Monomeric and oligomeric amine-borane σ-complexes of rhodium. Intermediates in the catalytic dehydrogenation of amine-boranes

Published on Web 09/28/2009

Monomeric and Oligomeric Amine-Borane σ-Complexes of

Rhodium. Intermediates in the Catalytic Dehydrogenation of

Amine-Boranes

Thomas M. Douglas,^†Adrian B. Chaplin,^†Andrew S. Weller,*^,†Xinzheng Yang,^‡

and Michael B. Hall*^,‡

Department of Chemistry, Inorganic Chemisty Laboratories, UniVersity of Oxford,

Oxford OX1 3QR, U.K., and Department of Chemistry, Texas A&M UniVersity, College Station,

Texas 77843

Received July 28, 2009; E-mail: andrew.weller@chem.ox.ac.uk; mbhall@tamu.edu

Abstract: A combined experimental/quantum chemical investigation of the transition metal-mediated

dehydrocoupling reaction of H₃B·NMe₂H to ultimately give the cyclic dimer [H₂BNMe₂]₂is reported.

Intermediates and model complexes have been isolated, including examples of amine-borane σ-complexes

of Rh(I) and Rh(III). These come from addition of a suitable amine-borane to the crystallographically

characterized precursor [Rh(η⁶-1,2-F₂C₆H₄)(PⁱBu₃)₂][BAr^F₄] [Ar^F) 3,5-(CF₃)₂C₆H₃]. The complexes [Rh(η²-

H₃B·NMe₃)(PⁱBu₃)₂][BAr^F₄] and [Rh(H)₂(η²-H₃B·NHMe₂)(PⁱBu₃)₂][BAr^F₄] have also been crystallographically

characterized. Other intermediates that stem from either H₂loss or gain have been characterized in solution

by NMR spectroscopy and ESI-MS. These complexes are competent in the catalytic dehydrocoupling (5

mol %) of H₃B·NMe₂H. During catalysis the linear dimer amine-borane H₃B·NMe₂BH₂·NHMe₂is observed

which follows a characteristic intermediate time/concentration proﬁle. The corresponding amine-borane

σ-complex, [Rh(PⁱBu₃)₂(η²-H₃B·NMe₂BH₂·NHMe₂)][BAr^F₄], has been isolated and crystallographically

characterized. A Rh(I) complex of the ﬁnal product, [Rh(PⁱBu₃)₂{η²-(H₂BNMe₂)₂}][BAr^F₄], is also reported,

although this complex lies outside the proposed catalytic cycle. DFT calculations show that the ﬁrst proposed

dehydrogenation step, to give H₂BdNMe₂, proceeds via two possible routes of essentially the same energy

barrier: BH or NH activation followed by NH or BH activation, respectively. Subsequent to this, two possible

low energy routes that invoke either H₂/H₂BdNMe₂loss or H₂BdNMe₂/H₂loss are suggested. For the

second dehydrogenation step, which ultimately affords [H₂BNMe₂]₂, a number of experimental observations

suggest that a simple intramolecular route is not operating: (i) the isolated complex [Rh(PⁱBu₃)₂(η²-

H₃B · NMe₂BH₂· NHMe₂)][BAr^F

] is stable in the absence of amine-boranes; (ii) addition of

4

H₃B·NMe₂BH₂·NHMe₂to [Rh(PⁱBu₃)₂(η²-H₃B·NMe₂BH₂·NHMe₂)][BAr^F₄] initiates dehydrocoupling; and (iii)

H₂BdNMe₂is also observed during this process.

NHC complexes.¹¹Colloidal-Rh has also been shown to be an

active catalyst,^12,13for which in situ EXAFS suggests that the

1. Introduction

The coordination chemistry of amine-boranes, prototypically

H₃B · NH₃, and their subsequent reactivity is an area of

signiﬁcant contemporary interest. This comes from the role that

metal complexes play in controlling kinetics and product

distributions for subsequent dehydrocoupling reactions to form

linear or cyclic oligomers and polymers alongside the concomi-

tant release of H₂. Control of these processes has relevance to

chemical hydrogen storage¹and the synthesis of new main-

group polymeric materials.²Transition metal catalysts utilized

for dehydrogenation include Cp₂Ti derivatives,^3,4Re-nitrosyls,⁵

Ir-pincer complexes,^6-8Ru-amino complexes,^9,10and Cu/Ni-

(4) Pun, D.; Lobkovsky, E.; Chirik, P. J. Chem. Commun. 2007, 3297–

3299.

(5) Jiang, Y.; Berke, H. Chem. Commun. 2007, 3571–3573.

(6) (a) Dietrich, B. L.; Goldberg, K. I.; Heinekey, D. M.; Autrey, T.;

Linehan, J. C. Inorg. Chem. 2008, 47, 8583–8585. (b) Rousseau, R.;

Schenter, G. K.; Fulton, J. L.; Linehan, J. C.; Engelhard, M. H.; Autry,

T. J. Am. Chem. Soc. 2009, 131, 10516–10524.

(7) Hebden, T. J.; Denney, M. C.; Pons, V.; Piccoli, P. M. B.; Koetzle,

T. F.; Schultz, A. J.; Kaminsky, W.; Goldberg, K. I.; Heinekey, D. M.

J. Am. Chem. Soc. 2008, 130, 10812–10820.

(8) Denney, M. C.; Pons, V.; Hebden, T. J.; Heinekey, D. M.; Goldberg,

K. I. J. Am. Chem. Soc. 2006, 128, 12048–12049.

(9) Kass, M.; Friedrich, A.; Drees, M.; Schneider, S. Angew. Chem., Int.

Ed. 2009, 48, 905–907.

^†University of Oxford.

^‡Texas A&M University.

(10) Blaquiere, N.; Diallo-Garcia, S.; Gorelsky, S. I.; Black, D. A.; Fagnou,

K. J. Am. Chem. Soc. 2008, 130, 14034–14045.

(1) (a) Marder, T. B. Angew. Chem., Int. Ed. 2007, 46, 8116–8118. (b)

Stephens, F. H.; Pons, V.; Baker, R. T. Dalton Trans. 2007, 2613–

2626. (c) Hamilton, C. W.; Baker, R. T.; Staubitz, A.; Manners, I.

Chem. Soc. ReV. 2009, 38, 279–293.

(11) Keaton, R. J.; Blacquiere, J. M.; Baker, R. T. J. Am. Chem. Soc. 2007,

129, 1844–1845.

(12) Jaska, C. A.; Manners, I. J. Am. Chem. Soc. 2004, 126, 9776–9785.

(13) Jaska, C. A.; Temple, K.; Lough, A. J.; Manners, I. J. Am. Chem.

Soc. 2003, 125, 9424–9434.

(2) Clark, T. J.; Lee, K.; Manners, I. Chem.sEur. J. 2006, 12, 8634–

8648.

(3) Clark, T. J.; Russell, C. A.; Manners, I. J. Am. Chem. Soc. 2006, 128,

9582–9583.

(14) Fulton, J. L.; Linehan, J. C.; Autrey, T.; Balasubramanian, M.; Chen,

Y.; Szymczak, N. K. J. Am. Chem. Soc. 2007, 129, 11936–11949.

9

15440 J. AM. CHEM. SOC. 2009, 131, 15440–15456

10.1021/ja906070r CCC: $40.75  2009 American Chemical Society

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

Scheme 1. Inner Sphere Dehydrogenation Mechanism

report of such species involvement in catalytic dehydrocoupling

reactions.²⁶There are no reported examples of group 9

complexes, although chelated systems of group 9 metals, related

to those with agostic C-H interactions,²⁷have been reported,^28-31

as have dihydridoborate complexes.³²Related σ-complexes of

three coordinate boron have also been synthesized.^33-37The

valence isoelectronic relationship between amine-boranes and

alkanes²⁵meansthatthecoordinationchemistryofσ-amine-borane

ligands also models that of transient transition-metal-alkane

complexes observed in solution.³⁸

Experimental and computational work on the mechanism of

transition-metal mediated dehydrocoupling of amine-boranes

indicate complex processes that are dependent on functional-

ization patterns on the amine-borane and the identity of the

metal center used.^1c,40For example, when using H₃B·NH₃with

Rh,¹³Ti,⁴and Ni-based¹¹systems, cyclic trimers and cross-

linked derivatives result (e.g., V-R and VI-R, Scheme 2), Ir-

pincer complexes give pentamers,⁸and bifunctional Ru catalysts

give poorly deﬁned polymeric material.⁹With H₃B·NMeH₂,

polymerization/oligermization/trimerization is reported,^6,10,13,41

active species might actually be smaller Rh₄clusters.¹⁴Transi-

tion-metal-mediated hydrolytic decomposition of amine-boranes

and their derivatives has also been reported.¹⁵

Recent insight into the role of the metal complex in these

transformations has revealed two possible general pathways for

dehydrogenation. The ﬁrst is inner sphere, in which the

amine-borane coordinates to the metal center and undergoes

H-transfer reactions aided by the metal by a variety of

mechanistic steps. Scheme 1 illustrates one plausible reaction

pathway for homogeneous systems that involves inner sphere

dehydrogenation, although exact details of the mechanism are

still being debated. Computational and experimental studies

indicate a number of mechanistic scenarios for the dehydroge-

nation step depending on the metal/ligand combination: NH

proton transfer to a coordinated ligand followed by transfer to

the metal (Ni-N-heterocyclic carbene),¹⁶intermolecular stepwise

transfer of NH then BH (Cp₂Ti-derivatives),¹⁷and concerted

NH/BH activation at the metal center (Ir-pincer complexes).¹⁸

Oxidative addition of the BH bond followed by NH ꢀ-elimina-

tion has also been suggested.¹¹All these routes implicate

σ-complexes of amine-boranes as intermediates in the reac-

tion.¹⁹The other alternative mechanistic scenario is that afforded

by bifunctional Ru catalysts that are suggested to work via an

outer-sphere dehydrogenation mechanism similar to that invoked

for transfer dehydrogenation of alcohols.^9,10The role of free,

dissociated, ligand in nickel N-heterocyclic carbene systems has

also been recently discussed.²⁰

(23) Yasue, T.; Kawano, Y.; Shimoi, M. Angew. Chem., Int. Ed. 2003, 42,

1727–1730.

(24) Kawano, Y.; Hashiva, M.; Shimoi, M. Organometallics 2006, 25,

4420–4426.

(25) Kawano, Y.; Yamaguchi, K.; Miyake, S. Y.; Kakizawa, T.; Shimoi,

M. Chem.sEur. J. 2007, 13, 6920–6931.

(26) Shimoi, M.; Kawano, Y.; Taeko, T. S. Presented at the 224th ACS

National Meeting, Boston, 2002.

(27) Brookhart, M.; Green, M. L. H.; Parkin, G. Proc. Natl. Acad. Sci.

U.S.A. 2007, 104, 6908–6914.

(28) Macias, R.; Rath, N. P.; Barton, L. Angew. Chem., Int. Ed. 1999, 38,

162–164.

(29) Ingleson, M.; Patmore, N. J.; Ruggiero, G. D.; Frost, C. G.; Mahon,

M. F.; Willis, M. C.; Weller, A. S. Organometallics 2001, 20, 4434–

4436.

(30) (a) Bajko, Z.; Daniels, J.; Gudat, D.; Hap, S.; Nieger, M. Organo-

metallics 2002, 21, 5182–5189. (b) Rankin, M. A.; MacLean, D. F.;

McDonald, R.; Ferguson, M. J.; Lumsden, M. D.; Stradiotto, M.

Organometallics 2009, 28, 74–83.

(31) Wagler, J.; Hill, A. F. Organometallics 2008, 27, 2350–2353.

(32) Westcott, S. A.; Marder, T. B.; Baker, R. T.; Harlow, R. L.; Calabrese,

J. C.; Lam, K. C.; Lin, Z. Polyhedron 2004, 23, 2665–2677.

(33) For examples of three coordinate boron σ complexes:(a) Muhoro,

C. N.; He, X.; Hartwig, J. F. J. Am. Chem. Soc. 1999, 121, 5033–

5046. (b) Hartwig, J. F.; Muhoro, C. N.; He, X.; Eisenstein, O.;

Bosque, R.; Maseras, F. J. Am. Chem. Soc. 1996, 118, 10936–10937.

(c) Montiel-Palma, V.; Lumbierres, M.; Donnadieu, B.; Sabo-Etienne,

S.; Chaudret, B. J. Am. Chem. Soc. 2002, 124, 5624–5625. (d) Crestani,

M. G.; Munoz-Hernandez, M.; Arevalo, A.; Acosta-Ramirez, A.;

Garcia, J. J. J. Am. Chem. Soc. 2005, 127, 18066–18073. (e) Alcaraz,

G.; Helmstedt, U.; Clot, E.; Vendier, L.; Sabo-Etienne, S. J. Am. Chem.

Soc. 2008, 130, 12878–12879. (f) Hebden, T. J.; Denney, M. C.; Pons,

V.; Piccoli, P. M. B.; Koetzle, T. F.; Schultz, A. J.; Kaminsky, W.;

Goldberg, K. I.; Heinekey, D. M. J. Am. Chem. Soc. 2008, 130, 10812–

10820.

Despite this intense recent interest, details of intermediate

species remain scarce,^4,8although materials that represent

catalyst deactivation products have been isolated.^4,7,8The

coordination chemistry of σ-amine-borane or phosphine-borane

complexes has been developed over the past decade by Shimoi

and co-workers and includes examples of group 6, 7, and 8

metals,^21-25but as far as we are aware there is only a brief

(15) For leading references see. (a) Yoon, C. W.; Sneddon, L. G. J. Am.

Chem. Soc. 2006, 128, 13992–13993. (b) Yan, J. M.; Zhang, X. B.;

Han, S.; Shioyama, H.; Xu, Q. Angew. Chem., Int. Ed. 2008, 47, 2287–

2289.

(34) (a) Lachaize, S.; Essalah, K.; Montiel-Palma, V.; Vendier, L.; Chaudret,

B.; Barthelat, J.-C.; Sabo-Etienne, S. Organometallics 2005, 24, 2935–

2943. (b) Hartwig, J. F.; Cook, K. S.; Hapke, M.; Incarvito, C. D.;

Fan, Y.; Webster, C. E.; Hall, M. B. J. Am. Chem. Soc. 2005, 127,

2538–2552.

(16) (a) Yang, X. Z.; Hall, M. B. J. Am. Chem. Soc. 2008, 130, 1798–

1799. (b) Yang, X. Z.; Hall, M. B. J. Organomet. Chem. 2009, 694,

2831–2838.

(17) Luo, Y.; Ohno, K. Organometallics 2007, 26, 3597–3600.

(18) Paul, A.; Musgrave, C. B. Angew. Chem., Int. Ed. 2007, 46, 8153–

8156.

(35) Alcaraz, G.; Clot, E.; Helmstedt, U.; Vendier, L.; Sabo-Etienne, S.

J. Am. Chem. Soc. 2007, 129, 8704–8705.

(19) We deﬁne a σ borane complex as one having an intermolecular

interaction between a B-H bond in a neutral ligand and a metal center.

It encompasses three and four coordinate borane species. Anionic

dihydridoborate and borohydride ligands are also known: Kubas, G. J.

Metal Dihydrogen and σ-Bond Complexes; Kluwer Academic/Plenum

Publishers: New York, 2001.

(36) Lam, W. H.; Shimada, S.; Batsanov, A. S.; Lin, Z.; Marder, T. B.;

Cowan, J. A.; Howard, J. A. K.; Mason, S. A.; McIntyre, G. J.

Organometallics 2003, 22, 4557–4568.

(37) Hesp, K. D.; Rankin, M. A.; McDonald, R.; Stradiotto, M. Inorg.

Chem. 2008, 47, 7471–7473.

(38) Lawes, D. J.; Geftakis, S.; Ball, G. E. J. Am. Chem. Soc. 2005, 127,

4134–4135.

(20) Zimmerman, P. M.; Paul, A.; Zhang, Z. Y.; Musgrave, C. B. Angew.

Chem., Int. Ed. 2009, 48, 2201–2205.

(39) Zimmerman, P. M.; Paul, A.; Zhang, Z. Y.; Musgrave, C. B. Inorg.

Chem. 2009, 48, 1069–1081.

(21) Shimoi, M.; Nagai, S.; Ichikawa, M.; Kawano, Y.; Katoh, K.; Uruichi,

M.; Ogino, H. J. Am. Chem. Soc. 1999, 121, 11704–11712.

(22) Kakizawa, T.; Kawano, Y.; Shimoi, M. Organometallics 2001, 20,

3211–3213.

(40) Pons, V.; Baker, R. T.; Szymczak, N. K.; Heldebrant, D. J.; Linehan,

J. C.; Matus, M. H.; Grant, D. J.; Dixon, D. A. Chem. Commun. 2008,

6597–6599.

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15441

A R T I C L E S

Douglas et al.

Scheme 2. Abridged Dehydrogenation and Dehydrocoupling Reaction Scheme for Amine-Boranes and the Suggested Involvement of

Linear Dimer III-R^17,18,39,40

Scheme 3. Representative Complexes Reported in This Paper^a

while H₃B · NMe₂H gives the cyclic dimer [H₂BNMe₂]₂

(IV).^{3,4,13,14,42,43}Detailed computational^16-18,20,39and experi-

mental studies⁴⁰suggest that metal-mediated dehydrogenation

leads ﬁrst to amino-boranes, H₂BdNR₂(II-R, Scheme 2), which

then dimerize or oligomerize to eventually form the observed

products. In some cases the formation of the linear dimer

H₃B·NR₂BH₂·NR₂H (III-R) is suggested as an early step in

these processes, which can come from reaction of H₂BdNR₂

with H₃B·NHR₂.^13,39,40Such a compound, which could be

bound to a metal or generated free under the reaction conditions,

could then proceed to give the ﬁnal products via further

dehydrocoupling pathways. The role of metal in this process

has not been ﬁrmly established, although the catalytic dehy-

drocoupling of III-Me using colloidal Rh to form dimeric

^a[BAr^F]^-anions are not shown.

[H₂BNMe₂]₂has been reported previously by Manners and co-

workers demonstrating its possible role as an intermediate.¹³

However, details of both the coordination chemistry of III-Me

and the role that the metal plays in its dehydrogenation have

not been reported. Related calcium complexes of the [HNBH-

NHBH]^2-dianion result from the reaction of calcium hydrides

with H₃B · NH₃.⁴⁴Manners and co-workers have shown that

the catalytic dehydrocoupling of H₃B · PHPh₂using Rh(I)

precatalysts gives H₃B · PPh₂BH₂· PPh₂H at 90 °C and the

cyclic products [Ph₂PBH₂]₃and [Ph₂PBH₂]₄at 120 °C. Monitor-

ing this higher temperature reaction shows a mixture of

H₃B · PPh₂BH₂· PPh₂H and the cyclic species, suggesting the

possible role of the linear oligomer as an intermediate.⁴⁵

As metal σ-complexes are implicated throughout the catalytic

cycle in inner-sphere mechanisms we have set out to deliberately

synthesize such species and study their coordination chemistry

and onward reactivity. Here we report upon the synthesis of

σ-complexes of a variety of amine-boranes based upon

H₃B·NMe₂H complexed with rhodium (Scheme 3) that are

either suggested as putative intermediates in the catalytic

dehydrocoupling or model complexes thereof. The approach we

use is to take a preformed cationic low-coordinate latent-“12-

electron” [Rh(PⁱBu₃)₂]⁺fragment with suitable amine-borane

ligands. Utilizing density functional theory (DFT), we also report

on computational work that evaluates possible mechanistic

routes in the dehydrocoupling of H₃B · NHMe₂to give

H₂BdNMe₂using these Rh systems (i.e., I to II). Although

none of the systems reported herein offer improvement on the

current best, in terms of turnover frequency for the dehydro-

coupling,^8-10this attenuated reactivity has the beneﬁt of

allowing for intermediates to be observed and thus gives insight

4

Scheme 4. Synthesis of the Precursor Complexes 1-3^a

^a[BAr^F]^-anions are not shown.

4

into mechanistic processes. Aspects of this work have been

communicated.⁴³

2. Results and Discussion

2.1. Synthesis of σ-Complexes of H₃B·NMe₃and

H₃B·NHMe₂. We have previously reported upon the synthesis

of the latent 12-electron complex [Rh(PⁱBu₃)₂][BAr^F] 1 (Scheme

4

4, Ar^F) C₆H₃(CF₃)₂) from straightforward H₂removal from

the 14-electron complex [Rh(H)₂(PⁱBu₃)₂][BAr^F] 3. For 3 a

4

crystal structure showed that the coordinate and electronic

unsaturation is relieved by supporting agostic C-H interac-

tions,⁴⁶and this structure is supported by DFT calculations

(Figure 1). Such interactions are also probably present in 1 in

the solid state, and the calculated structure shows two such

Rh· · ·H₃C contacts trans to the phosphines (Figure 1) making

a pseudo-square planar Rh(I) species. Noteworthy is that the

trans isomer of 1 is 3.1 kcal/mol less stable than the cis (gas

phase free energy). Both these complexes are well set up for

coordination of weak ligands such as amine-boranes as the

agostic interactions stabilize a latent coordination site. 1 reacts

with weak ligands such as dichloroethane or ﬂuorobenzene,

to form a simple adduct and an η⁶-arene complex, respec-

tively.⁴⁶In this contribution the even weaker solvent 1,2-

diﬂuorobenzene is the solvent of choice for the study of the

(41) Staubitz, A.; Soto, A. P.; Manners, I. Angew. Chem., Int. Ed. 2008,

47, 6212–6215.

(42) Sloan, M. E.; Clark, T. J.; Manners, I. Inorg. Chem. 2009, 48, 2429–

2435.

(43) Douglas, T. M.; Chaplin, A. B.; Weller, A. S. J. Am. Chem. Soc. 2008,

130, 14432–14433.

(44) Spielmann, J.; Jansen, G.; Bandmann, H.; Harder, S. Angew. Chem.,

Int. Ed. 2008, 47, 6290–6295.

(45) Dorn, H.; Singh, R. A.; Massey, J. A.; Nelson, J. M.; Jaska, C. A.;

Lough, A. J.; Manners, I. J. Am. Chem. Soc. 2000, 122, 6669–6678.

(46) Douglas, T. M.; Chaplin, A. B.; Weller, A. S. Organometallics 2008,

27, 2918–2921.

9

15442 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

As the dehydrocoupling of amine-boranes may proceed via

B-H/N-H activation for inner sphere and outer sphere reac-

tions (in whatever order), in the anticipation of ﬁrst isolating a

moderately stable amine-borane σ-complex, H₃B·NMe₃was

used initially, in which the quaternary nitrogen offers no protic

hydrogen. Metal complexes from groups 6,²¹7,²⁵and 8²⁴of

H₃B·NMe₃, as well as closely related complexes, have previ-

ously been reported by Shimoi and co-workers, but no later

transition metal complexes are known.

Addition of H₃B·NMe₃to a 1,2-F₂C₆H₄solution of 2

(generated in situ from 1) results in an immediate color change

from orange to deep purple. NMR data, ESI-MS, and a solid-

state structure identify this new product as [Rh(η²-

Figure 1. DFT optimized structures of the cationic portions of 1 and 3.

Bond lengths in Å. Bond angle in degrees.

H₃B·NMe₃)(PⁱBu₃)₂][BAr^F] (4) (Scheme 5). Air- and temper-

4

ature-sensitive crystals were grown from 1,2-F₂C₆H₄/pentane

solvent in excellent isolated yield (87%), and the solid-state

molecular structure is shown in Figure 3. The amine-borane

ligand is η²-coordinated through two Rh-H-B σ-bonds, and

the Rh(I) center is pseudo square planar: the sum of angles

around Rh1 ) 360.3°. The hydrogen atoms associated with the

boron B1 (H1A, H1B, H1C) were located on the Fourier

difference map. The Rh-B distance, 2.180(4) Å, is longer than

a Rh-B boryl, namely, (ⁱPr₃P)₂RhHCl(BO₂C₆H₄) [1.965(2) Å]³⁶

and (PMe₃)₄Rh(BO₂C₆H₄) [2.047(2) Å].⁵⁰It is signiﬁcantly

shorter than that observed in the Rh(I) chelated complexes

[Rh(cod)(η²-H₃B·PPh₂CH₂PPh₂)][BPh₄] [2.313(3) Å]²⁹and

Rh(PPh₃)₂(η²-H₃B-7-azaindol-7-ylborate)³¹but very similar

to that observed in the Rh(I) dihydridoborate complex Rh(η²-

H₂BC₈H₁₄)(PⁱPr₃)₂[2.170(2) Å]³²and the chlorohydridoborates

RuCp*(PMe₃)(η²-BH₃Cl) [2.122(8) Å]⁵¹and RuCp*(PⁱPr₃)-

(BH₂MesCl) [2.162(3) Å].³⁷Although within the experimental

limits of the X-ray diffraction experiment (3σ) the bridging

hydride ligands are not elongated compared to the terminal

hydride [viz. 1.23(3) vs B1-H1C 1.05(4) Å], there does appear

to be a trend that they become longer on coordination with the

metal, while the acute H-Rh-H angle of 68(2)° points toward

a complex that is best described as a Rh(I) center ligated with

a σ-borane that has a signiﬁcant Rh · · ·B interaction, rather than

an alternate Rh(III)-dihydride with a boryl ligand.^34,52The

observation of a signiﬁcantly quadrupolar broadened signal for

Figure 2. Molecular structure the cationic portion of 2: ball and stick

diagram. Minor disordered components and H atoms omitted for clarity.

isolation and reactivity of these sensitive complexes, which

decompose slowly in CH₂Cl₂. Dissolving 1 in this solvent

results in the corresponding η⁶-arene complex [Rh(η⁶-1,2-

F₂C₆H₄)(PⁱBu₃)₂][BAr^F] 2. As 2 represents the starting material

4

generated in situ for all subsequent reactivity, we have

characterized it by NMR spectroscopy, ESI-MS, and X-ray

crystallography. The solid-state structure of 2, although of low

quality (R₁) 9%), demonstrates that the 1,2-ﬂuorobenzene

ligand coordinates η⁶, similar to that observed for [Rh(η⁶-

1

the bridging hydrides in the H NMR spectrum (fwhm ) 210

FC₆H₅)(PⁱBu₃)₂][BAr^F]⁴⁶(Figure 2). The NMR data for 2 in

4

Hz) also demonstrates a signiﬁcant B-H interaction. The DFT

calculated structure for 4 closely mirrors those experimentally

observed in Figure 4.

1,2-F₂C₆H₄solution are in full accord with the structure,

although the coordinated arene resonances were obscured by

the protio solvent. Dissolving crystals of 2 in alternate solvents

results in the weakly bound arene ligand being displaced; for

example, in 1,2-dichloroethane the resulting adduct complex is

formed (Supporting Information).⁴⁶As far as we are aware 2 is

the ﬁrst example of a complex coordinated with 1,2-diﬂuo-

robenzene⁴⁷and is also a rare example of a ﬂuoroarene complex

of a late transition metal.^48,49Complex 2 is an excellent source

of a latent 12-electron [Rh(PⁱBu₃)₂]⁺fragment and is well set

up for reaction with amine-boranes, which is discussed next.

In solution the H₃B·NMe₃ligand remains coordinated to the

metal center, and ESI-MS shows a strong molecular ion at m/z

i

) 580.388 (calcd 580.382). In addition to signals due to Bu

and NMe₃groups, the ¹H NMR spectrum at room temperature

shows a quadrupolar broadened relative integral 3H signal at δ

1

-2.12, which sharpens in the H{¹¹B} NMR spectrum. This

integral 3-H signal indicates rapid exchange between terminal

and bridging hydrides at room temperature, as has been observed

for both η¹-σ-complexes^21,22,24and η²-chelated phosphine

boranes.^28,53,54Cooling to 190 K arrests this ﬂuxional process

and results in the observation of a very broad, integral 2-H signal

(47) Complexes of 1,2-iodoarenes are known, for example:Crabtree, R. H.;

Faller, J. W.; Mellea, M. F.; Quirk, J. M. Organometallics 1982, 1,

1361–1366.

(48) (a) Douglas, T. M.; Notre, J. L.; Brayshaw, S. K.; Frost, C. G.; Weller,

A. S. Chem. Commun. 2006, 3408–3410. (b) Palacios, M. D.; Puerta,

M. C.; Valerga, P.; Lledos, A.; Veilly, E. Inorg. Chem. 2007, 46, 6958–

6967. (c) Willems, S. T. H.; Budzelaar, P. H. M.; Moonen, N. N. P.;

de Gelder, R.; Smits, J. M. M.; Gal, A. W. Chem. Eur. J. 2002, 8,

1310–1320.

(50) Dai, C.; Stringer, G.; Marder, T. B.; Scott, A. J.; Clegg, W.; Norman,

N. C. Inorg. Chem. 1997, 36, 272–273.

(51) Kawano, Y.; Shimoi, M. Chem. Lett. 1998, 935–936.

(52) Hartwig, J. F.; De Gala, S. R. J. Am. Chem. Soc. 1994, 116, 3661–

3662.

(53) Merle, N.; Koicok-Kohn, G.; Mahon, M. F.; Frost, C. G.; Ruggerio,

G. D.; Weller, A. S.; Willis, M. C. Dalton Trans. 2004, 3883–3892.

(54) Merle, N.; Frost, C. G.; Kociok-Kohn, G.; Willis, M. C.; Weller, A. S.

Eur. J. Inorg. Chem. 2006, 4068–4073.

(49) Douglas, T. M.; Brayshaw, S. K.; Dallanegra, R.; Kociok-Kohn, G.;

Macregor, S. A.; Moxham, G. L.; Weller, A. S.; Wondimagegn, T.;

Vadivelu, F. Chem.sEur. J. 2008, 14, 1004–1022.

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15443

A R T I C L E S

Douglas et al.

Scheme 5. Synthesis of H₃B·NMe₃Complexes^a

^a[BAr^F]^-anions are not shown.

4

Table 1. Comparison of Rh-B Bond Length, δ(¹¹B) of Coordinated

Boron, and the Chemical Shift Difference (∆δ) Compared with

Free Ligand^a

^aFor the column marked with *, see Experimental Section for

chemical shifts of free amine-boranes in 1,2-F₂C₆H₄solvent.

and larger changes in chemical shift [∆δ(¹¹B)] compared to the

Rh(III) example (7).

Complex 4 is stable in 1,2-F₂C₆H₄solution, although it is

very air-sensitive and best handled in an argon-ﬁlled glovebox.

Addition of H₂to complex 4 results in an immediate color

change to yellow and the formation of the new complex

Figure 3. Molecular structure of the cationic portion of 4: ellipsoids

depicted at the 50% probably level. Minor disordered components and most

H atoms omitted for clarity. Selected bond lengths (Å) and angles (deg):

Rh1-B1, 2.180(4); Rh1-P1, 2.2412(10); Rh1-P2, 2.2149(10); B1-N1,

1.594(5); Rh1-H1A, 1.80(4); Rh1-H1B, 1.84(4); B1-H1A, 1.23(3);

B1-H1B, 1.23(3); B1-H1C, 1.05(4); P2-Rh1-P1, 97.35(4); P1-Rh1-B1,

135.18(11);P2-Rh1-B1,126.99(11);N1-B1-Rh1,129.2(3);Rh1-H1A-B1,

90(2); Rh1-H1B-B1, 88(2); H1A-Rh1-H1B, 68(2).

[Rh(H)₂(η²-H₃B·NMe₃)(PⁱBu₃)₂][BAr^F] 5. Complex 5 can also

4

be prepared by addition of H₃B·NMe₃to a 1,2-F₂C₆H₄solution

of 3 (Scheme 5). Complex 5 readily loses H₂to re-form 4 and

thus is best characterized under an atmosphere of H₂(1 atm),

by NMR spectroscopy and ESI-MS. These NMR data are very

similar to that for the H₃B·NHMe₂adduct (7) for which a single

crystal X-ray structural determination shows an η²-coordinated

amine-borane (vide infra). The presence of a Rh(III) center is

indicated by a reduced ¹⁰³Rh-³¹P coupling constant compared

1

with 4, δ 21.10 [d, J(RhP) 106 Hz]. The H NMR spectrum

shows a sharp, integral 2-H doublet of triplets at δ -18.81 due

to two equivalent hydride ligands, which becomes a doublet

on ³¹P decoupling. These coupling patterns demonstrate that the

hydrides are orientated cis to the phosphine ligands. A very

broad quadrupolar-broadened relative integral 3-H signal at δ

-0.47 is assigned to the B-H-Rh protons, indicating rapid

site exchange for the B-H protons, which becomes less broad

on decoupling ¹¹B. Cooling to 200 K did not freeze out the site

exchange between terminal and bridging hydrides, perhaps

suggesting a weaker Rh · · ·H-B interaction which would be

consistent with the strong trans inﬂuence of the hydrides (Table

1). The ¹¹B chemical shift also reﬂects this, with a very broad

signal observed at δ 5.1 shifted considerably less downﬁeld from

free amine-borane compared with 4 (Table 1).

Figure 4. DFT optimized structure of cationic portion of 4. Bond lengths

in Å. Bond angle in degrees. The Bu groups are not shown for clarity.

i

at δ -5.55 assigned to the Rh-H-B protons, now consistent

with the solid-state structure. We could not observe the

corresponding terminal B-H proton, which is presumably broad

and possibly obscured by the aliphatic phosphine protons.

Exchange probably occurs by a η²-η¹-η²process, as has been

calculated for exchange in η¹-amine-borane complexes⁵⁵and

also observed on the chemical time scale as hemilabile behavior

in chelated systems.^31,53,54A single environment is observed

in the ³¹P{¹H} NMR spectrum that has a large coupling constant

of 175 Hz consistent with a Rh(I) center. The ¹¹B NMR

spectrum displays a broad signal at δ 23.1, shifted 30.4 ppm

downﬁeld from H₃B·NMe₃, consistent with a signiﬁcant Rh · · ·B

interaction.^53,56Table 1 shows that for all the compounds

crystallographically characterized in this paper the ¹¹B chemical

shifts move downﬁeld signiﬁcantly on coordination. The Rh(I)

complexes (4, 10, and 12) show both shorter Rh · · ·B interactions

As a result of the ready H₂loss, we have not been able to

obtain crystalline material of 5; however, inspection of a space-

ﬁlling diagram of closely related 7 (Figure 5, vide infra)

indicates that steric pressure of the trans-axial phosphines and

the NMe₃group could promote H₂loss, favoring the pseudo-

square-planar Rh(I) structure in 4. This is not present in 7, which

does not lose H₂as readily (vide infra). Addition of H₃B·NMe₃

to d₂-3, [Rh(D)₂(PⁱBu₃)₂][BAr^F], under a partial pressure of

4

D₂(g), resulted in a washing of deuterium into the B-H positions

(55) Kawano, Y.; Kakizawa, T.; Yamaguchi, K.; Shimoi, M. Chem. Lett.

2006, 35, 568–569.

(56) Alcaraz, G.; Sabo-Etienne, S. Coord. Chem. ReV. 2008, 252, 2395–

2409.

9

15444 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

Figure 5. Molecular structure of the cationic portion of 7 (ellipsoids depicted at the 50% probably level, most H atoms omitted for clarity) and space ﬁlling

diagram (van der Waals radii with H₃B·NHMe₂highlighted in red). Selected bond lengths (Å) and angles (deg): Rh1-B1, 2.318(8); Rh1-P1, 2.307(2);

Rh1-P2, 2.302(2); Rh1-H0A, 1.42(6); Rh1-H0B, 1.41(6); B1-N1, 1.595(10); Rh1-H1A, 1.82(6); Rh1-H1B, 1.86(7); B1-H1A, 1.20(4); B1-H1B,

1.20(4); B1-H1C, 1.10(7); H1-H3B, 2.47; H1-H7A, 2.88; P1-Rh1-P2, 163.65(7); P1-Rh1-B1, 103.6(2); P2-Rh1-B1, 92.6(2); Rh1-B1-N1, 125.1(5);

Rh1-H1A-B1, 98(4); Rh1-H1B-B1, 96(4); H1A-Rh1-H1B, 61(3); H1A-B1-H1B, 104(5). H3B and H7A are placed in calculated positions.

Scheme 6. Synthesis of H₃B·NHMe₂Complexes^a

^a[BAr^F]^-anions are not shown.

4

of 5 and hydrogen into the Rh-D positions. Dissolved H₂and

HD are observed [δ 4.60; δ 4.56 J(HD) 42 Hz, respectively].

These observations suggest reversible B-H activation at the

metal center, and we suggest that this is most likely via a Rh(I)

complex such as 4 rather than a Rh(III) dihydride 5 as this

invokes a Rh(I)/Rh(III) couple rather than a Rh(III)/Rh(V). This

is supported by the rather low calculated B-H bond activation

barrier for the formation of an unstable boryl hydride intermedi-

ate from a Rh(I) species (cf. 6–13 versus 7–19, Scheme 11).

to those found for 4 suggesting a very similar structure, while

the NMR data for complex 7 are very close to those observed

for 5.

In solution at room temperature the BH₃group in complex 7

undergoes rapid exchange between terminal and bridging

hydrides that can be frozen out at low temperature (190 K) in

contrast to complex 5 where this ﬂuxional process cannot be

frozen out at low temperature. At 298 K an integral 3-H signal

is observed for the Rh-H-B protons, δ -0.77, and the hydride

ligands are observed as a doublet of triplets δ -17.42, displaying

cis-coupling to the two equivalent phosphines [δ(³¹P) 22.26;

J(RhP) 105 Hz]. On cooling (190 K) the ﬂuxional process is

halted, and a relative integral 2H signal is observed for the

Rh-H-B protons at δ -3.15. The solid-state structure of 7 is

shown in Figure 5 and shows a η²-coordinated amine-borane

on a Rh(III) center, with cis hydrides and trans phosphines.

The hydride ligands and BH protons were located in the ﬁnal

electron density difference map. The Rh · · · B distance,

[2.318(8) Å], is signiﬁcantly longer than that in 4 reﬂecting

a weaker interaction. It is similar to that observed in

[Rh(cod)(η²-H₃B · PPh₂CH₂PPh₂)][BPh₄], [2.313(3) Å].²⁹This

weaker interaction is mirrored in the ¹¹B NMR chemical shift,

δ 2.2 (Table 1). The H1A-Rh1-H1B angle is also more acute

[61(3)°], reﬂecting this weaker interaction. The NHMe₂group

Complexes 4 and 5 do not proceed to dehydrocouple, a

consequence of the lack of a N-H proton. By contrast,

addition of H₃B · NHMe₂to 1 or 3 in 1,2-F₂C₆H₄solution

results in a more reactive set of complexes which do proceed

on to dehydrocouple. Thus, addition of 2 equiv of

H₃B · NHMe₂to 1 resulted in an immediate color change to

give a short-lived purple complex, characterized as [Rh(η²-

H₃B·NHMe₂)(PⁱBu₃)₂][BAr^F] (6). 6 proceeds rapidly to give a

4

pale-yellow complex, which has been characterized by NMR

spectroscopy, ESI-MS, and X-ray crystallography as [Rh(H)₂(η²-

H₃B·NHMe₂)(PⁱBu₃)₂][BAr^F] (7) (Scheme 6). Concomitant

4

with this is the formation of [H₂BNMe₂]₂[δ(¹¹Β) 5.5; t, J(BH)

112 Hz].¹³Complex 6 thus is an intermediate in the dehydro-

coupling of H₃B·NHMe₂with the hydride ligands in the ﬁnal

product 7 presumably coming from the initial dehydrogenation.

Addition of smaller amounts (less than 2 equiv) of H₃B·NHMe₂

to 1 resulted in the formation of mixtures of 6, 7, 2, and

[H₂BNMe₂]₂in varying proportions, meaning that 6 could not

be isolated free of 7. 6 was longer-lived under these conditions

allowing for its characterization in solution by ESI-MS and in

situ NMR spectroscopy. The NMR data for 6 are very similar

i

is orientated so that the NH proton is directed toward the Bu

groups, ﬁtting snugly in a pocket made by the phosphine alkyl

groups. This no doubt minimizes steric pressures as the

N-H· · ·H-C distances are too long to be considered as strongly

attractive, namely, H1 · · ·H3B 2.47; H1 · · ·H7A 2.88 Å. As with

5, complex 7 loses H₂to re-form a Rh(I) complex 6, but as this

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15445

A R T I C L E S

Douglas et al.

J(BH) 112], Figures 6 and 7. In our original communication

we tentatively identiﬁed this as [H₂BNMe₂]₃,⁴³on the basis of

previous suggestions that this cyclic species is as an intermediate

in the dehydrogenation of H₃B·NHMe₂by “Cp₂Ti”³and Cu/

Ni-N-heterocyclic carbenes.¹¹This cyclic trimer would be

required to undergo B-N bond cleavage to eventually form

IV, which must be catalyzed, presumably by a transition metal,

as [H₂BNMe₂]₃has been reported to be stable at least up to its

melting point of 97° in the absence of a metal.⁵⁹However, we

now suggest that this intermediate is, in fact, the linear

diborazane H₃B·NMe₂BH₂·NHMe₂, III-Me. This linear dimer

was ﬁrst isolated in low yield by Hahn and Schaeffer⁶⁰and the

crystal structure reported by No¨th.⁶¹An improved synthesis by

Manners and co-workers has recently been reported.¹³In

addition they demonstrated that III-Me undergoes dehydro-

coupling to give IV using colloidal Rh catalysts, leading to the

suggestion that is was a viable intermediate in the conversion

of H₃B · NHMe₂to IV using a transition metal catalyst. The

protio analogue III-H, H₃B·NH₂BH₂·NH₃, has also been

calculated as a possible intermediate in hydrogen loss from

H₃B·NH₃,^39,40although calculations on the Ti-catalyzed dehy-

drocoupling of H₃B·NHMe₂suggest that metal mediated

dehydrocoupling of III-Me to form IV is not occurring.¹⁷The

¹¹B chemical shifts for independently prepared¹³III-Me in 1,2-

F₂C₆H₄solvent are observed at δ 2.5 [t, BH₂, J(BH) 112 Hz]

and δ -12.8 [q, BH₃, J(BH) 95 Hz], the former being that

observed for the intermediate species (Figure 7). The BH₃

resonance of III-Me is very close to that for the starting material

H₃B·NHMe₂[δ -12.8, t, J(BH) 98 Hz] in 1,2-F₂C₆H₄, and this

resonance in the ¹¹B NMR spectrum of the catalysis mixture is

rather broad (Figure 7) indicating overlapping signals. When

the reaction is quenched with MeCN this change in solvent

polarity results in a clear separation of these high-ﬁeld peaks

(Supporting Information). The key data that characterizes this

intermediate as III-Me in 1,2-F₂C₆H₄solvent comes from

Figure 6. Plot of ¹¹B concentration for the dehydrocoupling of H₃B·NHMe₂

(initial concentration ) 0.48 mol dm^-3) using 1 (298 K, 5 mol %, 1,2-

F₂C₆H₄, sealed system). [ ) [H₂BNMe₂]₂, the ﬁnal product; O )

H₃B·NMe₂BH₂·NHMe₂, III-Me. Reaction proceeds to 100% conversion

after 600 min.

is unstable itself this eventually results in a mixture of products.

The loss of H₂is much slower, requiring application of a

vacuum; this being consistent with the attenuated steric pressure

in 7 between the amine-borane and phosphine compared with

5. Complex 7 can also be accessed by addition of H₃B·NHMe₂

to 3. Addition of H₃B·NHMe₂to d₂-3, [Rh(D)₂(PⁱBu₃)₂][BAr^F],

4

resulted in washing of deuterium to the BH₃group consistent

with a low energy H/D exchange process. DFT optimized

structures of 6 and 7 are shown in Section 3 (Figure 16).

Complexes 6 and 7 represent intermediates on the dehydro-

genation cycle of H₃B·NHMe₂, and 4 and 5 are model

complexes for this cycle. All are accessed via addition of the

precursor amine-borane to 1 in 1,2-F₂C₆H₄solution. We next

investigated whether these Rh(I) or Rh(III) species turnover in

the catalytic dehydrocoupling of H₃B·NHMe₂.

2.2. Catalytic Dehydrogenation of H₃B·NHMe₂. Isolated

complexes 2 and 7 are active catalysts for the dehydrogenation

of H₃B·NHMe₂. In an open system under Ar the addition of

H₃B·NHMe₂to2in1,2-F₂C₆H₄solventresultsinamodest^4,8-11,42

overall turnover frequency (35 min, TOF 34 h^-1, 298 K, 5 mol

%, 100% conversion) to ultimately afford the cyclic dimer IV.

One equivalent of H₂is released during this process (by gas

buret). Addition of Hg did not inhibit catalysis. Repeating this

reaction in a sealed NMR tube resulted in a lower TOF (10 h,

TOF 2 h^-1) suggesting inhibition by H₂released during catalysis,

probably by favoring Rh(III) dihydride complex 3 or 7 rather

than a Rh(I) species. Under these attenuated conditions a time/

concentration plot (Figure 6) showed no evidence of sigmoidal

kinetics, which coupled with the lack of inhibition on addition

of Hg suggests nanoparticle formation is not occurring during

catalysis.¹²A small amount of H₂BdNMe₂, δ 38 [d, J(BH) 123

Hz]¹⁴was observed during catalysis that followed a time/

concentration proﬁle consistent with an intermediate.

H₂BdNMe₂can arise from initial dehydrogenation of

H₃B·NHMe₂or, as we demonstrate later, also from the

dehydrocoupling reaction of III-Me. Small amounts (less than

5%) of other amine-boranes are also observed including

HB(NMe₂)₂⁵⁷and (BH₂)₂NMe₂(µ-H).⁵⁸Using a catalyst loading

of 2.5 mol % dehydrocoupling proceeded at a considerably

slower rate (80 h, TOF 0.5 h^-1, sealed system, see Supporting

Information) but still to completion.

1

inspection of the H NMR spectra (Figure 8) that show, in

addition to H₃B·NHMe₂and IV, two resonances that can be

attributed to the N-methyl groups of III-Me, 2.57 (d, 6H,

NHMe₂), 2.45 (s, 6H, NMe₂), and the NH proton δ 5.25 (br s,

1H, NH). These resonances follow the same time/concentration

proﬁle as that for the intermediate in the ¹¹B NMR spectrum.

These observations show that III-Me is formed, rather than the

cyclic trimer [H₂BNMe₂]₃, as an intermediate which then

undergoes Rh-mediated dehydrocoupling to form IV. We return

to this later.

Monitoring the sealed system during catalysis by NMR

spectroscopy also identiﬁed a number of metal containing

species, including 7 (ca. 20%). Other species currently elude

deﬁnitive identiﬁcation. Toward the end of catalysis only two

compounds are observed: 7 and another that is currently only

characterized by solution techniques, which becomes the

dominant species after catalysis is complete and the concentra-

tion of H₃B·NHMe₂is close to zero. Addition of further

H₃B·NHMe₂to these solutions restarts catalysis. ³¹P{¹H} NMR

spectroscopy for this new complex suggests a Rh(III) center

1

[J(RhP) 104 Hz], while H NMR data indicate 2 × Rh-H [δ

-16.43, 2H], 2 × Rh-H-B [δ -2.11, 2H] groups, the latter

(57) No¨th, H.; Vahrenkamp, H. Chem. Ber. 1966, 99, 1049–67.

(58) Keller, P. C. J. Am. Chem. Soc. 1969, 91, 1231.

(59) Campbell, G. W.; Johnson, L. J. Am. Chem. Soc. 1959, 81, 3800–

3801.

Another species that shows characteristic intermediate time/

concentration dependence is also observed by ¹¹B NMR

spectroscopy in both the open and the closed systems, δ 2.5 [t,

(60) Hahn, G. A.; Schaeffer, R. J. Am. Chem. Soc. 1964, 86, 1503–1504.

(61) Noth, H.; Thomas, S. Eur. J. Inorg. Chem. 1999, 1373–1379.

9

15446 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

Figure 7. ¹¹B NMR spectrum after 100 min of the catalytic dehydrocoupling of H₃B·NHMe₂using 1 showing the species present. See Figure 6 for more

details.

Figure 8. ¹H NMR spectra of the N-methyl region during the catalytic dehydrogenation of H₃B·NHMe₂using 1 (298 K, 2.5 mol %, 1,2-F₂C₆H₄, sealed

system) over time, showing H₃B·NHMe₂, IV, and III-Me.

also being a nicely resolved 1:1:1:1 quartet [J(BH) 79 Hz] that

shows a reduced coupling constant compared to closely related

free H₂BdNⁱPr₂[J(BH) 126 Hz], indicative of a Rh · · H-B

interaction.⁵⁵No NH signals were identiﬁed. These data ﬁt an

empirical formula [Rh(H)₂(PⁱBu₃)₂(η²-H₂BdNMe₂)]⁺8. We

discount assignment as a complex of the product, IV, or

intermediate, III-Me, as independent synthesis of these com-

plexes gives very different NMR spectroscopic signatures (vide

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15447

A R T I C L E S

Douglas et al.

Scheme 7. Comparison of the Proposed Structures of 8 and 9^a

^a[BAr^F]^-anions are not shown.

4

Scheme 8. Synthesis of 10^a

^a[BAr^F]^-anions are not shown.

4

Figure 9. DFT optimized structure of the cationic portion of 8. Bond

lengths in Å.

62

infra). In support of this assignment, addition of H₂BdNCy₂

to 1 in 1,2-F₂C₆H₄solution results in a complex with similar

NMR spectroscopic characteristics to 8 that is characterized by

NMR spectroscopy as [Rh(H)₂(PⁱBu₃)₂(η²-H₂BdNCy₂)]⁺9.

Despite repeated attempts we have not been able to obtain a

crystalline material of this complex. Complex 9 shows a slightly

broadened, integral 2-H signal at δ -14.51 due to the Rh

hydrides, and a quadrupolar broadened, integral 2-H signal at

δ -1.71. No N-H signal was observed. The ¹¹B NMR spectrum

of 9 shows a very broad resonance at δ 35, which although

essentially unshifted from free amino-borane is signiﬁcantly

broader than free ligand.⁶²We saw no evidence for free

amino-borane. Although, as far as we are aware, other

complexes of amino-boranes have yet to be reported, com-

plexes of three coordinate mesitylborane have recently been

described and show an end on, η²-coordination mode for the

borane.³⁵We tentatively also assigned an end-on coordination

Figure 10. Molecular structure the cationic portion of 10: ellipsoids

depicted at the 50% probably level. Only one of the independent molecules

in the asymmetric unit is shown; most H atoms omitted for clarity. Selected

bond lengths (Å) and angles (deg): Rh1-B1, 2.166(8); Rh1-P1, 2.243(2);

Rh1-P2, 2.248(2); B1-N1, 1.565(11); N1-B2, 1.588(14); B2-N2,

1.621(14); Rh1-H1A, 1.74(3); Rh1-H1B, 1.74(3); Rh1-H2, 2.54; Rh1-N2,

3.89(8);B1-H1A,1.30(3);B1-H1B,1.30(3);B1-H1C,1.09(6);P1-Rh1-P2,

99.19(6); P1-Rh1-B1, 130.3(3); P2-Rh1-B1, 128.2(3); Rh1-B1-N1,

134.3(7); B1-N1-B2, 116.0(7); N1-B2-N2, 114.4(8); Rh1-H1A-B1,

90(2) Rh1-H1B-B1, 90(2); H1A-Rh1-H1B, 72(2); Rh1-H2-N2, 152.

H2 in calculated position.

1

mode for the amino-borane in 9 as this ﬁts the observed H

NMR data and also would minimize steric interactions with the

trans-phosphine groups. With the less bulky H₂BdNMe₂in 8

a side-on coordination motif might be more accessible. DFT

calculations predict that in 8, whose optimized structure is shown

in Figure 9, the ligand H₂BdNMe₂does not coordinate side-

on but only end-on through two Rh-H-B bridges. When the

ⁱBu₃P ligands were modeled by the sterically smaller Me₃P

ligands, a structure with H₂BdNMe₂coordinated side-on to Rh

could be determined. However, even for this simpliﬁed model,

the end-on structure for H₂BdNMe₂is about 15 kcal/mol more

stable (see Supporting Information for more details). These

results strongly indicate and end-on structures for both 8 and 9

(Scheme 7).

Using the same synthetic route as for the previous complexes,

addition of III-Me to a 1,2-F₂C₆H₄solution of 2 gives immediate

and quantitative formation (by NMR spectroscopy) of [Rh-

(PⁱBu₃)₂(η²-H₃B·NMe₂BH₂·NHMe₂)][BAr^F] 10 (Scheme 8). In

4

our hands complex 10 is very air-sensitive in the solid and

solution state and consequently difﬁcult to isolate and purify.

Temperature-sensitive crystalline material was grown at -30

°C. The hydrogen atoms associated with the Rh · ·H-B interac-

tion were located in the ﬁnal Fourier difference map, which

show the borane to be η²-bound to the metal center in the solid-

state structure (Figure 10). DFT calculations support their

positioning (Figure 11), as does NMR spectroscopy.

2.3. Synthesis of σ-Complexes of H₃B·NMe₂BH₂·NHMe₂

and [H₂BNMe₂]₂. Given that III-Me (a postulated intermediate)

and the ﬁnal product of the reaction, IV, can both potentially

coordinate to a metal center, we have investigated the indepen-

dent synthesis of complexes of these two amine-boranes with

the [Rh(PⁱBu₃)₂]⁺fragment.

In the solid state, complex 10 crystallizes with two indepen-

dent molecules in the asymmetric unit, with similar structural

metrics for both (see Supporting Information). III-Me binds to

the Rh(I) metal center via two B-H-Rh 3c-2e bonds and has

similar structural metrics to those observed for 4 (and 12, vide

infra) and chelated η²-phosphine borane complexes of Rh(I).²⁹

(62) Euzenat, L.; Horhant, D.; Ribourdouille, Y.; Duriez, C.; Alcaraz, G.;

Vaultier, M. Chem. Commun. 2003, 2280–2281.

9

15448 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

H₃B · NMe₂BH₂· NHMe₂)(PⁱBu₃)₂][BAr^F] 11 (40%). Leaving

4

this mixture for a further 2 h resulted in a change in composition

to 11 (35%) and 7 (65%) together with the formation of IV.

The NMR spectra of 11 show signals due to Rh-H-B (δ

-0.75, 2H) and Rh-H [δ -18.60, 2H, apparent quartet (dt), J

20 Hz] and a single ³¹P environment [δ 21.5, J(RhP) 107 Hz]

with a chemical shift and coupling constant similar to those of

5 and 7. Deﬁnitive signals due to 11 could not be assigned in

the ¹¹B NMR spectrum. ESI-MS shows a strong molecular ion

at m/z ) 625.457 in full accord with the formulation as a

dihydride (Figure 12). Although the mechanism of dehydro-

coupling for III-Me thus appears to be less than straightforward,

most importantly these observations demonstrate that III-Me

and transition metal complexes thereof are plausible intermedi-

ates on the dehydrocoupling pathway, underscoring the previous

experimental observations by Manners as well as our observation

here that III-Me is an intermediate. The observation of 7,

however, suggests that B-N bond cleavage has occurred during

the dehydrocoupling process, and this perhaps points to a

complicated and possibly multipathway process that might

invoke intermolecular activation. Indeed, computational studies

on the dehydrogenation of H₃B·NH₃suggest that the related

linear dimer H₃N·BH₂NH₂·BH₃is formed that then proceeds

via production of 2 equiv of H₂BdNH₂and H₂, the former

which then reacts with H₃B·NH₃in an autocatalytic reaction.³⁹

The metal-catalyzed dehydrogenation was not investigated in

this study. For III-Me, separate computational studies suggest

an off-metal dehydrogenation mechanism involves a relatively

high barrier (29.9 kcal mol^-1) and production of 2 equiv of

Me₂BdNH₂alongside H₂.¹⁷

Under catalytic conditions in a sealed NMR tube 10 and 2

are effective, but not spectacular, catalysts, for the dehydro-

coupling of III-Me (5 mol %, 1,2-F₂C₆H₄, sealed system, 70 h,

TOF 0.3 h^-1) to form IV (Figure 13), a comparable rate to that

observed for the overall dehydrogenation of H₃B·NHMe₂(2.5

mol %) under the same conditions (Supporting Information).

These data suggest that dehydrocoupling of III-Me is the rate

determining processes in the overall dehydrogenation of

H₃B·NHMe₂. Approximately 1 equiv of H₂per III-Me is

evolved during this catalytic process (gas buret). H₂BdNMe₂

is also formed during catalysis, suggesting a mechanism that

involves B-N cleavage as one pathway, adding further support

for B-N bond cleavage in the dehydrocoupling of the III-Me

under stoichiometric conditions.

Figure 11. DFT optimized structure of the cationic portion of 10. Bond

lengths in Å. Bond angle in degrees. The Bu groups are not shown for

clarity.

i

The Rh(I) is pseudo-square planar coordinated, with a rather

short Rh · · ·B distance [2.166(8)/2.202(9) Å], similar to that

found in 4. There appears to be a close interaction between

N-H2 and the Rh [e.g., 2.54 Å, 152°] and although not linear,

it is still probably best described as a 3c-4e Rh· · ·H-N

hydrogen bond.⁶³This interaction in 10 could represent an early

stage in the process of H-transfer to the metalsa mechanistic

step in the dehydrocoupling. Interestingly there is a also gradual

shortening of the B-N bonds as they become closer to the metal

center [1.565(11)/1.546(12)-1.621(14)/1.593(12) Å], unlike in

parent III-Me where they are all essentially the same

[1.589(2)-1.600(2) Å].⁶¹In solution (1,2-F₂C₆H₄) at room

temperature the BH₃group is observed as a broad, integral 3 H

signal at δ -2.31, demonstrating rapid exchange between

terminal and bridging hydrides. The BH₂group is not observed,

presumably obscured by the aliphatic peaks. The BH₃resonances

are shifted upﬁeld from that of observed free ligand, consistent

with coordination of the metal. The NH proton is observed at

δ 4.31 [cf. free ligand δ 5.45]. The ¹¹B{¹H} NMR spectrum

shows two broad resonances at δ 26.0 and δ 3.9, the former

assigned to the BH₃group, and is shifted by 38.8 ppm downﬁeld

[viz. 2.5 (BH₂), -12.8 (BH₃) in III-Me]. Although the ¹¹B

resonances are too broad to observe multiplicity in the coupled

spectrum which would aid assignment, selective decoupling of

the δ 26.0 peak results in a sharpening of the δ -2.31 resonance

in the ¹H NMR spectrum, identifying it as the BH₃signal. The

³¹P{¹H} NMR spectrum shows a single resonance with a large

[J(RhP) 173 Hz] coupling constant, consistent with a Rh(I)

center. ESI-MS shows a parent ion at m/z ) 623.444 (calcd

623.441). Cooling a CH₂Cl₂solution of 2 to 200 K halts the

ﬂuxional process that exchanges the terminal B-H protons on

B1 so that an integral 2H, Rh-H-B, signal is now observed at

δ -5.93 in the ¹H NMR spectrum. The corresponding terminal

B-H resonance was not observed. The N-H bond is unshifted.

These NMR data are fully consistent with the solid-state

structure.

Addition of IV to 3 is followed by immediate H₂loss and

the isolation deep-purple [Rh(PⁱBu₃)₂{η²-(H₂BNMe₂)₂}][BAr^F]

4

12 (Scheme 9) as the only product. Alternatively 12 can be

prepared by addition of IV to 1 in 1,2-F₂C₆H₄. The solid-state

structure is shown in Figure 14 and shows a η²-coordination

motif for the cyclic borane, similar to that observed for 4 and

10. 12 crystallizes as two independent salts in the asymmetric

unit cell, but there is no signiﬁcant variation between the two

discrete sets of molecules. The BH hydrogen atoms associated

with B1 were located. The Rh · · ·B distances, [2.161(6)/2.140(7)

Å], are comparable with that in 4 and 10. There appears to be

no major change in bond lengths and angles in [H₂BNMe₂]₂

compared to the free ligand,¹³although B-N bond lengths

associated with B1 are slightly shorter than those with B2 or

those found in the free ligand by ∼0.05 Å. The sum of angles

around Rh1 come to 360°. DFT calculations are in close accord

with the experimentally determined structure (Figure 15).

Interestingly in 1,2-F₂C₆H₄solution complex 10 does not

undergo further dehydrocoupling to form the cyclic oligomer

IV, being stable for 24 h in solution. This shows that 10 does

not proceed to dehydrocouple to give IV via a simple intramo-

lecular route. In support of a more complex process, addition

of two equivalents of III-Me to 1 in 1,2-F₂C₆H₄solution results

in the immediate formation of a number of organometallic

products: 7 (10%), 10 (50%), and a new complex tentatively

identiﬁed by NMR spectroscopy and ESI-MS as [Rh(H)₂(η²-

In solution the [H₂BNMe₂]₂ligand is static on the metal

1

(63) Brammer, L. Dalton Trans. 2003, 3145–3157.

center, and two BH₂environments are observed in the H

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15449

A R T I C L E S

Douglas et al.

Figure 12. ESI-MS of the addition of 2 equiv of III-Me to 1. Inset expands the 618-631 region. Complex 3 is not observed in the reaction mixture by

NMR spectroscopy and arises from loss of amine-borane from 7, 10, and 11 under the conditions of ESI-MS.

Figure 13. Concentration plot for the dehydrocoupling of III-Me (initial

concentration ) 0.48 mol dm^-3) using 1 (298 K, 5 mol %, 1,2-F₂C₆H₄,

sealed system). 0 ) III-Me, 4 ) IV.

Figure 14. Molecular structure of the cationic portion of 12: ellipsoids

depicted at the 50% probably level. One of the independent molecules in

the asymmetric unit; minor disordered component and most H atoms omitted

for clarity. Selected bond lengths (Å) and angles (deg): Rh1-B1, 2.161(6);

Scheme 9. Synthesis of Complex 12^a

Rh1-P1, 2.2457(14); Rh1-P2, 2.253(2); B1-N1, 1.559(8); B1-N2,

1.547(8); B2-N1, 1.596(9); B2-N2, 1.605(9); Rh1-H1A 1.87(5); Rh1-H1B

1.69(5); B1-H1A 1.21(3); B1-H1B 1.21(3); B2-H2A 1.09(3); B2-H2B

1.09(3); P1-Rh1-P2, 98.31(6); P1-Rh1-B1, 130.9(2); P2-Rh1-B1,

130.8(2); N1-B1-N2, 94.6(4); N1-B2-N2, 91.0(5); B1-N2-B2, 86.5(4);

B1-N1-B2, 86.4(4); Rh1-H1A-B1, 87(3); Rh1-H1B-B1, 95(3);

H1A-Rh1-H1B, 68(2).

^a[BAr^F]^-anion is not shown.

4

NMR spectrum. One is interacting with the metal, δ -5.07

[quartet, J(BH) 89 Hz], while the other is not, δ 2.90 [vbr].

The reduced coupling constant in the higher-ﬁeld signal

compared to free ligand [cf. 112 Hz] is consistent with

σ-coordination of the B-H bond with the metal center.^21,55

The ¹¹B NMR spectrum also shows two environments (δ 31.1,

1

5.4), and selective H{¹¹B} experiments show that the one at

δ 31.1 is due to the η²borane, is shifted 25.7 ppm downﬁeld

from free ligand, and also shows a reduced coupling constant,

[J(HB) 89 Hz],¹³while the other is unshifted, δ 5.4 [vbr], and

assigned to B2. Complex 12 comes from addition of IV to 3

and H₂loss, this presumably occurring due to steric pressure

between the trans phosphine ligands and the borane in the

Rh(III)-dihydride, as found for 5 and 7. With no conformational

ﬂexibility available to [H₂BNMe₂]₂the resulting dihydride

species is clearly too unstable to isolate. Addition of H₂to a

1,2-F₂C₆H₄solution of 12 results in an equilibrium mixture of

12 and 3/IV, consistent with this idea. Removal of the H₂

atmosphere re-established 12.

Figure 15. DFT optimized structure of the cationic portion of 12. Bond

lengths in Å. Bond angle in degrees. The Bu groups are not shown for

clarity.

i

3. Computational Study

With intermediates and model complexes in hand we now turn

to computational techniques to help elucidate possible mechanistic

9

15450 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

Scheme 10. Proposed Reaction Mechanism for the Dehydrogenation of H₃B·NMe₂H Using 1

pathways for the dehydrocoupling reaction using catalyst 1. This

study has been limited to the ﬁrst proposed dehydrogenation step

of H₃B·NMe₂H to afford H₂BdNMe₂given the potential complex-

ity (e.g., linear dimer formation, B-N bond cleavage reactions) of

the subsequent dehydrogenation steps to afford IV. Although in

1,2-F₂C₆H₄solution 2 is the organometallic precatalyst, the study

starts from 12-electron 1 (Figure 1) as the ﬁrst step in any

mechanism is likely to be dissociation of the bound arene solvent

from this 18-electron complex. Schemes 10 and 11 shows these

sequential processes and give a free-energy proﬁle for the intrinsic

reaction coordinate. Figure 16 shows the structure of the selected

intermediates. Supporting Information give geometries for all the

calculated structures reported here.

The initial dehydrogenation of H₃B·NMe₂H has been investi-

gated by following possible pathways that involve NH/BH or BH/

NH activation. Calculations indicate that there are two competitive

pathways that begin with the formation of the σ-complex 6 and

then bifurcate by either N-H or B-H bond activation. Rh(PⁱBu₃)₂

reacts with H₃B·NMe₂H to instantly form the adduct 6, which is

bound by -4.4 kcal/mol in solvent corrected free-energy. H-transfer

to the Rh center then occurs either through B-H bond activation,

transition state TS_6,13, or through N-H bond activation, transition

state TS_6,14, to form intermediate 13 or 14, respectively. Although

14 is 12.2 kcal/mol lower than 13 in free energy, the barrier to NH

transfer (TS_6,14) is much higher than BH oxidative addition (TS_6,13):

24.3 vs 10.4 kcal/mol, respectively. The second H transfer to Rh

occurs either from the nitrogen in 13 or from the boron in 14,

proceeding through transition states TS_13,15or TS_14,15respectively

to afford the Rh(I)/dihydrogen intermediate 15, the largest barrier

in this instance being TS_13,15(14.9 kcal/mol). In this catalytic

dehydrocoupling cycle of H₃B·NMe₂H to afford metal bound

H₂BdNMe₂the rate determining steps are two N-H bond

activation transition states TS_6,14and TS_13,15, which are at essentially

the same overall solvent corrected free energy, thus making both

pathways competitive with one another. An alternate higher energy

BH/NH bond activation pathway is discussed later.

The second part of the mechanistic pathway follows from 15

and requires H₂loss from the metal in 15. This has been studied

by following two possible pathways. For the ﬁrst, dihydrogen in

15 can be released through transition state TS_15,16with a barrier of

13.2 kcal mol^-1. Simultaneous to this release, the N atom is attracted

by Rh to form intermediate 16, with the H₂BdNMe₂ligand bonding

with the Rh(I) center through a three-center, two-electron Rh-H-B

bridging interaction and a weak Rh-N bond (2.281 Å). Catalyst 1

is then regenerated by release of H₂BdNMe₂from 16. Alternatively,

instead of H₂release from 15, H₂BdNMe₂can be released through

transition state TS_15,17, which is only 0.7 kcal/mol higher than TS_15,16

in free energy to form intermediate 17, which contains a dihydrogen

ligand. H₂loss from this regenerates 1. This close match in

calculated energy mean that pathways are equally accessible in

practice. The TS_16,1and TS_8,3(vide infra) could not be determined.

These ligand dissociation transition states are often problematic to

determine as the location of the transition state along the reaction

coordinate is poorly deﬁned. However, based on the somewhat

similar TS_15,17, these barriers are estimated to be about 14 kcal/

mol.

An alternative, productive, dehydrocoupling pathway involves

initial oxidative addition of the B-H bond, to ultimately give an

′′

alternate isomer to 13, 13 (TS_6,13′19.3 kcal/mol), in which the

phosphine ligands are orientated signiﬁcantly wider apart and more

trans-like than in 13, in which they are cis. This leads to an

intermediate almost isoenergetic with 13 but one that can now

proceed in a very different pathway to 13. N-H transfer from 13′′

leads to 8 with a reasonable barrier of 18.0 kcal/mol (TS_13′′,8) which

has an end-on coordination mode of H₂BdNMe₂and lies in a deep

potential-energy well. This then proceeds via loss of H₂BdNMe₂

to give 3 that then moves through 18 to give 17 via reductive bond

formation, followed by H₂loss to ﬁnally regenerate 1. The

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15451

A R T I C L E S

Douglas et al.

Scheme 11. Solvent Effect Corrected Free Energy Reaction Proﬁle^a

^aRefer to Scheme 10 and Supporting Information for the structures of the intermediates. The ﬁnal energies of 1 differ as the pathway through 7 does not

involve dehydrogenation of H₃B·NMe₂H.

Figure 16. Optimized geometric structures of the cationic portions of intermediates 6, 7, 13, 14, 15, 16, 17, and 19. Bond lengths in Å. The ⁱBu groups are

not shown for clarity.

transformation of 3 to 18 is an isomerization (TS_18,320.6 kcal/

mol), with 3 16.7 kcal/mol more favorable than 18; as when the

geometries are compared, 18 has a weaker Rh-H bond trans to

the Rh-P bond and a more acute P-Rh-P angle that results in

larger steric repulsions between the phosphines. The highest barrier

to this overall process is TS_18,3(20.6 kcal/mol) followed closely

by a step early in the reaction coordinate TS_6,13′19.3 kcal/mol.

Both of these are large but accessible. That 8 sits in such a deep

well but also has a relatively high barrier to access suggests

that it will pool slowly during catalysis and be a major,

thermodynamically favored product at the end. This is observed

experimentally. That 8 can also be formed from 3 mirrors the

synthetic route to form 9.

largest barrier to this is TS_13′′,19, 14.6 kcal/mol, which leads to 7

that sits in a relatively deep well (although not as deep as 8). These

energies ﬁt with the experimental observation that 7 is observed

during catalysis but will eventually be disfavored with respect to 8

at low substrate concentrations and long reaction times. The

pathway 1 to 7 to 1 via 19 is not a productive dehydrogenation

step, as overall it is simply addition and loss of H₃B·NMe₂H and

H₂. However, complex 7 does lie on a productive pathway as

deﬁned by 7-19-13′′-8-3-7, but as 7-19 is particularly high

energy (26.9 kcal/mol), this is not a kinetically signiﬁcant process

on the basis of these calculations.

Overall these calculations suggest that initial activation of

H₃B·NMe₂H to form metal-bound H₂BdNMe₂occurs either via

BH oxidative addition/NH transfer or NH activation/BH transfer.

This is followed by sequential H₂loss/H₂BdNMe₂loss or sequential

H₂BdNMe₂loss/H₂loss. These dissociation pathways have very

Complex 7 can be accessed via 13′′ by H₂addition to give 19

which has a dihydrogen and base-stabilized borylene ligand.

Concerted H₂cleavage and B-H bond formation leads to 7. The

9

15452 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

A R T I C L E S

similar calculated barriers, which are also close to the NH activation

barrier, suggesting that on the bench these competing pathways

are all energetically accessible. The pooling of 8 at the end of the

reaction, and 7 being observed during catalysis, sits with the relative

calculated thermodynamic stabilities of these complexes and the

reversibility of their formation, especially 7. The stepwise pathway

for dehydrogenation is similar (although the order different) to that

proposed for titanocene systems¹⁷but different from the concerted

mechanism using Ir-pincer complexes that invokes a Ir(III)/Ir(V)

process.¹⁸

performed by Elemental Microanalysis Ltd. H₃B · NMe₃and

H₃B·NHMe₂were purchased from Aldrich and sublimed twice

before use (5 × 10^-2Torr, 298 K). pcq ) partially collapsed quartet.

All ¹¹B NMR spectra show a singlet at ca. δ -6.0 due to the

[BAr^F]^-anion.

4

Chemical Shifts of Previously Reported Amine-Boranes

1

in 1,2-F₂C₆H₄(298 K). H₃B·NMe₃. H NMR (500 MHz, 1,2-

F₂C₆H₄): δ 2.27 [q, 3 H, BH₃, J(BH) 98], 2.63 (s, 9 H, NMe).

¹¹B{¹H} NMR (160 MHz, 1,2-F₂C₆H₄, 298 K): δ -7.3 [q, J(HB)

98].

1

H₃B·NHMe₂. H NMR (500 MHz, 1,2-F₂C₆H₄) δ 3.87 (s, 1H,

4. Conclusions

Br, NH), 2.55 [d, 6 H, NMe, J(HH) 5], 2.13 [pcq, 3 H, BH₃, J(BH)

96]. ¹¹B NMR (160 MHz, 1,2-F₂C₆H₄, 298 K): δ-12.8 [q, J(HB)

96].

We have reported a combined experimental/quantum chemical

investigation of the dehydrocoupling reaction of H₃B·NMe₂H

to ultimately give the cyclic dimer [H₂BNMe₂]₂. Intermediates

and model complexes have been isolated, including examples

of amine-borane σ-complexes of Rh(I) and Rh(III), and for

the ﬁrst time the isolation of a transition metal complex of the

oligomeric dimer H₃B·NMe₂BH₂·NHMe₂, which has been

shown to be an intermediate in the dehydrocoupling reaction.

DFT calculations show that the ﬁrst proposed dehydrogenation

step, to give H₂BdNMe₂, proceeds via sequential BH/NH

activation followed by two possible routes that invoke sequential

H₂loss/H₂BdNMe₂loss or H₂BdNMe₂loss/H₂loss. For the

second dehydrogenation step that ultimately affords [H₂BNMe₂]₂

the experimental observation that the isolated and crystal-

lographically characterized complex [Rh(PⁱBu₃)₂(η²-H₃B ·

H₃B·NMe₂BH₂·NHMe₂.¹³¹H NMR (500 MHz, 1,2-F₂C₆H₄):

δ 5.45 (s, 1H, NH), 2.55 [d, 6 H, NMeH, J(HH) 6], 2.45 [s, 6 H,

NMe], 2.16 [pcq, 2 H, BH₂, J(BH) 107], 2.04 [pcq, 3 H, BH₃, J(BH)

94]. ¹¹B NMR (160 MHz, 1,2-F₂C₆H₄, 298 K): δ 2.5 [t, BH₂, J(HB)

107], -12.8 [q, BH₃, J(HB) 94].

[H₂BNMe₂]₂.¹³¹H NMR (500 MHz, 1,2-F₂C₆H₄): δ 2.48 (s, 12

H, NMe), 2.97 [q, 4 H, BH₂, J(BH) 112]. ¹¹B{¹H} NMR (160 MHz,

1,2-F₂C₆H₄, 298 K): δ 5.5 [t, J(HB) 112].

[Rh(η⁶-1,2-F₂C₆H₄)(PⁱBu₃)₂][BAr^F] (2). [Rh(PⁱBu₃)₂][BAr^F]

4

(30 mg, 2.19 × 10^-2mmol) was dissolved in 1,2-F₂C₆H₄(2 cm³)

to give [Rh(η⁶-1,2-F₂C₆H₄)(PⁱBu₃)₂][BAr^F] as an orange solution.

4

Single crystals of 2 were obtained by diffusion of pentane into the

solution of 2 (22 mg, 68% yield). 2 was characterized by NMR

spectroscopy and ESI-MS in 1,2-F₂C₆H₄solution. Dissolving

crystals of 2 in C₂H₄Cl₂gave [Rh(C₂H₄Cl₂)(PⁱBu₃)₂][BAr^F] and

4

NMe₂BH₂·NHMe₂)][BAr^F] is stable in the absence of added

free 1,2-F₂C₆H₄.

4

¹H NMR (500 MHz, 1,2-F₂C₆H₄): δ 8.34 (s, 8H, BAr^F), 7.69

amine-borane suggests that a simple intramolecular route for

this dehydrocoupling process is not operating for this linear

dimer. Although the catalysts reported here demonstrate rather

poor turnover frequencies, it is exactly because of this that

intermediates can be observed and isolated under stoichiometric

conditions. Future studies will address the inﬂuence that the

phosphine and identity of the metal has on both the rate of

catalysis and the intermediates observed in the anticipation of

providing yet more information on this important transformation.

4

i

(s, 4H, BAr^F), 2.02 (m, 6H, Bu CH), 1.65 (m, 12H, Bu CH₂),

4

1.13 [d, J(HH) 6, 36H, CH₃]. ³¹P{¹H} NMR (202 MHz, 1,2-

F₂C₆H₄): δ 25.91 [d, J(RhP) 210]. ESI-MS (CH₂Cl₂, 100 °C)

positive ion {Rh(PⁱBu₃)₂}⁺: m/z ) 507.28 [M]⁺(calcd 507.28).

[Rh(η²-H₃B · NMe₃)(PⁱBu₃)₂][BAr^F] (4). Addition of H₃B ·

4

NMe₃(5 mg, 6.8 × 10^-2mmol, 2 equiv) to a solution of 1 (45 mg,

3.2 × 10^-2mmol) in 1,2-F₂C₆H₄(1 cm³) resulted in a rapid color

change from orange to deep purple; diffusion of pentane into the

solution gave 4 as dark purple crystals (40 mg, 87%).

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.33 (s, 8H, BAr^F),

4

7.69 (s, 4H, BAr^F), 2.96 (s, 9H, N-CH₃), 2.17 (br, 6H, CH), 1.79

5. Experimental Section

4

(br, 12H, CH₂), 1.24 [d, J(HH) 6, 36H, CH₃], -2.12 (br, 3H, BH₃).

³¹P{¹H} NMR (202 MHz, 1,2-F₂C₆H₄, 298 K): δ 35.94 [d, J(RhP)

175]. ¹¹B{¹H} NMR (160 MHz, 1,2-F₂C₆H₄, 298 K): δ 23.12 (br).

Selected ¹³C{¹H} NMR (126 MHz, CD₂Cl₂, 298 K): 53.37 (s,

N-CH₃), 39.45 (apparent t, J 13, P-CH₂), 26.85 (m, CH), 25.60

All manipulations, unless otherwise stated, were performed under

an atmosphere of argon, using standard Schlenk and glovebox

techniques. Glassware was oven-dried at 130 °C overnight and

ﬂamed under vacuum prior to use. Pentane was dried using a Grubbs

type solvent puriﬁcation system (MBraun SPS-800) and degassed

by successive freeze-pump-thaw cycles.⁶⁴CD₂Cl₂and 1,2-F₂C₆H₄

were distilled under vacuum from CaH₂and stored over 3 Å

molecular sieves. Na[BAr^F],⁶⁵[Rh(nbd)(PⁱBu₃)₂][BAr^F],⁴⁶

1

(s, CH₃). Selected H NMR (500 MHz, CD₂Cl₂, 190 K): δ -5.55

(vbr, 2H, η-BH₂), the other BH signal was not observed, presumably

as it was broad and obscured by the aliphatic resonances. ESI-MS

(1,2-F₂C₆H₄, 100 °C, 4.5 kV): m/z ) 580.388 (calcd for

[C₂₇H₆₆BNP₂Rh]⁺580.382). Microanalysis (C₅₉H₇₈B₂F₂₄NP₂Rh,

1443.71 g mol^-1): requires, C, 49.09; H, 5.45; N, 0.97.; found, C,

48.92; H, 5.25; N, 1.02.

4

[Rh(PⁱBu₃)₂][BAr^F] (1),⁴⁶[Rh(H)₂(PⁱBu₃)₂][BAr^F] (4),⁴⁶

4

[H₂BNMe₂]₂,¹³H₂BdNCy₂,⁶²and H₃B·NMe₂BH₂·NHMe₂¹³were

prepared by published literature methods or variations thereof. NMR

spectra were recorded on Varian 500 MHz and Bruker 500 MHz

spectrometers. Diﬂuorobenzene (center of downﬁeld multiplet) was

[Rh(H)₂(η²-H₃B·NMe₃)(PⁱBu₃)₂][BAr^F] (5). Addition of H₂(4

4

atm, 298 K/77 K ) 3.8) to a solution of [Rh(H₃B·NMe₃)(PⁱBu₃)₂]

used as reference for ¹H NMR spectra (1,2-F₂C₆H₄: δ ) 7.07). ³¹

P

[BAr^F] (10 mg, 6.9 × 10^-3mmol) in 1,2-F₂C₆H₄(0.3 cm³) resulted

4

NMR spectra were referenced against 85% H₃PO₄(external). ¹¹B

NMR spectra were referenced against BF₃·OEt₂(external). Chemi-

cal shifts are quoted in ppm and coupling constants in Hz. ESI-

MS were recorded on a Bruker MicroOTOF-Q instrument.⁶⁶The

correct isotope patterns were observed for all species reported, as

modeled by the Bruker MicroOTOF software. Microanalyses were

in a rapid color change from deep purple to pale yellow. 5 was

characterized in situ by NMR spectroscopy. Attempts to obtain

crystalline material of 6 were unsuccessful due to facile H₂loss to

give 4. ¹H and ³¹P{¹H} NMR spectra showed no change on cooling

to 190 K. ESI-MS showed only a peak corresponding to 4 [m/z )

580.38]. 5 could also be prepared by addition of TMAB to a solution

of 3 in 1,2-F₂C₆H₂.

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.33 (s, 8H, BAr^F),

(64) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;

Timmers, F. J. Organometallics 1996, 15, 1518–1520.

(65) Buschmann, W. E.; Miller, J. S.; Bowman-James, K.; Miller, C. N.

Inorg. Synth. 2002, 33, 83–85.

(66) Lubben, A. T.; McIndoe, J. S.; Weller, A. S. Organometallics 2008,

27, 3303–3306.

4

7.69 (s, 4H, BAr^F), 2.87 (s. 9H, N-CH₃), 2.08 (br, 6H, CH), 1.83

4

(br, 12H, CH₂), 1.14 [d, J(HH) 6, 36H, CH₃], -0.47 (br, 3H, BH₃),

-18.81 [dt, J(RhH) 23, J(PH) 17, 2H, Rh-H]. ³¹P{¹H} NMR (202

MHz, 1,2-F₂C₆H₄, 298 K): δ 21.10 [d, J(RhP) 106]. ¹¹B{¹H} NMR

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15453

A R T I C L E S

Douglas et al.

(160 MHz, 1,2-F₂C₆H₄, 298 K): δ 5.07 (br). Selected ¹³C{¹H} NMR

(126 MHz, CD₂Cl₂, 298 K): 56.52 (s, N-CH₃), 39.05 (m, P-CH₂),

26.61 (m, CH), 25.37 (s, CH₃).

(500.13 MHz, CD₂Cl₂, 190 K): δ -5.93 (br s, 2H, σ-BH), the

remaining BH and BH₂signals were not observed, presumably

obscured under aliphatic resonances. ESI MS (1,2-F₂C₆H₄, 100 °C,

4.5 kV): m/z ) 623.444 (calc for [RhC₂H₇₂P₂B₂N₂]⁺623.441).

[Rh(η²-H₃B · NHMe₂)(PⁱBu₃)₂][BAr^F] (6). Addition of

4

H₃B·NHMe₂(0.196 M in 1,2-F₂C₆H₄, 37 µL, 7.3 × 10^-3mmol, 1

equiv) to a solution of 1 (10 mg, 7.3 × 10^-3mmol) in 1,2-F₂C₆H₄

(1 cm³) resulted in a rapid color change from orange to deep purple.

6 was characterized in situ by NMR spectroscopy and ESI-MS. 6

rapidly (10 min) decomposes in solution to give 6 and other

[Rh(H)₂(η²-H₃B·NMe₂BH₂·NHMe₃)(PⁱBu₃)₂][BAr^F] (11). Ad-

4

dition of H₃B·NMe₂BH₂·NHMe₃(0.033 M in 1,2-F₂C₆H₄, 0.22

cm³, 7.3 × 10^-3mmol, 1 equiv) to a 1,2-F₂C₆H₄(0.5 cm³) solution

of [Rh(PⁱBu₃)₂][BAr^F] (10 mg, 7.29 × 10^-3mmol) resulted in a

4

rapid color change from orange to purple. The solution was im-

mediately characterized in situ by NMR spectroscopy and shown to

unidentiﬁed products. Addition of more than

1 equiv of

H₃B·NHMe₂results in much shorter lived 6 and the observation

contain a mixture of [Rh(η²-H₃B·NMe₂BH₂·NHMe₃)(PⁱBu₃)₂][BAr^F]

4

of dimeric [H₂BNMe₂]₂.

10 (50%), [Rh(H)₂(η²-H₃B·NMe₂BH₂·NHMe₃)(PⁱBu₃)₂][BAr^F] 11

4

(40%) and [Rh(H)₂(η²-H₃B·NHMe₂)(PⁱBu₃)₂][BAr^F] (10%) 7.

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.33 (s, 8H, BAr^F),

4

7.69 (s, 4H, BAr^F), 4.67 (br, 1H, NH), 2.88 (s, 6H, N-CH₃), 2.07

Leaving for 2 h resulted in observation of 7 (65%) and 11 (35%)

4

by NMR spectroscopy. [H₂BNMe₂]₂is also observed to be formed.

(br, 6H, CH), 1.79 (br, 12H, CH₂), 1.19 (m, 36H, CH₃), -2.13 (br,

3H, BH₃). ³¹P{¹H} NMR (202 MHz, 1,2-F₂C₆H₄, 298 K): δ 35.9

[d, J(RhP) 174]. ¹¹B{¹H} NMR (160 MHz, 1,2-F₂C₆H₄, 298 K): δ

19.31 (br). ESI-MS (1,2-F₂C₆H₄, 100 °C, 4.5 kV): m/z ) 566.366

(calcd for [C₂₆H₆₄BNP₂Rh]⁺566.369).

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.33 (s, 8H, BAr^F),

4

7.69 (s, 4H, BAr^F), 4.24 (br s, 1H, NH), 2.78 (m, 6H, N-CH₃),

4

2.58 (s, 6H, N-CH₃), 2.07 (m, 6H, CH), 1.81 (m, 12H, CH₂), 1.13

[d, J(HH) 6.7, 36H, CH₃], -0.75 (br s, 3H, BH₃), -18.60 [apparent

quartet, J 20, 2H, Rh-H]. BH₂signal could not be assigned as it

is probably obscured by the aliphatic resonances. ³¹P{¹H} NMR

(121.51 MHz, 1,2-F₂C₆H₄, 298 K): δ 21.50 [d, J(RhP) 107 Hz].

¹¹B{¹H} NMR (160.40 MHz, 1,2-F₂C₆H₄, 298 K): could not be

assigned, too broad. ESI MS (1,2-F₂C₆H₄, 100 °C, 4.5 kV): m/z )

625.457 (calcd for [RhC₂H₇₂P₂B₂N₂]⁺625.4). Also observed are

10 [m/z ) 623.444] and 7 [m/z ) 580.38] in approximately the

same ratio as in the solution NMR spectroscopy experiment.

[Rh(H)₂(η²-H₃B · NHMe₂)(PⁱBu₃)₂][BAr^F] (7). Addition of

4

H₃B·NHMe₂(0.406 M in 1,2-F₂C₆H₄, 36 µL, 1.4 × 10^-2mmol, 2

equiv) to a solution of 1 (10 mg, 7.3 × 10^-3mmol) in 1,2-F₂C₆H₄

(0.3 cm³) gave [Rh(H)₂(PⁱBu₃)₂(H₃B·NHMe₂)][BAr^F] (7) as a pale

4

yellow solution which was characterized in situ by NMR spectros-

copy and ESI-MS. Concomitant with the formation of 7,

[H₂BNMe₂]₂was also observed. A single crystal suitable for X-ray

diffraction studies was obtained by diffusion of pentane into a 1,2-

F₂C₆H₄solution at -35 °C.

[Rh(PⁱBu₃)₂{η²-(H₂BNMe₂)₂}][BAr^F] (12). Addition of

4

[H₂BNMe₂]₂(5 mg, 4.4 × 10^-2mmol, 3 equiv) to a solution of 1

(20 mg, 1.5 × 10^-2mmol) in 1,2-F₂C₆H₄(1 cm³) resulted in a

rapid color change from orange to deep purple. NMR spectroscopy

showed that the reaction was quantitative. Diffusion of pentane into

the solution gave 12 as deep purple crystals (11 mg, 51%). Addition

of H₂(4 atm, 298 K/77 K ) 3.8) to a solution of 12 in 1,2-F₂C₆H₄

gave a mixture of 3, 12, and [H₂BNMe₂]₂. Removal of the H₂

atmosphere re-established complex 12.

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.43 (s, 8H, BAr^F),

4

7.79 (s, 4H, BAr^F), 3.87 (s, 1H, NH), 2.97 (s, 6H, N-CH₃), 2.17

4

(br, 6H, CH), 1.91 (br, 12H, CH₂), 1.24 [d, J(HH) 6, 36H, CH₃],

-0.77 (br, 3H, BH₃), -17.42 [dt, J(RhH) 19.5, J(PH) 17.3, 2H,

Rh-H]. ³¹P{¹H} NMR (202 MHz, 1,2-F₂C₆H₄, 298 K): δ 22.26

[d, J(RhP) 105]. ¹¹B{¹H} NMR (160 MHz, 1,2-F₂C₆H₄, 298 K): δ

2.23 (br). Selected ¹H NMR (500 MHz, CD₂Cl₂, 190 K): δ -3.15

(vbr, 2H, η-BH₂), the other BH signal was not observed, presumably

as it was broad and obscured by the aliphatic resonances. ESI-MS

(1,2-F₂C₆H₄, 100 °C, 4.5 kV): m/z ) 568.385 (calcd for

[C₂₆H₆₆BNP₂Rh]⁺568.382).

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.33 (s, 8H, BAr^F),

4

7.69 (s, 4H, BAr^F), 2.90 (vbr, 2H, BH₂), 2.69 (s, 12H, N-CH₃),

4

2.18 (m, 6H, CH), 1.84 (m, 12H, CH₂), 1.27 [d, J(HH) 6, 36H,

CH₃], -5.07 (q, J(BH) 89, 2H, Rh-H₂B). BH assignments based

on ¹H{¹¹B-selective} experiments.¹³³¹P{¹H} NMR (202 MHz, 1,2-

F₂C₆H₄, 298 K): δ 33.38 [d, J(RhP) 170]. ¹¹B NMR (160 MHz,

1,2-F₂C₆H₄, 298 K): δ 31.14 [t, J(BH) 88, 1B, Rh-H₂B], 5.36 (vbr,

1B, BH₂). ESI-MS (1,2-F₂C₆H₄, 100 °C, 4.5 kV): m/z ) 621.419

(calcd for [C₂₈H₇₀B₂N₂P₂Rh]⁺621.426). Microanalysis

(C₆₀H₈₂B₃F₂₄N₂P₂Rh): requires, C, 48.54; H, 5.57; N, 1.89; found,

C, 48.43; H, 5.62; N, 1.79.

[Rh(H)₂(H₂BdNCy₂)(PⁱBu₃)₂][BAr^F] (9). Addition of H₂(4

4

atm, 298 K/77 K ) 3.8) to a solution of [Rh(nbd)(PⁱBu₃)₂][BAr^F]

4

(10 mg, 6.8 × 10^-3mmol) in 1,2-F₂C₆H₄(0.3 mL) gave 3 as a

pale yellow solution. H₂BdNCy₂(0.25 M in 1,2-F₂C₆H₄, 28 µL,

7.0 × 10^-3mmol, 1 equiv) was added and the resulting solution

characterized in situ by NMR spectroscopy and ESI-MS. Attempts

to isolate crystalline material of 9 suitable for microanalysis or X-ray

crystallography have failed to afford suitable material for analysis.

Reactions of [Rh(D)₂(PⁱBu₃)₂][BAr^F] with H₃BN ·HMe₂

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ 8.83 (s, 8H, BAr^F),

4

and H₃B·NMe₃. [Rh(nbd)(PⁱBu₃)₂][BAr^F] (10 mg, 6.84 × 10^-3

7.69 (s, 4H, BAr^F), 3.22 (m, 2H, N-CH), 2.50-0.79 (m, 38H,

4

mmol) was placed under D₂(4 atm) to give [Rh(D)₂(PⁱBu₃)₂][BAr^F]

CH₂/CH), 1.15 [d, J(HH) 6, 36H, CH₃], -1.71 (br, 2H, BH₂),

-14.51 (br, 2H, Rh-H). ³¹P{¹H} NMR (202 MHz, 1,2-F₂C₆H₄,

298 K): δ 22.36 [d, J(RhP) 99]. ¹¹B NMR (160 MHz, 1,2-F₂C₆H₄,

298 K): δ 35 (vbr). ESI-MS (1,2-F₂C₆H₄, 100 °C, 4.5 kV): m/z )

702.485 (calcd for [C₃₆H₈₀BNP₂Rh]⁺702.491).

4

as a pale yellow solution which was characterized in situ by NMR

spectroscopy. Excess H₃B·NMe₃(1 mg) was added to the solution

of [Rh(D)₂(PⁱBu₃)₂][BAr^F] and monitored in situ by ¹H NMR

4

spectroscopy. Initially a broad signal at approximately δ -0.6

consistent with coordinated BH₃was observed and no high ﬁeld

hydride signals. Over 10 min the BH₃signals reduce in intensity

with the concomitant appearance of high ﬁeld hydride signals at

approximately δ -18, H₂at δ 4.60, and HD at δ 4.56 J(HD) 42

Hz. The coordination of D₃B·NMe₃was conﬁrmed by ESI-MS.

ESI-MS calcd for [RhP₂C₂₇H₆₃D₃P₂B₁N₁]⁺583.401, obsvd; 583.406,

calcd.

[Rh(H₃B·NMe₂BH₂·NHMe₂)(PⁱBu₃)₂][BAr^F] (10). Addition

of H₃B·NMe₂BH₂·NHMe₃(2.5 mg, 2.19 × 10^-2⁴mmol) to a 1,2-

F₂C₆H₄(1 cm³) solution of 1 (30 mg, 2.19 × 10^-2mmol) resulted

in a rapid color change from orange to deep purple. In situ NMR

spectroscopy demonstrated complete conversion to 10. Diffusion

of pentane into the 1,2-F₂C₆H₄solution at -35 °C gave

[Rh(H₃B·NMe₂BH₂·NHMe₃)(PⁱBu₃)₂][BAr^F] as dark red crystals

4

Excess H₃B·NHMe₂(1 mg) was added to the solution of

(25 mg, 77%).

1

[Rh(D)₂(PⁱBu₃)₂][BAr^F] and monitored in situ by H NMR spec-

¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K) δ 8.33 (s, 8H, BAr^F),

4

7.69 (s, 4H, BAr^F), 4.31 (br s, 1H, NH), 2.70 (d, 6H, N-CH₃),

troscopy. A broad signal at approximately δ -0.6 consistent with

coordinated BH₃, hydride signals at approximately δ -18, H₂at δ

4.60, and HD at δ 4.56 [J(HD) 42 Hz] were all observed.

Catalytic Dehydrocoupling of H₃B·NHMe₂with 1 in a

Sealed NMR Tube. In a typical experiment H₃B·NHMe₂(8.6 mg,

1.5 × 10^-1mmol, 20 equiv) and 1 (10 mg, 7.3 × 10^-3mmol)

4

2.64 (s, 6H, N-CH₃), 2.19 (m, 6H, CH), 1.77 (m, 12H, CH₂), 1.23

[d, J(HH) 6.7, 36H, CH₃], -2.31 (br s, 3H, BH₃). The BH₂signal

could not be assigned. ³¹P{¹H} NMR (121.51 MHz, 1,2-F₂C₆H₄,

298 K): δ 37.46 [d, J(RhP) 173 Hz]. ¹¹B{¹H} NMR (160.40 MHz,

1,2-F₂C₆H₄, 298 K): δ 26.04 (br s), 3.89 (v br s). Selected ¹H NMR

9

15454 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Amine-Borane σ-Complexes of Rhodium

Table 2. Crystallographic Data

A R T I C L E S

2^a

4

7

10

12

formula

M

C₆₂H₇₀BF₂₆P₂Rh

1484.84

C₅₉H₇₈B₂F₂₄NP₂Rh

1443.69

C₅₈H₇₈B₂F₂₄NP₂Rh

1431.68

C₆₀H₈₄B₃F₂₄N₂P₂Rh

1486.57

C₆₀H₈₂B₃F₂₄NP₂Rh

1484.56

cryst syst

space group

a [Å]

b [Å]

c [Å]

monoclinic

P2₁/n

monoclinic

P2₁/c

monoclinic

C2/c

orthorhombic

Pna2₁

50.6131(4)

19.6787(2)

14.39940(10)

monoclinic

P2₁/n

21.7964(2)

13.1712(2)

25.5387(3)

113.7594(5)

6710.37(14)

4

1.470

0.412

5.10 e θ e 26.37

13607

12.97140(10)

20.0178(2)

26.8437(2)

93.4556(1)

6957.53(10)

4

1.378

0.391

5.14 e θ e 25.03

23229

21.0042(3)

18.2723(3)

37.9197(7)

102.2898(6)

14219.9(4)

8

1.337

0.382

5.10 e θ e 25.03

20424

14.2225(2)

50.9939(7)

19.5840(3)

91.5253(7)

14198.5(4)

8 (Z′ ) 2)

1.389

0.386

5.12 e θ e 25.03

22815

ꢀ [deg]

V [Å³]

Z

14341.8(2)

8 (Z′ ) 2)

1.377

0.382

5.14 e θ e 25.03

19425

density [g cm^-3

]

µ (mm^-1

)

θ range [deg]

reﬂns collected

R_int

0.0540

0.0175

0.0638

0.0228

0.075

no. of data/restr/param

R₁, wR₂[I > 2σ(I)]

GoF

13607/1596/1207

0.0906, 0.2499

1.023

12157/111/911

0.0542, 0.1387

1.020

12085/13/863

0.0862, 0.1909

1.038

19425/1564/2004

0.0563, 0.1357

1.076

22815/470/1997

0.0685, 0.1284

1.093

ﬂack x

0.48(3)

0.989, -0.446

largest diff. peak and hole [e Å^-3

]

2.089,^b-0.958

0.672, -0.536

0.826, -0.762

1.373,^c-0.606

b

^aThis is a low quality solution and is only included to show the gross structure. With the exception of one Fourier peak (2.089 eÅ^-3) at 1.74 Å

from Rh1, attributed to Fourier truncation errors, the max residual density peak is 1.48 e Å^-3

.

^cWith the exception of one Fourier peak (1.37 eÅ^-3) at

1.47 Å from Rh1, attributed to Fourier truncation errors, the max residual density peak is 0.62 e Å^-3

.

were placed into a New Era high pressure NMR sample tube. 1,2-

F₂C₆H₄(0.3 mL) was added by vacuum transfer. The reaction was

followed immediately in situ by ¹¹B{¹H} NMR spectroscopy. Under

these conditions the reaction went to completion by ¹¹B NMR

spectroscopy and gave two metal containing species, 7 and a

complex characterized in situ and tentatively assigned as

device [150(2) K];⁶⁷data were collected using COLLECT, and

reduction and cell reﬁnement was performed using DENZO/

SCALEPACK. Structures were solved by direct methods using

SIR92 (7, 12)⁶⁸and SIR2004 (4, 9, 10)⁶⁹and reﬁned full-matrix

least-squares on F²using SHELXL-97.⁷⁰All non-hydrogen atoms

were reﬁned anisotropically. H0a and H0b in 7 were located on

the Fourier difference map and their position reﬁned freely (isotropic

displacement parameters were ﬁxed to ride on Rh1). BH protons

in all compounds were located on the Fourier difference map

(isotropic displacement parameters were ﬁxed to ride on the parent

B atoms), with the exception of those on B2 and B12 in 10.

Restraints to BH bonds were applied as follows: B1-H1A )

B1-H1B in 4 and 7 (B1-H1C free to reﬁne); B1-H1A )

B1-H1B ) B11-H11A ) B11-H11B, Rh1-H1A ) Rh1-H1B

)Rh2-H11A)Rh2-H11B,B1-H1C)B11-H11C,H1A· · ·H1C

) H1B· · ·H1C, H11A· · ·H11C ) H11B· · ·H11C in 10; B1-H1A

) B1-H1B ) B3-H3A ) B3-H3B, B2-H2A ) B2-H2B )

B4-H4A ) B4-H4B, H2A · · ·H2B ) H4A· · ·H4B in 12. All other

hydrogen atoms were placed in calculated positions using the riding

model. Excessive thermal motion of some of the phosphine ligands

was treated by modeling the relevant phosphine substituents over

two sites and restraining their geometries. Rotational disorder of

CF₃groups on the anions was present in all of the structures. The

relevant ﬂuorine atoms were modeled over two sites, and the

geometries restrained were necessary. Problematic solvent disorder

in the solution of 7 was treated using the SQUEEZE algorithm.⁷¹

Restraints to the thermal parameters were applied where necessary

to maintain sensible values. Graphical representations of the

structures were made with XP⁷⁰and ORTEP3.⁷²

[Rh(H)₂(H₂BdNMe₂)(PⁱBu₃)₂][BAr^F] 8 by comparison with com-

4

plex 9. Selected ¹H NMR (500 MHz, 1,2-F₂C₆H₄, 298 K): δ -2.11

[bs q, J(BH) 79, 2H, BH₂], -16.43 (br, 2H, Rh-H). ³¹P{¹H} NMR

(202 MHz, 1,2-F₂C₆H₄, 298 K): δ 18.97 [d, J(RhP) 104]. As a result

of the low relative concentration of 8 in the catalysis mixture we

were unable to assign a ¹¹B resonance to complex 8.

Catalytic Dehydrocoupling of H₃B·NHMe₂with 1 in an

Open System. In a typical experiment H₃B·NHMe₂(0.196 M in

1,2-F₂C₆H₄, 0.75 mL, 0.146 mmol, 20 equiv) was added to a stirred

sample of 1 (10 mg, 7.3 × 10^-3mmol) in a Schlenk ﬂask open to

an argon atmosphere. The 50 µL samples were taken after 5, 15,

and 35 min. Samples were diluted with 1,2-F₂C₆H₄, frozen, and

analyzed by ¹¹B NMR spectroscopy immediately following warming

to RT. Complete conversion to [H₂BNMe₂]₂was achieved within

35 min, as measured by ¹¹B NMR. Under the same conditions no

change in rate was observed when mercury (0.1 mL) was added to

the reaction mixture. Addition of H₃B·NHMe₂(8.6 mg, 1.5 × 10^-1

mmol, 20 equiv) to the postcatalysis mixture resulted in the onset

of catalysis; complete conversion to [H₂BNMe₂]₂was again

achieved within 35 min. Following the reaction by hydrogen loss

using a water-ﬁlled gas-burrette showed that approximately 1 equiv

of H₂is lost per H₃B·NHMe₂.

Catalytic Dehydrocoupling of III-Me. In a typical experiment

III-Me (17 mg, 1.5 × 10^-1mmol, 20 equiv) and 1 (10 mg, 7.3 ×

10^-3mmol) were placed into a New Era high pressure NMR sample

tube. 1,2-C₆H₄F₂(0.3 mL) was added by vacuum transfer. The

reaction was followed in situ by ¹¹B{¹H} NMR spectroscopy at

298 K. Under these conditions the reaction gave IV in quantitative

yield in 70 h.

X-ray Crystallography. Relevant details about the structure

reﬁnements are given in Table 2. As all crystals were extremely

air-sensitive they were placed in dried and degassed perﬂuoropoly-

alkylether (viscosity 1600 cSt, ABCR) in a glovebox and then

mounted rapidly under a stream of argon. Data were collected on

an Enraf Nonius Kappa CCD diffractometer using graphite mono-

chromated Mo KR radiation (λ ) 0.71073 Å) and a low-temperature

Computational Details. All density functional theory calcula-

tions in this study were performed using the Gaussian 03 suite of

ab initio programs⁷³with the nonempirical Tao-Perdew-Staroverov-

(67) Cosier, J.; Glazer, A. M. J. Appl. Crystallogr. 1986, 19, 105–107.

(68) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. J. Appl.

Crystallogr. 1993, 26, 343–350.

(69) Burla, M. C.; Caliandro, R.; Camalli, M.; Carrozzini, B.; Cascarano,

G. L.; De Caro, L.; Giacovazzo, C.; Polidori, G.; Spagna, R. J. Appl.

Crystallogr. 2005, 38, 381–388.

(70) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112–122.

(71) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7–13.

(72) Farrugia, L. J. Appl. Crystallogr. 1997, 30, 565–565.

(73) Frisch, M. J.; et al. Gaussian 03, Revision E.01; Gaussian, Inc.:

Wallingford, CT, 2004 (suite of programs for ab initio calculation).

9

J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15455

A R T I C L E S

Douglas et al.

Scuseria⁷⁴(TPSS) functional and with the all-electron basis set

6-31G(d,p)⁷⁵(Def2-TZVP with ECP basis set for Rh).⁷⁶The

geometric structures of all complexes studied in this paper were

optimized as gas-phase as the cation only. Calculating the harmonic

vibrational frequencies at the optimized structures and noting the

number of imaginary frequencies conﬁrmed the nature of all

intermediates (no imaginary frequency) and transition states (only

one imaginary frequency), which also were conﬁrmed to connect

reactants and products by the intrinsic reaction coordinate (IRC)

calculations. The zero-point energy (ZPE) and entropic contribution

have been estimated within the harmonic potential approximation.

The enthalpies, H, and free energies, G, were calculated for T )

298.15 K. All relative energies are reported in kcal/mol. The effect

of solvent was taken into account by single point calculations on

gas-phase optimized structures using the integral equation formalism

polarizable continuum model (IEFPCM) for chlorobenzene (ε )

5.621 as a mimic of the 1,2-F₂C₆H₄solvent). The ﬁgures for

optimized 3D molecular structures displayed in text are drawn by

using the JIMP2 molecular visualization and manipulation pro-

gram.⁷⁷

Acknowledgment. This work was supported by Grants from

NSF (CHE-0518074, CHE-0541587, and DMS-0216275), The

Welch Foundation (A0648), the EPSRC (EP/E050743/1), and the

University of Oxford. Mr. Romaeo Dallanegra (Oxford) is thanked

for the low temperature NMR spectrum of 10. Dr. Simon Aldridge

(Oxford) is thanked for stimulating discussions and the gift of

H₂BdNCy₂.

Note Added after ASAP Publication. Errors were present in

the caption to Figure 13 and the last sentence of the ﬁrst paragraph

of the Experimental Section in the version published ASAP

September 28, 2009. The corrected version was published October

1, 2009.

Supporting Information Available: Reactivity of 2 with

ClCH₂CH₂Cl. Details of crystallographic analysis for the new

compounds. Plot of ¹¹B concentration for the dehydrocoupling

of H₃B·NHMe₂using 2 and MeCN-quenching experiment.

Complete reference 73, 3D molecular structures, atomic coor-

dinates, and absolute energies of all optimized stationary points

and transition states. This material is available free of charge

via the Internet at http://pubs.acs.org.

(74) Tao, J. M.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Phys.

ReV. Lett. 2003, 91, 146401.

(75) (a) Hehre, W. J.; Ditchﬁeld, R.; Pople, J. A. J. Chem. Phys. 1972, 56,

2257–2261. (b) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley,

J. S.; Gordon, M. S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982,

77, 3654–3665.

(76) (a) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297.

(b) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.

Theor. Chim. Acta 1990, 77, 123.

JA906070R

(77) (a) Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768. (b) Manson,

J.; Webster, C. E.; Hall, M. B. JIMP2, version 0.091; Texas A&M

University: College Station, 2006 (a free program for visualizing and

manipulating molecules).

9

15456 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Article Doi

Monomeric and oligomeric amine-borane σ-complexes of rhodium. Intermediates in the catalytic dehydrogenation of amine-boranes

DOI: 10.1021/ja906070r

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja906070r

Full text of DOI:10.1021/ja906070r

Products guided by the article

R&D Labs maybe for 1192372-93-2

Relevant to this article

Hot Product