Facile B-H bond activation of borane by stable carbenoid species

Article abstract of DOI:10.1021/ja401763c

Stable nucleophilic carbene compounds have recently been shown to be able to mimic in some instances the reactivity of metal fragments in the reaction of unactivated E-H bonds (E = H, R₃Si, NH₂, R₂P). However, the insertio

Full text of DOI:10.1021/ja401763c

Communication
pubs.acs.org/JACS
Facile B−H Bond Activation of Borane by Stable Carbenoid Species
Hadrien Heuclin,† Samuel Y.-F. Ho,†,‡ Xavier F. Le Goff,† Cheuk-Wai So,*,‡ and Nicolas Mezailles*,†,§
́
†Laboratoire Het
́ ́ ́ ́
eroelements et Coordination, Ecole Polytechnique, CNRS, Route de Saclay, 91120 Palaiseau, France
‡Division of Chemistry and Biological Chemistry, School of Physical and Mathematical Sciences, Nanyang Technological University,
Singapore 637371
S
* Supporting Information
Scheme 1. (a, b) Activation of NH3 and HBPin by CAACs;
ABSTRACT: Stable nucleophilic carbene compounds
(c) Orbitals Involved in Carbene−BH3 Activation
have recently been shown to be able to mimic in some
instances the reactivity of metal fragments in the reaction
of unactivated E−H bonds (E = H, R3Si, NH2, R2P).
However, the insertion into a B−H bond of the strongly
Lewis acidic BH3 molecule has never been observed at a
single C atom or even at a metal fragment. Our results
show that designed stable, highly electrophilic carbenoid
fragments in compounds 4 and 6 can achieve this reactivity
in a controlled manner. Density functional theory
calculations corroborated the experimental results on the
presently designed systems as well as the lack of reactivity
on nucleophilic carbenes.
ctivation of small molecules and unactivated bonds is one
of the greatest challenges of 21st century chemistry. For
A
many years, it has been thought that only transition metal
(TM) fragments are capable of activating such bonds because
of an appropriate double electron transfer involving donation
from the electron pair of the bond to the metal center and
back-donation from the metal into the corresponding
antibonding orbital. However, in 2005, Power and co-workers
proved that activation of H2 under mild conditions with heavier
group-14 metals (Ge, Sn) is feasible.1 Then, in 2006, the
Stephan group demonstrated that a system featuring a P/B
frustrated Lewis pair (FLP) could cleave dihydrogen heterolyti-
cally,2 opening ways for further transformations.3 Shortly
thereafter, the Bertrand group showed that stable singlet
carbenes can also act as mimics of TMs4 to activate a number of
small reactive molecules such as CO5 and white phosphorus.6
The more remarkable splittings of the H−H bond of H2, a N−
H bond of NH3 (Scheme 1a),7 and other unactivated E−H
bonds (E = Si, P) have also been reported.8 Furthermore, B−H
bond activation of the hydridic substrate pinacolborane
(HBPin) by stable singlet carbenes (Scheme 1b)8 and an
FLP9has been observed. This 1,1-addition at a single carbon
center is reminiscent of the oxidative addition of R2BH on TM
fragments, which has been extensively studied because it is a
key step in TM-catalyzed 1,2-addition over a π system. In
contrast, the 1,1-addition of a B−H bond of the much stronger
Lewis acid BH3 at a single metallic10 or organic center is not
known, although the 1,2-addition to a π system is one of the
most famous reactions in organic chemistry. The 1,1-activation
of BH3 at a C center is most difficult for several reasons. The
formation of two novel bonds, C−BH2 and C−H, from BH3
and a designed “C” fragment requires the efficient double
electron transfer mentioned above. However, in the case of
BH3, the preferred orbital interaction involves donation of the
lone pair at C to the empty orbital at B (Scheme 1c, left).
