Activation of a (cyclooctadiene) rhodium(i) complex supported by a chiral ferrocenyl phosphine thioether ligand for hydrogenation catalysis: A combined parahydrogen NMR and DFT study

Article abstract of DOI:10.1039/c3dt51429c

The reaction of [RhCl(P,S^tBu)(COD)] (1) or [Rh(P,S ^tBu)(COD)]BF₄ (2) where (P,S^tBu) is CpFe[η⁵-1,2-C₅H₃(PPh₂)(CH ₂S^tBu)] with H₂ in MeOH gives rise to COD hydrogenation and formation of a solvent-stabilized product. The formation of hydride species cannot be observed in view of a very rapid H/D exchange between H₂ and the solvent. Introduction of pyridine or acetonitrile slows down this exchange process and allows observation of diastereometric dihydride complexes, [Rh(P,S^tBu)(H)₂(L)₂]⁺, the stereochemistry of which was fully elucidated. The hydride site exchange rates have been derived from EXSY NMR experiments and used, with assistance from DFT calculation, to elucidate the isomerization and site exchange mechanisms.

Full text of DOI:10.1039/c3dt51429c

Dalton
Transactions
View Article Online
View Journal
PAPER
Activation of a (cyclooctadiene) rhodium(I) complex
supported by a chiral ferrocenyl phosphine thioether
ligand for hydrogenation catalysis: a combined
parahydrogen NMR and DFT study†
Cite this: DOI: 10.1039/c3dt51429c
Ekaterina M. Kozinets,a,b,c Marianna Fekete,d Oleg A. Filippov,c Natalia V. Belkova,*c
Elena S. Shubina,c Rinaldo Poli,*a,b,e Simon B. Duckett*d and Eric Manoury*a,b
The reaction of [RhCl(P,StBu)(COD)] (1) or [Rh(P,StBu)(COD)]BF4 (2) where (P,StBu) is CpFe[η5-1,2-
C5H3(PPh2)(CH2StBu)] with H2 in MeOH gives rise to COD hydrogenation and formation of a solvent-
stabilized product. The formation of hydride species cannot be observed in view of a very rapid H/D
exchange between H2 and the solvent. Introduction of pyridine or acetonitrile slows down this exchange
process and allows observation of diastereometric dihydride complexes, [Rh(P,StBu)(H)2(L)2]+, the stereo-
chemistry of which was fully elucidated. The hydride site exchange rates have been derived from EXSY
NMR experiments and used, with assistance from DFT calculation, to elucidate the isomerization and site
exchange mechanisms.
Received 31st May 2013,
Accepted 24th June 2013
DOI: 10.1039/c3dt51429c
www.rsc.org/dalton
mostly diphosphines.3–10 A notable example of this is provided
by IrI-catalyzed imine hydrogenation as used in the multi-ton
Introduction
Square planar d8 complexes of RhI and IrI are commonly scale industrial production of the herbicide metolachor.11,12
employed as catalysts in a variety of hydrogenation processes.
In one of our laboratories, chiral ferrocene-based phos-
The most popular examples include the so-called “Wilkinson’s phine thioether ligands (P,SR in Scheme 1)13 have been develo-
catalyst”, [RhCl(PPh3)3],1 and the cation [Ir(COD)(L)2]+ (L = ped and shown to be particularly efficient in terms of both
PPh3, PMePh2, py), the latter of which displays high activities activity and enantioselectivity for the hydrogenation of aro-
even with relatively hindered C–C double bonds.2 Complexes matic ketones when combined with [IrCl(COD)]2.14 The IrI
of the type [MCl(diene)]2 or [M(diene)2]+ [M = Rh or Ir], and coordination chemistry of these ligands has also been
particularly those of iridium, have also proven suitable as pre- described.15,16 Since the initial studies that were aimed at gen-
catalysts for the ionic hydrogenation of polar substrates such erating and characterizing the catalytically active species met
as ketones and imines in the presence of appropriate ligands, with difficulty, our attention has more recently turned to the
analogous (P,SR)-based rhodium complexes, inspired by
reports of the isolation and characterization of related dipho-
aCNRS, LCC (Laboratoire de Chimie de Coordination), 205 route de Narbonne,
sphine-based complexes at the pre-catalyst activation
BP 44099, F-31077 Toulouse Cedex 4, France CNRS.
stage.17–20 We have thus synthesized and characterized com-
E-mail: rinaldo.poli@lcc-toulouse.fr, eric.manoury@lcc-toulouse.fr;
plexes of the type [RhCl(P,SR)(diene)] and [Rh(P,SR)-
Fax: +33-561553003; Tel: +33-561333173
(diene)]+BF4 (diene = cyclooctadiene, COD; norbornadiene,
−
bUniversité de Toulouse, UPS, INPT, F-31077 Toulouse Cedex 4, France
cA. N. Nesmeyanov Institute of Organoelement Compounds, Russian Academy of
Sciences, Vavilov Street 28, 119991 Moscow, Russia. E-mail: nataliabelk@ineos.ac.ru;
Fax: +7-499-1355085; Tel: +7-499-1356448
NBD),21 as shown in Scheme 1. These complexes have been
demonstrated to act as both structural and functional mimics
of the analogous Ir systems, although they show lower catalytic
dCentre for Hyperpolarisation in Magnetic Resonance, Department of Chemistry,
University of York, Heslington, York, YO10 5NY, UK.
E-mail: simon.duckett@york.ac.uk; Fax: +44 (0)1904-322516;
Tel: +44 (0)1904-322564
eInstitut Universitaire de France, 103, bd Saint-Michel, 75005 Paris, France
†Electronic supplementary information (ESI) available: Sample preparation
details for the parahydrogen NMR experiments, selected NMR spectra, descrip-
tion of the EXSY experiments of hydride exchange, figures of DFT optimized geo-
metries and full table of Cartesian coordinates for the DFT optimized structures.
See DOI: 10.1039/c3dt51429c
Scheme 1
This journal is © The Royal Society of Chemistry 2013
Dalton Trans.

View Article Online
Paper
Dalton Transactions
activity and selectivity.22 We considered it of interest to investi- CHD2OD is δ 3.31 and for CDHCl2 is δ 5.37. Further details
gate the phenomenon of precatalyst activation by the use of can be found in the ESI.†
parahydrogen NMR.
Computational details
The parahydrogen method has been used extensively to
probe reaction mechanisms by enabling the detection of low
concentration species such as intermediates in catalysis.23,24 It
was first successfully employed by Weitekamp,25 and later
Eisenberg and Bargon,26 but is now being employed much
more widely as a consequence of the potential that hyperpolar-
ization methods offer to magnetic resonance imaging and
hence health care.27 When this approach is used in conjunc-
tion with DFT the synergy of the two methods becomes readily
apparent as hitherto unseen species are not only predicted but
firmly characterized in solution. Such studies have already
been used to rationalize a series of reactions involving ruthe-
nium clusters28–30 where they can play a key role by helping
with chemical shift and coupling assignments. Furthermore
they have established a definitive opportunity to examine the
role of electronic states in the oxidative addition of H2 to a
series of 16 electron ruthenium complexes31,32 and aided in
the understanding of hydrogenation catalysis by a range of pal-
ladium complexes,33,34 and even to detect unexpected CH
bond activation products in conjunction with the well-known
complex W(N2)2(dppe)2.35
Calculations were performed with the Gaussian09 package37
using the B3LYP38,39 and M0640 functionals under the DFT
approach. All carbon and hydrogen atoms were described with
the 6-31G(d,p) basis set, whereas the 6-31++G(d,p) basis set
was applied to the atoms of ligands involved in the bonding
with Rh (P, S atoms, hydride ligands, the OH group of metha-
nol and N atoms of pyridine and acetonitrile). Effective core
potentials (ECP) and its associated SDD basis set41–44 sup-
plemented with f-polarization functions (SDD(f))45 were
applied to the Rh and Fe atoms. Geometry optimizations were
performed without any ligand simplification for the cationic
[Rh(P,StBu)L2]+ and [Rh(P,StBu)(H)2L2]+ complexes (L = pyri-
dine, CH3OH, CH3CN).
Frequency calculations were performed for all optimized
complexes in the gas phase and reported without the use of
scaling factors. The nature of all the stationary points on the
potential energy surfaces was confirmed by a vibrational ana-
lysis.46 Transition state (TS) structures showed only one negative
eigenvalue in their diagonalized force constant matrices, and
their associated eigenvectors were confirmed to correspond to
the motion along the reaction coordinate under consideration
using the Intrinsic Reaction Coordinate (IRC) method.47
In this contribution, we address the stoichiometric reactiv-
ity of two representative compounds containing the same
(P,StBu) ligand, [RhCl(P,StBu)(COD)] (1) and [Rh(P,StBu)-
−
(COD)]+BF4 (2), towards H2 through the eyes of parahydrogen
NMR, complemented by DFT calculations.
Results and discussion
(a) NMR experiments
In CD3OD without additives. The addition of p-H2 to
d4-methanol solutions of either [RhCl(P,StBu)(COD)] (1) or
[Rh(P,StBu)(COD)]BF4 (2), over the temperature range
233–298 K, failed to result in the observation of any detectable
hydride containing species. There was, however, evidence for a
common slow reaction (ca. 18% conversion in 20 minutes at
243 K) which transformed the η2–η2-COD ligand into a κ3(σ:π)
cyclooct-4-enyl ligand and generated the new proton loss
complex, 3, (Scheme 2) in very small amounts. The κ3-ligand
in this complex is characterized by features at δ 4.62, 1.54 and
2.15, see Fig. S1 and S2.† This suggests that while slow H2
addition to the Rh centre occurs, rapid reaction transforms the
Experimental section
General
All the reactions described here and the complex purifications
were carried out under argon using a high vacuum line or
Schlenk line techniques. The Rh complexes [RhCl(P,StBu)-
(COD)] (1) and [Rh(P,StBu)(COD)]BF4 (2) were synthesized from
2-diphenylphosphino-(t-butylthiomethyl)ferrocene
and [RhCl(COD)]2 or [Rh(COD)2]BF4 according to the pub-
lished procedure.21
(P,StBu)
NMR experiments
NMR measurements were made using NMR tubes that were resulting dihydride into 3. This process is followed by reproto-
fitted with J. Young’s valves and solutions were degassed on a nation to form the corresponding alkene after hydride
high vacuum line prior to H2 addition. The samples were pre- migration as shown in Scheme 2. The liberation of either HCl
pared in a glovebox by addition of the specified deuterated sol- or HBF4 is proposed to give 3. Sola et al. have reported a
vents to the solid complex in the NMR tube. For the p-H2 related H2 addition to the complex [Ir(COD)(NCCH3)(PMe3)]-
induced polarization (PHIP) experiments, hydrogen enriched BF4, where the initial dihydride product [Ir(H)2(COD)(MeCN)-
in the para spin state was prepared by cooling H2 to 36 K over (PMe3)]BF4 (in CH2Cl2) reacts with MeCN to form the mono-
the paramagnetic catalyst Fe2O3 which was doped on hydride product [IrH(1-κ-4,5-η-C8H13)(NCCH3)2(PMe3)]BF4. In
silica.23,24,36 All the resulting NMR studies were then carried this case, these species could be isolated and fully character-
out with sample concentrations of approximately 4.6 mM and ized.48 We note that 3 is related to this Ir system by the formal
all spectra were recorded on a Bruker Avance III 400 NMR loss of a proton. Given that RhIII hydride complexes are known
spectrometer. 1H NMR chemical shifts are reported in ppm to be less stable than their IrIII analogues, this transformation
relative to the residual 1H signal of the solvent which for is not unexpected.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
When this reaction was carried out in CD2Cl2 using 2, reac-
tion with p-H2 was evident at temperatures down to 233 K.
Under these conditions weak signals for a new hydride con-
taining species were detected at δ −9.7 and −23.0. The high
field signal yielded a 31P coupling of 188 Hz which is indicative
of a trans 31P splitting. These two resonances were broad and
neither the JHH nor the JRhH couplings could be quantified. In
the corresponding 31P decoupled spectra the δ −23.0 signal
yields a JHH splitting of −9 Hz and a JRhH splitting of 20 Hz.
The signals for this species were too weak to enable its charac-
terisation but the hydride trans to phosphine deduction is con-
firmed by the large coupling and a species such as A in
Scheme 2 is possible. The addition of 1 μl of MeOH to this
solution suppressed the observation of these signals.
Two key observations outlined above (the coalescence of the
MeOH and H2 signals, and H/D exchange between these two
molecules) led us to speculate on the mechanism associated
with these phenomena, which are obviously related to the
same process, occurring rapidly on the NMR timescale. Two
possible pathways are indicated in Scheme 3. In pathway a, H2
oxidative addition yields a RhIII dihydride species 5 that would
be sufficiently acidic to release a proton to the solvent and
yield a RhI monohydride intermediate 6. Reversal of all steps
with the implication of deuteron incorporation is a conse-
quence of this exchange pathway. In the alternative pathway
b, H2 replaces a methanol ligand to yield an H2 complex (non-
classical dihydride) 7, which is then deprotonated by the
solvent to yield the same monohydride complex 6 as detailed
above. Both of these pathways can be imagined to occur via
initial reaction which places the hydride trans to S (as shown
in Scheme 3) or in the alternative position trans to P which is
not illustrated. Incidentally, the lack of observation of a
hydride resonance under these conditions does not exclude
the fact that the most stable species (at least at low temp-
eratures) is indeed a mono- or dihydride complex, since such a
species would exist primarily in the form of a deuteride and
would therefore be unobservable in the hydride region of the
1H NMR spectrum.
Scheme 2
At 253 K the generation of cyclooctene was revealed by its
characteristic 1H NMR resonance at δ 1.21, which rapidly
increases in intensity over the time scale of 17 minutes
(Fig. S2†). This demonstrates that the initial H2 addition to 1
or 2 proceeds at 253 K. When the temperature was raised to
263 K the formation of cyclooctane was also observed through
its characteristic singlet at δ 1.50. Hydride resonances were not
observed at any time during this reaction. However, the residual
OH signal of methanol and the H2 signal show a dramatic temp-
erature dependence, coalescing at 263 K, thereby suggesting the
rapid interchange of these sites, presumably through the for-
mation of transient, non-observable hydride species.
At 273 K the deuteration of the H2, forming HD, becomes
evident (Fig. S3†). In addition, the phenyl proton signals for
the phosphine become more complicated in appearance at
this point. This reaction monitoring demonstrated the
instability of 3 and suggested that the ultimate formation of a
bis(solvent) adduct, [Rh(P,StBu)(MeOH)2]+, 4, takes place even
though direct evidence for the production of this complex
could not be obtained. Further H2 oxidative addition to yield a
putative dihydride species 5 appears excluded by the absence
of hydride signals in this experiment (however, see additional
discussion below). Monitoring the reaction carried out with
regular H2 at 5 bar pressure and room temperature by
31P NMR did not result in the appearance of any new reso-
nance, in fact only the signal for unreacted starting material
could be seen even after 7 h. After 3 days the starting material
was completely consumed and the new NMR spectrum shows
very few small resonances, the major one being a doublet at
δ 43.9 (JPRh = 134.1 Hz). These data suggest that 4 is not stable.
Our attempts to isolate the final product of this reaction in
a crystalline form have so far not succeeded. Heller et al. have
recently reported the successful generation and isolation of
the complex [Rh(BINAP)(MeOH)2]+ by hydrogenation of
[Rh(BINAP)(COD)]+ or [Rh(BINAP)(NBD)]+ in methanol, although
no reaction intermediates were observed in those cases.18
Scheme 3
This journal is © The Royal Society of Chemistry 2013
Dalton Trans.

View Article Online
Paper
Dalton Transactions
In CD3OD in the presence of pyridine. When the analogous 31P centre that was located at δ 47.9 and appeared with a JRhP
reactions of either 1 or 2 with H2 were carried out in the pres- of 140.0 Hz. These resonances therefore arise from groups
ence of pyridine at low temperature (233–283 K), on the other within the same complex (8). In a similar manner the δ −14.3
hand, two common and dominant hydride containing pro- (JHH = −10 Hz, JRhH = 24 Hz) and δ −17.2 (JHH = −10 Hz, JRhH
=
ducts formed. For example, when a d4-methanol solution of 1 20 Hz) signals arise from a second species (9) where now the
was prepared which contained a 75-fold excess of pyridine connected 31P resonance appears at δ 45.0 with a JRhP of
(75 equiv.) and the resulting reaction with p-H2 was monitored 141.6 Hz. On the basis of the relative hydride resonance peak
by 1H NMR spectroscopy at 233 K, four PHIP enhanced areas it can be suggested that 8 and 9 are formed in a step
hydride signals were observed at δ −13.1, δ −14.3, δ −17.2 and which shows a kinetic selectivity of approximately 1 : 4 at
δ −17.6. These hydride ligand signals all contained anti-phase 233 K, if it is assumed that these resonances result from PHIP
features due to their formation from p-H2 that are associated derived magnetic states that are created with identical
with their common JHH coupling in addition to in-phase split- efficiency. When the temperature is raised to 273 K, the ratio
tings due to further couplings to a single rhodium and a increases to 1 : 13.
single phosphorus centre. A more complete NMR monitoring
Upon increasing the solution temperature to 283 K, no
of this reaction was undertaken after warming the sample to change in the relative size of the PHIP-enhanced signals was
263 K, during which process the NMR spectral features did not observed. However, upon warming to 298 K and beyond, the
change. The hydride region of the resulting 1H NMR spectrum hydride signals for 8 and 9 could no longer be observed.
is shown in Fig. 1a. The JPH coupling values are all consistent When this experiment was repeated at 263 K with normal dihy-
with cis hydride phosphorus ligand arrangements. In addition, drogen, the ratio of the hydride signals of 8 and 9 was 1 : 7.
1
in the corresponding H–31P HMQC NMR spectrum (Fig. 1b), When, on the other hand, the reaction was conducted at room
the δ −13.1 (JHH = −13 Hz, JRhH = 24 Hz) and δ −17.6 (JHH
=
temperature (ca. 15 min between bubbling and recording the
−13 Hz, JRhH = 19 Hz) signals proved to couple to a single NMR spectrum), only the presence of compound 9 could be
Fig. 1 NMR spectra showing characteristic resonances of 8 and 9 (as indicated) observed during reaction of 2 in d4-methanol with p-H2 and 75 equiv. of pyridine
at 263 K: (a) p-H2 enhanced 1H NMR spectrum showing the hydride region; (b) 1H–31P HMQC NMR spectrum collected using 15N labeled pyridine; (c) 15N labeled,
1H–15N HMQC NMR spectrum; (d) 1H–103Rh HMQC NMR spectrum (in (b) the inset boxes reflect vertical expansions of ×2 and ×8 relative to the baseline).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Table 1 NMR data for complexes 8 and 9 in MeOD at 283 K
Compound
δ 1H
δ
31P
δ
15N
δ
103Rh
8
9
−13.1, dd, JHH = −13 Hz; JRhH = 24 Hz, JRhP = 140.0 Hz
−17.6, dd, JHH = −13 Hz; JRhH = 19 Hz, JRhP = 140.0 Hz
−14.3, dd, JHH = −10 Hz; JRhH = 24 Hz, JRhP = 141.6 Hz
−17.2, dd, JHH = −10 Hz; JRhH = 20 Hz, JRhP = 141.6 Hz
47.9, d
45.0, d
267.6, t, JPN = 50 Hz; JNH = 20 Hz
266.2, t, JPN = 50 Hz; JNH = 20 Hz
−7390, d JRhP = 131 Hz
−7268, d JRhP = 131 Hz
detected, reflecting the greater thermodynamic preference for
this isomer.
In order to probe the ligand arrangement in compounds 8
and 9 further, a 15N labelled pyridine sample was examined.
Now, the two phosphorus signals associated with these two
complexes exhibit additional and identical extra splittings of
50 Hz due to the presence of a resolved trans 31P–15N coupling.
1
In the corresponding H–15N HMQC NMR spectrum (Fig. 1c),
the hydride signals at δ −17.2 and −17.6 showed strong corre-
lation peaks to 15N resonances at δ 266.2 and δ 267.6 respecti-
vely. In addition, a hydride-15N splitting of 20 Hz was
exhibited by the two low field hydride resonances of 8 and 9.
Rhodium signals have also been detected at δ −7268 for 8 and
δ −7389 for 9 through the recording of a 1H-103Rh HMQC spec-
trum as shown in Fig. 1d.
Scheme 4
This information therefore confirms that there are two pyri-
dine ligands attached to the metal centre in 8 and 9, which are
located trans to one hydride and to the phosphine donor.
Given the bidentate nature of the P,S ligand it can be further
concluded that the second hydride ligand in both 8 and 9 is
trans to sulfur. All the NMR data associated with 8 and 9
resulting from this study are summarized in Table 1. The simi-
larity of these NMR data suggests that these two products are
simply diastereoisomers of one another, differentiated by the
ferrocene ligand orientation as shown in Scheme 4. This
product geometry indicates that the initial H2 oxidative
addition takes place over the S–Rh–C axis.
When an nOe experiment was recorded to probe the
hydride site interchange process undergone by these com-
plexes at 273 K with the 75-fold excess of pyridine, several
exchange processes were observed (as revealed in Fig. 2). These
include a mutual hydride site interchange within 9 which pro-
ceeds with an exchange rate constant of 0.785(6) s−1 and the
interconversion of 9 into 8 on a slower timescale that places
the moving hydride into either site of 8 with rates of 0.167(2)
s−1 where Hc becomes Ha and Hd becomes Hb and 0.160(3) s−1
where Hc becomes Hb and Hd becomes Ha. Hence there is a
limited selectivity in this process. Isomer 8 converts into 9 on
a faster timescale where the observed rate constant is 1.10(1)
s−1 for the Ha→d and Hb→c transformations and 1.19(1) s−1 for
the Ha→c and Hb→d transformations. The experimental rate
constant for hydride site interchange in 8 is zero. The overall
scheme of hydride site exchange rates is summarized in
Scheme 5. Table S1 (ESI†) shows the list of constraints that
were used for the calculation of the exchange rate constants.
There was no evidence for hydride exchange into free H2 or
MeOH from 8 or 9 in these experiments which are limited
by the timescale of NMR relaxation, and contrary to the
Fig. 2 Hydride ligand exchange data for the interconversion of 8 and 9 in the
presence of 75-fold excess of py, as probed through the selective excitation of
(a) Ha of 8 and (b) Hc of 9, over the defined observation period; the observation
■
), Hb
◆
●
▲
points are listed as Ha
(
(
), Hc ( ) and Hd ( ) as defined in Scheme 4. The
solid lines correspond to simulated changes that yield the rate constants in the
text.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans.

View Article Online
Paper
Dalton Transactions
and 6.78. In contrast, the 31P signal of the minor isomer 8
proved to connect with 1H signals at 8.03 and 7.5 in an HMQC
measurement. Furthermore, nOe data confirmed that the
signal at δ −13.1 in 8 connect with 1H signals at δ 8.42 and
8.03. The δ 8.42 resonance therefore corresponds to an ortho-
pyridine signal. This information has therefore confirmed the
assignment shown in Scheme 4. The structure corresponding
to 9 has therefore the hydride ligand trans to pyridine located
on the same side as the ferrocenyl group. When a long range
1H–31P experiment was recorded, further proton signals were
located at δ 7.6, 4.8 and 4.2 in 9.
Scheme 5 The three numbers on each arrow are the exchange rates in s−1
relative to the solution with an 8-fold, 75-fold and 173-fold excess of py.
experiment run in pure CD3OD there was no coalescence
between the solvent and H2 resonances. We note, however,
that the hydride sites of 8 and 9 are partially deuterated in
these experiments. This readily shows up in a 31P-decoupled
HMQC measurement as isotopically perturbed signals at
δ −13.1, −14.3, −17.2 and −17.6 in the corresponding 1H NMR
spectrum for the Rh(H)(D) partners and at δ −14.3 and δ −17.2
for the corresponding 31P signals. Hence, the electronic effect
of pyridine coordination stabilizes the dihydride species
against deprotonation (cf. Scheme 3).
The hydrogenation of the cyclooctadiene ligand is readily
1
evident in these p-H2 enhanced H NMR spectra from 253 K.
Two sets of polarised signals appear at δ 1.53 and 1.50 due to
the CH2 protons of the hydrogenated components of COE. A
further resonance is evidenced at δ 1.41 for the backbone
signals as described above. In these 1H NMR spectra the
corresponding cyclooctane signal appears at δ 1.2 and forms
very slowly at 273 K.
When the reaction of the chloride derived precursor 1 with
normal dihydrogen is monitored in an analogous experiment
with only a 15-fold excess of pyridine (instead of 75 in the
experiment described above), 5 appears as the first formed
hydride containing product at 273 K. The problem with this
measurement is, however, that it relies on detecting a weak
signal from magnetisation that is at thermal equilibrium.
When this reaction is repeated with p-H2, weak PHIP-enhanced
hydride signals are now seen for a further two species, 10 and
11, besides those of 8 and 9 in the resulting NMR spectra at
273 K (Fig. 3). The hydride signals of 10 appear at δ −12.3 and
−18.7 with Rh–H couplings of 25.1 Hz and 26.8 Hz, respecti-
vely, and a common JHH coupling of −5.9 Hz. The hydride
ligand signals from species 11 are, however, much weaker than
those of 10 and appear at δ −15.0 (JRhH = 20, JHH = −9 Hz) and
δ −17.8 (JRhH = 15, JHH = −9 Hz). The relative ratio of the
hydride signals for 10, 8, 9, and 11 proved to be
0.9 : 7.0 : 39.6 : 1 at 273 K. In a series of 2D measurements the
δ −12.3 hydride signal proved to connect with a δ 34.3 31P
While the deuterium label incorporation into the hydride
sites is relatively slow, it precludes the measurement of precise
rate data as would be required for the assembly of an Eyring
plot. Nonetheless, the interconversion between these species
has been defined as occurring without H2 loss. This is
reflected in the fact that strong PHIP is only seen when 1 or 2
are being converted into 8 and 9. When we monitor these pro-
cesses with differing amounts of pyridine, changing the Rh/
pyridine ratio from 1 : 8 to 1 : 75 and 1 : 173, we see no change
in the relative hydride signal intensities of 8 and 9. There is,
however, a significant effect on the hydride site interchange
rate constants.
