(POP)Rh pincer hydride complexes: Unusual reactivity and selectivity in oxidative addition and olefin insertion reactions

Article abstract of DOI:10.1039/c3sc50380a

We report on the synthesis and reactivity of rhodium complexes featuring bulky, neutral pincer ligands with a "POP" coordinating motif, ^tBuxanPOP, ^iPrxanPOP, and ^tBufurPOP ( ^tBuxanPOP = 4,5-bis(di-tert-butylphosphino)-9,9-dimethyl-9H-xanthene; ^iPrxanPOP = 4,5-bis(diisopropylphosphino)-9,9-dimethyl-9H-xanthene; ^tBufurPOP = 2,5-bis((di-tert-butylphosphino)methyl)furan). The (POP)Rh complexes described in this work are, in general, more reactive than their (PNP)Rh and (PCP)Rh analogues, which allows for the generation of several new species under relatively mild conditions. Thus, monomeric (POP)RhCl complexes oxidatively add H₂ to form (POP)Rh(H)₂Cl, from which the coordinatively unsaturated hydride complexes (POP)Rh(H) ₂⁺ and (^tBuxanPOP)Rh(H) can be obtained. In the case of the new ligand ^tBufurPOP, a major kinetic product of the reaction with H₂ is, surprisingly, the trans dihydride, i.e. trans-(^tBufurPOP)Rh(H)₂Cl; this is most likely attributable to reversible decoordination of one of the pincer coordinating groups, followed by addition of H₂ to a highly reactive three-coordinate species. Ethylene is hydrogenated by (^tBuxanPOP) Rh(H)₂⁺ at 25 °C, but propylene is not, even at elevated temperatures. Ethylene undergoes insertion into the Rh-H bond of ( ^tBuxanPOP)RhH; this reaction is reversible, allowing for an experimental determination of the equilibrium constant for this hydrometalation. The less bulky ^iPrxanPOP ligand affords a dihydride complex which functions as a modestly active alkane dehydrogenation catalyst, the first such example for a cationic pincer complex of any metal.

Full text of DOI:10.1039/c3sc50380a

Chemical Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
(POP)Rh pincer hydride complexes: unusual reactivity
and selectivity in oxidative addition and olefin insertion
reactions†
Cite this: Chem. Sci., 2013, 4, 3683
*
Michael C. Haibach, David Y. Wang, Thomas J. Emge, Karsten Krogh-Jespersen
and Alan S. Goldman
*
We report on the synthesis and reactivity of rhodium complexes featuring bulky, neutral pincer ligands
with a “POP” coordinating motif, tBuxanPOP, iPrxanPOP, and tBufurPOP (tBuxanPOP ¼ 4,5-bis(di-tert-
butylphosphino)-9,9-dimethyl-9H-xanthene; iPrxanPOP ¼ 4,5-bis(diisopropylphosphino)-9,9-dimethyl-9H-
xanthene; tBufurPOP ¼ 2,5-bis((di-tert-butylphosphino)methyl)furan). The (POP)Rh complexes described
in this work are, in general, more reactive than their (PNP)Rh and (PCP)Rh analogues, which allows
for the generation of several new species under relatively mild conditions. Thus, monomeric (POP)
RhCl complexes oxidatively add H2 to form (POP)Rh(H)2Cl, from which the coordinatively unsaturated
hydride complexes (POP)Rh(H)2 and (tBuxanPOP)Rh(H) can be obtained. In the case of the new ligand
+
tBufurPOP, a major kinetic product of the reaction with H2 is, surprisingly, the trans dihydride, i.e. trans-
(
tBufurPOP)Rh(H)2Cl; this is most likely attributable to reversible decoordination of one of the pincer
coordinating groups, followed by addition of H2 to a highly reactive three-coordinate species. Ethylene
is hydrogenated by (tBuxanPOP)Rh(H)2 at 25 ꢀC, but propylene is not, even at elevated temperatures.
Ethylene undergoes insertion into the Rh–H bond of (tBuxanPOP)RhH; this reaction is reversible, allowing
+
Received 8th February 2013
Accepted 1st July 2013
for an experimental determination of the equilibrium constant for this hydrometalation. The less bulky
DOI: 10.1039/c3sc50380a
iPrxanPOP ligand affords
a dihydride complex which functions as a modestly active alkane
www.rsc.org/chemicalscience
dehydrogenation catalyst, the first such example for a cationic pincer complex of any metal.
temperature the preferred olen metathesis catalysts have
limited lifetimes. Partly for this reason, we are interested in the
Introduction
Since their introduction by Shaw in 1976,1 complexes featuring
pincer ligands have received much attention for their applica-
tions in fundamental organometallic chemistry, catalysis, and
materials chemistry.2 One particularly attractive application is in
catalytic dehydrogenation, where pincer–Ir and Ru complexes
occupy “privileged” status. For the most part, pincer–Ru chem-
istry has found applicability in the dehydrogenation of func-
tionalized substrates3 and pincer–Ir chemistry in alkane
dehydrogenation.4 (For a notable exception, see recent work by
Roddick on alkane dehydrogenation by (CF3PCP)Ru(H).5) We
have reported a tandem catalytic system for alkane metathesis
that uses PCP- and POCOP-ligated Ir catalysts (Fig. 1) for dehy-
drogenation and Schrock-type W and Mo olen metathesis
catalysts.6 Currently, one bottleneck to improving catalytic
activity in this system is catalyst interoperability; the pincer–Ir
development of catalysts for alkane transfer dehydrogenation
that operate at lower temperatures.
Fig. 1 (PCP)M and (POCOP)M fragments.
ꢀ
catalysts operate best at temperatures above 150 C, at which
Department of Chemistry and Chemical Biology, Rutgers, The State University of New
Jersey, New Brunswick, New Jersey 08903, USA. E-mail: kroghjes@rutgers.edu; alan.
goldman@rutgers.edu
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c3sc50380a
Fig. 2 Selected PNP pincer complexes of rhodium.
Chem. Sci., 2013, 4, 3683–3692 | 3683
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Chemical Science
Edge Article
Fig. 3 Some previously reported (POP)Rh complexes.
Since alkane transfer dehydrogenation is approximately ther- oxidative addition thermodynamics were detailed recently by
moneutral there is no intrinsic barrier to development of a low- Heinekey and Brookhart for the related (PONOP)Rh(H2)+
temperature alkane dehydrogenation catalyst. Earlier work in our complex.14 Most recently, Brookhart has prepared rhodium
group ledtothe development ofefficient Rh-basedprecatalysts for complexes of anionic PNP ligands derived from pyrrole and
thermochemical alkane dehydrogenation, (PMe3)2RhClL and carbazole.15
[(PMe3)ꢀ2Rh(m-Cl)]2, which operated at relatively low temperature
Considering the lesser trans-inuence of ethers versus N-
(ca. 50 C).7 This class of complexes requires H2 atmosphere to ligating groups, we decided to investigate the relatively unex-
continuously regenerate
a
catalytically active species, plored (POP)Rh class of complexes. In 1980, Timmer et al.
(PMe3)2RhCl(H)2, by cleavage of either the Rh–L bond or the di(m- investigated the reaction of O(CH2CH2PtBu2)2 (POetherP) with a
chloride) bridge. The presence of H2 results in loss of multiple variety of rhodium precursors.16 While these workers were
moles of acceptor (through simple hydrogenation) per mol of unable to form complexes such as (POetherP)RhCl or (POetherP)
substrate dehydrogenated, thereby limiting practical applica- Rh(olen)+, the pincer complex [(POetherP)Rh(CO)]OTf (Fig. 3)
tions. However, the fast kinetics of these systems are remarkable, was isolated and characterized by NMR and IR. This was the sole
particularly in view of the fact that most of the rhodium is in an report of a pincer (POP)Rh complex until recently, when several
inactiveout-of-cycle stateatanytime;thissuggestedthatwemight groups reported pincer rhodium complexes of Xantphos. Weller
attempt to emulate the stereoelectronic features of the reported Xantphos (PhxanPOP) rhodium complexes (Fig. 3)
(PMe3)2RhCl fragment in a pincer–Rh complex, with the hope of bearing a P(cyclo-C5H9)3 ligand and found the Xantphos ligand
creatingananaloguewhichdoesnotsufferfromdimer formation. equilibrating between fac and mer coordination modes.17 Haynes
Initial efforts towards this goal focused on the (PCP)Rh reported that (PhxanPOP)Rh iodide complexes catalyzed meth-
framework;8 however, these complexes showed very low activity anol carbonylation18 while Hartwig reported that the pincer
for alkane dehydrogenation.9 We soon realized that oxidative complex (Me2NxanPOP)Rh(NCMe)+ catalyzes the intramolecular
addition to (PCP)Rh was much less favorable than addition to hydroamination of simple aminoalkenes.19 In this article we
the catalytically active (PMe3)2RhCl fragment.9 In order to report the synthesis and novel reactivity of several neutral and
understand the factors that govern the thermodynamics of H–H cationic (POP)Rh complexes, particularly hydride complexes,
and C–H addition to these and related 3-coordinate complexes using a range of bulky, neutral POP ligands, and we discuss their
more generally, we have undertaken a combined theoretical and relevance to rhodium catalyzed C–H activation.
experimental study.10 We found that in (PXP)Rh and Ir pincer
complexes, the trans inuence of the X-bound coordinating
group exerts a very strong inuence on oxidative addition
Results and discussion
thermodynamics. Strongly sigma-donating ligands were found
to strongly disfavor oxidative addition, in contrast with the
conventional view that electron-rich metal centers favor oxida-
tive addition. The magnitude of this effect is easily large enough
that a PXP ligand with decreased trans inuence relative to the
aryl group of PCP could be expected to afford a (PXP)Rh complex
that could undergo addition of alkane C–H bonds and possibly
act as a catalyst for alkane dehydrogenation.10
In light of the recent work reported with Xantphos derivatives,
we decided to select tBuxanPOP as a starting point for this study
(Fig. 4). tBuxanPOP features the bulky (dimerization-inhibiting)
and cyclometalation-resistant PtBu2 groups that lend stability to
the (PCP)Ir catalysts.4 While it is commercially available,
tBuxanPOP has seen limited application since its synthesis in
PNP complexes of rhodium have been investigated by several
groups, particularly in the last few years (Fig. 2). Milstein repor-
ted the synthesis of the cationic (PNpyP)Rh(C2H4)+ complex and
derivatives in his seminal report on the pyridine-based PNP
ligand.11 Rhodium complexes of the anionic PNamidoP ligand,12
have been synthesized and investigated in depth by Ozerov.
Weller recently reported a variety of (PNpyP)Rh(L)+ complexes;13
notably, the (PNpyP)Rh fragment was found not to oxidatively
add H2, instead forming the dihydrogen complex. Similar Fig. 4 Bulky, neutral POP ligands studied in this work.
3684 | Chem. Sci., 2013, 4, 3683–3692
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
(1)
2005 by Perrio.20 For reasons discussed below, we also studied
the analogous iPrxanPOP (Fig. 4) recently prepared by Asensio
and Esteruelas in their study of (POP)Os and (POP)Ru
complexes.21
Given the high degree of crowding engendered by the
xanthene ligand backbone, we also designed a furan-based
ligand tBufurPOP.22 Surprisingly, no phosphine pincer ligands
using a furan backbone have been reported previously in the
literature although 4,6-bis(phosphino)dibenzofurans23 have
been investigated, including 4,6-bis(diisopropylphosphino)
dibenzofuran, reported by Asensio and Esteruelas.21,24 PhfurPOP
(
is claimed in two 1980s patents dealing explicitly with bidentate
ligands although no synthetic procedure is given.25) Aer some
fruitless experimentation with attempted lithiation or free-
radical bromination of 2,5-dimethylfuran, we arrived at a viable
synthetic route to tBufurPOP, shown in eqn (1). Commercially
available 2,5-bis(hydroxymethyl)furan is converted to 2,5-bis-
(chloromethyl)furan according to Cook’s protocol.26 Phosphi-
nation occurs under mild conditions (owing to the high reactivity
of the bis(chloromethyl)furan), and aer deprotonation of the
bis(hydrochloride) salt, tBufurPOP is obtained as an air-sensitive
solid in acceptable yield and high purity.
Fig. 6 Initial metalation attempts and unexpected formation of (tBufurPOP)
Rh(H)2Cl.
hydrogenating the NBD ligands (Fig. 6). Unfortunately, heating
a solution of tBuxanPOP and [Rh(NBD)Cl]2 under H2 above 40 ꢀC
resulted in the formation of rhodium black within minutes; no
metalated product was observed by 31P NMR spectroscopy.
The same problem was not encountered with tBufurPOP,
however, probably due to faster initial coordination with this
more exible ligand. Thus, reuxing tBufurPOP and [Rh(NBD)Cl]2
in toluene for 48 hours under H2 led to the formation of a meta-
lated product in 54% yield. Surprisingly, the product was not the
expected (tBufurPOP)RhCl species but rather cis-(tBufurPOP)
Rh(H)2Cl, as indicated by 1H and 31P NMR spectroscopy.
When the analogous reaction was carried out with the PNpyP
ligand, (PNpyP)RhCl (ref. 11) was obtained in quantitative yield,
suggesting that the stability of the tBufurPOP-ligated H2 addition
product derives from the reduced trans inuence of furPOP
relative to PNpyP. Nozaki has shown that cis-(PNpyP)Ir(H)2Cl
forms from PNpyP and [Ir(COE)2Cl]2 under similar conditions.28
Hence, the (tBufurPOP)Rh system seems closer in character to
the PNpyP iridium complex in this respect than to the previously
reported PNpyP rhodium systems.
With the t-Bu2P-substituted POP ligands in hand, we inves-
tigated their reactivity with Rh(I) precursors. Initial attempts at
metalation of tBuxanPOP with [Rh(NBD)Cl]2 (NBD ¼ norborna-
diene) resulted in no complexation, even at elevated tempera-
tures. tBufurPOP reacted with [Rh(NBD)Cl]2 at 25 ꢀC to yield the
undesired bimetallic complex (m2-tBufurPOP)[Rh(NBD)Cl]2
which was crystallographically characterized (Fig. 5). Timmer
obtained a similar complex in the reaction of POetherP with
[Rh(COD)Cl]2.16
As in our group’s modication of the synthesis of (tBuPCP)
Ir(H)Cl (ref. 27) we attempted to use H2 to drive the reaction by
Using the more labile [Rh(COE)2Cl]2 precursor in lieu of
[Rh(NBD)Cl]2, we could obtain both (tBuxanPOP)RhCl and
(
tBufurPOP)RhCl in high yield (Fig. 7). Both are monomeric
pincer complexes as revealed by their X-ray structures (Fig. 8).
The slightly more acute O–Rh–P angles in (tBufurPOP)RhCl
relative to (tBuxanPOP)RhCl (average of 82.2ꢀ vs. 83.6ꢀ) reect
the difference between the ve- and six-member oxygen--
containing rings. In both cases the coordination environment is
almost perfectly planar (the sum of the four cis-L–Rh–L angles is
361ꢀ for both complexes).
Both (tBuxanPOP)RhCl and (tBufurPOP)RhCl add H2 at room
temperature to afford the corresponding (POP)Rh(H)2Cl
Fig. 5 X-ray structure of (m2-tBufurPOP)[Rh(NBD)Cl]2.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 3683–3692 | 3685

View Article Online
Chemical Science
Edge Article
In contrast with the tBuxanPOP complex, when an atmosphere
of H2 is added to a toluene solution of (tBufurPOP)RhCl at 25 ꢀC,
1H and 31P NMR spectroscopy reveal full conversion to a 7:3
mixture of cis:trans dihydride isomers (see Experimental section
in ESI†). DFT calculations (see Computational Methods in ESI†)
show that the cis isomer is considerably more stable thermody-
namically (by ca. 23 kcal molꢁ1) and indeed, heating the mixture
for 12 h at 110 ꢀC in toluene results in complete conversion to the
cis isomer. Net trans-addition of H2 has been reported with (PNP)
Ir(Ph) by Milstein29 and with (PONOP)Ir(CH3) by Brookhart.30
Milstein has proposed that addition to (PNP)Ir(Ph) proceeds via
initial proton transfer from the PNP ligand to the metal center to
form the dearomatized pincer complex (PNP*)Ir(H)(Ph), fol-
lowed by H2 addition across the metal and PNP* ligand (trans to
the hydride ligand) to form trans-(PNP)Ir(H)2(Ph). The PONOP
ligand cannot undergo a similar deprotonation; hence proton-
ation to give [(PNP)Ir(H)(CH3)]X was proposed. Subsequent
coordination of H2 is then followed by deprotonation by Xꢁ to
Fig. 7 Synthesis of monomeric (POP)RhCl complexes.
complexes in high yield (Fig. 9). (tBuxanPOP)Rh(H)2Cl is
obtained as the cis dihydride isomer, as evident from the signals
due to inequivalent hydrides and t-Bu groups in the H NMR
spectrum as well as its X-ray structure (Fig. 10).
1
Fig. 8 X-ray structures of (tBuxanPOP)RhCl and (tBufurPOP)RhCl.
Fig. 9 Oxidative addition of H2 by (POP)RhCl complexes.
3686 | Chem. Sci., 2013, 4, 3683–3692
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
+
Fig. 12 Formation of the cationic hydrides (POP)Rh(H)2
.
Both complexes feature a dt at ca. d ꢁ21 ppm in the 1H NMR
spectrum, attributable to the equivalent hydride ligands, with
similar values of JPH (ca. 11 Hz) and particularly high values of
JRhH (ca. 40 Hz). These parameters are all indicative of classical
dihydride complexes with strong Rh–H bonds, in contrast with
the analogous neutral PCP9,32 and POCOP,14 and cationic PNP13
and PONOP14 rhodium dihydrogen complexes, all of which have
pincer ligands with central coordinating groups more strongly
sigma-donating (stronger trans-inuence) than the POP oxygen
atoms. Indeed, the addition of H2 to the [(tBuxanPOP)Rh]+ frag-
ment is calculated (DFT) to be extremely exothermic; DH ¼ ꢁ37.2
kcal molꢁ1. This compares with a value of ꢁ27.0 kcal molꢁ1
calculated for H2 addition to the archetypal pincer-ligated
dehydrogenation catalyst, the iridium complex (tBuPCP)Ir.
The t-Bu groups of the respective POP complexes each afford a
single signal in the 1H NMR spectra indicating their equivalency
(see ESI†). An X-ray structure of [(tBuxanPOP)Rh(H)2]SbF6 (Fig. 13)
did not locate the hydride ligands with high certainty, but
conrmed that no solvent or other ligand was coordinating to Rh.
The chemical shi and coupling constant of the hydride
ligands are not signicantly affected by variations of solvent,
including acetone, THF, or uorobenzene, further supporting
the conclusion that the complexes are coordinatively unsatu-
rated dihydrides in these solutions as well as in the solid state.
Fig. 10 X-ray structure of cis-(tBuxanPOP)RhH2Cl.
form the product. In the present case, we have neither a ligand
capable of undergoing Milstein-type dearomatization, nor a
protic solvent presumably required for the Brookhart mecha-
nism. A possible explanation is that (tBufurPOP)RhCl is hemi-
labile under the reaction conditions, undergoing reversible
decoordination of one of the coordinating groups, most likely
the weakly bound furan group (although dissociation of a
phosphino group is also possible). Addition of H2 to a three-
coordinate (k2-tBufurPOP)RhCl intermediate (Fig. 11) is expected
to be much faster than addition to the four-coordinate k3
complex,31 and would result in a ve-coordinate species to which
the furan oxygen could recoordinate to afford the trans adduct.
Initial formation of the trans adduct (ca. 30% of the kinetic
distribution) might also occur via this three-coordinate inter-
mediate, or via a conventional mechanism of concerted cis
addition to the four-coordinate (k3-tBufurPOP) complex, or a
combination of these two possibilities.
With convenient accessto (tBuxanPOP)Rh(H)2Cland(tBufurPOP)
Rh(H)2Cl we targeted formation of the (POP)RhH2 cationic
+
complexes which are isoelectronic to the catalytically active
complexes (PCP)IrH2.4,6 Both (tBufurPOP)Rh(H)2+ and (tBuxanPOP)
+
H2O, however, does coordinate; trans-(tBuxanPOP)Rh(OH2)(H)2
+
exhibits a broad singlet at d ꢁ9 ppm in the 1H NMR spectrum,
Rh(H)2 can be formed in high yield by chloride abstraction
with AgBF4 or AgSbF6 in acetone, THF, or CH2Cl2 (Fig. 12).
suggesting that the hydrides are rapidly exchanging with the
Fig. 11 Proposed mechanism for trans H2 addition to (tBufurPOP)RhCl, and potential mechanism for isomerization.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 3683–3692 | 3687

View Article Online
Chemical Science
Edge Article
(norbornene) (0.01–0.2 M), even at temperatures up to 100 ꢀC, in
alkane or uorobenzene solution. For comparison, (tBuPCP)IrH2
is readily dehydrogenated by NBE or TBE at ambient or near-
ambient temperatures respectively.33 Likewise (tBuxanPOP)-
RhH2 and (tBufurPOP)RhH2 underwent no reaction with
+
+
1-hexene (0.2 M) for over 48 hours at 100 C (whereas (tBuPCP)
ꢀ
IrH2 is dehydrogenated immediately upon mixing with 1-hexene
at room temperature). However, we found that [(tBuxanPOP)
RhH2]+ reacts with ethylene to yield the cationic [(tBuxanPOP)
Rh(C2H4)]+ (conrmed by independent synthesis from tBux-
+
anPOP and Rh(C2H4)2(solv.)2 (ref. 34)) and ethane at room
temperature within 1 hour. Conversely, [(tBuxanPOP)Rh(C2H4)]+
also reacts with H2 at room temperature to regenerate the dihy-
dride complex, reaching completion aer ca. 12 hours (Fig. 15).
Propene, like 1-hexene, TBE and NBE, also fails to react with
(
tBuxanPOP)Rh(H)2+. This exclusive selectivity for hydrogenation
Fig. 13 X-ray structure of the cation of [(tBuxanPOP)Rh(H)2]SbF6.
of ethylene (probably largely due to steric factors, but perhaps
also attributable to the more favorable thermodynamics of
ethylene hydrogenation35) is to our knowledge without close
precedent.
Given the failure of both (tBuxanPOP)Rh(H)2+ and (tBufurPOP)-
Rh(H)2+ to react with either TBE or 1-olens, it is not surprising
that neither is active for catalytic transfer-dehydrogenation of
alkanes using these olens. The use of ethylene as a hydrogen
acceptor was also unsuccessful, also unsurprisingly in view of
its ability to coordinate much more strongly than more bulky
olens.
+
In view of the unusual lack of reactivity of (tBufurPOP)Rh(H)2
and (tBuxanPOP)Rh(H)2 towards olens bulkier than ethylene
+
we investigated the less bulky iPrxanPOP ligand, reported
recently by Asensio and Esteruelas.21 cis-(iPrxanPOP)Rh(H)2Cl
could be prepared analogously to the tBuxanPOP complex,
although the product appears to be in slow equilibrium with a
species that is possibly dimeric (or oligomeric). The monomeric
cis-(iPrxanPOP)Rh(H)2Cl complex exhibits two dtd signals at d
ꢁ17.46 and ꢁ20.07 in the 1H NMR spectrum, similar to the
signals at d ꢁ17.02 and ꢁ20.51 observed for (tBuxanPOP)-
Rh(H)2Cl. The suspected dimer or oligomer exhibits a very
complex signal in both the 1H and 31P NMR spectra, which
could not be effectively decoupled. We observed no trans-H2
addition to (iPrxanPOP), suggesting that this phenomenon
derives from the exibility of the methylene-linked backbone of
Fig. 14 X-ray structure of the cation of [trans-(tBuxanPOP)Rh(H)2(OH2)]SbF6.
H2O protons; the structure of this adduct was determined
crystallographically (Fig. 14).
The cationic dihydrides are potential precursors to the cor-
responding 14-electron (POP)Rh+ complexes; accordingly we
investigated their reaction with common hydrogen acceptors.
Probably due in large part to the high endothermicity of H2 loss
(calculated at 37.2 kcal molꢁ1 as noted above) neither
(
tBufurPOP), rather than from the steric environment at the
metal center.
The cationic dihydrides discussed above show limited solu-
bility in alkane solvents at lower temperatures. Toward the goal
of evaluating activity for alkane dehydrogenation in mind, we
+
+
(
tBuxanPOP)Rh(H)2 nor (tBufurPOP)Rh(H)2 reacted with the
commonly used acceptors TBE (t-butylethylene) or NBE
+
Fig. 15 Reactions of (tBuxanPOP)Rh(H)2 with ethylene and (tBuxanPOP)Rh(ethylene)+ with dihydrogen.
3688 | Chem. Sci., 2013, 4, 3683–3692
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
+
Fig. 16 Preparation of (iPrxanPOP)Rh(H)2
.
therefore employed the more lipophilic B(C6F5)4 anion for this
complex. Treatment of the mixture of cis-(iPrxanPOP)Rh(H)2Cl
and the putative dimer with LiB(C6F5)4$OEt2 in CD2Cl2 led to
+
the clean formation of (iPrxanPOP)Rh(H)2 which exhibited a
sharp dt at ꢁ23.93 ppm with JPH ¼ 15 Hz, and JRhH ¼ 44 Hz (2H),
in good agreement with those values obtained with (tBuxanPOP)
Rh(H)2+ (Fig. 16).
Unlike either of the tBuPOP complexes, (iPrxanPOP)Rh(H)2+ is
catalytically active in the transfer dehydrogenation of COA.
Under our most commonly employed conditions, TBE (0.2 M) as
an acceptor and 1 mM catalyst in COA, formation of 47 mM COE
+
is observed aer 10 h at 150 C. (iPrxanPOP)Rh(H)2 is, to our
ꢀ
knowledge, the rst cationic pincer complex reported to cata-
lyze alkane dehydrogenation.
A four-coordinate rhodium(I) monohydride
Fig. 18 X-ray structure of (tBuxanPOP)RhH.
In part interested by the fact that (POP)Rh(H)3 would be
isoelectronic to Nozaki’s (PNP)Ir(H)3catalyst28 for hydrogena-
tion of CO2, we investigated the dehydrohalogenation of
(POP)Rh(H)2Cl complexes. The reaction of (tBuxanPOP)Rh(H)2Cl
with KOtBu in benzene leads to a color change from orange to
red within seconds.
therefore assigned as the monohydride (tBuxanPOP)RhH (Figs. 17
and 18), which is a rare example of a 4-coordinate Group 9 metal
hydride.22,36 The only other example of such a rhodium complex
The 1H NMR spectrum of the red product solution indicates to our knowledge, (NHCPCP)RhH, was recently reported by Fry-
that all 4 t-Bu groups are equivalent, and a single dt signal inte- zuk.37 This complex was found to be thermally unstable above
grating to 1H is observed in the hydride region. The selectively 70 C,37whereas (tBuxanPOP)RhH shows no appreciable decom-
ꢀ
hydride-coupled 31P signal appears as a dd. These NMR data are
all indicative of a product with a single hydride ligand trans to the
POP oxygen atom. X-rayanalysis of the productconrmed thatno
additional ligands are bound to the Rh center; the product is
position in solutions at 150 ꢀC for several hours.
No evidence for addition of H2 is observed upon the addition
of 1 atm H2 to a C6D6 solution of (tBuxanPOP)RhH at room
temperature. This result stands in marked contrast with the
behavior of the chloride analog, (tBuxanPOP)RhCl, which readily
adds H2, with no observable concentration of four-coordinate
species remaining in solution, as discussed above. Accordingly,
cis H2 addition to (tBuxanPOP)Rh(X) is calculated (DFT) to be
favorable with DGꢀ
¼ ꢁ5 kcal molꢁ1 for X ¼ Cl, but unfa-
298K
vorable for X ¼ H with DGꢀ
¼ +5 kcal molꢁ1. This large
298K
difference in reactivity represents another case where oxidative
addition is strongly disfavored by increased sigma-donating
ability of an ancillary ligand.
Fig. 17 Formation of a square-planar rhodium monohydride, (tBuxanPOP)RhH.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 3683–3692 | 3689

View Article Online
Chemical Science
Edge Article
Addition of 1 atm ethylene to a solution of (tBuxanPOP)RhH
results in insertion into the Rh–H bond to give (tBuxanPOP)RhEt
within ca. 1 minute (eqn (2)). The methylene protons of the ethyl
1 atm propene
ð
tBuxanPOPÞRhH
NR
C
(5)
ƒƒƒƒƒƒƒƒƒƒƒ!
C6D6; 25 ꢀCꢁ100 ꢀ
2
ligand appear as a qd (d 2.21; JRhH ¼ 2.4 Hz, 3JHH ¼ 7.5 Hz).
(
tBuxanPOP)RhCl + n-BuLi / (tBuxanPOP)Rh(n-Bu) + LiCl (6)
1atm C2H4
tBuxanPOP)Rh(n-Bu) underwent b-H elimination to give
tBuxanPOP)RhH (eqn (7)), with a half-life on the order of 1 hour,
ð
tBuxanPOPÞRhH
ð
tBuxanPOPÞRhðC2H5Þ
(2)
ƒƒƒƒƒƒƒƒƒ!
(
C6D6; 25 ꢀC; 1 min
(
the reaction proceeding to give complete conversion to the
hydride within 18 hours. Thus, the failure of (tBuxanPOP)RhH to
insert higher olens is due to thermodynamics, not kinetics.
Addition across the double bond of a terminal alkene is
generally less favorable than addition to ethylene (e.g.
DH ¼ ꢁ29.9 kcal molꢁ1 for hydrogenation of propene or
1-butene, vs. –32.6 kcal molꢁ1 for ethylene35). Given our deter-
A similar insertion reaction, in the case of the Rh(III)
complex (PPh3)2Rh(H)Cl2 was reported in 1967 by Wilkinson.38
The resulting (PPh3)2Rh(Et)Cl2 (eqn (3)) was found to be
stable to air.
(PPh3)2Rh(H)Cl2 + C2H4 / (PPh3)2Rh(C2H5)Cl2
(3)
mination of DGinsertion
¼ ꢁ2.0 kcal molꢁ1 for ethylene
298K
In contrast, the Rh(I) complex (tBuxanPOP)RhEt species
loses ethylene when pumped dry, to reform (tBuxanPOP)RhH.39
This system thus afforded an unusual opportunity to deter-
mine the thermodynamics of insertion of an olen into a
late-metal hydride bond. The thermodynamics of olen
insertion into M–H bonds are of signicance in the context of
alkane dehydrogenation, olen hydrogenation, and numerous
transition-metal catalyzed additions of H–X to olen. Yet
reports of absolute thermodynamic data for olen insertion
into metal–hydride bonds are rare. The only other examples
of which we are aware, reported by Chirik and Bercaw, are
for insertion into a Zr–H bond of bis(cyclopentadienyl)-
ZrH2 complexes.40 For (C5Me5)2ZrH2 it was found that DG ¼
ꢁ4.6 kcal molꢁ1 for insertion of isobutene and ꢁ6.9 kcal
molꢁ1 for 1-butene; for the slightly less crowded (C5Me5)-
(C5Me4H)ZrH2 the respective values are ꢁ8.5 kcal molꢁ1 and
insertion, this difference of ca. 2.7 kcal molꢁ1 can largely
explain the great difference in reactivity between ethylene and
higher olens toward (tBuxanPOP)RhH. Note that the thermo-
dynamics of insertion in this system are much less favorable
than those for the bis(cyclopentadienyl)ZrH2 complexes noted
above40 (ꢁ6.9 kcal molꢁ1 and ꢁ10.6 kcal molꢁ1 for 1-butene
insertion). Future studies will address the question of electronic
and/or steric components of this difference.
(
tBuxanPOP)Rh(n-Bu) / (tBuxanPOP)RhH + butenes
(7)
While (tBuxanPOP)RhH does not react with 1-hexene to give
an observable insertion product, it is highly active as an olen
isomerization catalyst (eqn (8)). For example, in a solution of
1 mmol (4 mM) (tBuxanPOP)RhH in 2 mmol 1-hexene (neat),ꢀan
initial TOF of 460 hꢁ1 for isomerization is obtained at 50 C.
Aer 16 hours, nearly 2000 TON were obtained. Moderate
selectivity for formation of trans-2-hexene is observed.
ꢁ10.6 kcal molꢁ1
.
Keq¼0:033 M
*
ð
tBuxanPOPÞRhðC2H5Þ )
ð
tBuxanPOPÞRhH þ C2H4
ð
tBuxanPOPÞRhHð0:05 mol %Þ
1-hexene
ðneatÞ
C6D6; 25 ꢀC; 12 h
96% internal hexenes
ðca: 60% trans-2-hexeneÞ
initial TOF ¼ 460 hꢁ1
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
50 ꢀC; 16 h
(4)
(8)
A solution of (tBuxanPOP)RhEt in C6D6 was prepared under
ethylene atmosphere in a J-Young NMR tube. The solution was
then frozen in pentane/liquid N2 (ꢁ126 C) and the headspace
Summary
ꢀ
was evacuated on a high-vacuum line. The solution was thawed
and an NMR spectrum was immediately (ca. 1 min) recorded,
showing ca. 10% conversion into the (tBuxanPOP)RhH complex
and ethylene. Aer 12 h, the reaction had reached equilibrium
with [(tBuxanPOP)RhEt] ¼ 3.8 mM, [(tBuxanPOP)RhH] ¼ 16 mM,
and [C2H4] ¼ 7.8 mM, corresponding to Keq ¼ 0.033 M for
deinsertion or Keq ¼ 30.5 Mꢁ1 for insertion of ethylene into the
In conclusion, we have synthesized several (POP)Rh complexes
and explored their reactivity including oxidative addition of
hydrogen and hydrometalation of olens. Using the bulky,
neutral ligands, tBuxanPOP or tBufurPOP, monomeric pincer
(POP)RhCl species can be obtained. These four-coordinate
complexes add H2 to afford the cis dihydrides as the thermo-
dynamic products but (tBufurPOP)RhCl gives a trans dihydride
as a kinetic product; this likely proceeds via reversible decoor-
dination and addition to a highly reactive three-coordinate
Rh–H bond, corresponding to DGinsertion 298K ¼ ꢁ2.0 kcal molꢁ1
.
(
tBuxanPOP)RhH did not react with 1-hexene or propylene to
intermediate. (tBufurPOP)RhH2 and (tBufurPOP)RhH2 were
synthesized. Although they are isoelectronic to complexes
(PCP)IrH2 which are well known to catalyze alkane dehydroge-
nation,4 these cationic rhodium complexes showed no such
reactivity. The less crowded (iPrxanPOP)RhH2+, however, is
+
+
yield any observable rhodium-containing products at tempera-
tures ranging from ambient to 100 ꢀC (eqn (5)). To determine if
this failure to observe insertion products was due to kinetic or
thermodynamic factors we reacted
(
tBuxanPOP)RhCl with
approximately one half equivalent of n-BuLi, at 25 ꢀC. Over the
course of ca. 1 hour the reaction proceeded to completion
(consuming all the n-BuLi), generating (tBuxanPOP)Rh(n-Bu)
(eqn (6)), the homologue of (tBuxanPOP)RhEt.
found to catalyze cyclooctane/TBE transfer-dehydrogenation.
+
(
iPrxanPOP)RhH2 is the rst cationic pincer complex repor-
ted to catalyze alkane dehydrogenation and it offers a relatively
3690 | Chem. Sci., 2013, 4, 3683–3692
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
rare example of alkane dehydrogenation by a second row metal
complex. Notably, this activity has been achieved not by
increasing electron density on the metal center, but by the use
of a very poorly sigma electron-donating group at the central
coordinating position of the pincer ligand, i.e. an ether oxygen
as compared with nitrogen donors or the anionic aryl groups
found in pincer ligands of the PCP type. Presumably related to
this is the observation that, unlike cationic PNP or neutral PCP
rhodium fragments, the (POP)Rh unit oxidatively cleaves H2
rather than forming a dihydrogen complex. Indeed, it appears
that the thermodynamics of H2 oxidative addition (calculated as
exothermic by ca. 37 kcal molꢁ1) may be too favorable for
optimal catalytic activity, and that the thermodynamic difficulty
of removing hydrogen from (iPrxanPOP)RhH2+ limits the rate of
5 B. C. Gruver, J. J. Adams, S. J. Warner, N. Arulsamy and
D. M. Roddick, Organometallics, 2011, 30, 5133–5140.
6 A. S. Goldman, A. H. Roy, Z. Huang, R. Ahuja, W. Schinski
and M. Brookhart, Science, 2006, 312, 257–261.
7 (a) J. A. Maguire and A. S. Goldman, J. Am. Chem. Soc., 1991,
113, 6706–6708; (b) J. A. Maguire, A. Petrillo and
A. S. Goldman, J. Am. Chem. Soc., 1992, 114, 9492–9498; (c)
K. Wang, M. E. Goldman, T. J. Emge and A. S. Goldman, J.
Organomet. Chem., 1996, 518, 55–68.
8 S. Nemeh, C. Jensen, E. Binamira-Soriaga and W. C. Kaska,
Organometallics, 1983, 2, 1442–1447.
9 W. Xu, G. P. Rosini, M. Gupta, C. M. Jensen, W. C. Kaska,
K. Krogh-Jespersen and A. S. Goldman, Chem. Commun.,
1997, 2273–2274.
catalytic transfer-dehydrogenation. Work is currently underway 10 D. Y. Wang, Y. Choliy, K. Krogh-Jespersen, J. F. Hartwig and
on the application of new POP ligands with greater trans
inuence, which may more closely emulate the electronics of
the (PMe3)2RhCl system.41
A. S. Goldman, Abstracts of Papers, 240th ACS National
Meeting, Boston, MA, United States, INOR-5, 22–26 August
2010.
In addition, we report the formation of the stable four- 11 D. Hermann, M. Gandelman, H. Rozenberg, L. J. W. Shimon
coordinate hydride (tBuxanPOP)RhH. This complex reversibly
and D. Milstein, Organometallics, 2002, 21, 812–818.
hydrometalates ethylene, thus allowing an unusual direct 12 (a) L. Fan, B. M. Foxman and O. V. Ozerov, Organometallics,
determination of the thermodynamics of insertion of olen into
a metal–hydrogen bond.
2004, 23, 326–328; (b) L. Fan, L. Yang, C. Guo, B. M. Foxman
and O. V. Ozerov, Organometallics, 2004, 23, 4778–4787; (c)
O. V. Ozerov, C. Guo, L. Fan and B. M. Foxman,
Organometallics, 2004, 23, 5573–5580; (d) O. V. Ozerov,
C. Guo, V. A. Papkov and B. M. Foxman, J. Am. Chem. Soc.,
2004, 126, 4792–4793.
Acknowledgements
This work was supported by NSF through the CCI Center for
Enabling New Technologies through Catalysis (CENTC), CHE- 13 A. B. Chaplin and A. S. Weller, Organometallics, 2011, 30,
1205189 and CHE-0650456. M.C.H. is a graduate training fellow 4466–4469.
in the NSF IGERT for Renewable and Sustainable Fuels. We 14 M. Findlater, K. M. Schultz, W. H. Bernskoetter,
thank Johnson-Matthey for a gi of rhodium salts, and Penn A
A. Cartwright-Sykes, D. M. Heinekey and M. Brookhart,
Kem for a generous supply of 2,5-bis(hydroxymethyl)furan. We
Inorg. Chem., 2012, 51, 4672–4678.
also wish to acknowledge Chen Cheng, Dr Damien Guironnet, 15 M. Brookhart, C. Chang, B. G. Kim, A. Goldman and
and Prof. Maurice Brookhart (University of North Carolina –
D. Y. Wang, Abstracts of Papers, 243rd ACS National
Meeting, San Diego, CA, United States, INOR-1007, 25–29
March 2012.
´
Chapel Hill), and Dr Montserrat Olivan, and Prof. Miguel A.
Esteruelas (Universidad de Zaragoza) for helpful discussions
and sharing unpublished work.
16 K. Timmer, D. H. M. W. Thewissen and J. W. Marsman, Recl.
Trav. Chim. Pays-Bas, 1988, 107, 248–255.
17 R. Dallanegra, A. B. Chaplin and A. S. Weller,
Organometallics, 2012, 31, 2720–2728.
References
1 C. J. Moulton and B. L. Shaw, J. Chem. Soc., Dalton Trans., 18 G. L. Williams, C. M. Parks, C. R. Smith, H. Adams,
1976, 1020–1024.
A. Haynes, A. J. H. M. Meijer, G. J. Sunley and S. Gaemers,
2 (a) G. van Koten and D. Milstein in Organometallic Pincer
Organometallics, 2011, 30, 6166–6179.
Chemistry In Top. Organomet. Chem., Springer GmbH, 2013, 19 L. D. Julian and J. F. Hartwig, J. Am. Chem. Soc., 2010, 132,
vol. 40, p. 356; (b) D. Morales-Morales and C. Jensen in The 13813–13822.
Chemistry of Pincer Compounds, Elsevier, Amsterdam, 2007; 20 C. Mispelaere-Canivet, J.-F. Spindler, S. Perrio and P. Beslin,
(c) M. Albrecht and G. van Koten, Angew. Chem., Int. Ed.,
2001, 40, 3750–3781.
3 (a) C. Gunanathan and D. Milstein, Acc. Chem. Res., 2011, 44,
588–602; (b) C. Gunanathan and D. Milstein, Top. Organomet.
Tetrahedron, 2005, 61, 5253–5259.
21 G. Asensio, A. B. Cuenca, M. A. Esteruelas, M. Medio-Simon,
M. Olivan and M. Valencia, Inorg. Chem., 2010, 49, 8665–
8667.
Chem., 2011, 37, 55–84; (c) D. Srimani, E. Balaraman, 22 In the course of this work we became aware of independent
B. Gnanaprakasam, Y. Ben-David and D. Milstein, Adv.
Synth. Catal., 2012, 354, 2403–2406, S2403/2401-S2403/2421.
4 (a) J. Choi, A. H. R. MacArthur, M. Brookhart and
A. S. Goldman, Chem. Rev., 2011, 111, 1761–1779; (b)
M. C. Haibach, S. Kundu, M. Brookhart and A. S. Goldman,
Acc. Chem. Res., 2012, 45, 947–958.
work by Esteruelas and co-workers which included the
synthesis of iPrfurPOP as well as the synthesis of
(
iPrxanPOP)RhCl, cis-(iPrxanPOP)RhH2Cl and iPrxanPOP)
(
RhH. For the iPrxanPOP complexes see: M. A. Esteruelas,
M. Olivan and A. Velez, Inorg. Chem., 2013, 52, 5339–
5349.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 3683–3692 | 3691

View Article Online
Chemical Science
Edge Article
23 (a) M. W. Haenel, D. Jakubik, E. Rothenberger and 32 W. C. Kaska, S. Nemeh, A. Shirazi and S. Potuznik,
G. Schroth, Chem. Ber., 1991, 124, 1705–1710; (b) Organometallics, 1988, 7, 13–15.
E. M. Vogl, J. Bruckmann, M. Kessler, C. Kruger and 33 K. B. Renkema, Y. V. Kissin and A. S. Goldman, J. Am. Chem.
M. W. Haenel, Chem. Ber., 1997, 130, 1315–1319; (c) Soc., 2003, 125, 7770–7771.
M. Kranenburg, d. B. Y. E. M. van, P. C. J. Kamer, 34 R. R. Schrock and J. A. Osborn, J. Am. Chem. Soc., 1976, 98,
L. P. W. N. M. van, K. Goubitz and J. Fraanje,
Organometallics, 1995, 14, 3081–3089.
24 M. A. Esteruelas, N. Honczek, M. Olivan, E. Onate and
M. Valencia, Organometallics, 2011, 30, 2468–2471.
25 (a) H. Kumobayashi, H. Taketomi and S. Akutagawa in
Isomenthone, Takasago Perfumery Co., Ltd., application: JP,
Japan, 1980, p. 10; (b) H. Kumobayashi, H. Taketomi and
2134–2143.
35 H. Y. Afeefy, J. F. Liebman and S. E. Stein, Neutral
Thermochemical Data, in NIST Chemistry WebBook, NIST
Standard Reference Database Number 69, ed. P. J.
Linstrom and W. G. Mallard, National Institute of
Standards and Technology, Gaithersburg MD, 20899,
http://webbook.nist.gov, retrieved 26 January 2013.
S. Akutagawa in Isomerization of piperitol, Takasago 36 M. D. Millard, C. E. Moore, A. L. Rheingold and
Perfumery Co., Ltd., application: JP, Japan, 1980, p. 10. J. S. Figueroa, J. Am. Chem. Soc., 2010, 132, 8921–8923.
26 A. N. Cammidge, M. J. Cook, K. J. Harrison and 37 B. K. Shaw, B. O. Patrick and M. D. Fryzuk, Organometallics,
N. B. McKeown, J. Chem. Soc., PerkinTrans. 1, 1991, 3053–3058. 2012, 31, 783–786.
27 K. Zhu, P. D. Achord, X. Zhang, K. Krogh-Jespersen and 38 M. C. Baird, J. T. Mague, J. A. Osborn and G. Wilkinson, J.
A. S. Goldman, J. Am. Chem. Soc., 2004, 126, 13044–13053. Chem. Soc. A, 1967, 1347–1360.
28 R. Tanaka, M. Yamashita and K. Nozaki, J. Am. Chem. Soc., 39 In an even sharper contrast with the Wilkinson system of
2009, 131, 14168–14169.
eqn (3), Milstein and co-workers have reported that
ethylene eliminates irreversibly from (tBuPCP)IrEtI to give
(
M. E. Van der Boom, C. L. Higgitt and D. Milstein,
Organometallics, 1999, 18, 2413–2419.
29 (a) E. Ben-Ari, G. Leitus, L. J. W. Shimon and D. Milstein, J.
Am. Chem. Soc., 2006, 128, 15390–15391; (b) M. A. Iron,
E. Ben-Ari, R. Cohen and D. Milstein, Dalton Trans., 2009,
9433–9439.
tBuPCP)IrHI; the kinetics of this system were investigated:
30 M. Findlater, W. H. Bernskoetter and M. Brookhart, J. Am. 40 P. J. Chirik and J. E. Bercaw, Organometallics, 2005, 24, 5407–
Chem. Soc., 2010, 132, 4534–4535. 5423.
31 K. Tatsumi, R. Hoffmann, A. Yamamoto and J. K. Stille, Bull. 41 M. C. Haibach and A. S. Goldman, manuscript in
Chem. Soc. Jpn., 1981, 54, 1857–1867. preparation.
3692 | Chem. Sci., 2013, 4, 3683–3692
This journal is ª The Royal Society of Chemistry 2013