Comparatively, the donation to the BH σ* orbital is less
efficient (Scheme 1c, right). Moreover, the BH σ orbital is a
weak donor, and the orbital overlap with the empty p orbital at
C is poor. Overall, the formal oxidative addition at a single “C”
center is thus typically an endergonic process (ΔG° > 0), and
even the most reactive systems presented above failed to split
the BH bond of BH3. Instead, the Lewis acid−base adduct was
obtained with cyclic alkyl amino carbenes (CAACs), and ring
opening was observed with N-heterocyclic carbenes
(NHCs).8,11
Strategies for B−H bond activations have thus relied on
suppressing the strong Lewis acidity of BH3 in a preliminary
step via the formation of strong Lewis acid−base adducts such
as phosphine−borane, amine−borane, or NHC−borane
adducts.12 In particular, it has been shown that only the
transiently generated dichlorocarbene reacted with L→BH3
adducts (L = tertiary amine or phosphine).13
To favor the double electron transfer, we reasoned that the
nucleophilic character of a carbene-like species should be
decreased whereas the electrophilic character should be
Received: February 18, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja401763c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
increased compared with the stable carbenes such as NHCs and
CAACs. Our group has been working for some years on the use
of geminal dianions such as 3 (Scheme 2) as carbene precursors
formation of a single new species, 6, was confirmed by 31P
NMR spectroscopy (a doublet at δP = 45.5 ppm and a broad
signal centered at δP = 30 ppm). Although 6 was stable at −35
°C for over a month, it reacted further at room temperature
(vide supra). Yellow single crystals were thus grown from Et2O
at −35 °C and analyzed by X-ray diffraction (XRD). This
analysis confirmed the formation of carbenoid 6 (Figure 1).
Scheme 2. Activation of BH3 by Carbenoid Species
Figure 1. Crystal structure of 6 with 50% thermal ellipsoids. H atoms
on the phenyl rings and on the solvent molecules have been omitted
for clarity. Selected bond lengths (Å) and angles (deg): C1−P1,
1.722(3); C1−P2, 1.728(3); C1−Cl1, 1.775(3); P1−B1b, 1.89(3);
P2−S1b, 2.0526(8); S1b−Li1, 2.42(1); Li1−O1, 1.999(5); Li1−O2,
1.972(6); P1−C1−P1, 126.7(2); P2−C1−Cl1, 118.5(2); P1−C1−
Cl1, 114.5(2).
for TMs as well as lanthanides and actinides.14 We have shown
experimentally and corroborated by density functional theory
(DFT) calculations that the very highly nucleophilic character
brought by the two lone pairs at C is decreased by very efficient
electron transfer to the strongly accepting phosphorus
substituents. It was thus hypothesized that oxidation of such
a dianion would result in the formation of a carbene fragment
with reduced nucleophilic character but also with significant
electrophilic character because of a lack of significant
stabilization of the empty p orbital. In 2007, the oxidation of
dianion 3 with hexachloroethane was studied, and the first
example of a carbenoid species that is stable at room
temperature (compound 4 in Scheme 2) was obtained.15 In
line with our hypothesis, 4 was shown to be a competent
precursor of a Pd carbene complex upon reaction with an
electron-rich Pd(0) precursor. This first carbenoid was used in
the present work to probe the intermolecular reactivity with
BH3. A second carbenoid/carbene species, 6, derived from the
dianionic derivative 5 (Scheme 2)16 was envisaged to probe
thermodynamically more favorable intramolecular reactivity. In
this contribution, we present a novel example of very rare
isolable, stable carbenoid species as well as intramolecular and
intermolecular 1,1-additions of a B−H bond of BH3 at a C
center. DFT calculations providing mechanisms of these
unprecedented activations by the aforementioned carbenoid
species as well as comparisons with the stable CAAC and NHC
systems are also presented.
Low-temperature 31P NMR spectra confirmed the assignment.
The most important feature of 6 is the separation of Li+ and
Cl−, which explains the stability of 6 at such a high temperature
compared with other carbenoid species.17 It resembles
compound 4 in this sense. The Cl atom is bound to the C
center, whereas the Li cation is bound to the S atom and also to
the BH3 moiety via interactions with two B−H bonds. The
lower thermal stability of 6 (stable up to −20 °C) compared
with 4 (stable at room temperature) is obviously due to the
weaker coordination of Li+ by the BH3 moiety in 6 compared
with the S atom in 4. Following the evolution of 6 in solution
revealed the expected reactivity. Two new products (7a and
7b) were identified by NMR spectroscopy. Both compounds
were characterized by two sets of well-resolved multiplets at δP
= 46.1 and 45.9 ppm (3:2 ratio) as well as a broad signal at δP =
19 ppm. Also, both compounds featured one H atom on the
P−C−P bridge, as evidenced by the 1H NMR spectrum, which
contained two highly coupled signals at δH = 3.37 and 2.96 ppm
that simplified upon 31P decoupling and appeared as the
expected multiplets because of coupling with one B center.
