E., Matteson,
D. S., Eds.; Verlag: Stuttgart, 2004; Vol. 6, pp 485-512.
(3) (a) Baguley, P. A.; Walton, J. C. Angew. Chem., Int. Ed. 1998, 37,
3072–3082. (b) Studer, A.; Amrein, S. Synthesis 2002, 835–849. (c)
Darmency, V.; Renaud, P. Top. Curr. Chem. 2006, 263, 71–106.
9
11256 J. AM. CHEM. SOC. 2009, 131, 11256–11262
10.1021/ja904103x CCC: $40.75  2009 American Chemical Society

NHC-Boryl Radicals
A R T I C L E S
Scheme 1. Postulated Chain Mechanism for Reduction of
Xanthates by N-Heterocyclic Carbene Boranes
Figure 1. Structures of N-heterocyclic carbene boranes used in xanthate
reductions.
are still rather strong,9 so the derived boryl radicals (for example,
R3N-BH2•) can also abstract hydrogen atoms from suitable
substrates. Such complexes are useful in applications of “polarity
reversal catalysis” of radical reactions.10
Based on a structural analysis and calculations of B-H bond
dissociation energies (BDEs), we recently suggested the N-
heterocyclic carbene boranes would serve as radical hydrogen
atom donors.11 This suggestion was supported experimentally
by reductive deoxygenation of a collection of secondary
xanthates (Barton-McCombie reaction12) with stable NHC-
boranes 1,3-bis-(2,6-diisopropylphenyl)imidazol-2-ylidene bo-
rane 1a and 2-phenyl-1,2,4-triazol-3-ylidene borane 1b (Figure
1).
1). Fragmentation of 4a,b provides alkyl radical 5 (step 2), which
in turn abstracts hydrogen from the starting NHC-borane 1a,b
to complete the chain cycle (step 3). If this mechanism is correct,
then it should be possible to characterize the borane product
6a,b (step 2) and to measure a rate constant (kH) for the radical
hydrogen transfer reaction (step 3).
We suggested11a that this transformation occurs by a radical
chain reaction that is analogous to the reduction of xanthates
by other radical hydrogen atom donors such as tin and silicon
hydrides.12 This implies the intermediacy of heretofore unknown
NHC-boryl radicals. Here we describe the isolation and
characterization of the expected boron-derived product from
reaction of a secondary xanthate with an NHC-boryl radical.13
We also describe an unexpected boron product isolated from
the reaction of a primary xanthate with the NHC-boryl radical.
We provide the first rate constant measurements for hydrogen
abstraction from an NHC-borane by an alkyl radical, and this
in turn allows an estimate of the BDE of the boron-hydrogen
bond. Finally, we observe the postulated NHC-boryl radical in
these transformations directly by EPR spectroscopy. Taken
together, the results confirm the existence of NHC-boryl radicals
and secure the radical chain mechanism for xanthate reduction.
We expected from the preliminary studies11a that the radical
hydrogen transfer reaction would have a rate constant e105 M-1
s-1 because a hexenyl radical probe substrate gave only the
cyclized product and because deuterium labeling with an NHC-
BD3 reagent was only moderately efficient (70%). Accordingly,
we selected the opening of a cyclobutylcarbinyl radical14 as an
appropriate clock reaction15 to time the radical hydrogen
transfer. We started by using the clock substrate to investigate
the boron-derived products.
Boron-Derived Reaction Products. A flask containing a
benzene solution of undecylcyclobutyl carbinyl xanthate 2a,
NHC-borane 1a, and triethylborane (1 equiv each, Scheme 2)
was exposed to air by piercing its stopper with a needle. After
24 h, the starting xanthate was consumed. The mixture was
charged to silica gel, and initial elution with dichloromethane
provided a nonpolar fraction consisting of directly reduced and
ring opened products16 (see below for the discussion of the
reduced products). Subsequent elution with dichloromethane/
methanol (8/1) provided pure boron-derived product, (meth-
ylthiocarbonylthio)borane 6a, as a white solid, mp 201-202
°C.
Results and Discussion
Mechanistic Framework. The postulated propagation steps
for the reduction of xanthates by NHC-boranes are shown in
Scheme 1.12 NHC-boryl radical 3a,b adds to the CdS bond of
the xanthate 2 (possibly reversibly) to provide adduct 4a,b (step
(7) (a) Dang, H. S.; Roberts, B. P. Tetrahedron Lett. 1992, 33, 6169–
6172. (b) Dang, H. S.; Diart, V.; Roberts, B. P.; Tocher, D. A. J. Chem.
Soc., Perkin Trans. 2 1994, 1039–1045. (c) Roberts, B. P.; Steel, A. J.
J. Chem. Soc., Perkin Trans. 2 1994, 2411–2422.
With 6a in hand, we repeated the previously reported11a
reduction of secondary xanthate 2b but now with an eye toward
isolating the boron product rather than the deoxygenation
product. Indeed, 6a was produced in 66% yield after flash
chromatography. Likewise, reduction of the xanthate derivative
of 2-octanol 2c by a similar preparative procedure provided the
same product 6a in 70% isolated yield (see Supporting Informa-
tion). Evidently, 6a is the boron product formed in the reductions
of secondary xanthates by 1a.
(8) (a) Barton, D. H. R.; Jacob, M. Tetrahedron Lett. 1998, 39, 1331–
1334. (b) Lucarini, M.; Pedulli, G. F.; Valgimigli, L. J. Org. Chem.
1996, 61, 1161–1164. (c) Baban, J. A.; Roberts, B. P. J. Chem. Soc.,
Perkin Trans. 2 1988, 1195–1200.
(9) Rablen, P. R. J. Am. Chem. Soc. 1997, 119, 8350–8360.
(10) Roberts, B. P. Chem. Soc. ReV. 1999, 28, 25–35.
(11) (a) Ueng, S.-H.; Makhlouf Brahmi, M.; Derat, E.; Fensterbank, L.;
Lacoˆte, E.; Malacria, M.; Curran, D. P. J. Am. Chem. Soc. 2008, 130,
10082–10083. (b) Walton, J. C. Angew. Chem., Int. Ed. 2009, 48,
1726–1728.
(12) (a) Crich, D.; Quintero, L. Chem. ReV. 1989, 89, 1413–1432. (b)
Chatgilialoglu, C.; Ferreri, C. Res. Chem. Intermed. 1993, 19, 755–
775. (c) Zard, S. Z. In Radicals in Organic Synthesis; Renaud, P.,
Sibi, M. P., Eds.; Wiley-VCH: Weinheim, 2001; Vol. 1, pp 90-108.
(13) Barton and Jacob described the isolation of the analogous product
from a phosphine-borane reduction, Bu3P-BH2SC(O)SMe. See
footnote 14 in ref 8a.
(14) Jin, J.; Newcomb, M. J. Org. Chem. 2007, 72, 5098–5103.
(15) Newcomb, M. In Radicals in Organic Synthesis; Renard, P., Sibi, M.,
Eds.; Wiley-VCH: Weinheim, 2001; Vol 1, pp 317-336.
(16) GC analysis of this experiment showed a 76% yield of 18 along with
5% of 20.
9
J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009 11257

A R T I C L E S
Ueng et al.
Scheme 2. Isolation of Boron-Derived Product 6a
Figure 2. Structure of 6b and 128.4 MHz 11B NMR spectrum showing
evidence for its formation in the reaction of 1b with 2b after 40 min at 70
°C.
Scheme 3. Isolation of a Second Type of Boron-Derived Product
11
(Methylthiocarbonylthio)borane 6a was characterized by the
usual battery of spectroscopic methods (1H and 13C NMR, IR,
MS), and full data are contained in the Supporting Information.
The 11B NMR spectrum of 6a at 298 K exhibited a broad singlet
at -25.6 ppm in CDCl3 and at -25.2 ppm in C6D6 and
C6D5CD3. Upon heating a sample in toluene-d8 to 80 °C,17 the
resonance sharpened to a triplet, JB-H ) 105 Hz. Contrast these
data to the 11B NMR spectrum of the precursor 1a, which
exhibits a quartet at -34.5 ppm, JB-H ) 90 Hz (C6D6).
In addition to surviving flash chromatography unscathed,
(methylthiocarbonylthio)borane 6a was recovered unchanged
after heating at 150 °C in mesitylene-d12 for 17 h. The compound
was also stable to heating with AIBN but slowly degraded on
exposure to Et3B and air. For example, the yield of 6a in the
reaction with the xanthate derivative of undecylcyclobutyl-
carbinol decreased to 28% when the mixture was exposed to
air for 3 days prior to isolation. This degradation may explain
our early difficulties in isolating the boron-derived products from
reductions with 1a.
Attempts to characterize the boron-derived product from the
reduction of a secondary xanthate with 1b were only partially
successful. In a reaction between 1b and 2b conducted in an
NMR tube (C6D6/CD3CN, 9/1, 70 °C), we observed a new triplet
in the 11B NMR spectrum whose chemical shift was in the region
expected for 6b (-24.2 ppm). However, after a brief increase
in intensity over 40 min, this peak began to fade. Figure 2 shows
the appearance of the spectrum at the 40 min point, while the
complete time course is in the Supporting Information. After
90 min, the intensity of the new peak had decreased consider-
ably, and several other overlapping resonances began to
appear.18
the following rate constant and experiments on the radical 3a
derived from 1a.
A second class of boron product was obtained from the
reduction of primary xanthate 8 as shown in Scheme 3. Standard
reaction of 1a (2 equiv) and 8 (1 equiv) initiated by Et3B and
air provided the reduced product 10 (55%), the starting borane
1a (59%), the (methylthiocarbonylthio)borane 6a formed to-
gether with 10 (19%), and a new oxycarbonothioylthioborane
11 (9%).19
The oxycarbonothioylthioborane 11, hereafter called a borane
xanthate, was also a stable white solid, mp 117-119 °C,
exhibiting a broad resonance in the 11B NMR spectrum at -24.4
ppm in CDCl3. Other spectral data of 11 are summarized in the
Supporting Information.
Borane 6a, a dithiocarbonate with BsS and CsS single
bonds and a CdO double bond, is the expected product of a
Barton/McCombie reaction resulting from cleavage of the CsO
single bond of the precursor. In contrast, 11 is an O-alkyl
xanthate with BsS and CsO single bonds and a CdS double
bond that results from cleavage of the CsS single bond of the
Apparently, the boron-derived product from 1b is much less
stable than that derived from 1a. Because of that, we focused
(17) Beall, H.; Bushweller, C. H. Chem. ReV. 1973, 73, 465–486.
(18) An attempt to isolate CsO cleavage product 6b from this mixture by
chromatography did not succeed. However, we did isolate a small
amount (∼5%) of an impure sample whose major component was
tentatively assigned by NMR analysis to be the C-S cleavage product,
NHCsBH2SC(dS)OCH(iPr)(CH2)5OBn, where NHC is 1,2,4-triazol-
3-ylidene.
(19) The yield of recovered 1a is based on the starting amount of 1a, while
all other yields are based on the starting amount of 8.
9
11258 J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009

NHC-Boryl Radicals
A R T I C L E S
Figure 5. Xanthate 2a and PTOC 15 precursors of substituted cyclobu-
tylcarbinyl radical probe 16.
a methyl (or ethyl) radical and boron xanthate 11. Pathway “b”
is favored relative to “a” because a weaker single bond is
cleaved (C-S versus C-O), but it is disfavored because it
provides both a less stable radical (Me• versus 1°-alkyl 14) and
a less stable molecule (6a has a CdO bond while 11 has a CdS
bond). The approximately equal yields of 6a/10 and 11 in the
AIBN-initiated reaction of S-methyl xanthate 8 suggest that
radical 13 partitions to 6a and 11 with a ratio of ∼1/1 at 80 °C.
The C-S path “b” becomes favored when radicals of compa-
rable stability are produced, as shown by the predominance of
11 over 12 in the reduction of S-ethyl analogue 9.
Figure 3. X-ray crystal structures of 6a and 11.
The C-S cleavage pathway “b” has precedent in Barton-
McCombie reactions20 and is also reminiscent of group transfer
reactions of carbon-substituted xanthates.21 But in such cases
the radical resulting from fragmentation of the C-S bond is
usually more stable than that resulting from fragmentation of
the C-O bond.
Rate Constant Measurements. Having isolated the boron-
derived product from cyclobutylcarbinyl xanthate 2a, we initially
tried using the same substrate for rate constant measurements
on radical 16 (Figure 5). Reductions progressed well at high
concentrations (0.2-0.4 M) and temperatures (rt and above)
typically used for preparative experiments.11a However, as the
concentrations and temperatures were reduced to generate
standard kinetic plots, the chains did not propagate well and
the starting xanthate was not consumed efficiently. Although
using xanthate 2a for a full kinetic study was not successful,
we did still learn from preliminary analyses of reduced products
that radical 16 was a suitable clock.
Accordingly, we turned to an alternative precursor to radical
16, the PTOC ester 15.22a PTOC esters are photolabile when
irradiated with visible light, and they react to completion even
when chain propagation steps are inefficient. In principle, they
are better precursors for an indirect kinetic study, but the use
of a PTOC ester radical precursor in a kinetic study introduces
another complication in that PTOC esters can react with carbon-
centered radicals to give alkyl pyridyl sulfide products.22b
In the case of an indirect kinetic study using radical 16 derived
from 15, the reactions shown in Scheme 4 are possible. Radical
16 ring opens to acyclic radical 17 with a known rate constant
kr (path 1),22a is trapped by the NHC-borane 1a to give
cyclobutane 18 (path 2), or reacts in a “self-trapping” reaction
with precursor 15 to give the cyclobutylcarbinyl pyridyl sulfide
19 (path 3). The acyclic radical 17 formed by ring opening of
16 is trapped by the NHC-borane 1a to give 20 (path 4) or
reacts with the PTOC precursor 15 to give sulfide 21 (path 5).
Small amounts of cisoid ring opened products (cis-isomers of
20 and 21) are expected, but the amount of cis-4-hexadecene
Figure 4. C-O (a) and C-S (b) fragmentation pathways provide the two
boron products 6a and 11.
precursor. This latter class of product has also been observed
in tin hydride xanthate cleavages.20
Knowing that the yields of boron-derived products might be
decreased by prolonged exposure to Et3B and air, we conducted
a similar experiment at 80 °C between 8 and 1a with initiation
by AIBN. After 2 h, the mixture was cooled and the products
were separated by flash chromatography to provide boron
products 6a (33%) and 11 (33%), reduced product 10 (36%),
and starting materials, xanthate 8 (12%) and NHC-borane 1a
(54%).
To further favor the formation of the C-S cleavage product
over the C-O cleavage one, we conducted a reduction of S-ethyl
xanthate derivative 9 initiated by AIBN as above. This produced
11 in 41% isolated yield. In contrast, only traces of 12 were
formed (<5%).
The structures of both 6a and 11 were confirmed by X-ray
crystallography, and diagrams of the two structures oriented to
feature the sulfur substituent on boron are shown in Figure 3.
(Other diagrams that better show the N-heterocyclic carbene
substituent are presented in the Supporting Information.)
Interestingly, the BH2S-C(O) bond of 6a is syn, while the
corresponding BH2S-C(S) bond of 11 is anti.
The different boron products arise from two different modes
of fragmentation of adduct radical 13, as shown in Figure 4.20
Path “a” is a C-O cleavage that provides radical 14 (precursor
to 10) and 6a, while path “b” is a C-S cleavage that provides
(21) Quiclet-Sire, B.; Zard, S. Z. Top. Curr. Chem. 2006, 264, 200–236.
(22) (a) Jin, J.; Newcomb, M. J. Org. Chem. 2008, 73, 4740–4742. (b)
Newcomb, M.; Kaplan, J. Tetrahedron Lett. 1987, 28, 1615–1618.
(23) Newcomb, M. Tetrahedron 1993, 49, 1151–1176.
(20) Barton, D. H. R.; Crich, D.; Löbberding, A.; Zard, S. Z. Tetrahedron
1986, 42, 2329–2338.
9
J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009 11259

A R T I C L E S
Ueng et al.
Scheme 4. Evolution of Radical Probe 16 to Directly Trapped (18,
19) and Ring Opened (20, 21) Products
Figure 6. Relative Arrhenius plot for rearrangement of radical 16 and its
trapping by NHC-borane 1a.
conducted at 28 °C with 0.01 or 0.02 M 1a, the rate constant
for trapping was kH ) (3.4 ( 1.0) × 104 M-1 s-1 where the
error is at 1σ.
The temperature-dependent function for the H-atom transfer
reaction of donor 1a was obtained from a series of reactions
conducted over the temperature range -75 to 50 °C. From 10
reactions run with 0.01 M 1a, we obtained the relative Arrhenius
function in eq 2, which is shown graphically in Figure 6. This
relative Arrhenius function was subtracted from the Arrhenius
function for ring opening of radical 16, given in eq 3,22a to give
the Arrhenius function for radical trapping by NHC-borane 1a
in eq 4. In these equations, all errors are listed at 1σ, and the
activation energies are in kcal/mol. Solving the Arrhenius
function in eq 4 for 28 °C gives kH ) 3.8 × 104 M-1 s-1, in
excellent agreement with the average value determined from
studies at this temperature.
found in the reaction of 16 with tin hydride was previously
shown to be inconsequential.22a
The rate constant kH for reaction of the NHC-borane with
radical 16 can be determined from the concentration of the
H-atom donor, the product mixture ratio, and the rate constant
for ring opening of radical 16 (kr) via eq 1, where [X-H]aver is
the average concentration of the trapping reagent over the course
of the reaction.23 Because reactions were conducted with excess
trapping agent, using the average concentration of X-H in a
pseudo-first-order kinetic expression introduces an insignificant
error in comparison to using a second-order kinetic expression.23
The reaction of the PTOC ester 15 with radical 16 reduces
the amount of probe radical from the reaction mixture, but the
yield of product 19 is not of consequence for the kinetic
determination. The amount of pyridyl sulfide 21 obtained is
important for the kinetic determination, however, because it is
produced from the ring-opened radical 17. Rate constants for
ring opening of 16 (kr) were previously determined at various
temperatures, and an Arrhenius function for the ring-opening
reaction is available.22a
log(kr/kH) ) (3.61 ( 0.29) - (6.62 ( 0.34)/2.3RT (2)
log(kr) ) (13.2 ( 0.4) - (13.5 ( 0.6)/2.3RT
log(kH) ) (9.6 ( 0.5) - (6.9 ( 0.7)/2.3RT
(3)
(4)
From the Arrhenius function in eq 4, the H-atom transfer
reaction of NHC-borane 1a displays a “normal” log A term for
a second-order H-atom transfer. That is, the entropic demand
for the reaction is typical for a simple bimolecular reaction and
similar to the entropic demand in, for example, H-atom transfer
from a group 14 metal hydride.2 In contrast, H-atom transfer
from methanol in the Lewis acid complex Et3B•CH3OH had an
unusually small log A term (high entropy demand) indicating a
highly organized transition state or an unfavorable equilibrium
prior to the rate-determining step.14,22a Thus, NHC-borane 1a
reacts with the characteristics of a covalent molecule that donates
the H-atom in a single elementary process, which is consistent
with the postulated mechanism of reaction.
Because the H-atom transfer from NHC-borane 1a competes
with “self-trapping” by the PTOC ester for both the cyclic
radical 16 and the acyclic radical 17, it is possible in principle
to use those competitions to estimate the rate constant for the
reaction of 1a. From the results at 28 °C listed in the Supporting
Information, this approach leads to a conclusion that the rate
constant for H-atom transfer from 1a is 2-3 times greater than
that found from the clock reaction approach. This difference
might result because the estimated rate constant for the PTOC
self-trapping reaction is too large.22b But irrespective of the
origin of the difference, the latter result is not consistent with
the original attempted kinetic studies. If the rate constant for
reaction of 1a was on the order of 105 M-1 s-1, then the
kH ) (kr/[X-H]aver) × ([18]/([20] + [21]))
(1)
Kinetic studies were conducted in deoxygenated benzene or
toluene by adding NHC-borane 1a to a temperature-equilibrated
solution containing PTOC ester 15 and irradiating the stirred
solution with a tungsten-filament bulb. The reaction mixtures
were checked periodically by TLC, and samples were analyzed
when PTOC ester 15 was no longer detected. Yields of products
18-21 were determined by GC, and the total yields of these
products for reactions with 0.01 to 0.02 M 1a were in the range
65-98%. Because the amounts of alkyl pyridyl sulfides 19 and
21 formed were significant, we isolated these products from
bulk reactions and identified them by 1H and 13C NMR
spectroscopy and high resolution mass spectrometry as described
in the Supporting Information.
Detailed kinetic results are in the Supporting Information.
From the ratios of products as determined by GC, the rate
constant for trapping by NHC-borane 1a was determined via
eq 1. We conducted multiple reactions at 28 °C and a set of
variable temperature studies to generate a relative Arrhenius
function (see below). For a set of seven trapping reactions
9
11260 J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009

NHC-Boryl Radicals
A R T I C L E S
Table 1. Comparison of NHC-Boryl EPR Parameters with Those
of Model Radicals
T/K or DFT
basis set
a(11B)
G
a(2N)
G
a(2HR)
G
a(other)
G
Radical
g-factor
3a
300
2.0028
7.3 4.0
5.3 3.1
11.4
1.0 (2H)
model NHCa EPR-iiib
-10.0 -1.2 (2H)
9.6 1.4 (9H)
16.8 43.6 (31P)
c
Me3N-BH2
Et3P-BH2
280
183
2.0022 51.3 1.4 (1N)
2.0020 17.6
c
a N,N′-Diphenylimidazol-2-ylidene-BH2•.b B3LYP/EPR-iii//B3LYP/
6-31G(d). c See ref 26.
Figure 7. Solution EPR spectrum of NHC-BH2• radical 3a. Top: First
derivative experimental spectrum at 300 K in t-BuPh. Bottom: Computer
simulation with parameters noted in Table 1.
Figure 8. Sketches of the SOMOs of an amine-boryl radical (left) and
radical 3a.
5-hexenyl radical should not have cyclized completely in the
presence of 1a.11a
values for a model NHC-BH2• radical (Table 1) provide further
support.25 Thus, the photochemically generated t-BuO• radicals
efficiently abstract H-atoms from 1a to generate 3a. The EPR
spectroscopic observation of 3a provides strong support for the
mechanism of Scheme 1.
In an effort to benchmark the rate constant measurement for
1a, we also briefly investigated reactions of alkyl radicals with
two amine boranes (Me3N•BH3 and pyridine•BH3) and phos-
phine boranes (n-Bu3P•BH3 and Ph3P•BH3). These reagents
proved to be poor hydrogen donors and did not provide enough
of the directly reduced products to allow determination of a
rate constant. With kH < 104 M-1 s-1, they are clearly less
reactive than 1a toward alkyl radicals.
Amine-boryl radicals such as Me3N-BH2• and phosphine-
boryl radicals have been extensively studied by Roberts and
others.7,8,10 EPR spectra show them to be σ-type radicals,
pyramidal at boron.26 The much smaller a(11B) and larger a(N)
hfs of 3a show that this radical is planar at boron and that the
unpaired electron is delocalized into the NHC ring. In essence,
the NHC-boryl radical is not a σ-type radical but is instead a
π-type radical quite like benzyl.27 The sketches of the SOMOs
of amine-boryl radical Me3N-BH2• and the NHC-boryl radical
3a are compared in Figure 8 to illustrate this difference.
The DFT computation for the model N,N′-diphenylimidazol-
2-ylidene-BH2• radical was in full agreement with this conclu-
sion. Formation of such a resonance-stabilized radical is likely
an important factor in lowering the BDE of the B-H bond and
the transition state energy of H-atom transfer reactions of NHC-
boranes. This agrees with Rablen’s original conclusion,9 based
on a theoretical study of Lewis base complexed boranes, that
the extent of lowering of B-H BDEs is greater when the spin
density is delocalized away from boron and onto the associated
Lewis base. Further, because of the substantial structural
differences, the chemistry of NHC-boryl radicals may differ
from the known chemistry of amine- and phosphine-boryl
radicals, and new reactivity patterns are foreseen.
The availability of the rate constant kH allows a preliminary
estimate of the bond dissociation energy (BDE) by applying an
Evans-Polanyi relationship;24 the estimated boron-hydrogen
BDE of 1a is ∼88 kcal/mol. Though this estimate could be in
error by several kcal/mol, it is considerably higher than the
calculated BDE for 1a of 80 kcal/mol.11a Nonetheless, the BDE
of 1a is still very much lower than experimental or high-level
computed values of B-H BDEs for borane (BH3, ∼105 kcal/
mol) and its ammonia and water complexes (102-104
kcal/mol).5,9 The estimated BDE of 1a is also below the
computed BDE of phosphine borane (92-93 kcal/mol),9 and
indeed the competition experiments clearly show that 1a is a
better H-donor than Bu3P•BH3 toward a secondary alkyl radical.
EPR Experiments. Having provided strong circumstantial
evidence for the existence of NHC-boryl radicals by product
studies and rate constant measurements, we next examined the
possibility of directly detecting 3a in solution by EPR spec-
troscopy. We selected di-tert-butyl peroxide (DTBP) to generate
the NHC-boryl radical because of the convenience of generating
t-BuO• by UV irradiation. The t-BuO• radical is EPR “silent”,
so we expected that clean spectra would be obtained provided
that t-BuO• promptly abstracted a hydrogen atom from NHC-
borane 1a.
Conclusions
In summary, we have isolated and characterized both the
boron dithiocarbonate product resulting from standard C-O
cleavage of primary and secondary xanthates and a boron
xanthate product resulting from C-S cleavage during the
reduction of primary xanthates. These compounds are the first
NHC-boranes with boron-sulfur bonds. The first rate constants
When a deaerated solution of 1a and DTBP in tert-
butylbenzene was UV irradiated in the resonant cavity of a 9
GHz EPR spectrometer, the spectrum shown in Figure 7 was
obtained. The spectrum was very well simulated with the
hyperfine splittings shown in Table 1. The simulated spectrum
is also shown in Figure 7.
(25) For the B3LYP/EPR-iii//B3LYP/6-31G(d) basis set, see: Barone, V.
In Recent AdVances in Density Functional Theory; Chong, D. P., Ed.;
World Scientific Publishing Co.: Singapore, 1996.
The hyperfine splittings (hfs) are entirely appropriate for
radical 3a, and comparison of these with the DFT computed
(26) (a) Paul, V.; Roberts, B. P.; Robinson, C. A. S. J. Chem. Res. (S)
1988, 264–265. (b) Kirwan, J. N.; Roberts, B. P. J. Chem. Soc., Perkin
Trans. 2 1989, 539–550.
(24) Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin Trans. 2 1994, 2155–
(27) Tumanskii, B.; Sheberla, D.; Molev, G.; Apeloig, Y. Angew. Chem.
Int. Ed. 2007, 46, 7408–7411.
2162.
9
J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009 11261

A R T I C L E S
Ueng et al.
Hz, 4H), 2.07 (s, 3H), 1.34 (d, J ) 6.9 Hz, 12H), 1.15 (d, J ) 6.9
Hz, 12H); 13C NMR (75 MHz, CDCl3): δ 191.9, 145.4, 133.3,
130.4, 124.0, 123.0, 28.9, 25.5, 22.6, 13.6; 11B NMR (96.3 MHz,
CDCl3): δ -25.6 (br s); 11B NMR (96.3 MHz, C7D8): δ -25.2 (br
s, w1/2 ) 288 Hz); 11B NMR (96.3 MHz, C7D8, 353 K): δ -25.1
(t, J ) 105 Hz); LRMS (EI) m/z: 508 ([M]+, 0.4), 493 ([M - CH3]+,
2.4), 447 (11), 433 (97), 399 (100), 389 (20), 357 (26), 186 (15),
75 (35); HRMS (EI) calcd. for C28H3811BN2OS2 ([M - CH3]+):
493.2519, found 493.2517.
for H-atom transfer reactions of the NHC-borane 1a with a
secondary alkyl radical clock have been determined. At ambient
temperature, the reaction has a rate constant of 4 × 104 M-1
s-1, and the Arrhenius function for the reaction demonstrates a
typical entropic term indicating a simple H-atom transfer in the
rate-limiting step. Finally, we observed the NHC-boryl radical
by EPR spectroscopy, and spectral analysis suggests that it is a
π-radical, perhaps more analogous to the benzyl radical than
to apparently similar amine- and phosphine-boryl σ-radicals.
These results support the intermediacy of the NHC-boryl radical
and secure the radical mechanism for the reduction of xanthates
with NHC-boranes.
The NHC-borane 1a reacts nearly 2 orders of magnitude less
rapidly than the prototypical tin hydride reagent Bu3SnH and 1
order of magnitude less rapidly than (TMS)3SiH.2 However,
NHC-borane 1a reacts 2 orders of magnitude faster than the
prototypical silane Et3SiH,2 which cannot be used in conven-
tional carbon radical chain reactions. Accordingly, it occupies
a valuable position on the scale of hydrogen atom donors,
capable of allowing relatively slow radical cyclizations or
additions to occur at high reagent concentrations.
The analysis of BDEs presents a similar picture. At 88 kcal/
mol, NHC-borane 1a has a considerably weaker bond to H than
Et3Si-H (95.1 kcal/mol). Its bond is stronger than (TMS)3SiH
(84 kcal/mol) and Bu3SnH (78.6 kcal/mol) and comparable to
Bu3GeH (88.6 kcal/mol).
Finally, with the existence NHC-boryl radicals firmly estab-
lished, the door is open for their further study. The ready
availability and good stability of NHC-boranes provides a
convenient source of NHC-boryl radical precursors that will
facilitate this study.
[1,3-Bis(2,6-diisopropylphenyl)imidazol-2-ylidene]-[6-(benzyl-
oxy)hexyloxycarbonothioylthio]borane (IPr-IMD-BH2-S-C(dS)-
O(CH2)6OBn) (11). AIBN-initiated free radical reaction: A solution
of 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene borane (1a)
(IPr-IMD-BH3)11a (161.0 mg, 0.4 mmol), O-6-(benzyloxy)hexyl
S-methyl xanthate (8) (59.7 mg, 0.2 mmol), and AIBN (16.4 mg,
0.1 mmol) in deoxygenated benzene (2 mL) was refluxed for 2 h.
After cooling to room temperature, the solvent was removed in
vacuo, and the residue was purified by flash column chromatography
(elution with hexane/ethyl acetate ) 98:2, hexane/ethyl acetate )
9:1, hexane/ethyl acetate ) 8:2) to give 11 (45 mg, 33%) as a white
solid and 6a (34 mg, 33%). A mixture of (hexyloxymethyl)benzene
(10) and starting xanthate (8) (3/1) (21 mg, 48%) and recovered
starting IPr-IMD-BH3 (1a) (87 mg, 54%) were also isolated.
Crystals of 11 suitable for X-ray analysis were grown from CH2Cl2/
hexane, and the complete data are contained in the accompanying
cif file. Mp 117-119 °C; IR (ATR, cm-1): νmax 2963, 2930, 2868,
1
2399, 1468, 1178, 1072, 1023; H NMR (400 MHz, CDCl3): δ
7.46 (t, J ) 7.6 Hz, 2H), 7.34 (d, J ) 7.6 Hz, 4H), 7.33-7.26 (m,
5H), 7.08 (s, 2H), 4.49 (s, 2H), 4.19 (t, J ) 6.7 Hz, 2H), 3.43 (t,
J ) 6.7 Hz, 2H), 2.56 (septet, J ) 6.8 Hz, 4H), 1.65-1.50 (m,
4H), 1.34 (d, J ) 6.8 Hz, 12H), 1.33-1.19 (m, 4H), 1.14 (d, J )
6.8 Hz, 12H); 13C NMR (100 MHz, CDCl3): δ 221.4, 145.4, 138.7,
133.3, 130.4, 128.3, 127.6, 127.4, 124.0, 123.0, 72.8, 72.6, 70.4,
29.6, 28.9, 28.1, 25.8, 25.7, 25.5, 22.7; 11B NMR (128.4 MHz,
CDCl3): δ -24.4 (br s); HRMS (EI) calcd. for C41H5811BN2O2S2
([M + H]+): 685.4033, found: 685.4034.
Experimental Section
[1,3-Bis(2,6-diisopropylphenyl)imidazol-2-ylidene]-(methylthio-
carbonylthio)borane (IPr-IMD-BH2-S-C(dO)-SCH3) (6a), from
Reaction of Undecylcyclobutylcarbinyl Xanthate 2a. Triethylborane
(1 M solution in hexane, 0.15 mL, 0.15 mmol) was added to the
solution of 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene borane
(1a) (IPr-IMD-BH3)11a (58.4 mg, 0.15 mmol) and undecylcyclobu-
tylcarbinyl xanthate 2a (48.0 mg, 0.15 mmol)14 in benzene (0.85
mL). The solution was exposed to air by piercing the septum stopper
with a needle. The mixture was stirred at room temperature for
20 h. The reaction mixture was loaded onto a silica gel column (h
) 20 cm, d ) 4 cm) for flash chromatography (elution with CH2Cl2,
CH2Cl2/MeOH ) 8:1) to give compound 6a (58.0 mg, 79%) as a
white solid. A lower yield (28%) was obtained with a longer
reaction time (3 days).
Acknowledgment. This work was supported by grants from
the U.S. National Science Foundation (CHE-0601857 to M.N. and
CHE-0645998 to D.P.C.) and the French Agence Nationale de la
Recherche (ANR, BLAN0309 Radicaux Verts, and 08-CEXC-011-
01 Borane). We thank Dr. Lise-Marie Chamoreau (IPCM, Paris)
for solving a crystal structure. D.P.C. thanks l’e´tat et la re´gion Ile-
de-France for a Chaire Blaise Pascal and ANR for a “chaire
d’excellence”. A.S. thanks the University of Pittsburgh for a
Graduate Excellence Fellowship. We dedicate this paper to Profes-
sor Keith U. Ingold on the occasion of his 80th birthday.
Supporting Information Available: Procedures and charac-
terization of all new compounds not in the experimental section,
details of kinetic experiments, copies of spectra of products,
and cif files of the crystal structures. This material is available
free of charge via the Internet at http://pubs.acs.org.
Crystals of 6a for X-ray analysis were grown from methanol at
20 °C, and the complete data are contained in the accompanying
cif file. Mp 201-202 °C; IR (thin film, cm-1): νmax 3165, 3132,
2963, 2927, 2870, 2458 (B-H), 2390 (B-H), 1631 (CdO), 1471,
1
1036, 855; H NMR (300 MHz, CDCl3): δ 7.49 (t, J ) 7.8 Hz,
2H), 7.29 (d, J ) 7.8 Hz, 4H), 7.08 (s, 2H), 2.55 (septet, J ) 6.9
JA904103X
9
11262 J. AM. CHEM. SOC. VOL. 131, NO. 31, 2009