Alkyl dehydrogenation in a Rh(I) complex via an isolated agostic intermediate

Article abstract of DOI:10.1039/b816739g

A well-characterised 14-electron rhodium phosphine complex, [Rh(P ⁱPr₃)₃][BAr^F₄], which contains a β-CH agostic interaction, is observed to undergo spontaneous dehydrogenation to afford [Rh(Pⁱ

Full text of DOI:10.1039/b816739g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Alkyl dehydrogenation in a Rh(I) complex via an isolated agostic
intermediatew
Adrian B. Chaplin,a Amalia I. Poblador-Bahamonde,b Hazel A. Sparkes,c
Judith A. K. Howard,*c Stuart A. Macgregor*b and Andrew S. Weller*a
Received (in Cambridge, UK) 24th September 2008, Accepted 31st October 2008
First published as an Advance Article on the web 19th November 2008
DOI: 10.1039/b816739g
A well-characterised 14-electron rhodium phosphine complex,
[Rh(PiPr3)3][BArF4], which contains a b-CH agostic interaction,
is observed to undergo spontaneous dehydrogenation to afford
[Rh(PiPr3)2(PiPr2(C3H5))][BArF4]; calculations on a model sys-
tem show that while C–H activation is equally accessible from
the b-CH agostic species or an alternative c-CH agostic isomer,
subsequent b-H-transfer can only be achieved along pathways
originating from the b-CH agostic form.
metallacyclobutane formation (C–H activation only8,9) which
arises from an alternative g-agostic interaction.
Transition metal-mediated alkane dehydrogenation is an impor-
tant methodology for the selective transformation of alkanes.1
The putative intermediates for such reactions are alkane sigma
complexes, which then undergo successive C–H activation and
b-H-elimination. For the intramolecular dehydrogenation of
alkyl groups it is accepted that C–H activation is usually
preceded by a Mꢀ ꢀ ꢀHC agostic interaction (Scheme 1);2 however,
well-defined examples of such complexes that subsequently
undergo alkyl dehydrogenation are, to the best of our know-
ledge, unknown. Indeed, as far as we are aware, there is only one
example of a complex where an agostic interaction undergoes
C–H activation for which both the agostic and C–H activated
product have been crystallographically characterised,3 and only a
few examples where these tautomers are directly observed in
solution.4
Scheme 2
Reaction of [Rh(BINOR-S)(PiPr3)][BArF4] (2)10 with 2
equivalents of PiPr3 in C6H5F solution results in reductive
elimination of BINOR-S and the formation of 1 in quantitative
yield by NMR spectroscopy. Alternatively, 1 can be formed by
the addition of PiPr3 to [Rh(C6H5F)(PiPr3)2][BArF4] (3) in
C6H5F (Scheme 2). Complex 1 is highly fluxional in solution
at room temperature, displaying one phosphine environment
by 31P NMR spectroscopy which shows coupling to 103Rh
(d 47.1, d, JRhP = 173 Hz). This fluxional behaviour is not frozen
out, even upon cooling to 200 K, where the 31P{1H} NMR
1
spectrum shows a very broad signal. The H NMR spectrum
shows a featureless hydride region at all temperatures. In the
solid statez (Fig. 1) the structure displays a distorted square
planar geometry in which a b-CH agostic interaction from an
isopropyl group occupies the fourth coordination site, show-
ing a relatively short Rh1–C1a distance [2.494(12) A, located
Rh1–H1a 1.91(9) A]11 and an acute Rh1–P1–C1a angle
Scheme 1
We report here the isolation of a ‘‘T-shaped’’ 14-electron
rhodium phosphine complex, [Rh(PiPr3)3][BArF4] [1, ArF
=
3,5-C6H4(CF3)2], that contains an unusual b-CH agostic inter-
action from the isopropyl phosphine ligand and undergoes
intramolecular dehydrogenation. We also demonstrate, using
computational methods, that the b-agostic interaction in 1 is
important in defining the ultimate product of the reaction:
dehydrogenation (C–H activation/b-elimination5–7
) versus
a Department of Inorganic Chemistry, University of Oxford,
UK OX1 3QR. E-mail: andrew.weller@chem.ox.ac.uk
b School of EPS – Chemistry, Heriot-Watt University, Edinburgh,
UK EH14 4AS. E-mail: S.A.Macgregor@hw.ac.uk
Fig. 1 Complex 1; ellipsoids are depicted at the 50% probability
level. The anion, most H atoms and minor component (4) are omitted
for clarity. Key bond lengths (A) and angles (1): Rh1–P1, 2.249(2);
Rh1–P2, 2.395(2); Rh1–P3, 2.268(2); P1–Rh1–P2, 149.91(6);
P1–Rh1–P3, 104.24(6); P2–Rh1–P3, 105.47(6); Rh1–C1a, 2.494(12);
Rh1–H1a, 1.91(9); Rh1–P1–C1a, 73.8(4); C1a–C2a, 1.540(13);
C1a–C3a, 1.540(13).
c Department of Chemistry, University Science Laboratories, South
Road, Durham, UK DH1 3LE. E-mail: j.a.k.howard@durham.ac.uk
w Electronic supplementary information (ESI) available: Full experi-
mental, characterisation and crystallographic details. Energies and
Cartesian coordinates of all computed species; full reference for
Gaussian 03. See DOI: 10.1039/b816739g/
ꢁc
This journal is The Royal Society of Chemistry 2009
244 | Chem. Commun., 2009, 244–246

View Article Online
[73.8(4)1]. The RhP3 core is planar [sum of angles 359.6(1)1].
Complex 1 is a rare example of a formally 14-electron late-
transition metal complex,9–13 and is closely related to
[Rh(PPh3)3][ClO4]13 and Rh(tBu2PCH2PtBu2)(CH2Me3).12a
Isopropyl phosphine groups have been shown to coordinate
through g-agostic interactions,9,10,14 but as far as we are aware
1 is the first structurally characterised example of a b-agostic
interaction with this ligand motif. Given the fluxional beha-
viour of 1, even at 200 K, we have not been able to determine
experimentally whether this interaction dominates in solution,
or rapidly interconverting b- or g-interactions are present. We
did not observe a clear C–H stretch in the IR spectrum that
could be assigned to the agostic C–H bond.
Fig. 2 Complex 4; ellipsoids are depicted at the 50% probability
level, the anion and most H atoms have been omitted for clarity. Key
bond lengths (A) and angles (1): Rh1–P1, 2.291(2); Rh1–P2, 2.395(2);
Rh1–P3, 2.347(2); P1–Rh1–P2, 153.93(6); P1–Rh1–P3, 99.10(6);
P2–Rh1–P3, 106.63(5); Rh1–C1, 2.183(6); Rh1–C2, 2.240(6);
Rh1–P1–C1, 63.6(2); C1–C2, 1.399(8); C1–C3, 1.489(8).
Complex 1 spontaneously,6,16 but slowly (t1/2 B 1.5 h),
undergoes dehydrogenation of one of the isopropyl substi-
tuents in solution at room temperature (C6H5F solvent).
This results in the generation of an equilibrium mixture of
[Rh(PiPr3)2(Z3-PiPr2(C3H5))][BArF ] (4) and [RhH2(PiPr3)3][BArF ]
To address these issues we have employed density functional
theory calculations17 to study the reactivity of the model
4
4
(5)—see ESIw for full details including a solid-state struc-
ture—the latter of which arises from reaction of 1 with
liberated H2 (1:4:5 = 1.0:0.95:0.35) (Scheme 3). 5 is closely
related to previously reported [IrH2(PiPr2Ph)3][BArF4].15 Due
to this process and the structural similarity of 4, isolation of
1 by crystallisation at low temperature occurs concomitant
with small amounts of 4 (1:4 B 4:1 on the basis of integrals in
i
cation, [Rh(PH2 Pr)(PH3)2]+, 10. Three forms of 10 were
located, 10b and 10c, with b- and g-CH agostic interactions,
respectively (Fig. 3), as well as an anagostic form, 10an. For the
simple model system 10c is more stable than 10b, reflecting the
lesser degree of strain associated with the five-membered
Rh–P–C–C–H ring in this case. Although this is contrary to
what is observed in the solid-state structure our model does
not take into account the effects of ligand bulk which can play
a significant role in determining the nature of an agostic
interaction.15 In addition solid-state packing effects may play
a role in favouring one agostic form over the other. 10b and 10c
are linked via 10an (E = 2.9 kcal molꢂ1), the highest point
along this pathway being loss of the b-CH agostic via a TS at
5.9 kcal molꢂ1. 10an corresponds to a very shallow minimum
and forms 10c with a negligible barrier (not shown in Fig. 3).
The low overall activation energies for interconversion
between 10b and 10c are consistent with the highly fluxional
behaviour observed in the NMR spectra of 1.
1
the H NMR spectrum and X-ray diffraction). The dehydro-
genation/hydrogenation equilibrium in solution is shifted
towards 4 by successive removal of hydrogen through freeze–
vacuum–thaw cycles, while addition of the hydrogen acceptor
t-butylethene (tbe) results in the clean formation of 4 (86%
isolated yield). Addition of H2 to 1 or 4 rapidly generates 5,
the latter presumably via 1, showing that the dehydrogenation is
reversible. In CH2Cl2 or THF dehydrogenation of 1 also occurs
alongside decomposition to unidentified products.
C–H activation TS structures were located from both 10b and
10c and the computed activation barriers of 11.9 kcal molꢂ1
and 11.5 kcal molꢂ1
, respectively, indicate that these
processes are equally accessible. C–H activation is also
endothermic and should be reversible. Importantly, the
accessibility of the initial C–H activation does not depend
significantly on the nature of the CH agostic bond present in
the reactant.
In contrast, b-H-transfer in the C–H activated species 60b
and 60c depends markedly on which intermediate is involved.
Scheme 3
The solid-statez structure of isolated 4 is shown in Fig. 2 and
confirms dehydrogenation of one of the isopropyl groups to
form an Z3-coordinated vinyl phosphine. The rhodium has an
approximately square planar geometry, and the featureless
high-field region of the 1H NMR spectrum of 4 (200 K)
supports the lack of a hydride ligand. The molecular geometry
and structural metrics of the Z3-ligand are similar to those
reported previously for vinyl phosphines.6,7
Mechanistically, phosphine dissociation during the dehydro-
genation reaction is discounted as no inhibition is observed in
the presence of excess phosphine (10 equivalents), while no
dehydrogenation is observed for 3 in C6H5F. These observa-
tions suggest successive C–H activation and b-elimination in
1 followed by loss of dihydrogen (Scheme 1). Two questions
thus arise: does C–H activation occur at the b or g position, and
how does this affect the ultimate outcome of the reaction
(dehydrogenation versus cyclometallation)?
Fig. 3 Reaction profiles (kcal molꢂ1) for C–H activation in 10b and 10c.
Chem. Commun., 2009, 244–246 | 245
ꢁc
This journal is The Royal Society of Chemistry 2009

View Article Online
ginating from the b-agostic form. Therefore only the product
of C–H activation at the b-position can lead to productive
dehydrogenation.
Notes and references
z Crystallographic data. 1: C59H74.6BF24P3Rh, M = 1446.42, mono-
clinic, P21/n (Z = 4), a = 13.1039(6) A, b = 28.652(1) A, c =
17.9634(8) A, b = 106.228(1)1. V = 6475.6(5) A3, T = 120(2) K,
13 255 unique reflections [R(int) = 0.0395]. Final R1 = 0.0455
[I 4 2s(I)]. 4: C59H73BF24P3Rh, M = 1444.80, monoclinic, P21 (Z = 6),
a = 19.3219(2) A, b = 17.7292(2) A, c = 29.0445(3) A, b =
96.6182(4)1. V = 9883.2(2) A3, T = 150(2) K, 31 319 unique
reflections [R(int) = 0.0396]. Final R1 = 0.0506 [I 4 2s(I)].
Fig. 4 Reaction profiles (kcal molꢂ1) for b-H-transfer in 60b.
1 (a) C. M. Jensen, Chem. Commun., 1999, 2443; (b) R. H. Crabtree,
J. Chem. Soc., Dalton Trans., 2001, 2437; (c) A. S. Goldman,
A. H. Roy, Z. Huang, R. Ahuja, W. Schinski and M. Brookhart,
Science, 2006, 312, 257.
2 (a) G. J. Kubas, Metal Dihydrogen and s-Bond Complexes, Kluwer
Academic/Plenum Publishers, New York, 2001; (b) M. Brookhart,
M. L. H. Green and G. Parkin, Proc. Natl. Acad. Sci. U. S. A.,
2007, 104, 6908.
For metallacyclophosphabutane 60c, we were unable to locate
a TS for b-H-transfer to Rh. Scans based on the Rh-bH
distance led to a steady increase in energy to over 30 kcal
molꢂ1 above 60c. TS structures were located but these were
shown to be for a Rh-assisted 1,2 H shift resulting in isomeri-
sation to 60b (E = +31.1 kcal molꢂ1, see ESIw). For 60b,
however, a number of low energy b-H-transfer pathways were
characterised, two of which are shown in Fig. 4. In order to
complete the dehydrogenation process a cis-dihydride must be
formed upon b-H-transfer so that H2 reductive elimina-
tion can be accessed. In Pathway 1 this is achieved by
3 N. M. Scott, R. Dorta, E. D. Stevens, A. Correa, L. Cavallo and
S. P. Nolan, J. Am. Chem. Soc., 2005, 127, 3516.
4 (a) A. C. Albeniz, G. Schulte and R. H. Crabtree, Organometallics,
1992, 11, 242; (b) B. Rybtchinski, L. Konstantinovsky, L. J.
W. Shimon, A. Vigalok and D. Milstein, Chem.–Eur. J., 2000, 6,
3287; (c) D. Buccella and G. Parkin, J. Am. Chem. Soc., 2006, 128,
16358; (d) A. Toner, J. Matthes, S. Gruendemann, H. Limbach,
B. Chaudret, E. Clot and S. Sabo-Etienne, Proc. Natl. Acad. Sci.
U. S. A., 2007, 104, 6945; (e) L. Mole, J. L. Spencer, N. Carr and
A. G. Orpen, Organometallics, 1991, 10, 49.
initial b-H-transfer from 60b to form trans-dihydride 70trans
,
.
(E = +16.7 kcal molꢂ1) followed by isomerisation to 70cis
5 M. Baya, M. L. Buil, M. A. Esteruelas and E. Onate, Organometallics,
2004, 23, 1416–1423.
The cis–trans isomerisation TS is the highest point along
Pathway 1 (E = +25.2 kcal molꢂ1) and corresponds to a
barrier of 19.6 kcal molꢂ1 relative to 10b. Alternatively,
isomerisation of square-pyramidal 60b (where H is apical)
occurs prior to b-H-transfer. The lowest energy mechanism
of this type, Pathway 2, proceeds via a PH3 apical isomer
(60b(PH3), E = +8.9 kcal molꢂ1) from which b-H-transfer
leads to 70cis. The isomerisation TS is the highest point along
Pathway 2 (E = +14.3 kcal molꢂ1) equating to a barrier of
only 8.7 kcal molꢂ1 relative to 10b. A second isomerisation/
b-H-transfer route via isomer 60b(PH3) (with {PH2} apical)
was also defined, Pathway 3. This was energetically intermedi-
ate with regard to Pathways 1 and 2 with an overall barrier of
13.4 kcal molꢂ1, the highest TS being for b-H-transfer at
+19.0 kcal molꢂ1. Full details are given in the ESIw.
To complete the dehydrogenation, reductive elimination of
H2 from 70cis is required and a TS for this process was located
at +15.1 kcal molꢂ1. For Pathway 2 this is the highest point
in the overall process, although for Pathways 1 and 3 this
occurs earlier in the profile (either cis–trans isomerisation or
b-H-transfer, respectively). The model products, 40 + H2,
have a computed relative energy of +9.0 kcal molꢂ1, although
the entropy associated with H2 dissociation means that the free
energy of the products is only +2.9 kcal molꢂ1 above 10b,
consistent with the reversibility of the dehydrogenation.
In conclusion we report a ‘‘14-electron’’ T-shaped Rh(I)
complex with a supporting b-agostic interaction from an
isopropyl phosphine that spontaneously undergoes dehydro-
genation (C–H activation followed by b-H-transfer). Calcula-
tions show that while both g- and b-agostic interactions can
undergo reversible C–H activation to give metallacycle inter-
mediates, subsequent H-transfer is only accessible when ori-
6 P. B. Glaser and T. D. Tilley, Organometallics, 2004, 23, 5799.
7 A. J. Edwards, M. A. Esteruelas, F. J. Lahoz, A. M. Lopez, E. Onate,
L. A. Oro and J. I. Tolosa, Organometallics, 1997, 16, 1316.
8 (a) D. L. Thorn, Organometallics, 1998, 17, 348; (b) N. Thirupathi,
D. Amoroso, A. Bell and J. D. Protasiewicz, Organometallics,
2005, 24, 4099.
9 M. J. Ingleson, M. F. Mahon and A. S. Weller, Chem. Commun.,
2004, 2398.
10 S. Brayshaw, J. Green, G. Kociok-Kohn, E. Sceats and
¨
A. S. Weller, Angew. Chem., Int. Ed., 2006, 45, 452.
11 W. Baratta, C. Mealli, E. Herdtweck, A. Ienco, S. A. Mason and
P. Rigo, J. Am. Chem. Soc., 2004, 126, 5549.
12 (a) H. Urtel, C. Meier, F. Eisentrager, F. Rominger, J. P. Joschek
and P. Hofmann, Angew. Chem., Int. Ed., 2001, 40, 781; (b) P. H.
M. Budzelaar, N. N. P. Moonen, R. de Gelder, J. M. M. Smits and
A. W. Gal, Eur. J. Inorg. Chem., 2000, 753; (c) W. Baratta,
S. Stoccoro, A. Doppiu, E. Herdtweck, A. Zucca and P. Rigo,
Angew. Chem., Int. Ed., 2003, 42, 105; (d) A. Y. Verat, M. Pink,
H. Fan, J. Tomaszewski and K. G. Caulton, Organometallics,
2008, 27, 166, and references therein.
13 Y. W. Yared, S. L. Miles, R. Bau and C. A. Reed, J. Am. Chem.
Soc., 1977, 99, 7076.
14 H. Aneetha, M. Jimenez-Tenorio, M. C. Puerta, P. Valerga,
V. N. Sapunov, R. Schmid, K. Kirchner and K. Mereiter,
Organometallics, 2002, 21, 5334.
15 A. C. Cooper, E. Clot, J. C. Huffman, W. E. Streib, F. Maseras,
O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1999, 121,
97–106.
16 (a) T. M. Douglas and A. S. Weller, New J. Chem., 2008, 32,
966; (b) T. M. Douglas, S. K. Brayshaw, R. Dallanegra,
G. Kociok-Kohn, S. A. Macgregor, G. Moxham, A. S. Weller,
¨
T. Wondimagegn and P. Vadivelu, Chem.–Eur. J., 2008, 14, 1004;
(c) T. M. Douglas, H. Le Notre, S. K. Brayshaw, C. G. Frost and
A. S. Weller, Chem. Commun., 2006, 3408.
17 Calculations employed Gaussian 03 with the BP86 functional. Rh
and P centres were described with Stuttgart RECPs and basis sets
with polarisation on P. 6-31G** basis sets were used for C and H
atoms. All energies include a correction for zero-point energies. See
ESIw for full details.
ꢁc
This journal is The Royal Society of Chemistry 2009
246 | Chem. Commun., 2009, 244–246