Moreover, the 13C NMR spectrum showed a C−B coupling
pattern for the corresponding carbon in the two compounds.
The rest of the NMR data suggested that these two products
were structurally very similar. These compounds were crystal-
lized from the crude mixture, and XRD analysis on each
compound proved the formation of diastereomeric pairs. The
structures of compounds 7a and 7b are presented in the
Supporting Information (SI). Both 7a and 7b are six-membered
P2C2B2 rings; 7a adopts a boat conformation, whereas 7b
adopts a chair conformation. More interesting is the
In a first approach, the more favorable intramolecular
reactivity was probed. The oxidation of dianion 5 was studied
in diethyl ether (Scheme 2), as this highly reactive species
decomposes in most organic solvents, including tetrahydrofur-
an (THF) (quantitative deprotonation). The total conversion
of 5 was observed after only 15 min at room temperature. The
B
dx.doi.org/10.1021/ja401763c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
mechanism of their formation (Scheme 2). They most likely
result from a common reaction intermediate, 7′, which is
formed by intramolecular insertion of the carbenoid center into
a B−H bond of the BH3 moiety. The dimerization of 7′ is
thermodynamically favored by the formation of two strong P→
B bonds.
Motivated by this result, we then focused our attention on
the more challenging intermolecular reaction involving the
room-temperature-stable carbenoid 4. Most satisfyingly,
addition of 2 equiv of BH3·SMe2 to a solution of 4 in Et2O
resulted in the formation of compound 8 after one night at
room temperature (Scheme 2). Compound 8 was characterized
by a singlet at δP = 52 ppm in the 31P{1H} NMR spectrum. A
resonance corresponding to the H atom on the P−C−P bridge
was seen at δH = 3.62 ppm. This resonance exhibited a
characteristic H−B coupling pattern upon 31P decoupling.
Crystallization of 8 resulted in the formation of a dimer, 9. As
for compounds 7a and 7b, dimerization of 8 is favorable
because of the formation of two S→B bonds. It is the only
example to date of a P2S2C2B2 eight-membered ring.
Figure 2. Free energy profile for the reaction between carbenoid 4
(modeled as A) and BH3.
Mechanistic investigations were carried out by means of
deuterium labeling experiments. Most importantly, the reaction
of 4 with BH3·SMe2 in C6D6 did not lead to any deuterated
products, such as “C(D)(BH2)” species. Moreover, the reaction
with commercial BD3·THF (containing ca. 2% BD2H) in Et2O
led to the formation of the corresponding “C(D)(BD2)” species
8-d3 (singlet at δP = 49 ppm in the 31P{1H} NMR spectrum) as
well as a very minor species (singlet at δP = 50 ppm in the 31P
NMR spectrum) corresponding to the “C(D)(BDH)”/“C(H)-
(BD2)” species in the expected 98/2 ratio. These experiments
support the B−H insertion at C rather than a radical process
involving abstraction of H from the solvent.
DFT calculations were performed to provide further insights
into the mechanisms of the transformations and the require-
ments for B−H activation of BH3 by metal-free organic
fragments. Optimizations were conducted on model com-
pounds, in which the Et2O solvent was modeled by Me2O. Full
details on the calculations are given in the SI. The two
mechanisms bear some resemblance, especially from the point
where BH3 is coordinated to the central C atom. Thus, only the
“intermolecular” mechanism involving compound A is
presented here. The “intramolecular” process is presented in
the SI.
A low-energy path was found, in accord with the
experimental results. The computed energies for the mecha-
nism are shown in Figure 2. The transformation of A into E is
highly exergonic (ΔG = −59.5 kcal/mol). In the first step, the
coordination of BH3 to the carbon center of A is favorable by
only 14.6 kcal/mol, as expected for a rather weak nucleophilic
center. In the second step, from carbenoid B, the Li−Cl
interaction develops because of a facile rotation about the C−P
bond (TSBC was calculated to be 14.8 kcal/mol higher than B),
forming compound C. Most remarkably, the key insertion step
then proceeds with a very low activation barrier of 1.7 kcal/mol
(TSCD) and a large exergonicity (ΔG = −45.7 kcal/mol).
The structure of TSCD (Figure 3) deserves some comments.
First, the C−Cl distance of 2.342 Å is very long compared with
those in A (1.781 Å) and D (1.889 Å), as this bond is broken in
this step. The C−B bond at 1.550 Å is classical for Lewis acid−
base adducts such as NHC−BH3. In TSCD, the B−H bond
featuring the H atom to be transferred is only slightly elongated
(1.271 Å vs 1.200 Å av for the two other BH bonds), as
expected for an early transition state. Finally, the C−B−H angle
Figure 3. View of TSCD with distances in Å. For clarity, only Cipso of
the phenyl groups and pertinent H atoms are shown.
of 78.06° is very acute and points an interaction of the H atom
with the electrophilic C center. Natural bond order (NBO)
charge analysis revealed that the migrating H in TSCD bears a
positive charge (0.16 vs 0.02 and −0.01 for the two other H
atoms of BH3), while it is almost neutral (−0.01) in C and
becomes more positive in D (0.30). Finally, as expected, the
dimerization to form E is slightly favorable because the newly
formed RBH2 species D is a Lewis acid. Overall, TSCD was
found to be 11.2 kcal/mol higher than compound B. As a
comparison, the insertion step in the “intramolecular” process
(see the SI) proceeds with a barrier of 10.7 kcal/mol, which is
also readily overcome at room temperature.
To provide a better understanding of the difference between
the reactivity of these two stable carbenoid species and that of
stable nucleophilic carbenes, calculations on similar BH
insertions with CAAC and NHC compounds were performed
(Figure 4). In the case of the NHC, the Lewis acid−base
adduct is very strong and the electrophilicity of the carbene is
very weak, which makes not only the activation energy
prohibitively high (ΔG⧧ = 44.5 kcal/mol) but also renders
the reaction strongly endergonic (ΔG = 31.8 kcal/mol). The
increased electrophilicity of the CAAC compared to the NHC
decreases the activation barrier to a great extent (ΔG⧧ = 12.8
kcal/mol), but the reaction is nevertheless endergonic (ΔG =
11.1 kcal/mol), even when coordination with a solvent
C
dx.doi.org/10.1021/ja401763c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(3) For a review, see: Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed.
2010, 49, 46.
(4) For a review, see: Martin, D.; Soleilhavoup, M.; Bertrand, G.
Chem. Sci. 2011, 2, 389.
(5) Lavallo, V.; Canac, Y.; Donnadieu, B.; Schoeller, W. W.; Bertrand,
G. Angew. Chem., Int. Ed. 2006, 45, 3488.
(6) (a) Masuda, J. D.; Schoeller, W. W.; Donnadieu, B.; Bertrand, G.
Angew. Chem., Int. Ed. 2007, 46, 7052. (b) Masuda, J. D.; Schoeller, W.
W.; Donnadieu, B.; Bertrand, G. J. Am. Chem. Soc. 2007, 129, 14180.
(c) Back, O.; Kuchenbeiser, G.; Donnadieu, B.; Bertrand, G. Angew.
Chem., Int. Ed. 2009, 48, 5530.
(7) Frey, G. D.; Lavallo, V.; Donnadieu, B.; Schoeller, W. W.;
Bertrand, G. Science 2007, 316, 439.
(8) Frey, G. D.; Masuda, J. D.; Donnadieu, B.; Bertrand, G. Angew.
Figure 4. Free energy profiles for the activations of BH3 by (left)
NHC and (right) CAAC.
Chem., Int. Ed. 2010, 49, 9444.
(9) Dureen, M. A.; Lough, A.; Gilbert, T. M.; Stephan, D. W. Chem.
Commun. 2008, 4303.
(10) Bonanno, J. B.; Henry, T. P.; Wolczanski, P. T.; Pierpont, A. W.;
Cundari, T. R. Inorg. Chem. 2007, 46, 1222.
molecule is taken into account (ΔG = 4.2 kcal/mol), because of
their pronounced nucleophilic character.
In conclusion, we have shown that stable carbenoid species
fill the void in the activation of strong unactivated E−H bonds
[E = H, group 14 element (Si), group 15 element (N, P)],
allowing the activation of the E−H bond for E = group 13
element (B) by non-TM fragments. In these compounds, the
balance between the high electrophilicity and the reduced
nucleophilicity provides both a low-energy path as well as the
appropriate driving force for the 1,1-addition of a B−H bond at
a single carbon center. Finally, the highly electrophilic character
of the carbenoid species 4 and 6 presented here, which are
stable at room temperature and −20 °C, respectively, allows the
development of reactivity complementary to that of ubiquitous
stable carbene compounds.