The rate constant for the mutual hydride site exchange in 9
falls to 0.29 s−1 with an 8-fold excess of pyridine while it is
0.39 s−1 with a 173-fold excess. The kinetic effect of pyridine is
therefore complex, first promoting the process and then inhi-
biting it. A similar trend is observed for the Hd→a (and Hc→b
)
rate from 0.06 s−1 through 0.16 s−1 to 0.02 s−1 and for the
Hc→a (and Hd→b) rate from 0.14 s−1 through 0.17 s−1 to
0.12 s−1, though the effect here is less dramatic. In contrast,
the Ha→c (Hb→d) process shows a rate increasing with [py]
from 0 s−1 through 1.19 s−1 to 1.23 s−1 while the rate of the
Ha→d (Hb→d) process shows the opposite trend from 1.47
through 1.10 to 0.02 s−1. We can conclude therefore that pyri-
dine plays a role in these processes. All observed hydride site
exchange rate constants are collected in Table S3.†
The hydride signal at δ −17.2 shows nOe connections to
peaks at δ 8.71, 8.23, 4.67, 3.84 and 1.21, which are due to the
ortho proton of a pyridine ligand, the ortho proton of the phos-
phine, the ferrocenyl group and the tBu group in 9. The
second hydride signal in 9 which resonates at δ −14.3 shows
through space interactions with protons that give rise to reso-
nances at δ 8.71, 8.31, 8.23 and 6.68. The extra signal at δ 8.31
is therefore the ortho proton of a second pyridine ligand. In
the corresponding 1H–31P HMQC, the 31P centre which
resonates at δ 45.0 connects to two aromatic signals at δ 8.23
Fig. 3 NMR spectra showing characteristic hydride resonances of 8, 9, 10 and
11 (as indicated) observed during reaction of 1 in d4-methanol with p-H2 and
45-fold excess of pyridine at 273 K.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Scheme 6
resonance where JRhP = 140 Hz, while the δ −15.0 hydride
signal connects with a 31P resonance at δ 51.4 that exhibits a
JRhH coupling of 148 Hz. The relative intensity of the signals
for 10 and 11 proved to fall as the excess of pyridine is
increased. When labelled pyridine is employed, both of the
sets of hydride signals show limited broadening but no
trans-15N coupling is evident. We therefore assign 10 and 11 to
chloride containing [Rh(H)2(P,SR)(pyridine)Cl] as shown in
Scheme 6. This deduction is further supported by the fact that
while the relative intensities of the signals for 8 and 9 seem to
remain constant regardless of the [py], those for 10 and 11
drop as the [py] increases.
Fig. 4 p-H2 enhanced 1H NMR spectrum showing the hydride region of 12
observed during reaction of 1 in d4-methanol with p-H2 and acetonitrile at
273 K.
When a series of EXSY measurements were undertaken to
examine the dynamic behaviour of this complex reaction
system, interconversion between 10, 8 and 9 was evident with
the signals for 11 proving to be too weak to monitor.
Scheme 7
proved to couple to a single 31P centre located at δ 47.6 which
exhibited a JRhP splitting of 148 Hz. These results suggest that
hydrogenation of 1 in the presence of acetonitrile selectively
forms the MeCN analogue (12) of 8 (Scheme 7). There are a
number of similarities between these data and those of 8. For
example, all the couplings suggest a cis PH ligand arrangement
and the chemical shifts are reflective of hydride ligands trans
to sulfur and nitrogen. The coordination of MeCN has there-
fore the same effect as that of pyridine in reducing the hydride
acidity and reducing the speed of the H/D exchange between
H2 and the solvent.
Hydrogenation studies. We have also explored the activity of
these systems as hydrogenation catalysts. When a d4-methanol
sample of 1 containing 20 μl (75 equiv.) of pyridine and
diphenylacetylene was examined at 258 K, limited hydrogen-
ation was evident and strongly enhanced signals for 8, 9, 10
and 11 were visible which exceed those in the analogous
experiment without the alkyne. Under these conditions at
258 K the ratio of the p-H2 enhanced hydride signals of 8 and
9 was now 1 : 3.8. In contrast, when a similar sample was exam-
ined without pyridine, more rapid, but still slow, hydrogen-
ation of diphenylacetylene occurs; signals for both cis and
trans stilbene are evident. These signals appear in an OPSY
measurement which was used to detect only protons that were
previously located in a single molecule of p-H2. No hydride
signals were observed during these measurements.
When the pyridine excess was 75 fold, the experimental rate
constant for the He→c and He→d processes (10 → 9) were in-
distinguishable at 31 s−1. In contrast the corresponding rate
constant for He→a and He→b (10 → 8) was zero, as was mutual
He–Hf interchange (10). These values decrease from 31 s−1 to
20 s−1 to 3 s−1 as the pyridine excess falls from 75 fold through
42 fold to 15 fold. Hence this process is [py] dependent. Con-
cerning the 9 → 10 process, the Hc→e rate constant is 0.41 s−1
,
and the Hc→f rate has a similar value. The corresponding rates
for the 9 → 8 process are slightly smaller than those found
when 2 is employed (Hc→a and Hc→b are again similar at 0.1
s−1 vs. 0.16 s−1 for 2). The new rate constants for 8 → 9 with
hydride position retention are zero when 1 is the precursor
and 1.1 s−1 when 2 is employed. 8 does, however, form 10 with
rate constants 1.3 s−1 for Ha→f and zero for Ha→e. The observed
hydride site exchange rate constants at 273 K measured on
sample 1 for different pyridine concentrations are reported in
Table S2.†
In CD3OD in the presence of acetonitrile. In a further study,
a d4-methanol sample of 1 containing 20 μL of acetonitrile was
prepared. No reaction of 1 with H2 was evident until 273 K. At
this point, the corresponding 1H NMR spectrum contained
two PHIP polarized hydride peaks at δ −14.5 and −17.6. These
hydride signals again appear as simple anti-phase doublets
with additional phosphorus and rhodium couplings (Fig. 4).
The signal at δ −14.5 exhibits a JHH coupling of 11 Hz, a JHP
coupling of 20 Hz and a JRhH coupling of 21 Hz while the
corresponding splittings of the δ −17.6 signal are 12, 17 Hz
and 20 Hz respectively. In the corresponding 1H–31P HMQC
NMR spectrum, the δ −14.5 signal and the δ −17.2 signal
When hydrogenation of the more reactive substrate phenyl-
acetylene by 1 was examined in d4-methanol using p-H2, PHIP
polarized signals could be readily seen for the styrene product
at 263 K when no pyridine was present. No hydride signals still
This journal is © The Royal Society of Chemistry 2013
Dalton Trans.

View Article Online
Paper
Dalton Transactions
are visible however in this experiment. In contrast, when a
The hypothetical replacement of the COD ligand in
10-fold excess of phenylacetylene and acetonitrile was added [Rh(P,StBu)(COD)]+ (2) with two L donors to yield IL (without
there was evidence for both the formation of 12 and the hydro- COD hydrogenation) was found at the B3LYP level as endoergic
genation of phenylacetylene.
when L = MeOH (ΔE = 15.4 kcal mol−1, ΔG = 24.7 kcal mol−1
)
and more favourable for the other two ligands (ΔE/ΔG are
−6.4/+4.4 kcal mol−1 for L = py and −1.5/+6.0 for L = MeCN).
However, in consideration of the energy gain of the COD
hydrogenation process (ΔE = −34.4 or ΔG = −18.5 kcal mol−1
for the hydrogenation to cyclooctene; ΔE = −64.3 or ΔG =
−33.1 kcal mol−1 for the hydrogenation to cyclooctane), all the
underlying reactions become favourable.
(b) DFT calculations
On the basis of the presented experimental evidence, H2 is
able to hydrogenate the COD ligand in compounds 1 and 2 in
MeOH, presumably yielding a [Rh(P,StBu)(MeOH)2]+ product
(4). However, any further oxidative addition of H2 to yield a
putative dihydride product [Rh(H)2(P,StBu)(MeOH)2]+ (5) would
1
The oxidative addition of H2 was also found to be energeti-
cally favourable for all three ligand systems, as shown by the
data collected in Table 2. It should be noted that while the gas
phase free energies are positive, the computed values do not
take into account stabilisation from solvation and the partial
quenching of translational and rotational modes in the con-
densed phase. Furthermore, the hydride complexes were
observed at low temperatures, where the detrimental effect of
the positive TΔS contribution is smaller (for instance, the
calculated ΔG for the oxidative addition to Ipy to yield Vpy at
the M06 level decreased from +6.0 to +4.9 kcal mol−1 when
applying the thermochemical corrections at 253 K instead of
298 K) and indeed the hydride resonances were lost upon
warming the NMR tubes to higher temperatures (vide supra).
For the pyridine system, the calculations have also been
carried out at the M06 level, showing a slightly less favourable
process. We note that the oxidative addition process for L =
MeOH is predicted by these calculations to be more favourable
than for the other two ligands. Therefore, failure to observe
not be evidenced by H NMR if there was rapid H/D exchange
as suggested. On the other hand, further H2 oxidative addition
takes place in the presence of py or MeCN to produce observa-
ble dihydridorhodium(III) species. A first question addressed
by the computation tool was to explore the relative stability of
the [Rh(P,StBu)L2]+ (L = MeOH, py, MeCN) complexes and their
ability to oxidatively add H2, as well as defining the relative
stability of all possible product isomers, see Scheme 8. The
associated calculations were carried out using either the
B3LYP functional or the M06 functional.
1
the corresponding hydride resonances by H NMR during the
experiments appears indeed to be attributable to the accumu-
lation of deuteride species by rapid H/D exchange. The dihy-
dride complex is, however, at least sufficiently accessible to
allow the H/D process to take place as suggested in Scheme 3.
The computational results also suggest that the two
isomers resulting from the H–H addition across the S–Rh–L
axis are energetically preferred independent of L, in agreement
with the assignment of the NMR spectra. The stereochemistry
of structure IV corresponds to that of 8 and 12 and the stereo-
chemistry of V corresponds to that of 9. Structure Vpy is slightly
less stable than IVpy at the B3LYP level, but more stable at the
M06 level in agreement with the NMR evidence. For L = MeCN
and MeOH, the calculations also indicate greater stability for
isomer VL even at the B3LYP level, whereas the NMR assign-
ment indicates that the observed compound 12 has the same
configuration as IV.
Scheme 8
Table 2 Relative gas phase energies E (free energies G in parentheses) in kcal mol−1 for the products of H2 oxidative addition to IL
L
Functional
IIL
IIIL
IVL
VL
MeOH
Py
Py
B3LYP
B3LYP
M06
−8.5(+2.6)
−4.5(+7.4)
+4.5(+15.2)
−4.2(+5.6)
−10.5(+1.3)
−4.9(+7.9)
+2.2(+14.7)
−4.9(+6.0)
−14.1(−2.5)
−10.1(+1.5)
−2.5(+7.8)
−8.5(+1.7)
−17.6(−6.5)
−10.0(+2.1)
−3.5(+6.0)
MeCN
B3LYP
−10.9(−0.1)
Dalton Trans.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
A second question addressed by the DFT study is the mech- dissociation, [(P,StBu)Rh(H)2(py)]+, is located at 29.7 kcal
anism for the H-site exchange in the two observed isomers for mol−1 from Vpy, while the isomeric non-classical complex
the pyridine systems (8 and 9, or IVpy and Vpy). The simplest [(P,StBu)Rh(H2)(py)]+ is more stable at only 18.3 kcal mol−1
way to imagine the isomerisation process, given the known from Vpy. However, the most stable dihydride complex result-
strong trans labilizing effect of the hydride ligands, involves ing from dissociation of the thioether arm (B in Scheme 9) is
dissociation of a neutral donor placed trans to a hydride ligand only 12.4 kcal mol−1 from Vpy, with A being only slightly
(either the sulphur donor of the bidentate P,S ligand or a pyri- higher at +14.7 kcal mol−1 and the nonclassical isomers C and
dine molecule) and rearrangement of the coordination sphere, D are even more stabilized at +11.8 and +8.3 kcal mol−1
,
as shown in Scheme 9 for the specific case of thioether dis- respectively. Therefore, the isomerization pathway shown in
sociation. The putative 5-coordinate intermediate A would Scheme 9 was fully explored. The results are summarized in
then rearrange to its diastereoisomer B (the same chirality at Fig. 5. Rearrangement of the classical to the nonclassical dihy-
ferrocene and inverted chirality at rhodium) by either con- drides occurs via localized transition states TSAC and TSBD at
certed hydride migration and rotation of the P,S ligand or via +18.4 and +16.2 kcal mol−1. A more direct pathway from IVpy
tautomerization to the dihydrogen complex C followed by P,S to C and from Vpy to D could not be found. The P,S ligand
rotation to yield the rotamer D. Hydride site exchange can be rotation pathway involved in the interconversion of C and D
envisaged by rotation of the dihydrogen ligand in the inter- was not straightforward to investigate for the location of a
mediates C and D. The observation of such a ligand exchange stationary point, but a manual scan of the dihedral N–Rh–P–C
process by NMR is not inconsistent provided the lifetime of angle (see Fig. S4†) led to the identification of a new local
the rapidly relaxing dihydrogen form is short.23,24
minimum corresponding to a new rotational intermediate E
All calculations of this pathway were carried out with the and two maxima, the highest point being 4.3 kcal mol−1
M06 functional, since this is expected to better handle the higher than D (+12.6 kcal mol−1 from Vpy). The search for a
long range dispersion interactions involved in the ligand dis- more direct pathway from A to B was unsuccessful. The barrier
sociation processes. They show that pyridine dissociation is for site exchange was calculated for D and turned out rather
less favourable than the P,S ligand thioether arm dissociation. low (2.5 kcal mol−1 or +10.8 kcal mol−1 from Vpy), lower than
The most favourable dihydride complex resulting from py the barriers required to go back to the stable classical dihy-
dride isomers.
The results in Fig. 5 are in agreement with certain experi-
mental observations. Isomerization from Vpy (9) to IVpy (8)
requires transiting over the TSAC barrier, whereas the site
exchange in 9 may occur by simple access of the nonclassical
intermediate and facile H2 rotation, followed by the reverse
pathway to Vpy through the lower TSBD barrier. This agrees
with the faster site exchange relative to the 9 → 8 process.
During the reverse isomerization of IVpy (8) to Vpy (9), on the
other hand, once the highest TSAC barrier is passed and the
intermediate
C is generated, the isomerization process
through P,S ligand rotation and the lower energy TSBD occurs
faster than the reverse generation of IVpy, consistent with the
fact that site exchange for 8 is not observed.
Fig. 5 Energy profile (gas phase energies in kcal mol−1) of the isomerization
Scheme 9
pathway of Scheme 9.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans.

View Article Online
Paper
Dalton Transactions
A most interesting experimental result is that the transform- excess pyridine). This phenomenon is presumably linked to
ation of the minor isomer (8) into the major one (9) occurs pre- the need of a strong base promoter for the catalytic action of
ferentially as Ha→c and Hb→d at low [py] but preferentially as these compounds and of the iridium analogues in the ionic
Ha→d and Hb→c at high [py] (see Scheme 5). The former situ- hydrogenation of polar unsaturated substrates.14,22
ation, according to Scheme 9, would be compatible with a con-
certed pathway directly converting A to B without transiting
through the nonclassical intermediate (arrows with a question
mark in Scheme 9) through a transition state at a lower energy
than TSAC and TSBD. As stated above, such a pathway could not
Acknowledgements
be located but we cannot exclude that it exists. At greater pyri- The York authors would like to thank the EPSRC (grant no.
dine concentrations, this pathway may be blocked by reversible EP/G009546/1) for funding. The Toulouse and Moscow authors
coordination of pyridine to A and B while perhaps the excess are grateful to the CNRS (Centre National de la Recherche
pyridine could also assist in a faster collapse of the two Scientifique) and the RFBR (Russian Foundation for Basic
hydrides into the H2 ligand (for instance through the effect of Research) for support through the GDRI (Groupe de Recherche
H-bonding).
Internationale) “Homogeneous catalysis for Sustainable Devel-
Reversal of the relative rates of site exchange requires a new opment” and RFBR grants no. 12-03-93112, 12-03-31326, and
isomerisation pathway, resulting in selective migration of the 12-03-33018. We are also grateful to the IUF (Institut Universi-
axial H ligand in the square pyramidal geometry of A and B taire de France) for additional funding and to the CINES
without moving the second H ligand in the equatorial plane. (Centre Informatique National de l’Enseignement Supérieur)
One attractive possibility for this transformation is a selective and the CICT (Centre Interuniversitaire de Calcul de Toulouse,
deprotonation of the axial site in A by excess pyridine, which is project CALMIP) for granting free computational time. EMK
the strongest base present in solution, to yield a square planar thanks the Embassy of France in Moscow for a PhD grant.
monohydride intermediate [RhH(κ1-P,S)(py)2], followed by
reprotonation at the opposite face of the square plane to afford
B. This pathway is closely related to one of the proposed path-
ways for the H/D exchange between H2 and solvent in pure
CD3OH (through intermediates 5 and 6 of Scheme 3). No calcu-
Notes and references
lations were carried out, however, to confirm the feasibility of
this pathway. The fact that this strong effect is not observed
during the transformation of the minor isomer into the major
one (8 → 9) is consistent with the need to overcome the
smaller TSBD barrier in the first step to achieve the rapid site
scrambling. Hence, if the transition state of the putative site-
conserving concerted pathway is lower than TSAC but higher
than or comparable to TSBD, then the site exchange would
remain operative even at low pyridine concentration. For the
same reason, a very large concentration of pyridine does not
afford a selective site inversion because collapse to the non-
classical intermediate promotes nonselective deprotonation.
1 J. A. Osborn, F. H. Jardine, J. F. Young and G. Wilkinson,
J. Chem. Soc. A, 1966, 1711–1732.
2 R. H. Crabtree, H. Felkin and G. E. Morris, J. Organomet.
Chem., 1977, 141, 205–215.
3 S.-M. Lu, X.-W. Han and Y.-G. Zhou, Adv. Synth. Catal.,
2004, 346, 909–912.
4 S.-F. Zhu, J.-B. Xie, Y.-Z. Zhang, S. Li and Q.-L. Zhou, J. Am.
Chem. Soc., 2006, 128, 12886–12891.
5 T. Imamoto, N. Iwadate and K. Yoshida, Org. Lett., 2006, 8,
2289–2292.
6 M. N. Cheemala and P. Knochel, Org. Lett., 2007, 9, 3089–
3092.
7 G. Hou, F. Gosselin, W. Li, C. McWilliams, Y. Sun,
M. Weisel, P. D. O’Shea, C.-Y. Chen, I. W. Davies and
X. Zhang, J. Am. Chem. Soc., 2009, 131, 9882–9883.
8 W. J. Tang, J. Tan, L. J. Xu, K. H. Lam, Q. H. Fan and
A. S. C. Chan, Adv. Synth. Catal., 2010, 352, 1055–1062.
9 W. Tang, Y. Sun, L. Xu, T. Wang, Q. Fan, K.-H. Lam and
A. S. C. Chan, Org. Biomol. Chem., 2010, 8, 3464–3471.
Conclusions
The present contribution has explored the precatalyst acti-
vation phase for hydrogenation processes carried out in
alcohol solvents with [Rh(diene)(LL′)]+ systems, using [Rh-
(COD)(P,StBu)]+ as a model compound. The COD ligand is 10 D. Cartigny, T. Nagano, T. Ayad, J.-P. Genet, T. Ohshima,
removed by hydrogenation to yield a putative [Rh(P,StBu)-
(CH3OH)2]+ complex which promotes
very rapid H/D
K. Mashima and V. Ratovelomanana-Vidal, Adv. Synth.
Catal., 2010, 352, 1886–1891.
a
exchange between H2 and the solvent, possibly via a dihydride 11 R. Dorta, D. Broggini, R. Stoop, H. Ruegger, F. Spindler and
species [Rh(P,StBu)(H)2(CH3OH)2]+ that is accessible according
A. Togni, Chem.–Eur. J., 2004, 10, 267–278.
to the DFT calculations. Addition of L (pyridine or MeCN) 12 H.-U. Blaser, Adv. Synth. Catal., 2002, 344, 17–31.
slows down this exchange, allowing the observation of dia- 13 L. Routaboul, S. Vincendeau, J.-C. Daran and E. Manoury,
steromeric dihydride species at low temperature. Evidence has Tetrahedron: Asymmetry, 2005, 16, 2685–2690.
also been obtained for equilibrium deprotonation of these cat- 14 E. Le Roux, R. Malacea, E. Manoury, R. Poli, L. Gonsalvi
ionic dihydride complexes in the presence of strong bases (e.g.
and M. Peruzzini, Adv. Synth. Catal., 2007, 349, 309–313.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
15 R. Malacea, J.-C. Daran, S. B. Duckett, J. P. Dunne, 34 J. Lopez-Serrano, S. B. Duckett and A. Lledós, J. Am. Chem.
C. Godard, E. Manoury, R. Poli and A. C. Whitwood, Dalton
Trans., 2006, 3350–3359.
16 R. Malacea, E. Manoury, L. Routaboul, J.-C. Daran, R. Poli,
J. P. Dunne, A. C. Withwood, C. Godard and S. B. Duckett,
Eur. J. Inorg. Chem., 2006, 1803–1816.
Soc., 2006, 128, 9596–9597.
35 B. Eguillor, P. J. Caldwell, M. C. R. Cockett, S. B. Duckett,
R. O. John, J. M. Lynam, C. J. Sleigh and I. Wilson, J. Am.
Chem. Soc., 2012, 134, 18257–18265.
36 S. B. Duckett and C. J. Sleigh, Prog. Nucl. Magn. Reson. Spec-
trosc., 1999, 34, 71–92.
17 A. Preetz, H. J. Drexler, C. Fischer, Z. Dai, A. Borner,
W. Baumann, A. Spannenberg, R. Thede and D. Heller, 37 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
Chem.–Eur. J., 2008, 14, 1445–1451.
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers,
K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand,
K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene,
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
GAUSSIAN 09 (Revision A.01), Gaussian, Inc., Wallingford
CT, 2009.
18 A. Preetz, C. Fischer, C. Kohrt, H. J. Drexler, W. Baumann
and D. Heller, Organometallics, 2011, 30, 5155–5159.
19 E. Alberico, W. Baumann, J. G. de Vries, H. J. Drexler,
S. Gladiali, D. Heller, H. J. W. Henderickx and L. Lefort,
Chem.–Eur. J., 2011, 17, 12683–12695.
20 C. Fischer, S. Schulz, H. J. Drexler, C. Selle, M. Lotz, M. Sawall,
K. Neymeyr and D. Heller, ChemCatChem, 2012, 4, 81–88.
21 E. M. Kozinets, O. Koniev, O. A. Filippov, J.-C. Daran,
R. Poli, E. S. Shubina, N. V. Belkova and E. Manoury,
Dalton Trans., 2012, 41, 11849–11859.
22 E. M. Kozinets, N. V. Belkova, E. S. Shubina, R. Poli and
E. Manoury, Russ. Chem. Bull., 2013, 750–756.
23 S. B. Duckett and R. E. Mewis, Acc. Chem. Res., 2012, 45,
1247–1257.
24 R. A. Green, R. W. Adams, S. B. Duckett, R. E. Mewis,
D. C. Williamson and G. G. R. Green, Prog. Nucl. Magn.
Reson. Spectrosc., 2012, 67, 1–48.
25 C. R. Bowers and D. P. Weitekamp, J. Am. Chem. Soc., 1987, 38 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
109, 5541–5542.
26 T. C. Eisenschmid, R. U. Kirss, P. P. Deutsch,
39 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens.
Matter, 1988, 37, 785–789.
S. I. Hommeltoft, R. Eisenberg, J. Bargon, R. G. Lawler and 40 Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120,
A. L. Balch, J. Am. Chem. Soc., 1987, 109, 8089–8091. 215–241.
27 K. Golman and J. S. Petersson, Acad. Radiol., 2006, 13, 932– 41 D. Andrae, U. Haussermann, M. Dolg, H. Stoll and
942. H. Preuss, Theor. Chim. Acta, 1990, 77, 123–141.
28 D. Blazina, S. B. Duckett, P. J. Dyson, B. F. G. Johnson, 42 U. Haussermann, M. Dolg, H. Stoll, H. Preuss,
J. A. B. Lohman and C. J. Sleigh, J. Am. Chem. Soc., 2001,
123, 9760–9768.
P. Schwerdtfeger and R. M. Pitzer, Mol. Phys., 1993, 78,
1211–1224.
29 D. Blazina, S. Duckett, P. Dyson and J. Lohmann, Chem.– 43 W. Kuchle, M. Dolg, H. Stoll and H. Preuss, J. Chem. Phys.,
Eur. J., 2003, 9, 1046–1061. 1994, 100, 7535–7542.
30 M. A. M. Al-Ibadi, S. B. Duckett and J. E. McGrady, Dalton 44 T. Leininger, A. Nicklass, H. Stoll, M. Dolg and
Trans., 2012, 41, 4618–4625. P. Schwerdtfeger, J. Chem. Phys., 1996, 105, 1052–1059.
31 D. Schott, C. J. Sleigh, J. P. Lowe, S. B. Duckett, R. J. Mawby 45 A. W. Ehlers, M. Boehme, S. Dapprich, A. Gobbi,
and M. G. Partridge, Inorg. Chem., 2002, 41, 2960–2970.
32 D. Schott, P. Callaghan, J. Dunne, S. B. Duckett, C. Godard,
J. M. Goicoechea, J. N. Harvey, J. P. Lowe, R. J. Mawby,
A. Hoellwarth, V. Jonas, K. F. Koehler, R. Stegmann,
A. Veldkamp and G. Frenking, Chem. Phys. Lett., 1993, 208,
111–114.
G. Müller, R. N. Perutz, R. Poli and M. K. Whittlesey, Dalton 46 J. Fritsch and G. Zundel, J. Phys. Chem., 1981, 85, 556–561.
Trans., 2004, 3218–3224.
33 J. Lopez-Serrano, S. B. Duckett, S. Aiken, K. Q. Almeida
47 R. Kramer and G. Zundel, J. Chem. Soc., Faraday Trans.,
1990, 86, 301–305.
Lenero, E. Drent, J. P. Dunne, D. Konya and 48 M. Martin, E. Sola, O. Torres, P. Plou and L. A. Oro, Organo-
A. C. Whitwood, J. Am. Chem. Soc., 2007, 129, 6513–6527.
metallics, 2003, 22, 5406–5417.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans.