Operationally unsaturated pincer/rhenium complexes form metal carbenes from cycloalkenes and metal carbynes from alkanes

Article abstract of DOI:10.1021/ja062327r

Operationally unsaturated (i.e., 16/18-electron) (PNP^R)Re(H) ₄, where PNP^R is N(SiMe₂CH₂PR ₂)₂, is reactive at 22°C with cyclic olefins. The first observed products are generally (PNP^R)Re(H)₂- (cycloalkylidene), with hydrogenated olefin as the product of hydrogen abstraction from the tetrahydride. The tetrahydride complex with R = ¹Bu generally fails to react (too bulky), that with R = cyclohexyl suffers a (controllable) tendency to abstraction of 3H from one ring, forming an η³-cyclohexenyl compound, and that with R = ⁱPr generally gives the richest bimolecular reactivity. The cyclic monoolefins studied show distinct reactivity, C6 giving first the carbene and then coordinated cyclohexadiene, C5 giving carbene, then diene, and then η⁵-C₅H₅, C8 giving carbene and then η²-cyclooctyne, and C12 giving an η³-allyl. Norbornene gives a π-complex of the norbornene in thermal equilibrium with its carbene isomer; at 90°C, hydrocarbon ligand Cα-Cβ bond cleavage occurs to give, for the first time, a carbyne complex from an internal olefin. Two compounds synthesized here have the formal composition (PNP^R)Re + olefin , and each of these is capable of dehydrogenating the methyl group of a variety of alkanes at 110°C to form (PNP)ReH(≡CR).

Full text of DOI:10.1021/ja062327r

Published on Web 04/13/2007
Operationally Unsaturated Pincer/Rhenium Complexes Form
Metal Carbenes from Cycloalkenes and Metal Carbynes from
Alkanes
Oleg V. Ozerov,† Lori A. Watson,‡ Maren Pink, and Kenneth G. Caulton*
Contribution from the Department of Chemistry, Indiana UniVersity,
Bloomington, Indiana 47405-7102
Received April 5, 2006; E-mail: caulton@indiana.edu
Abstract: Operationally unsaturated (i.e., 16/18-electron) (PNPR)Re(H)4, where PNPR is N(SiMe2CH2PR2)2,
is reactive at 22 °C with cyclic olefins. The first observed products are generally (PNPR)Re(H)2-
(cycloalkylidene), with hydrogenated olefin as the product of hydrogen abstraction from the tetrahydride.
The tetrahydride complex with R ) tBu generally fails to react (too bulky), that with R ) cyclohexyl suffers
a (controllable) tendency to abstraction of 3H from one ring, forming an η3-cyclohexenyl compound, and
i
that with R ) Pr generally gives the richest bimolecular reactivity. The cyclic monoolefins studied show
distinct reactivity, C6 giving first the carbene and then coordinated cyclohexadiene, C5 giving carbene,
then diene, and then η5-C5H5, C8 giving carbene and then η2-cyclooctyne, and C12 giving an η3-allyl.
Norbornene gives a π-complex of the norbornene in thermal equilibrium with its carbene isomer; at 90 °C,
hydrocarbon ligand CR-Câ bond cleavage occurs to give, for the first time, a carbyne complex from an
internal olefin. Two compounds synthesized here have the formal composition “(PNPR)Re + olefin”, and
each of these is capable of dehydrogenating the methyl group of a variety of alkanes at 110 °C to form
(PNP)ReH(tCR).
Scheme 1
Introduction
Transformations of organic molecules on transition-metal
centers are a focal point of organo-transition-metal chemistry.
Of particular interest to us are the transformations that occur
on strongly π-basic metal centers, capable of rebuilding an
organic molecule into a π-acidic ligand complementary to the
metal. A recent paper1 details how the strongly π-basic M(OSit-
Bu3)3 (M ) Nb, Ta) isomerizes olefins to alkylidenes. We have
previously reported2,3 that the highly π-basic Re center supported
by the PNP ligands, (R2PCH2SiMe2)2N-, effects the rearrange-
ment of acyclic alkenes into hydrido carbyne complexes.
(PNPR)ReH4 complexes 1 (Scheme 1) react rapidly with 3 mol
of acyclic alkenes H2CdCHR at 22 °C to quantitatively produce
(PNPR)ReH(tCCH2R) and 2 mol of the corresponding alkane
H3CCH2R. These experimental results, together with the sup-
porting DFT study, demonstrate the strong preference of Re in
the PNP complexes for the formation of hydrido carbyne
products. In particular, we found that the hydrido carbyne
structure is preferred over the isomeric carbene, η2-alkene, vinyl
hydride, or vinylidene dihydride complexes. Our observed
production of a hydrido carbyne from (PNPR)ReH4 and alkene
requires making and breaking of C-H, Re-C, and Re-H bonds
(but not C-C bonds!) and can only occur in hydrocarbons where
the metal can migrate to a terminal carbon purely by C-H
cleavage. With this background, we were intrigued by the
possible outcome of the reaction of (PNPR)ReH4 with cycloalk-
enes.
† Department of Chemistry, Brandeis University, 415 South St., Waltham,
MA 02454.
‡ Department of Chemistry, Earlham College, 801 National Rd. W.,
Richmond, IN 47374.
The ultimate focus of our interest was on the C-H activation
of alkanes.4-16 The elements of the hydrido + carbyne set of
(1) Hirsekorn, K. F.; Veige, A. S.; Marshak, M. P.; Koldobskaya, Y.;
Wolczanski, P. T.; Cundari, T. R.; Lobkovsky, E. B. J. Am. Chem. Soc.
2005, 127, 4809.
(2) Ozerov, O. V.; Watson, L. A.; Pink, M.; Caulton, K. G. J. Am. Chem. Soc.
2004, 126, 6363.
(3) Ozerov, O. V.; Huffman, J. C.; Watson, L. A.; Caulton, K. G. Organo-
metallics 2003, 22, 2539.
(4) Jones, W. D. Inorg. Chem. 2005, 44, 4475.
(5) Jones, W. D.; Vetter, A. J.; Wick, D. D.; Northcutt, T. O. ACS Symp. Ser.
2004, 885, 56.
(6) Morales-Morales, D.; Redon, R.; Yung, C.; Jensen, C. M. Inorg. Chim.
Acta 2004, 357, 2953.
9
10.1021/ja062327r CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 6003-6016
6003

A R T I C L E S
Ozerov et al.
Scheme 2
ligands might be formally derived (eq 1a) from an alkane by
formal loss of “2H”. Thus, we envisioned that, with an
appropriate hydrogen acceptor A (eq 1b), conversion of alkanes
to hydrido carbynes can be accomplished on the (PNP)Re center.
To make the new Re-carbon bonds in the reaction in Scheme
1, the Re center in (PNP)ReH4 (1) must lose the hydrides, which
in this case is accomplished by hydrogenation of excess acyclic
olefin substrate. To make new Re-C bonds in reactions with
substrates that do not possess H2-accepting functionalities, a
different approach must be taken. Either the formation of new
bonds to Re is thermodynamically favorable enough to expel
free H2 or an additional, sacrificial H2 acceptor incapable of
carbyne formation must be introduced. We have reported17 on
examples of both types of behavior in activation of pyridine
substrates where cyclooctene suitably served as a sacrificial H2
acceptor. Here we report the dehydrogenation of alkanes to
hydrido carbynes with various cycloalkenes as hydrogen ac-
ceptors. We also describe the details of the fascinatingly
divergent, ring-size-dependent reactivity of the (PNP)ReH4
complexes with cyclic alkenes involving hydrogen migration,
C-H activation, and even C-C bond scission. The influence
of the phosphine substituent is also described. Some aspects of
this chemistry have been communicated.18
electronic influence of iPr and Cy. The major difference is that
1b, upon removal of hydrides from the Re center, is prone to
undergo intramolecular C-H activation (vide infra).
Reactivity of 1a,b with Cycloalkenes. Solvent Choice. As
will be discussed later, some of the species formed in the
reactions of cycloalkenes with 1a,b are quite reactive, even with
many common solvents. Deuterated solvents such as C6D6 also
can undergo H/D exchange with these reactive Re compounds.
We found that cyclohexane, cyclooctane, and hexamethyldisi-
loxane were effectively inert as solvents.
Results
Influence of R in PNPR on the Rate. We have previously
reported syntheses of compounds 1a,b (Scheme 1) that bear
different substituents on P. While the (PNPtBu) analogue
(“1tBu”) has served to uncover some alluring Ru chemistry,19
t
t
the Bu analogue has proven to be too hindered sterically. Bu
was recovered unchanged after thermolysis (160 °C, 24 h) in
mesitylene in the presence of excess cyclopentene, and it was
also unreactive toward cyclohexene and cyclooctene. Nor-
t
bornene did react with Bu slowly, but the reaction was found
Reactions with Cyclohexene and C-H Activation of the
PNPCy Ligand (Schemes 2 and 3). The preliminary results of
the studies of the reactivity of 1 with cyclohexene have been
communicated.18 At ambient temperature, 1a and 1b react with
excess cyclohexene in under 1 h to produce 2a or 2b,
respectively (with concomitant production of cyclohexane). The
solid-state structure of 2b (see below) is unusual as it possesses
a â-agostic cyclohexylidene ligand and the geometry about the
carbene carbon is strongly distorted toward a T-shape. Solution
spectroscopic data suggest a similar â-agostic interaction in 2a.
The formation of 2a from 1a was quantitative when at least 2
equiv of cyclohexene was used, and no intermediates were
observed during the course of this reaction at 22 °C. In the case
of 1b, reaction with 2 equiv of cyclohexene at 22 °C led to a
mixture of 2b and the purple-blue cyclohexenyl hydride 13
(Scheme 3). Activation of three C-H bonds in cyclohexyl-
substituted phosphines to give cyclohexenyl and cyclohexene
complexes has been documented in several related systems.20
We have now identified 13 as the impurity we previously
observed2 in the reactions of 1b with acyclic alkenes. In those
instances, the formation of 13 could be suppressed by using an
to lead to unselective decomposition. We have previously found
that Bu reacts with linear alkenes much more slowly than 1a
or 1b; we ascribed that to the prohibitive steric bulk of the
(PNPtBu) ligand. The rates of reaction of 1a and 1b are
qualitatively similar, reflecting the comparable steric and
t
(7) Morales-Morales, D.; Lee, D. W.; Wang, Z.; Jensen, C. M. Organometallics
2001, 20, 1144.
(8) Jensen, C. M. Chem. Commun. 1999, 2443.
(9) Lee, D. W.; Kaska, W. C.; Jensen, C. M. Organometallics 1998, 17, 1.
(10) Gupta, M.; Hagen, C.; Flesher, R. J.; Kaska, W. C.; Jensen, C. M. Chem.
Commun. 1996, 2083.
(11) Gupta, M.; Hagen, C.; Kaska, W. C.; Cramer, R. E.; Jensen, C. M. J. Am.
Chem. Soc. 1997, 119, 840.
(12) Xu, W.-w.; Rosini, G. P.; Krogh-Jespersen, K.; Goldman, A. S.; Gupta,
M.; Jensen, C. M.; Kaska, W. C. Chem. Commun. 1997, 2273.
(13) Renkema, K. B.; Kissin, Y. V.; Goldman, A. S. J. Am. Chem. Soc. 2003,
125, 7770.
(14) Krogh-Jespersen, K.; Czerw, M.; Summa, N.; Renkema, K. B.; Achord, P.
D.; Goldman, A. S. J. Am. Chem. Soc. 2002, 124, 11404.
(15) Kanzelberger, M.; Singh, B.; Czerw, M.; Krogh-Jespersen, K.; Goldman,
A. S. J. Am. Chem. Soc. 2000, 122, 11017.
(16) Liu, F.; Pak, E. B.; Singh, B.; Jensen, C. M.; Goldman, A. S. J. Am. Chem.
Soc. 1999, 121, 4086.
(17) Ozerov, O. V.; Pink, M.; Watson, L. A.; Caulton, K. G. J. Am. Chem. Soc.
2004, 126, 2105.
(18) Ozerov, O. V.; Watson, L. A.; Pink, M.; Caulton, K. G. J. Am. Chem. Soc.
2003, 125, 9604.
(19) Ingleson, M. J.; Yang, X.; Pink, M.; Caulton, K. G. J. Am. Chem. Soc.
2005, 127, 10846.
(20) Esteruelas, M. A.; Lopez, A. M. Organometallics 2005, 24, 3584.
9
6004 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007

Pincer/Rh Complexes Form Metal Carbenes and Carbynes
A R T I C L E S
Figure 1. ORTEP plot of 13 (50% thermal ellipsoids). Omitted for
clarity: all H atoms except ReH, the second independent molecule, the
pentane molecule in the lattice. Selected bond distances (Å) and angles
(deg): Re1a-N1a, 2.031(3); Re1a-C27a, 2.181(3); Re1a-C28a, 2.200-
(3); Re1a-C29a, 2.363(3); Re1a-P1a, 2.3763(8); Re1a-P2a, 2.3004(8);
N1a-Re1a-P1a, 87.16(8); N1a-Re1a-P2a, 87.40(8); P1a-Re1a-P2a,
174.52(3); C27a-C28a-C29a, 111.6(3).
Figure 2. ORTEP plot of 2b (50% thermal ellipsoids). Omitted for
clarity: the CH2 groups on all P atoms and all H atoms except those on
C36 and the one (of two actually present) on Re. Selected bond distances
(Å) and angles (deg): Re1-C31, 1.889(4); Re1-N1, 2.182(3); Re1-P1,
2.4091(9); Re1-P2, 2.3715(1); C31-Re1-N1, 136.90(15); N1-Re1-C36,
102.36(13), P1-Re1-P2, 161.30(4); Re1-C31-C32, 147.2(3); Re1-C31-
C36, 100.4(3); N1-Re1-P1, 79.88(9).
Scheme 3
PCy2 phosphorus. There is no significant difference in the Re-
PCy2 distance here from the Re-PiPr2 groups in compound 8.
(PNPCy)Re(H)2[dC(CH2)5], 2b. Only one of the two (in-
equivalent) hydrides known to be present from the NMR spectra
and from the diamagnetism of the compound (e.g., observation
of 31P NMR signals) was found by X-ray diffraction; only it is
illustrated in Figure 2, but there is ample room for the second
hydride cis to both N1 and H1R. -(P1-Re-P2) is the typical
transoid value (161.30(4)°) of PNP in a mer arrangement, but
the angle N1-Re-C31 is large, 136.90(15)°, allowing ample
space for an agostic interaction with C36-H36b. Relevant
distances from Re to C36 and H36b are 2.635(4) Å (cf. the
norbornene values below) and 2.38 Å (H position not refined).
This is accomplished by deforming -(Re-C31-C36) (100.4-
(3)°) much below that of Re-C31-C32 (147.2(3)°). The Re-
C31 distance, 1.889(4) Å, is consistent with a multiple bond,
although this is a rare example21 of a carbene with a â-agostic
interaction (cf. â-agostic alkyls). The C-C bond lengths and
C-C-C angles in the cyclohexylidene ring are unexceptional,
and the molecular mirror symmetry deduced from the NMR
spectra requires rapid exchange between H36a and H36b as the
agostic donors. The Re-N1 distance, 2.182(3)Å, is longer than
that in 1b (2.063(2)Å), indicative of diminished π-donation from
N to Re as a consequence of the agostic donation.
excess of the alkene. Similarly, a large (20-fold) excess of
cyclohexene eliminates 99% of the formation of 13 (on a <1 h
time scale).
While pure 2a is stable in solution at ambient temperature
for at least several hours, solutions of 2b readily evolve into a
2:1 mixture of 13 and 1b. In the presence of excess cyclohexene,
pure 3, a cyclohexadiene complex, is ultimately produced from
2b in solution even though excess cyclohexene retards the
transformation kinetically. However, 2b is stable in the solid
state at ambient temperature, which allowed for its characteriza-
tion by X-ray crystallography and elemental analysis. Ther-
molysis (120 °C, 30 min) of a solid sample of 2b in a vacuum
transfer apparatus resulted in the production of a mixture of
cyclohexane and cyclohexene as volatile products and 13 and
1b as the organometallic residue (see the Experimental Section
[Cy2PCH2SiMe2NSiMe2CH2PCy(C6H8)]ReH, 13. This mol-
ecule crystallizes with two molecules in the asymmetric unit,
but the two show no significant differences. The two are
numbered analogously to facilitate comparison in the Supporting
Information. Figure 1 shows how the conformation of one C6
ring allows the “meta” and “para” carbons, which each have
been singly dehydrogenated, to bind to the metal: P is axial on
the C6 ring, which contrasts to its equatorial site in every other
compound reported here. This η3-allyl binding also makes
-(Re-P-Cipso) very small (99.2-99.6°). The Re-N distances
are short (<2.04 Å), and the Re-P distance to the dehydroge-
nated, allyl bound C6 ring is shorter (by >0.08 Å) than to the
(21) Feng, S. G.; White, P. S.; Templeton, J. L. J. Am. Chem. Soc. 1992, 114,
2951.
9
J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007 6005

A R T I C L E S
Scheme 4
Ozerov et al.
Scheme 5
carbynes (12c, 12g, and others) were described previously.2 The
new hydrido carbynes 12 prepared in this work display very
similar spectroscopic features and a red color. Addition of
cyclohexene before thermolysis eliminated 1a from the products,
but it also increased the 3:12 ratio. Thermolysis of 2a in heptane
in the presence of a large excess of cyclohexene completely
suppressed the formation of 12 (only 3 was formed). Ther-
molysis of 3 (colorless) in heptane or ethylbenzene at 130 °C
did not lead to any NMR-detectable change or any color change.
When solutions of 3 in heptane or ethylbenzene were briefly
heated at 165 °C, the red color typical of 12 was observed and
the corresponding hydrido carbynes 12d and 12h were identified
by subsequent NMR analysis. However, at 165 °C the onset of
unselective decomposition of 12 occurs, and we have not been
able to convert 3 into 12 in high yield.
for details). Compound 13 is thermally stable in alkane or arene
solvents, although it readily undergoes H/D exchange with C6D6
at 22 °C.
3 is a colorless compound that dissolves well in aromatic
(and sparingly in aliphatic) solvents. The 31P NMR resonance
of 3 is broad at 22 °C and decoalesces into two broad peaks at
low temperature. This presumably reflects a ground-state
orientation of the cyclohexadiene which leaves inequivalent P,
but with a modest barrier for rotation of the diene ring. Only
one hydride signal (δ ) -8.33, br dt, JHH ) 7 Hz, JHP ) 16
Hz) could be clearly observed. The presence of the other hydride
(presumably obscured by the aliphatic resonances) is inferred
from the observed H-H coupling and the similarity with the
cyclopentadiene species 5b (Scheme 3), for which a structure
in the solid state was determined (vide infra).
Compounds 1a-c are produced by Mg reduction of (PNPR)-
ReOCl222 in an ethereal solvent under a H2 atmosphere. When
the reduction of 1b, (PNPCy)ReOCl2, is performed in ether under
an atmosphere of pure Ar in an attempt to form merely “(PNP)-
Re”, mixtures with ratios of 3:1 to 10:1 of 13 and 1b are formed
(by 31P NMR in situ) in a few different experiments (see the
Experimental Section). The stoichiometric ratio that may be
expected is 2:1 (corresponding on average to the elements of
(PNPCy)Re), and it is not immediately clear how an excess of
hydrogen-poor 13 is produced. Compound 13 readily reacts with
H2 to give 1b in the time of mixing. Reduction of (PNPR)-
ReOCl2 under Ar followed by hydrogenolysis is a convenient
way of synthesizing 1b on a large scale. However, these studies
show that any attempt to synthesize simply “(PNPCy)Re” (or
its ether complex) by magnesium reduction in Et2O has a
different outcome, one based on intramolecular attack by the
highly reduced, electron-rich rhenium.
Thermolysis (90 °C, 1 h) of 1a with g3 equiv of cyclohexene
in cyclohexane led to the exclusive formation of 3. 2a is an
intermediate in this reaction. Isolated samples of 2a transform
into 3 upon thermolysis in cyclohexane in the presence of g1
equiv of cyclohexene. When 1a was thermolyzed in the presence
of <3 equiv of cyclohexene or when a solution of pure 2a was
thermolyzed in cyclohexane, a mixture of 1a and 3 was formed
after the thermolysis. The cyclohexadiene ligand in 3 (as well
as cyclohexylidene in 2a) comes from cyclohexene and not the
cyclohexane solvent as evidenced by the invariance of the
reaction outcome in cyclooctane or cyclo-C6D12 solvents.
Alkane Dehydrogenation with 2a. Compound 2a is the
formal equivalent of (PNP)Re and cyclohexane and thus is
attractive for the alkane reaction of eq 1a. Thermolysis of 2a in
heptane or neohexane led to a mixture of the corresponding
hydrido carbynes 12d and 12g, together with 3 (the product of
competitive C6 ring dehydrogenation in 2a), the hydrogen
acceptor product 1a (Scheme 4), and presumably cyclohexane.
The structure and spectroscopic properties of several hydrido
1
The H NMR resonances of the cyclohexadiene ligand in 3
are somewhat broad. Spin saturation transfer experiments were
consistent with the process shown in Scheme 5 that accounts
for the site exchange between six of the hydrogens of the
cyclohexadiene ligand. Similar observations were made in the
case of the (R3P)2ReH3(η4-diene) complexes.23 The reversible
cyclohexadiene/dihydride-cyclohexene/monohydride transfor-
mation exchanges the two endo hydrogens of the cyclohexadiene
ligand with one of the hydrides. The same process also
exchanges the remaining six hydrogens of the cyclohexadiene
ligand (two exo hydrogens from the CH2 groups and the four
C(sp2)-H hydrogens). Irradiation of the hydride signal at -8.33
ppm had no effect on the cyclohexadiene ligand resonances and
vice versa, so this hydride does not participate in the exchange
process.
The activation barrier for the formation of the hydrido
cyclohexenyl intermediate 14 is likely lowered by the π-donating
ability of N. While 3 is a saturated, 18-electron complex, 14 is
a 16-electron complex without counting the π-donation from
N. We have previously demonstrated3,24 that the π-donation from
N in (PNP)Re systems can stabilize operationally unsaturated
compounds. Consequently, we expect that 14 is also stabilized
by π-donation and thus more accessible energetically. Indeed,
we have isolated other closely related allyl hydride complexes
(23) Baudry, D.; Ephritikhine, M.; Felkin, H.; Zakrzewski, J. J. Organometallics
1984, 272, 391.
(24) Ozerov, O. V.; Watson, L. A.; Pink, M.; Baik, M.-H.; Caulton, K. G.
Organometallics 2004, 23, 4934.
(22) Ozerov, O. V.; Gerard, H. F.; Watson, L. A.; Huffman, J. C.; Caulton, K.
G. Inorg. Chem. 2002, 41, 5615.
9
6006 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007

Pincer/Rh Complexes Form Metal Carbenes and Carbynes
A R T I C L E S
Thermolysis of pure 5b in C6D6 produced a mixture of 13
and 6b; the formation of either product was irreversible (i.e.,
the mixture does not change proportion with further heating).
6b can be isolated in high yield if 5b or 1b is thermolyzed in
the presence of a large excess of cyclopentene (cosolvent). 6a
and 6b were identified primarily on the basis of the similarity
of the key solution NMR data with those of the known CpRe-
(H)2(PR3)2 compounds.23 Compounds of the type CpRe(H)2-
(PR3)2 have a four-legged stool geometry, with hydrides being
the opposite “legs”. The flexible backbone of the η2-PN(H)PiPr
ligand in 6 should not lead to a geometry about Re that is
significantly different in CpRe(H)2(PR3)2. A single resonance
arising from the five Cp hydrogens was observed at ca. 4.5 ppm,
as well as a triplet hydride resonance at ca. -12 ppm with a
2JHP ) 41 Hz characteristic of the CpRe(H)2(PR3)2 complexes.
6a and 6b are thermally stable in solution in alkane and arene
solvents.
Figure 3. ORTEP plot of 5b (50% thermal ellipsoids). Omitted for
clarity: all H atoms except ReH, the methylene carbons of cyclohexyl
groups. Selected bond distances (Å) and angles (deg): Re1-N1, 2.212(2);
Re1-P1, 2.4193(4), Re1-P2, 2.3979(4); C32-C33, 1.430(2); C33-C34,
1.440(2); C34-C35, 1.439(2); P1-Re1-P2, 139.51(1); N1-Re1-P1,
76.12(4).
5a,b each gave rise to a single broad 31P NMR resonance
and somewhat broad resonances of the cyclopentadiene ligand.
(13 and 9a, vide infra). For (R3P)2ReH3(η4-diene) it was also
proposed23 that the diene fragment slowly rotates about the Re-
centroid axis. Such a slow rotation is likely in 3 as well and
may be responsible for the observed low-temperature inequiva-
lence of the two P sites. In addition, we have observed
broadening of the observed 1H and 31P resonances at low
temperatures as a result of the flexing motion of the PNP
backbone in other PNP complexes. The combination of three
dynamic processes (Re-H/C-H exchange, diene rotation, and
PNP backbone flexing) in 3 is responsible for the observed
dynamic features of the NMR spectra.
Cyclopentene (Schemes 2 and 3). Upon mixing of 1a with
a 4-fold excess of cyclopentene at 22 °C (in cyclohexane-d12),
a mixture of 4a, 5a, and 6a was initially observed. 4a could
not be isolated and was only identified on the basis of the
selected NMR resonances in solution that resembled those of
2a. A lone singlet was observed in the 31P{1H} NMR spectrum
for each (4a, δ 49.4), and it became a triplet upon selective
decoupling of only the alkyl hydrogens. 5a was the dominant
product after 5 h at 22 °C, and thermolysis of this solution
produced 6a in a >95% yield. 1b reacted similarly with
cyclopentene at 22 °C, except that 13 was also produced. As
with cyclohexene and 1b, the formation of 13 could be
suppressed by using a large excess of cyclopentene. 5b could
be isolated as an analytically pure crystalline solid in high yield,
including an X-ray-quality single crystal.
(PNPCy)Re(H)2(C5H6), 5b. The X-ray structure in Figure 3
shows that the bulky PCy2 groups bend away from C5H6 and
toward the small hydride H2r. Nevertheless, the amide nitrogen
is coplanar (angle sum is 359.98°) with its three substituents.
The two trans angles in the Re/P2/N/H1r unit are nearly equal
(-(N1-Re-H1r) ) 139.9(9)° and -(P1-Re-P2) ) 139.51-
(1)°, the latter unusually small for this pincer ligand), and
-(N1-Re-H2r) is cisoid, 76.0(9)°. The distances from Re to
the diene terminal carbons (C32 at 2.2963(6) Å and C35 at
2.2608(15) Å) are clearly longer than to the diene internal
carbons (C33 at 2.2095(16) Å and C34 at 2.1697(15) Å), and
C5H6 is oriented in a way which leaves the phosphorus nuclei
inequivalent, in agreement with the 31P NMR spectrum. The
Re-N1 distance, 2.212(2) Å, is very long, consistent with this
amide not participating in N f Re π-donation, but also
influenced by the trans influence of H1r.
1
One of the H NMR hydride signals resonated at ca. -8 ppm,
while the other was apparently broad and/or obscured by the
aliphatic resonances. Presumably, 5 undergoes the same dynamic
processes as 3 (vide supra). We selected 3, rather than 5a or
5b, for the spin saturation transfer and low-temperature NMR
studies because of the ease of isolation in a pure state and
1
simpler H NMR spectrum (vs 5b).
Cyclooctene. Cyclooctene reacts (Schemes 2 and 3) with 1a,b
more slowly than do cyclohexene and cyclopentene. Initially,
the formation of the cyclooctylidene 7 was observed, which
further evolves either exclusively into the cyclooctyne/dihydride
8 (from 1a) or exclusively into 13 (from 1b and g3 equiv of
cyclooctene). 7a and 7b could not be isolated and were only
characterized by their selected NMR resonances that resemble
those of 2 and 4.
(PNPiPr)Re(H)2(cyclooctyne). The synthesis of (PNPiPr)Re-
(H)2(cyclooctyne), 8, was reported earlier, but not discussed in
detail there.17 Thermolysis of 1a in the presence of g3 equiv
of cyclooctene in cyclooctane, cyclohexane, or hexamethyld-
isiloxane led to the exclusive formation of 8. 8 is thermally
stable (<100 °C, 24 h) in these solvents and is light pink in
both solution and the solid state (the faint pink color is retained
after three recrystallizations). 8 reacts with H2 to produce
cyclooctane and 1a. It is best formed after treatment of (PNPiPr)-
Re(H)4 with g3 mol of cyclooctene at 90 °C. Two moles of
cyclooctane is produced from the hydride ligands lost. Proton
and 31P{1H} NMR spectra at 22 °C are consistent with a
molecule with mirror symmetry, the mirror plane bisecting the
P-Re-P angle. The hydrides are observed to be inequivalent,
as are the related substituents on the PNPiPr ligand: CH2, SiMe2,
i
and Pr. DFT calculations reported earlier on (PNPH)Re(H)2-
(HCCH) establish that the Cs isomer is 27.6 kcal/mol (∆G°298
)
1
more stable than the C2V isomer. Both H and 13C{1H} NMR
spectra at 22 °C show Cs symmetry of the cyclooctyne ring.
The 31P{1H} NMR spectrum at -76 °C shows that the broad
9
J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007 6007

A R T I C L E S
Ozerov et al.
Scheme 6
Figure 4. ORTEP plot of 8 (50% thermal ellipsoids). Omitted for clarity:
Me groups of isopropyls, all H atoms including two ReH atoms, second
independent molecule. Selected bond distances (Å) and angles (deg): Re1a-
C19a, 2.026(7); Re1a-C20A, 2.022(7); Re1a-N1a, 2.158(6); Re1a-P1a,
2.4023(18); Re1a-P2a, 2.3872(17); C19a-C20a, 1.320(10); P1a-Re1a-
P2a, 145.78(6); N1a-Re1a-P1a, 78.93(16); N1a-Re1a-C19a, 117.4(3).
singlet at 22 °C has decoalesced into two signals, thus breaking
1
the molecular mirror symmetry implied by the 22 °C H and
13C NMR spectra; this we attribute to the CtC vector not
exactly eclipsing the P-Re-P plane, as established (see below)
in two crystallographically independent molecules in the solid.
In addition, PiPr2/cyclooctyne congestion can lead to the 31P
NMR inequivalence by favoring a non-mirror-symmetric con-
formation at low temperature; this conformation includes lack
of mirror symmetry of the two fused rings of the (PNP)Re
substructure.
compounds similar to 12 form as a result of attack on the ligand
or solvent or C-C cleavage in the eight-membered ring of the
solvent of the cyclooctyne ligand. This chemistry is much more
promising when the hydrocarbon solvent contains a methyl
group, as will now be described.
This sample crystallized (Figure 4) with two independent
molecules in the unit cell. A least-squares fit shows that the
two molecules agree to within 3σ for the core of the Re/ligand
molecule; differences involve single bonds at the molecular
i
periphery: the conformation of one Pr group and the CH2
Thermolysis (24 h, 110 °C) of 8 or of 1a + 3 equiv of
cycloooctene in several hydrocarbon solvents (Scheme 6)
resulted in liberation of cyclooctane and the formation of the
solvent-derived hydrido carbyne products 12. In general, these
reactions were near quantitative (NMR evidence) when the
substrate (solvent) contained an ethyl group. Methyl-substituted
arenes apparently did produce the hydrido carbynes (12i-k),
but only in ca. 50% yield (the balance is other unidentified
products). The carbyne products can only form if the substrate
contains a methyl group; the reactive transient is apparently
incapable of C-C bond cleavage. While cyclohexane and
cyclooctane were not activated to any detectable degree,
thermolysis of 8 in cyclopentane resulted in a clean formation
of the cyclopentadienyl complex 6a. Interestingly, thermolysis
of 8 in methylcyclopentane produced both the carbyne-contain-
ing (12m) and the Cp-containing (15) products. The higher
reactivity of the cyclopentane vs larger cycloalkanes is generally
in line with the differences in C-H activation of cycloalkanes
described by others.26,27 2,3-Dimethylbutane was not activated,
instead producing a mixture of unidentified products similar to
that produced in larger cycloalkanes. Two substrates offered a
choice of different methyl groups within the same molecule.
Activation of 3-methylpentane (product 12f) and of neohexane
(product 12g) was selective for the less hindered methyl group.
groups of the cyclooctyne. The Re-N distances are of inter-
mediate length (∼2.15 Å), the Re-C distances (2.007(7)-2.026-
(7) Å) are shorter than to the norbornene carbons (∼2.20 Å;
see below) and the alkyne C-C distance (1.320(10)-1.328(9)
Å) is shorter than that (1.455(2) Å) in coordinated norbornene.
-(P-Re-P) is small (∼145.6°), consistent with repulsion from
the cyclooctyne, whose CtC vector the ReP2 plane eclipses.
This sterically unfavorable alkyne conformation must be
electronically dictated and involves competitive π-donation by
amide N and the alkyne-filled π-orbital. The 13C chemical shift,
190.4 ppm, approaches the region characteristic of alkyne 4e
donation.25 The coordinated alkyne has severely bent C(sp3)-
C(sp)-C(sp) angles: 129.1(7)-135.1(7)°. The amide nitrogen
is planar. Although the hydrides were located in a difference
map, they could not be refined to chemically reasonable
positions and thus permit no conclusion.
Prolonged (>24 h) thermolysis of 8 at 130 °C in cyclooctane
led to decomposition to unknown products. Several dozen(!)
new resonances appeared in the 31P NMR spectrum. Interest-
ingly, the color of the resultant solution was red, reminiscent
of the color of the hydrido carbynes 12. In addition, hydride
peaks in the 1H NMR spectrum of this mixture and resonances
in the 31P NMR spectrum appeared in the same region as those
of 12. We were not able to characterize the products of this
decomposition conclusively. It is possible, however, that some
(26) Crabtree, R. H.; Mihelcic, J. M.; Quirk, J. M. J. Am. Chem. Soc. 1979,
101, 7738.
(25) Compare, however, Wells, M. B.; White, P. S.; Templeton, J. L.
Organometallics 1997, 16, 1857.
(27) Crabtree, R. H.; Mellea, M. F.; Mihelcic, J. M.; Quirk, J. M. J. Am. Chem.
Soc. 1982, 104, 107.
9
6008 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007

Pincer/Rh Complexes Form Metal Carbenes and Carbynes
A R T I C L E S
Scheme 7
Monitoring the reactions at early stages revealed that small
amounts of 1a, cyclooctane, and cyclooctene were produced
during the thermolysis of pure 8 in alkanes. This is similar to
the observations we made in the activation of alkylpyridines
by 8.17
Cyclododecene. Cyclododecene reacted (Scheme 2) slowly
with 1a (reaction with 1b was not explored) to ultimately give
two isomers of 9. Our attempts at separating the two isomers
by recrystallization or at obtaining a suitable crystal for an X-ray
diffraction study were not successful. The solution NMR data
are consistent with the proposed cyclododecene/hydride for-
mulation. The validity of the proposed structure is reinforced
by the similarity (color, allylic, and hydridic NMR resonances)
with the structurally authenticated P-cyclohexene/hydride 13.
The single hydride and the three allylic hydrogens for each
Figure 5. ORTEP plot of 10a (50% thermal ellipsoids). Omitted for
clarity: Me groups of isopropyls, all H atoms except ReH and the hydrogens
on the bridgehead carbon of norbornene. Selected bond distances (Å) and
angles (deg): Re1-N1, 2.122(1); Re1-P1, 2.3924(4); Re1-P2, 2.3657-
(4); Re1-C19, 2.191(1); Re1-C20, 2.215(1); C19-C20, 1.455(2); P1-
Re1-P2, 154.70(1); N1-Re1-P1, 80.24(4).
of 10/11 are dark blue-purple, and the solids obtained by
crystallization from such solutions were of the same color as
well. 10 was the major isomer for both PNP ligands; e.g., the
ratio of 10a to 11a was ca. 8:1. This ratio was constant for
different batches and remained the same after multiple crystal-
lization/redissolution sequences. Because of the difficulty in
isolation of 10b/11b, we concentrated on the characterization
of 10a/11a. The NMR resonances of 10a and 11a were distinct,
indicating that exchange between 10a and 11a is slow on the
NMR time scale at 22 °C but fast on the time scale of
experimental handling. The material obtained by cooling a
solution of 10a/11a was highly crystalline, allowing for a facile
selection of an X-ray-quality crystal. Visual microscopic inspec-
tion revealed the presence of crystals of only one morphology
and color (blue-purple). By analogy with 2, one may expect
that 11a should be of a color closer to orange rather than blue-
purple. The intense blue-purple color is probably that of 10a,
and in solution it may conceal the lighter color of 11a. It seems
possible that only 10a crystallizes out of the solution and rapidly
equilibrates into a 10a/11a mixture upon redissolution. Both
10a and 11a display two inequivalent upfield hydride reso-
nances. The chemical shifts and the broadness of the hydrides
of 11a resemble those of 2. While the hydride resonances of
10a are not as broad as those of 11a, they show no coupling
fine structure at 22 °C. Upon cooling to -40 °C, the H-P
coupling becomes resolved. Consistent with the proposed
formulations, the NMR spectra of 10a indicate Cs symmetry
with a single 31P resonance observed, while 11a is C1-symmetric
by NMR and displays an AB pattern in the 31P{1H} NMR
spectrum; this lack of symmetry is caused by the chirality in
the norbornylidene moiety. The NMR resonances of 11a are
somewhat broad at 22 °C. The alkylidene nature of 11a was
confirmed by the observation of a downfield (δ 281 ppm) 13C
resonance corresponding to the R-carbon. Unfortunately, as-
signment of the other 13C resonances of 11a was not possible
owing to the overlap and low symmetry and concentration.
(PNPiPr)Re(H)2(norbornene). The structure (Figure 5) has
idealized mirror symmetry, which leaves the hydrides inequiva-
lent (and cis) and the two phosphorus atoms equivalent, both
in agreement with NMR observations. The amide nitrogen is
coplanar with its three substituents (angle sum 359.46°). The
angles between N and the norbornene olefinic carbons C19 and
1
isomer of 9 were clearly identified in the H NMR spectrum,
as well as the allylic resonances in the 13C NMR spectrum. The
hydride signal of each isomer appears as a doublet of doublets
at ca. -12 ppm owing to the coupling to two inequivalent
phosphines. Accordingly, an AB pattern for each isomer was
observed in the 31P{1H} NMR spectrum at ca. 30 ppm (trans
2JPP ) 241 or 243 Hz). Selective decoupling experiments
confirm the proposed assignment in each isomer. All operation-
ally unsaturated (PNP)Re compounds possess distinct colors,
and 9 and 13 have similar blue-purple color in solution and the
solid state. The exact difference between the two isomers of 9
is not clear from the solution NMR studies. The commercial
cyclododecene is a mixture of cis- and trans-isomers. It is
possible that isomerism of the two allyl substituents accounts
for the presence of two isomers of 9. However, the two isomers
of 9 do not have to duplicate the population of the two different
isomers of cyclododecene because 1a (and likely other Re-H-
containing intermediates) was previously shown2 to rapidly
isomerize olefins (e.g., stilbene). The two isomers of 9
equilibrate, but on a time scale too slow to be the ring
reorientation (“rotation”) observed for the cyclohexadiene ring
of 3.
Compound 9 reacted with heptane similarly to the cyclooctyne
complex 8. Thermolysis of 9 in heptane (110 °C) produced 12d
in >98% yield (Scheme 7). Monitoring this reaction at early
stages revealed the intermediate formation of 1a. Compound 9
is thus an effective formal source of (PNP)Re, and the allyl +
hydride serves as a 2H acceptor in heptane dehydrogenation
(eq 1a).
Norbornene. Norbornene reacted (Schemes 2 and 3) rapidly
at 22 °C in alkane solvents with 1a(b) to give a mixture of
isomeric 10a(b) and 11a(b). As with other cycloalkenes, in the
case of 1b, a large excess of norbornene was necessary to
suppress cyclohexyl dehydrogenation (i.e., the formation of 13).
Isolated samples of 10a/11a were stable at 22 °C in an alkane
solution, whereas dissolution of solid samples of 10b/11b
(recrystallized at -30 °C from solutions containing an excess
of norbornene) resulted in the formation of a mixture of 10b,
11b, 13, 1b, and norbornane that evolved into a 2:1 mixture of
13 and 1b (and also norbornane) over 24 h at 22 °C. Solutions
9
J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007 6009

A R T I C L E S
Ozerov et al.
Table 1. Product Yields in Competition between Hydrocarbons for
(PNP)Re(H)2(cyclooctyne)
C20 are much larger (138.60(5)° and 138.62(5)°, respectively)
than 90°, thereby opening a place for the norbornene CH2. The
Re-C25 and Re-H25 distances, 3.044 and 2.52 Å, respectively,
are consistent with an agostic interaction, which accords with
reagents
product mole ratio
n-hexane/cyclopentane
n-heptane/cyclopentane
n-hexane/ethylbenzene
n-heptane/ethylbenzene
12c:6a ) 73:27
12d:6a ) 73:27
12c:12h ) 52:48
12d:12h ) 53:47
1
one H NMR chemical shift observed at -1.51 ppm (i.e., as
anticipated for H interacting with a transition metal).28 The Re-
N1 distance, 2.122(1) Å, is of intermediate length, consistent
with an agostic interaction completing an 18-electron count but
in competition with N f Re π-donation. The olefinic C19-
C20 distance, 1.455(2) Å, is long,29-31 consistent with strong
π-basicity of the (PNPiPr)Re center.
competition with cyclopentane is consistent with these linear
alkanes being reactive preferentially at primary (methyl) C-H
bonds. Experiments c and d confirm this and show that aryl
C-H bonds are not competitive in terms of formation of
irreversible product(s) and that the transient is indiscriminate
(i.e., highly reactive) between n-alkane and ethylbenzene methyl
group C-H bonds. On a per-bond basis, the reactivity ratios
are benzyl CH3 (8.0) > alkyl CH3 (4.5) > CH2 (1.0).
DFT Computational Study. The questions we hoped to
answer with these computations are the following: (a) What
are the thermodynamics of synthesis of the cyclooctyne
complex, and how necessary is accompanying cyclohexane
hydrogenation? (b) Is cyclooctane production (eq 2) required
for favorable alkane dehydrogenation to carbyne, or is cy-
clooctene (eq 3)
Thermolysis (90 °C) of 10a/11a in the presence of 5 equiv
of norbornene led rather unexpectedly to the hydrido carbyne
product 12a (ca. 90%). 12a is an isomer of 10a and 11a, formed
via scission of a CR-Câ bond in 11a, our first observation of
C-C bond cleavage. 12a is trapped in the reaction mixture as
its norbornene adduct 12a*. The coordinated norbornene can
be removed in vacuo at 110 °C. The NMR data for 12a and
12a* are similar to those of the other hydrido carbynes 12 and
their olefin adducts. For 12a, the chemical shifts of the 31P NMR
(δ 60.8 ppm) and Re-H 1H NMR (δ -10.57 ppm) resonances
are especially similar to those of the most structurally similar
1
12m (31P NMR, δ 60.6 ppm; H NMR, δ -10.58 ppm). 12a
possesses a chiral center, but it is fairly remote from the PNP
(PNP)Re(H)2(cyclooctyne) +
1
ligand. Consequently, it is evident from the H and 13C NMR
H3CR f (PNP)ReH(tCR) + (CH2)8 (2)
that the symmetry of 12a is C1 (four Si-Me resonances in the
1H and 13C NMR spectra), but the chemical shift difference for
several other diastereotopic nuclei is too small to be resolved.
Benzene C-H Cleavage. Because of our frequent observa-
tion of isotope exchange of reactive transients produced here
with d6-benzene solvent (and thus our frequent use of d12-
cyclohexane as an NMR solvent), we have briefly studied this
behavior. In fact, (PNPiPr)Re(H)2[dC(CH2)5], 2a, is 1/3 con-
verted in 10 min at 22 °C in C6H6 to a single compound,
proposed to be (PNPiPr)Re(H)3(Ph). This is apparently an
equilibrium state, since the mole ratio of 2a to this product
remains unchanged over an additional 24 h. Consistent with
the expectation that cyclohexene is being displaced in this
reaction, dissolving 2a in benzene containing 10 equiv of
cyclohexene showed no production of (PNPiPr)Re(H)3(Ph).
C-H Bond Competition Experiments. To evaluate the
relative reactivity of primary (from n-alkanes) vs secondary
(from cycloalkane) vs aromatic C-H bonds, several direct
competition experiments were performed between 8 and mix-
tures of equimolar amounts of two hydrocarbons. These were
carried out to completion at 110 °C, and the product distribution
was assayed by 31P{1H} NMR spectroscopy. These experiments
will not paint a detailed picture of C-H bond selectivity, which
is undoubtedly highly complex.32 Results for the chosen
hydrocarbons are shown in Table 1. Given that the products
are formed under irreversible conditions, these are taken to be
kinetic ratios characteristic of the reactive transient. The near
identity (experiments a and b) of hexane and heptane in the
(PNP)Re(H)2(cyclooctyne) +
H3CR f (PNP)ReH(tCR) + H2 + cyclo-C8H14 (3)
sufficient? (c) Are there any mechanisms whose intermediates
have such a high energy that they can be excluded? (d) How
do reaction thermodynamics help to understand the lack of
equilibrium production of phenyl complexes from benzene with
(PNP)Re(H)4, but detectable phenyl products from (PNPiPr)Re-
(H)2(dC(CH2)5)? (e) What is special about the norbornene
substrate that causes η2-olefin and carbene to be in near
thermoneutral equilibrium, and why is this the only case where
C-C bond scission is observed? (f) Is the norbornylidene isomer
â-agostic? (g) What are the thermodynamics of production of
(PNP)ReH(η3-allyl) + free H2 vs the (PNP)ReH2 (carbene) or
η2-olefin isomers? (h) Is (PNP)Re(H)2(dC(CH3)R) â-agostic,
or is this a feature specific to cyclic carbenes?
Features we have already established from our analogous
study of acyclic alkenes focus on the interconversion between
monosubstituted carbenes (PNP)Re(H)2[dCH(R)] and the hy-
drido carbynes (PNP)ReH(tCR), so these are complementary
to the present paper, in which conversion to carbyne encounters
a new and demanding feature: C-C bond scission. Previously
established features (Scheme 8) are the following:2,24 (1)
(PNPH)Re(H)2, while highly reactive (eq b), is too endergonic
(eq a) by H2 loss from (PNPH)Re(H)4 to be an intermediate in
the reactions of unsaturated hydrocarbons with the latter at 22
°C. (2) (PNPH)Re(H)2(η2-HCCH) is more stable as the less
symmetric isomer (eq c); this is also true for the analogous η2-
H2CdCH2 adduct. (3) H2 loss from (PNPH)Re(H)3(tCCH3) is
favorable (eq d), an effect attributed to product stabilization by
π-donation from amide N to Re. (4) Dehydrogenation of
(PNPH)Re(H)2(dC(H)CH3) is also favorable (eq e). The â-ago-
stic methyl C-H donor contributes to this rotamer being more
stable than that with the carbene plane rotated 180°. (5) Ethylene
(28) For a thoughtful discussion of the uniqueness of norbornene in this regard,
see: Budzelaar, P. H. M.; Moonen, N. N. P.; DeGelder, R.; Smits, J. M.
M.; Gal, A. W. Eur. J. Inorg. Chem. 2000, 753.
(29) Yamamoto, Y.; Ohno, T.; Itoh, K. Organometallics 2003, 22, 2267.
(30) Gorski, M.; Kochel, A.; Szymanska-Buzar, T. Organometallics 2004, 23,
3037.
(31) Pasquali, M.; Floriani, C.; Gaetani-Manfredotti, A.; Chiesi-Villa, A. J. Am.
Chem. Soc. 1978, 100, 4918.
(32) Vetter, A. J.; Flaschenriem, C.; Jones, W. D. J. Am. Chem. Soc. 2005,
127, 12315.
9
6010 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007

Pincer/Rh Complexes Form Metal Carbenes and Carbynes
A R T I C L E S
Scheme 8
Scheme 9
conversion to carbyne and free H2 is (just barely) favorable,
but hydrogenation of available ethylene can dramatically
improve the reaction free energy (eqs f and g). (6) Dehydro-
genation of ethane by (PNPH)Re(H)4 is unfavorable (eq h).
However, if it were possible to make a species more reactive
than (PNPH)Re(H)4, this challenging reaction might be ac-
complished. The present paper deals with this goal. The
computational method is detailed in the Experimental Section.
Scheme 10 a
Equation a of Scheme 9 shows that the alkyne isomer is less
stable than the alkylidyne, so the cyclooctyne complex contains
additional unused chemical potential. However, eq b shows that
the alkyne complex lacks the ability to dehydrogenate ethane
(or a longer chain alkane, as in the present experimental work)
until the ethylene released is hydrogenated to alkane (add eq c
to eq b). Equation d shows that the somewhat more “hydrogen
rich” cyclohexylidene complex (vs alkyne complex) does have
the chemical potential to dehydrogenate ethane, even when free
H2 is the product.
Equation e shows that the η2-norbornene isomer is more stable
than the norbornylidene isomer, in agreement with experiment,
and comparison to eq f shows that the norbornyl fragment makes
the carbene isomer more thermodynamically accessible (i.e., less
unfavorable) than for ethylene. The cyclic C6 case (eq i) actually
reverses the relative stability of these two isomers, accomplish-
ing something that was not true of RuHCl(PiPr3)2 as the π-base.
Comparison of eq g with eq d shows that the norbornylidene
and the cyclohexylidene complexes are comparable in their
thermodynamic potential for ethane dehydrogenation. Conver-
sion of (PNP)Re(H)4 by equimolar propylene to (PNP)ReH-
(η3-allyl) + 2H2 is nearly thermoneutral, but this observed
conversion (with cyclododecene) will be further stabilized by
hydrogenation of olefin to alkane; this explains the observed
production of the unique allyl product 9. Equation h shows that
C-C bond rupture is favorable, in accord with experiment.
a ∆G°298 in kilocalories per mole.
Another provisional way to rank the chemical potential in
the three alkane dehydrogenation precursors employed here is
shown in Scheme 10 (∆G°298 values given in kilocalories per
mole). Since the product is, in each case, an olefin and the same
metal species, the most positive standard free energy change
involves the “most stable” reagent complex. By this criterion,
the highest chemical potential is the alkyne complex and the
lowest the η3-allyl. Calculations show that (PNP)Re(H)2(HCCH)
lacks the thermodynamic potential to convert cyclohexane to
coordinated cyclohexylidene (eq 7, ∆G°298 given in kilocalories
per mole).
(PNP)Re(H)2(HCtCH) +
(CH2)6 +10.58 (PNP)Re(H)2[dC(CH2)5] + C2H4 (7)
Selected Structural Features. Those structures calculated
(on a simplified (H2PCH2SiH2)2N- model) for which an X-ray
structure has been determined showed agreement (see the
9
J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007 6011

A R T I C L E S
Ozerov et al.
Table 2. Selected Structural Features (Å) of DFT-Optimized Species [(H2PCH2SiH2)2N]Re(ligand)
Re
−N
Re
−C
Re‚‚‚Cag
Re‚‚‚Hag
C−C
(PNPH)Re(H)2(η2-norbornene)
(PNPH)Re(H)2(HCtCH)
2.129
2.131
2.183
2.158
2.136
2.030
2.212, 2.212
2.013, 2.013
1.897
1.904
2.226, 2.249
2.178, 2.244, 2.378
2.991
2.430
1.443
1.315
(PNPH)Re(H)2(cyclohexylidene)
(PNPH)Re(H)2(norbornylidene)
(PNPH)Re(H)2(η2-cyclohexene)
(PNPH)Re(H)(C3H5)
2.671
2.861
3.075
2.471
2.848
2.410
1.430
1.437, 1.402
Scheme 11
against â-H migration to which late-transition-metal N(alkyl)2
ligands are vulnerable.35 The Fryzuk ligand also permits
installing variable steric protection at the two phosphorus donor
centers to prevent the dimerization which could destroy the
unsaturation we hoped to maintain. Finally, we wanted not only
the higher reducing power of a 5d metal, but one which has a
lower oxidation state than the +2 of our Ru chemistry. We
therefore settled on (PNPR)Re as our goal. Isolated (PNP)ReI
complexes remain inaccessible. We also have no evidence that
this low oxidation state is ever reached in the reactions we have
studied. It is likely that ReIII is the lowest oxidation state
accessible in this chemistry. This only reinforces the notion of
(PNP)ReI being an exceptionally reactive, reducing fragment.
Furthermore, even though (PNP)ReI compounds many not have
been involved at all, it is convenient to analyze the results from
the point of view of the thermodynamic preferences of the
(PNP)ReI fragment. Magnesium reduction of (PNPR)ReOCl2
proceeded best in the presence of H2, which then represented a
compromise in that we produce (PNPR)Re(H)4. To reach the
desired lower formal oxidation state, we explored scavenging
H from (PNPR)Re(H)4 using sacrificial olefins, which then
become the substrate for the more unsaturated transient “(PNPR)-
Re” moiety. In fact, operationally unsaturated (PNPR)Re(H)4
reacts under mild conditions with olefins, which are indeed
hydrogenated and also converted to rearranged hydrocarbon
fragments which, being more π-acidic than olefins, are more
suited to accommodate the strong π-basicity (reducing power)
of the (PNPR)Re moiety: carbenes, carbynes, a cycloalkyne,
conjugated dienes, an allyl, and even a cyclopentadienyl. At
90 °C, a C-C bond of a coordinated norbornyl skeleton is
cleaved on the basis of the evident thermodynamic driving force
to form a carbyne ligand. These isomerizations of olefins to
non-heteroatom-stabilized carbenes are of interest in understand-
ing olefin metathesis catalyst generation when no obvious
carbene precursor is supplied and show that a sufficiently
reducing metal complex can isomerize an olefin substrate to
the necessary carbene complex,1 provided a hydride ligand is
available. In the norbornene case we describe, the carbene
complex forms on mixing at 22 °C and is in thermal equilibrium
with its olefin isomer.
Supporting Information) sufficient to create confidence in those
calculated structures where an experimental determination is
lacking. For example, (PNPH)Re(H)2(HCtCH) has the alkyne
CtC bond eclipsing the P-Re-P plane and Re-C and C-C
distances (2.01 and 1.32 Å) which compare well with those in
Figure 4. The calculation also captures well the unusually small
-(P-Re-P) (150.3° vs 145.8° in Figure 4).
Cyclohexene and norbornene bind to the (PNP)Re(H)2
fragment similarly to η2-olefin complexes, judging by their
Re-C bonds and leaving the C-C distance shorter (by 0.12
Å). The optimized geometry for coordinated cyclohexene puts
one H on C4 (i.e., not an allylic H) with a close contact to Re.
The η3-allyl complex has the allyl group oriented as observed
in the X-ray structure of 13.
As shown in Table 2, the agostic CH‚‚‚Re interaction is much
weaker or essentially absent in coordinated norbornylidene vs
coordinated cyclohexylidene. This is inversely reflected in the
Re-N distances (i.e., increased N f Re π-donation when the
agostic disappears). This may be related to the fact that the
agostic donor in the norbornylidene is a (less flexible) bridge-
head C-H. This destabilization of the norbornylidene may be
a factor favoring the thermodynamics of C-C cleavage for the
caged C7 compound.
Summary
The reaction chemistry we reported earlier33 for [RuHCl(Pi-
Pr3)2]2 showed that an unsaturated species devoid of strong
π-acid ligands (which diminish the π-basicity of the metal
center) had the ability to isomerize olefins carrying a donor
substituent (e.g., OR or NR2) into Fischer carbenes bound to
Ru. This Ru also doubly dehydrogenated tetrahydrofuran and
N-alkyltetrahydropyrroles into their corresponding metal carbene
complexes.33 However, [RuHCl(PiPr3)2]2 lacked the ability to
convert an olefinic hydrocarbon to a non-heteroatom-substituted
carbene complex. A 5d metal is generally more reducing than
its 4d analogue, and this simple idea rationalizes why Os(H)3Cl-
(PiPr3)2 (this was used in place of the unknown OsHCl(PiPr3)2)
transforms the above substrates more dramatically, sacrificing
the donor substituent from carbon to form a saturated carbyne
isomer of the ruthenium products (Scheme 11).
A brief study shows that the tBu substituents in (PNPtBu)Re-
(H)4 apparently shield the metal so effectively that internal
olefins cannot react. Cyclohexyl substituents pendent on the
reactive rhenium transient are effectively “cannibalized”, via
triple dehydrogenation, but this ligand degradation reaction can
be minimized by feeding excess olefin substrate, so that the
bimolecular reaction becomes at least competitive. Because the
cyclohexyl carbons that are attacked are those farthest from
phosphorus (a fact consistent with intramolecular attack and
requisite flexibility to get these C-H bonds back to the reactive
In seeking still more reducing power to accomplish trans-
formations of non-donor-substituted hydrocarbons, we turned
to a stronger π-donor substituent to replace chloride; for this,
we chose amide. The pincer ligand class pioneered by Fryzuk34
combined this donor with protection (via the SiMe2 substituents)
(33) Ferrando-Miguel, G.; Coalter, J. N., III; Gerard, H. F.; Huffman, J. C.;
Eisenstein, O.; Caulton, K. G. New J. Chem. 2002, 26, 687.
(34) Fryzuk, M. D.; Montgomery, C. D. Coord. Chem. ReV. 1989, 95, 1.
(35) Kuhlman, R.; Folting, K.; Caulton, K. G. Organometallics 1995, 14, 3188.
9
6012 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007

Pincer/Rh Complexes Form Metal Carbenes and Carbynes
A R T I C L E S
Experimental Section
center of the molecule), an effective antidote is to employ the
otherwise sterically similar isopropyl substituent in place of
cyclohexyl. While steric factors play a large role in (PNPR)Re
chemistry, the norbornyl species 10 and 11 show considerable
tolerance of substrate bulk, and adduct 12a*, containing two
norbornyl-derived units, is especially noteworthy. The frequently
formed 13 is an especially sterically efficient way to satisfy the
electronic demands of the (PNP)Re fragment.
General Considerations. All manipulations were performed using
standard Schlenk techniques or in an argon-filled glovebox unless
otherwise noted. Aliphatic and aromatic hydrocarbons (except nor-
bornene) were dried over and distilled or vacuum transferred from NaK/
benzophenone/18-crown-6. Norbornene was dried and vacuum trans-
ferred from Na/K alloy into a sufficiently wide mouthed vessel that
allowed retrieval of solid norbornene with a spatula. The preparation
of compounds 1a,2 1b,3 1c,2 2b3, and 817 was described previously. 1H
NMR chemical shifts are reported in parts per million relative to the
peak for protio impurities in the deuterated solvents. 31P spectra are
referenced to external standards of 85% H3PO4 (at 0 ppm). NMR spectra
Certain of the molecules we have produced from cyclic
olefins, which are thus frustrated from reaching their thermo-
dynamic minimum, (PNP)ReH(tCR), have been examined as
possible reagents for alkane conversion to unsaturated fragments.
Each successful example has the ability to accept 2H from the
alkane, as predicted in eq 1a. Most studied is the cyclooctyne
complex 8. It effectively converts methyl groups to the carbyne
complexes 12, with ethyl groups reacting faster than (hindered)
methyl groups and benzylic methyls showing no major advan-
tage. The lack of formation of any η6-arene complexes is
suggested to be due to the PNP ligand high electron donor
number and also its steric bulk. The driving force for binding
the smaller η5-C5H4R (R ) H or CH3) is evidently strong,
however, since it ultimately forms in every reaction of cyclo-
pentane or methylcyclopentane and then by H migration to the
pincer amide nitrogen, generating a pendent secondary amine;
this amide nitrogen is thus not only an electron pump/acceptor
when PNP is bound η3 to a metal, but it is also a reactive
functionality, involved in bond breaking and making. The allyl
monohydride complex 9 also shows an ability to dehydrogenate
n-heptane. Prerequisite for (PNP)ReLx species to serve as alkane
dehydrogenation reagents is thus the ability to expel the ligands
Lx, especially as an alkane, and that Lx cannot rearrange to a
carbyne without C-C bond scission.
The idea of “stripping” H ligands from an LnMHm species
with a sacrificial olefin (e.g., tBuHCdCH2) has a long tradition
as a way to create a transient capable of attacking unreactive
substrates (e.g., alkanes, N2) and lies at the core of catalytic
alkane dehydrogenation, based on the core 14-electron fragment
(pincer) IrI operating above 200 °C; there, mechanistic details
have been more carefully established36 than in our present survey
of general reagent features and substrate capability. Here, as
there, it is necessary to design the supporting ligand to be
resistant to intramolecular attack by the reactive metal center,
and analogous cases where a cyclohexyl on phosphorus is
dehydrogenated under “hydrogen-poor conditions” (i.e., excess
sacrificial olefin) have been reported.37-39 Our work here has
identified additional general principles, as well as unusual
structural features of some intermediates and final products. In
sum, these establish the (PNPR)Re fragment as one with the
ability to attack unreactive H-C(sp3) bonds, to make an η2-
olefin/carbene isomerization nearly thermoneutral, and even to
cleave C-C bonds. It is accessible to reactants at 22 °C because
of the operationally unsaturated character of (PNP)Re(H)4, and
the H ligands there represent cooperative leaving groups.
1
were recorded with a Varian Gemini 2000 (300 MHz, H; 121 MHz,
31P; 75 MHz, 13C), a Varian Unity Inova instrument (400 MHz, H;
1
162 MHz, 31P; 101 MHz, 13C), or a Varian Unity Inova instrument
(500 MHz, 1H; 126 MHz, 13C). Elemental analyses were performed
by CALI, Inc. (Parsippany, NJ). ORTEP plots were generated using
Ortep-3 for Windows.40 Full details of compounds not shown below
are available as Supporting Information.
Computational Details. All calculations were performed with the
Gaussian 98 package41 at the B3PW9142 level of theory, with
unrestricted wave functions used for all triplet-state calculations. Basis
sets used included LANL2DZ for Re and Si and 6-31G** for C, N,
and H.43 The basis set LANL2DZ is the Los Alamos National
Laboratory ECP plus a double-ú valence on Re, P, and Si;44 additional
d polarization functions45 were added to all phosphorus and silicon
atoms in all DFT calculations. All optimizations were performed with
C1 symmetry, and all minima were confirmed by analytical calculation
of frequencies, which were also used to compute zero-point energy
corrections without scaling.
(PNPiPr)ReH2(dC(CH2)5) (2a). (PNPiPr)ReH4 (1a) (60 mg, 103
µmol) was dissolved in 0.5 mL of cyclohexene. The mixture was
allowed to stand for 30 min, during which time the color became orange.
The volatiles were removed in vacuum, and the oily residue was
dissolved in cyclohexane-d12 for NMR characterization. The product
thus obtained was 99% pure by NMR, but all attempts to crystallize it
1
failed. H NMR (C6D12): δ 2.53 (br m, 2H, CH2), 1.55-1.65 (8H,
CH2 and PCH overlapping), 1.35-1.45 (CH2 overlapping with C6D11H),
2.65 (m, JHH ) 7 Hz, 2H, PCH), 1.15 (AB dvt, JHH ) 15 Hz, JHP ) 4
Hz, 2H, PCH2Si), 1.00-1.12 (four apparent q (dvt) overlapping, for
each 8 Hz, 6H, PCHCH3), 0.65 (AB dvt, JHH ) 15 Hz, JHP ) 4 Hz,
2H, PCH2Si), 0.15 (s, 6H, SiCH3), 0.10 (s, 6H, SiCH3), -6.7 (v br,
1H, ReH), -10.0 (v br, 1H, ReH). 31P{1H} NMR (C6D12): δ 47.4 (s).
31P NMR (C6D12), selectively decoupled from alkyl hydrogens: δ 47.4
(t, 12 Hz). 13C{1H} NMR (C6D12): δ 264.9 (t, 9 Hz, RedC), 57.4 (s,
RedCCH2), 30.2 (t, 9 Hz, PCH), 29.9 (s, CH2), 29.5 (t, 14 Hz, PCH),
27.3 (s, CH2), 25.7 (s, CH2), 20.3 (s, PCHCH3), 19.9 (s, PCHCH3),
19.0 (br s, 2 s overlapping, PCHCH3), 15.9 (br s, PCH2Si), 15.6 (s,
RedCCH2), 5.6 (s, SiCH3), 5.4 (br s, SiCH3). 13C NMR (selected
resonances, C6D12): δ 57.4 (t, JCH ) 127 Hz, RedCCH2), 15.6 (t, JCH
) 117 Hz, RedCCH2).
Thermolyses of 2a. (i) In Heptane + 0 equiv of Cyclohexene. 2a
was prepared in situ from 1a (52 mg, 89 µmol) and cyclohexene (150
µL, 1480 µmol) in heptane. The volatiles were removed, and the residue
(free from cyclohexene) was redissolved in 0.6 mL of heptane. NMR
(40) Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565.
(41) Gaussian 98, revision A.7: Frisch, M. J.; et. al. See the Supporting
Information for the full citation.
(42) Becke, A. D. Phys. ReV. 1988, A38, 3098. Becke, A. D. J. Chem. Phys.
1993, 98, 1372. Becke, A. D. J. Chem. Phys. 1993, 98, 5648. Perdew, J.
P.; Wang, Y. Phys. ReV. B 1992, 45, 13244.
(36) Zhu, K.; Achord, P. D.; Zhang, Z.; Krogh-Jespersen, K.; Goldman, A. S.
J. Am. Chem. Soc. 2004, 126, 13044.
(37) Mohammad, H. A. Y.; Grimm, J. C.; Eichele, K.; Mack, H.-G.; Speiser,
B.; Novak, F.; Quintanilla, M. G.; Kaska, W. C.; Mayer, H. A. Organo-
metallics 2002, 21, 5775.
(38) Baya, M.; Buil, M. L.; Esteruelas, M. A.; Onate, E. Organometallics 2004,
23, 1416.
(39) Borowski, A. F.; Sabo-Etienne, S.; Christ, M. L.; Donnadieu, B.; Chaudret,
B. Organometallics 1996, 15, 1427.
(43) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(44) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. Wadt, W. R.; Hay,
P. J. J. Chem Phys. 1985, 82, 284. Hay, P. J.; Wadt, W. R. J. Chem. Phys.
1985, 82, 299.
(45) Ho¨llwarth, A.; Bo¨hme, M.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.; Jonas,
V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys.
Lett. 1993, 208, 237.
9
J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007 6013

A R T I C L E S
Ozerov et al.
analysis indicated a >99% purity of 2a. This solution was allowed to
stand at 22 °C for 24 h, but only traces of 1, 12d, and 3 were observed.
After 1 h at 90 °C a 27:60:0:13 ratio of 1a, 12d, 2a, and 3 was observed.
Observation of (PNPiPr)ReH2(dC(CH2)7) (7a). 1a (30.0 mg, 51.5
µmol) was dissolved in 0.6 mL of heptane and treated with cyclooctene
(67 µL, 515 µmol). The evolution of the reaction mixture was followed
by 31P NMR. After 45 min at 22 °C, 7a was observed (14%) along
with the remaining 1a. After 18 h at 22 °C, 8 was the major component
(70%) along with a small amount of 7a. After the mixture was heated
at 90 °C for 2 h, complete conversion to 8 was observed. NMR data
for 7a follow. 31P{1H} NMR (C7H16): δ 45.8 (s). 31P NMR, with
selective decoupling of aliphatic hydrogens (C7H16): δ 45.8 (dd).
Observation of (PNPCy)ReH2(dC(CH2)7) (7b). 1b (41 mg, 53
µmol) was dissolved in 0.6 mL of heptane and treated with 0.1 mL of
cyclooctene (0.77 mmol). The evolution of the reaction mixture was
(ii) In Heptane + 8 equiv of Cyclohexene. 1a (30.0 mg, 51.5 µmol)
and cyclohexene (52.2 µL, 515 µmol) were mixed in 0.6 mL of heptane.
NMR analysis indicated quantitative conversion to 2a. This solution
was heated at 90 °C for 2 h, resulting in a 15:85 mixture of 12d and
3. Further heating (90 °C) did not change this ratio.
(iii) In Heptane + 2 equiv of Cyclohexene. 1a (39.6 mg, 68.0 µmol)
and cyclohexene (27.5 µL, 272 µmol) were mixed in 0.6 mL of heptane.
NMR analysis indicated quantitative conversion to 2a. This solution
was heated at 110 °C for 40 min, resulting in a 43:57 mixture of 12d
and 3. Further heating (110 °C) did not change this ratio.
1
followed by 31P and H NMR. After 1.5 h at 22 °C, 7b was observed
(20%) along with unreacted 1b (80%). The mixture was allowed to
stand at 22 °C, and over 2 d, conversion to 13 was observed. Extensive
H/D scrambling involving the C6D6 NMR solvent, cyclooctene, and
(iv) In Neohexane + 1.3 equiv of cyclohexene. 1a (39.6 mg, 68.0
µmol) and cyclohexene (22.6 µL, 224 µmol) were mixed in 0.6 mL of
neohexane. NMR analysis indicated quantitative conversion to 2a. This
solution was heated at 110 °C for 18 h, resulting in a 8:92 mixture of
12g and 3.
1
the PNP ligand was registered. NMR data for 7b follow. H NMR
(C6D6, selected resonances): δ 0.47 (s, 6H, SiCH3), 0.42 (s, 6H, SiCH3),
-6.08 (br, 1H, ReH), -7.96 (br, 1H, ReH). 31P{1H} NMR (C6D6): δ
35.6 (br s).
(PNPiPr)ReH2(cyclooctyne) (8) was reported previously.17
Reaction of 2a with C6H6. 1a (20.0 mg, 34.3 µmol) was dissolved
in 0.1 mL (ca. 1 mmol) of cyclohexene. The reaction mixture turned
orange. The mixture was allowed to stand for 1 h, and then the volatiles
were removed. The residue was dissolved in cyclohexane, and then
the volatiles were removed. The residue was dissolved in C6H6, and
the evolution of this solution was monitored by 31P NMR. After 10
min at 22 °C the mixture consisted of a 35:65 mixture of the tentatively
identified (PNPiPr)ReH3Ph [31P{1H} NMR (C7D8): δ 57.1 (s)] and 2a.
This corresponds to Keq ≈ 1 × 10-3 (∆G ≈ +4 kcal/mol) for the
displacement of C6H10 by C6H6. The mixture was allowed to stand for
24 h at 22 °C. The ratio of (PNPiPr)ReH3Ph to 2a did not change
significantly, and traces of 5a, 1a, and a few unidentified compounds
were observed. Dissolution of 2a (from 50.0 mg (86.8 µmol) of 1a) in
benzene in the presence of excess (87 µL or 860 µmol) cyclohexene
does not lead to formation of (PNPiPr)ReH3Ph; only 2a was present.
Upon dissolution of 2a in toluene, a broad resonance (or overlapping
resonances) at δ ca. 57 ppm was also observed in the 31P NMR
spectrum.
Decomposition of 8. 8 (30 mg, 44 µmol) was heated in 0.6 mL of
cyclooctane for 48 h. After this time, 31P NMR analysis indicated near
complete disappearance of 8. More than 30 resonances were observed
by 31P NMR in the 20-90 ppm region. Thermolysis of 8 (30 mg) in
(a) cyclohexane, (b) 2,3-dimethylbutane, and (c) a 5:1 mixture of
cyclohexane and SiMe4 for 48 h at 110 °C led to essentially the same
mixtures.
(PNPiPr)ReH(cyclododecenyl) (9). 1a (61 mg, 105 µmol) and
cyclododecene (0.200 mL, 1.05 mmol, mixture of cis- and trans-
isomers) were dissolved in 0.6 mL of hexamethyldisiloxane and heated
for 3 h at 90 °C. The deep-blue solution was transferred to a 10 mL
flask with a Kontes stopcock and dried in vacuo at 90 °C for 15 min.
The oily residue was dissolved in 1 mL of a SiMe4/pentane mixture
and placed in the freezer at -30 °C for 2 days. The blue crystalline
precipitate was isolated by decantation, washed with cold SiMe4, and
dried in vacuo. Yield: 56 mg (72%). Ten minutes after dissolution at
22 °C, NMR analysis shows a 10:1 mixture of isomers; after <1 h at
22 °C, the mixture equilibrates to a ca. 5:1 ratio. Anal. Calcd (Found)
for C30H66NP2ReSi2 (9): C, 48.35 (48.26); H, 8.93 (8.88); N, 1.88
(1.79).
(PNPCy)ReH2(dC(CH2)5) (2b) was reported previously.18
Thermolysis of 2b. (i) In Solution. A crystalline sample of 2b (15
mg, 18 µmol) was dissolved in heptane and heated at 90 °C for 30
min. Subsequent NMR analysis detected a 1:2 mixture of 1b and 13.
Data for the Major Isomer of 9. 1H NMR (C6D6): δ 4.11 (m, 1H,
allyl CH), 2.99 (m, 1H, allyl CH), 2.81 (m, 1H, allyl CH), 2.49 (br d,
12 Hz, 1H), 2.24 (quintet, 6 Hz, 1H), 1.3-2.05 (several m, C12 ring
CH2 groups and PCH groups), 1.23 (dd, 8 Hz, 13 Hz, 3H, PCHCH3),
1.15 (dd, 8 Hz, 13 Hz, 3H, PCHCH3), 1.09 (dd, 8 Hz, 13 Hz, 3H,
PCHCH3), 1.05 (2 dd overlapping, 6H, PCHCH3), 0.99 (2 dd overlap-
ping, 6H, PCHCH3), 0.75 (dd, 8 Hz, 13 Hz, 3H, PCHCH3), 0.63 (s,
3H, SiCH3), 0.59 (s, 3H, SiCH3), 0.28 (s, 3H, SiCH3), 0.12 (s, 3H,
SiCH3), -11.96 (dd, 14 Hz, 20 Hz, 1H, ReH). PCH2Si proton signals
are concealed by other aliphatic resonances. 31P{1H} NMR (C6D6): δ
30.0 (AB d, 241 Hz), 26.1 (AB d, 241 Hz). 13C{1H} NMR (C6D6): δ
71.6 (s, CH of the C12 ring), 46.7 (s, CH of the C12 ring), 42.8 (s, CH
of the C12 ring), 36.8 (s), 32.6 (d, 26 Hz, PCH), 33.5 (s), 31.3 (d, 25
Hz, PCH), 29.8 (s), 29.4 (s), 28.4 (s), 25.0-27.5 (several s, PCH and
CH2 or the C12 ring), 21.1 (s, PCHCH3), 20.1 (s, PCHCH3), 19.44 (s,
PCHCH3), 19.39 (s, PCHCH3), 19.30 (s, PCHCH3), 18.9 (s, PCHCH3),
18.4 (s, PCHCH3), 17.2 (s, PCHCH3), 13.5 (s, PCH2Si), 11.3 (s, PCH2-
Si), 7.9 (s, SiCH3), 6.6 (s, SiCH3), 5.0 (2 s overlapping, SiCH3).
Data for the Minor Isomer of 9. 1H NMR (C6D6): δ 4.22 (m, 1H,
allyl CH), 3.69 (m, 1H, allyl CH), 3.38 (m, 1H, allyl CH), 0.49 (s, 3H,
SiCH3), 0.41 (s, 3H, SiCH3), 0.34 (s, 3H, SiCH3), 0.31 (s, 3H, SiCH3),
-11.91 (dd, 14 Hz, 20 Hz, 1H, ReH). Other resonances are obscured
and cannot be reliably assigned. 31P{1H} NMR (C6D6): δ 29.4 (AB d,
243 Hz), 28.6 (AB d, 243 Hz). 13C{1H} NMR (C6D6): δ 67.2 (s, CH
of the C12 ring), 46.1 (s, CH of the C12 ring), 42.1 (s, CH of the C12
(ii) In the Solid State. A crystalline sample of 2b (15 mg, 18 µmol)
was heated (120 °C, 30 min, in vacuo) in a vacuum transfer apparatus,
on the other end of which a J. Young tube with 0.6 mL of C6D6 was
attached and frozen in liquid nitrogen to trap the volatile products. The
solid turned red-purple upon thermolysis. The trap was thawed and
analyzed by 1H NMR spectroscopy. A mixture of cyclohexene and
cyclohexane in a 55:45 ratio was observed. The purple-red residue was
dissolved in C6D6 and analyzed by 31P NMR spectroscopy. A 63:37
mixture of 13 and 1b was measured. If all of 2b underwent a thermal
loss of the C6 fragment in a 55:45 ratio of C6H10 to C6H12, then a
theoretical ratio of 61:39 of 13 to 1b would have been expected in the
residue.
(K2-P,P-PN(H)PCy)ReH2(η5-Cp) (6b). (PNPCy)ReH4 (1b) (200 mg,
260 µmol) was dissolved in 2 mL of heptane, and cyclopentene (0.3
mL, 3.5 mmol) was added to it. The mixture was heated at 90 °C for
3 h, and then the resultant colorless solution was placed into the -30
°C freezer. The next day, colorless crystalline material was collected
by decantation of the solvent. Yield: 0.186 g (89%). 1H NMR (C6D6):
δ 4.67 (s, 5H, η-C5H5), 1.20-2.08 (several m, 48H, PC6H11, PCH2Si),
0.30 (s, 12 H, SiMe), -12.11 (t, 2H, 41 Hz, ReH2). 13C{1H} NMR
(C6D6): δ 75.0 (η-C5H5), 45.5 (m, PCH), 30.7 (s, CH2 of Cy), 29.7 (s,
CH2 of Cy), 28.0 (t, 5 Hz, CH2 of Cy), 27.8 (t, 5 Hz, CH2 of Cy), 27.4
(s, CH2 of Cy), 4.9 (s, SiMe). 31P{1H} NMR (C6D6): δ 16.8 (s). Anal.
Calcd (Found) for C35H68NP2ReSi2 (6b): C, 52.08 (52.17); H, 8.49
(8.80); N, 1.74 (1.67).
9
6014 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007

Pincer/Rh Complexes Form Metal Carbenes and Carbynes
A R T I C L E S
ring), multiple singlets in the 16-36 ppm region cannot be reliably
assigned, 10.7 (s, PCH2Si), 9.8 (s, PCH2Si), 7.8 (s, SiCH3), 6.4 (s,
SiCH3), 5.9 (2 s overlapping, SiCH3).
12a* to 12a, and then the residue was dissolved in C6D6 for NMR
characterization.
1
Data for 12a. H NMR (C6D6): δ 1.4-2.5 (several m, PCH and
3-methylcyclopentyl), 1.31 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.21
(apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.08 (apparent q (dvt), 8 Hz,
6H, PCHCH3), 1.06 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 0.94 (d,
3H, 8 Hz, 3-H3C(cyclopentyl)), 0.82 (AB dvt, JHH ) 14 Hz, JHP ) 4
Hz, 2H, PCH2Si), 0.78 (AB dvt, JHH ) 14 Hz, JHP ) 4 Hz, 2H, PCH2-
Si), 0.370 (two s overlapping, 6H, SiCH3), 0.324 (s, 3H, SiCH3), 0.319
(s, 3H, SiCH3), -10.57 (t, 15 Hz, ReH). 31P{1H} NMR (C6D6): δ 60.8
(s). 13C{1H} NMR (C6D6): 281.7 (br t, 12 Hz, RetC), 62.2 (s, Ret
CCH), 41.1 (s, methylcyclopentyl), 34.6 (s, methylcyclopentyl), 34.0
(s, methylcyclopentyl), 30.7 (s, methylcyclopentyl), 29.2-29.8 (4
overlapping m, PCH), 21.0 (t, 5 Hz, 1C, PCHCH3), 20.8 (t, 5 Hz, 1C,
PCHCH3), 20.7 (s, methylcyclopentyl), 19.2 (br dd, 2C, PCHCH3), 19.1
(br t, 2C, PCHCH3), 17.2 (br s, 2C, PCHCH3), 10.4 (br s, PCH2Si),
7.5 (s, 1C, SiCH3), 7.4 (s, 1C, SiCH3), 6.3 (br dd, 2C, SiCH3).
Alkane Activation by 9. 1a (15.0 mg, 25.8 µmol) and cyclododecene
(16.4 µL, 85.0 µmol) were dissolved in 0.6 mL of heptane. This mixture
was heated and periodically monitored by NMR. The ratio of 1a to
12d to 9 observed after (a) 2 h at 22 °C was 99:0:1, (b) that after 1.5
h at 90 °C was 41:3:56, (c) that after 18 h at 90 °C was 13:63:24, and
(d) that after 3 h at 110 °C was 2:98:0. The volatiles were removed
1
from the mixture, and the red residue was dissolved in C6D6. H, 13C,
and 31P NMR data were collected and were identical to 12d obtained
by other methods.
(PNPiPr)ReH2(norbornene) (10a) and (PNPiPr)ReH2(norbornylidene)
(11a). 1a (30 mg, 52 µmol) was dissolved in 0.6 mL of pentane and
treated with norbornene (34 mg, 360 µmol). The mixture immediately
became deep blue-purple, and NMR analysis showed essentially
quantitative conversion to the products. The volatiles were removed in
a vacuum, and the residue was redissolved in 0.5 mL of Me3SiOSiMe3
and placed in the freezer at -30 °C overnight. The blue-purple crystals
were separated by decantation, washed with cold SiMe4, and dried in
vacuo. Yield: 31 mg (89%). Upon dissolution, a mixture of 10a and
11a was observed by NMR (ca. 8:1 ratio) in solution. Anal. Calcd
(Found) for C25H56NP2ReSi2 (10a/11a): C, 44.48 (44.32); H, 8.36
(8.42); N, 2.07 (1.97).
1
Data for 12a*. H NMR (C6D6, selected resonances): δ 0.26 (s,
6H, SiCH3), 0.19 (s, 6H, SiCH3). 31P{1H} NMR (C6D6): δ 44.0 (s).
13C{1H} NMR (C7H14, selected resonance): 270.3 (t, 12 Hz, RetC).
(PNPiPr)ReH(tCCH2Ph) (12h). 8 (25 mg, 36 µmol) was dissolved
in 0.6 mL of ethylbenzene in a J. Young tube. The tube was inserted
into a 110 °C oil bath. After 24 h the conversion was essentially
quantitative by NMR. The volatiles were removed from the solution
in vacuo, leaving behind a viscous red oil. It was redissolved in C6D6
for full NMR characterization. 1H NMR (C6D6): δ 7.44 (d, 8 Hz, 2H,
o-Ph), 7.16 (t, 8 Hz, 2H, m-Ph), 7.04 (t, 8 Hz, 1H, p-Ph), 2.85 (m, 2H,
RetCCH2), 1.91 (m, JHH ) 7 Hz, 2H, PCH), 1.75 (m, JHH ) 7 Hz,
2H, PCH), 1.18 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.11 (apparent
q (dvt), 8 Hz, 6H, PCHCH3), 1.03 (apparent q (dvt), 8 Hz, 6H,
PCHCH3), 0.96 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 0.83 (AB dvt,
JHH ) 15 Hz, JHP ) 4 Hz, 2H, PCH2Si),), 0.73 (AB dvt, JHH ) 15 Hz,
JHP ) 4 Hz, 2H, PCH2Si), 0.39 (s, 6H, SiCH3), 0.33 (s, 6H, SiCH3),
-9.69 (t, 14 Hz, 1H, ReH). 31P{1H} NMR (C6D6): δ 60.34 (s). 13C-
{1H} NMR (C6D6): δ 273.0 (t, 12 Hz, RetC), 136.2 (s, arom C),
128.8 (s, arom CH), 127.8 (s, arom CH), 126.2 (s, arom CH), 56.4 (s,
RetCCH2), 29.5 (t, 13 Hz, PCH), 29.0 (t, 11 Hz, PCH), 20.1 (s,
PCHCH3), 18.9 (s, PCHCH3), 18.8 (s, PCHCH3), 17.3 (s, PCHCH3),
10.5 (s, PCH2Si), 7.2 (s, SiCH3), 6.3 (t, 2 Hz, SiCH3).
Data for the Major Isomer (Norbornene Complex 10a). 1H NMR
(C7D8): δ 3.01 (br s, 2H), 2.67 (br s, 2H), 2.65 (m, JHH ) 7 Hz, 2H,
PCH), 1.80 (m, JHH ) 7 Hz, 2H, PCH), 1.59 (AB dvt, JHH ) 15 Hz,
JHP ) 4 Hz, 2H, PCH2Si), 1.51 (br, 4H), 1.25 (apparent q (dvt), 8 Hz,
6H, PCHCH3), 1.11 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.08
(apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.06 (apparent q (dvt), 8 Hz,
6H, PCHCH3), 0.64 (AB dvt, JHH ) 15 Hz, JHP ) 4 Hz, 2H, PCH2Si),
0.39 (s, 6H, SiCH3), 0.25 (s, 6H, SiCH3), -1.51 (d, 8 Hz, 1H), agostic
1
CH, -12.7 (v br, 1H, ReH), -15.8 (v br, 1H, ReH). H NMR (C7D8,
-40 °C, selected resonances): δ -12.65 (t, 10 Hz, 1H, ReH), -15.82
(t, 21 Hz, 1H, ReH). 31P{1H} NMR (C7D8): δ 44.4 (s). 13C{1H} NMR
(C7D8): δ 48.9 (s, norbornene), 38.8 (t, 6 Hz, norbornene), 34.9 (s,
norbornene), 33.1 (s, norbornene), 32.6 (t, 12 Hz, PCH), 28.7 (t, 9 Hz,
PCH), 20.2 (s, PCHCH3), 19.7 (s, PCHCH3), 18.7 (s, PCHCH3), 18.6
(s, PCHCH3), 16.1 (br s, PCH2Si), 5.9 (s, SiCH3), 4.5 (br s, SiCH3).
Data for the Minor Isomer (Norbornylidene Complex 11a),
Selected Resonances. 1H NMR (C7D8, -40 °C): δ -5.98 (br s, ReH),
-13.04 (br s, ReH); at 22 °C the hydride peaks are too broad to be
detected. 13C{1H} NMR (C7D8): δ 281.0 (br, RedC). 31P{1H} NMR
(C7D8): δ 51.6 (AB d, 187 Hz), 47.1 (AB br d, 187 Hz).
Activation of C6HnMe6-n. Three portions of 8 (each 25 mg, 36
µmol) were dissolved in 0.6 mL of toluene, p-xylene, and mesitylene,
respectively. These solutions were thermolyzed at 110 °C for 24 h and
then analyzed by 31P NMR. In each case, the solution became brown-
red and a 31P NMR resonance assignable to an aryl-substituted carbyne
emerged (ca. 40% for 12i, ca. 50% for 12j, ca. 60% for 12k). Upon
partial decoupling of aliphatic hydrogens, these 31P NMR resonances
were observed as doublets. The highly lipophilic nature of the resultant
mixtures did not allow us to isolate the products in a pure state.
Data for 12i. 31P{1H} NMR (C7H8): δ 59.3 (s).
(PNPCy)ReH2(norbornene) (10b) and (PNPCy)ReH2(norbornylidene)
(11b). 1b (44 mg, 56 µmol) was dissolved in 0.6 mL of C6D6 and
treated with norbornene (35 mg, 380 µmol). The mixture turned purple
immediately. Extensive H/D scrambling occurred as evidenced by the
broad multiplets in the 31P{1H} NMR spectrum and the low intensity
of the hydride signals. Because of the great congestion in the 1H NMR
spectrum, only selected resonances could be assigned. Removal of
volatiles (including norbornene) and dissolution of the residue in toluene
led to decomposition to a mixture of 13 and 1b, complete in 24 h at 22
Data for 12j. 31P{1H} NMR (p-xylene): δ 59.4 (s).
Data for 12k. 31P{1H} NMR (mesitylene): δ 59.3 (s).
(PNPiPr)ReH(tC(cyclohexyl)) (12l). 8 (25 mg, 36 µmol) was
dissolved in 0.6 mL of methylcyclohexane in a J. Young tube. The
tube was inserted into a 110 °C oil bath. After 24 h the conversion
was essentially quantitative by NMR. The volatiles were removed from
the solution in vacuo, leaving behind a viscous red oil. It was
1
°C (the end ratio of 13 to 1b was 2:1). NMR data for 10b follow. H
NMR (C6D6): δ 0.49 (s, 6H, SiMe), 0.32 (s, 6H, SiMe), -1.47 (d, 8
Hz, 1H), agostic CH, -12.7 (v br, 1H, ReH), -15.9 (v br, 1H, ReH).
31P{1H} NMR (C6D6): δ 35.8 (br s). A minor (ca. 15%) isomer (11b)
was also detected by 31P NMR at δ 40.1 (br m). Anal. Calcd (Found)
for C37H72NP2ReSi2 (10b/11b): C, 53.20 (52.98); H, 8.69 (8.51); N,
1.68 (1.69).
(PNPiPr)ReH(tC(3-methylcyclopentyl)) (12a) and (PNPiPr)ReH-
(tC(3-methylcyclopentyl))(norbornene) (12a*). 1a (40 mg, 68 µmol)
was dissolved in 0.5 mL of heptane, and norbornene (32 mg, 340 µmol)
was added to it. The mixture was heated for 2 h at 90 °C. NMR
indicated complete conversion to 12a*. The volatiles were removed in
vacuo, the oily residue was heated at 110 °C in vacuo for 2 h to convert
1
redissolved in C6D6 for full NMR characterization. H NMR (C6D6):
δ 2.17 (m, JHH ) 7 Hz, 2H, PCH), 2.11 (m, 1H, RetCCH), 1.76 (m,
JHH ) 7 Hz, 2H, PCH), 1.0-1.7 (several m of the CH2 groups of
cyclohexyl), 1.32 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.21 (apparent
q (dvt), 8 Hz, 6H, PCHCH3), 1.08 (apparent q (dvt), 8 Hz, 6H,
PCHCH3), 1.06 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 0.83 (AB dvt,
JHH ) 14 Hz, JHP ) 4 Hz, 2H, PCH2Si), 0.75 (AB dvt, JHH ) 14 Hz,
JHP ) 4 Hz, 2H, PCH2Si), 0.38 (s, 6H, SiCH3), 0.33 (s, 6H, SiCH3),
-10.33 (t, 15 Hz, 1H, ReH). 31P{1H} NMR (C6D6): δ 61.13 (s). 13C-
{1H} NMR (C6D6): δ 283.8 (t, 11 Hz, RetC), 60.9(s, RetCCH), 31.6
9
J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007 6015

A R T I C L E S
Ozerov et al.
(s, CH2 of cyclohexyl), 29.8 (t, 11 Hz, PCH), 29.3 (t, 13 Hz, PCH),
26.8 (s, CH2 of cyclohexyl), 26.4 (s, CH2 of cyclohexyl), 20.9 (s,
PCHCH3), 19.3 (s, PCHCH3), 19.0 (s, PCHCH3), 17.2 (s, PCHCH3),
10.3 (s, PCH2Si), 7.5 (s, SiCH3), 6.3 (br s, SiCH3).
JHP ) 4 Hz, 2H, PCH2Si), 0.37 (s, 6H, SiCH3), 0.33 (s, 6H, SiCH3),
-10.58 (t, 15 Hz, 1H, ReH). 31P{1H} NMR (C6D6): δ 60.6 (s). 31P
NMR, with selective decoupling of aliphatic hydrogens (C6D6): δ 60.6
(d).
(PNPiPr)ReH(tC(cyclopentyl)) (12m) and (K2-P,P-PN(H)PiPr)-
ReH2(η5-C5H4Me) (15). 1a (26 mg, 45 µmol) was dissolved in 0.5
mL of methylcyclopentane, and 0.2 mL cyclooctene was added to it.
The mixture was heated at 110 °C for 18 h. A 77:23 mixture of 12m
and 15 formed (NMR evidence in situ). The volatiles were removed in
vacuo, and the residue was redissolved in C6D6 for NMR characteriza-
tion. Extensive overlap precluded assignment of all the 1H NMR
resonances in this mixture. Therefore, only selected NMR data are
reported.
Data for 15. H NMR (C6D6): δ 4.58 (br t, 2H, CH of η5-C5H4-
1
Me), 4.39 (br t, 2H, CH of η5-C5H4Me), 0.24 (s, 12H, SiCH3), -12.30
(t, 2H, 41 Hz, ReH). 31P{1H} NMR (C6D6): δ 27.1 (s). 31P NMR, with
selective decoupling of aliphatic hydrogens (C6D6): δ 27.1 (t).
Acknowledgment. This work was supported by the Depart-
ment of Energy.
Supporting Information Available: CIF files for X-ray
structures and drawings of DFT-geometry-optimized structures.
This material is available free of charge via the Internet at
http://pubs.acs.org.
1
Data for 12m. H NMR (C6D6): δ 1.30 (apparent q (dvt), 8 Hz,
6H, PCHCH3), 1.21 (apparent q (dvt), 8 Hz, 6H, PCHCH3), 1.02-
1.14 (two apparent q (dvt) overlapping, 12H, PCHCH3), 0.83 (AB dvt,
JHH ) 14 Hz, JHP ) 4 Hz, 2H, PCH2Si), 0.76 (AB dvt, JHH ) 14 Hz,
JA062327R
9
6016 J. AM. CHEM. SOC. VOL. 129, NO. 18, 2007