(11) Kuhn, N.; Henkel, G.; Kratz, T.; Kreutzberg, J.; Boese, R.;
Maulitz, A. H. Chem. Ber. 1993, 126, 2041−2045.
(12) (a) Ueng, S.-H.; Brahmi, M. M.; Derat, E.; Fensterbank, L.;
Laco
10082. (b) Curran, D. P.; Solovyev, A.; Brahmi, M. M.; Fensterbank,
L.; Malacria, M.; Lacote, E. Angew. Chem., Int. Ed. 2011, 50, 10294.
̂
te, E.; Malacria, M.; Curran, D. P. J. Am. Chem. Soc. 2008, 130,
̂
(c) Tian, R.; Mathey, F. Chem.Eur. J. 2012, 18, 11210.
(13) (a) Bedel, C.; Foucaud, A. Tetrahedron Lett. 1993, 34, 311.
́
́
́ ̋ ́ ̋
(b) Keglevich, G.; Ujszaszy, K.; Szollosy, A.; Ludanyi, K.; Toke, L. J.
̈
Organomet. Chem. 1996, 516, 139. (c) Monnier, L.; Delcros, J. G.;
Carboni, B. Tetrahedron 2000, 56, 6039. (d) Staubitz, A.; Robertson,
A. P. M.; Sloan, M. E.; Manners, I. Chem. Rev. 2010, 110, 4023.
(14) (a) Liddle, S. T.; Mills, D. P.; Wooles, A. J. Chem. Soc. Rev. 2011,
40, 2164. (b) Heuclin, H.; Fustier, M.; Auffrant, A.; Mez
Org. Chem. 2011, 7, 596.
(15) Cantat, T.; Jacques, X.; Ricard, L.; Le Goff, X. F.; Mez
Le Floch, P. Angew. Chem., Int. Ed. 2007, 46, 5947.
(16) Heuclin, H.; Fustier-Boutignon, M.; Ho, S. Y.-F.; Le Goff, X. F.;
So, C. W.; Mezailles, N. Organometallics 2013, 32, 498.
́
ailles, N. Lett.
ASSOCIATED CONTENT
* Supporting Information
Experimental details, crystal data, and details of the theoretical
calculations. This material is available free of charge via the
Internet at http://pubs.acs.org.
́
ailles, N.;
■
S
́
(17) (a) Kobrich, G. Angew. Chem., Int. Ed. Engl. 1972, 11, 473.
̈
(b) Niecke, E.; Becker, P.; Nieger, M.; Stalke, D.; Schoeller, W. W.
Angew. Chem., Int. Ed. Engl. 1995, 34, 1849. (c) Boche, G.; Lohrenz, J.
C. Chem. Rev. 2001, 101, 697.
AUTHOR INFORMATION
Corresponding Author
CWSo@ntu.edu.sg; mezailles@chimie.ups-tlse.fr
■
Present Address
§N.M.: Laboratoire Het
́ ́ ́
ochimie Fondamentale et Appliquee,
er
́
Paul Sabatier, CNRS, 118 Route de Narbonne,
Universite
31062 Toulouse, France.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors are grateful to CalMip (CNRS, Toulouse, France)
for calculation facilities. The CNRS, the Ecole Polytechnique,
and the MOE (AcRF Tier 1, RG 22/12) are thanked for
supporting this work. H.H. thanks the Ecole Polytechnique for
a Ph.D. fellowship. S.Y.-F.H. gratefully acknowledges Nanyang
Technological University and the Ecole Polytechnique for a
joint Ph.D. scholarship. N.M. and C.-W.S. thank the Merlion
Program (Project 4.05.10) for logistical support. We thank Dr.
D. Bourissou for helpful discussions.
REFERENCES
■
(1) (a) Spikes, G. H.; Fettinger, J. C.; Power, P. P. J. Am. Chem. Soc.
2005, 127, 12232. (b) Power, P. P. Nature 2010, 463, 171.
(2) Welch, G. C.; San Juan, R. R.; Masuda, J. D.; Stephan, D. W.
Science 2006, 314, 1124.
D
dx.doi.org/10.1021/ja401763c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX