A rhenium-cyclohexane complex with preferential binding of axial C-H bonds: A probe into the relative ability of C-H, C-D, and C-C bonds as hyperconjugative electron donors?

Article abstract of DOI:10.1002/anie.200600313

Which bonds bind better? According to NMR experiments with the alkane complex [CpRe(CO)₂(cyclohexane)] at low temperatures, complexation with an axial proton is preferred over that through an equatorial one by about 1.5 kJ mol^-1. Cal

Full text of DOI:10.1002/anie.200600313

Communications
Alkane Ligands
DOI: 10.1002/anie.200600313
A Rhenium–Cyclohexane Complex with
ꢀ
Preferential Binding of Axial C H Bonds: A
ꢀ
ꢀ
Probe into the Relative Ability of C H, C D, and
ꢀ
C C Bonds as Hyperconjugative Electron
Donors?**
Douglas J. Lawes, Tamim A. Darwish, Timothy Clark,*
Jason B. Harper,* and Graham E. Ball*
Dedicated to Professor Gernot Frenking
on the occasion of his 60th birthday
Formation of alkane s complexes with agostic-like interac-
ꢀ
tions is widely accepted as being a prerequisite for C H
activation by transition-metal centers and has therefore
received considerable attention by experimental and theo-
retical chemists.[1–3] One important aspect of s-complex or
agostic interactions is their specificity. Do, for instance, such
interactions prefer h2-C,H, h1-H, h2-H,H, or h3-H,H,H binding
(Scheme 1)[1b] and can this preference be influenced by
electronic differences between the hydrogen atoms?
We now describe experimental (NMR) and theoretical
investigations of an alkane complex, [CpRe(CO)2-
(cyclohexane)] (1), in which competition exists between
ꢀ
axial and equatorial C H bonds for binding to the metal
Scheme 1. Possible binding modes for s-complex or agostic interac-
tions.
[*] Prof. Dr. T. Clark
Computer-Chemie-Centrum
Friedrich-Alexander-Universität Erlangen-Nürnberg
Nägelsbachstrasse 25, 91052 Erlangen (Germany)
Fax: (+49) 9131-8526565
E-mail: Tim.Clark@chemie.uni-erlangen.de
D. J. Lawes, T. A. Darwish, Dr. J. B. Harper, Dr. G. E. Ball
School of Chemistry
University of New South Wales
Sydney NSW 2052 (Australia)
Fax: (+61) 2-9385-6141
E-mail: j.harper@unsw.edu.au
g.ball@unsw.edu.au
[**] We acknowledge Prof. W. Adcock (Flinders University) and Prof. M.
Paddon-Row (UNSW) for early communication of results and
helpful discussions, and the Australian Research Council for
funding and an Australian Postgraduate Award to D.J.L. The work at
Erlangen was supported by the Deutsche Forschungsgemeinschaft
as part of SFB583 “Redox-active metal complexes: Control of
reactivity via molecular architecture”.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
4486
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4486 –4490

Angewandte
Chemie
center. Given the role of alkane complexes as intermediates
0.49 ppm (Figure 1a). These resonances are due to the
ꢀ
in the C H activation reaction, knowledge of the selectivity of
alkane coordination sites may prove relevant to directing the
formation of [CpRe(CO)2(cyclohexane)] (1). The experiment
was performed using either 60:40 [D12]pentane/cyclohexane
or 8:82:10 [D12]pentane/isobutene/cyclohexane to allow very
ꢀ
site selectivity of C H activation products. Complex 1 allows
the order of relative hyperconjugative donor strengths of
ꢀ
ꢀ
ꢀ
bonds, C H > C D > C C, to be probed experimentally. Our
experiments and calculations suggest that the interaction of
the alkane with the metal is primarily with one hydrogen atom
of the methylene unit and that the axially complexed isomer is
preferred.
Some of us have previously reported the direct NMR
observation of an alkane complex in solution, [CpRe(CO)2-
(cyclopentane)].[3a,b] This complex was generated through
exposure of a photolabile rhenium precursor, [CpRe(CO)3],
dissolved in cyclopentane to UV/visible light at 180 K
[Eq. (1)]. A resonance is observed at d = ꢀ2.32 ppm and
Figure 1. Signals due to the protons of a bound methylene unit in
a,b) [CpRe(CO)2(cyclohexane)] (1), c) a mixture of 1 and
hn, 180 K
½CpReðCOÞ Š
½CpReðCOÞ ðcyclopentaneފ
ð1Þ
ƒƒƒƒƒƒƒƒƒ!
3
2
[CpRe(CO)2([2,2,3,3,5,5,6,6-D8]cyclohexane-h2-C1,H1)] ([D8]-1), and
d) [CpRe(CO)2([13C6]cyclohexane)] ([13C6]-1) in their 1H NMR spectra at
173 K.[5] The spectra in (a), (c), and (d) have line broadening of 5 Hz
and the spectrum in (b) is resolution-enhanced with Gaussian multi-
plication. All spectra shown were obtained for 8:82:10 [D12]pentane/
isobutene/cyclohexane solvent mixtures.
cyclopentane, ꢀCO
assigned to the protons of the bound methylene unit. The fact
that a single resonance is observed, coupled with a highly
shielded 13C resonance and a large equilibrium isotope shift
accompanying deuterium incorporation,[3c] suggests a fast
exchange between two equivalent complexes in which one of
ꢀ
the C H bonds in this methylene unit is complexed with the
low temperatures to be used with the otherwise high-melting
cyclohexane. Any bound [D12]pentane (formation of
[CpRe(CO)2([D12]pentane)]) is not observed in the
1H NMR spectra as it is fully deuterated. When isobutane is
also present as solvent, a resonance due to formation of the
isobutane complex [CpRe(CO)2(isobutane-h2-C1,H1)] (d =
ꢀ2.13 ppm (d, 3J(H,H) = 6 Hz)) is also seen.
metal center. A rapid, mutual exchange between two alter-
native h2-C,H structures renders the two protons equivalent,
and the net result is that each proton is directly complexed
with the metal center half of the time.
More recently, we have demonstrated that three isomeric
pentane complexes, [CpRe(CO)2(pentane-h2-Cx,Hx)], (x = 1–
3) are observable in which the pentane ligand may bind using
either of the methylene or the methyl moieties.[3c] Using
partially deuterated, bound methyl groups in [CpRe-
(CO)2(pentane-h2-C1,H1)], it was demonstrated that: 1) one
hydrogen (deuterium) atom is directly complexed with the
metal center at any given instant and there is a rapid exchange
of which hydrogen (deuterium) atom is engaged in this direct
interaction; 2) binding of hydrogen is preferred over that of
deuterium (KH/D = 2.02); 3) the observed shielding in the
1H NMR spectrum is directly related to the fraction of time a
hydrogen atom spends directly complexed with the metal
Repeating the experiment using uniformly 13C-labeled
cyclohexane (Figure 1d) confirms this molecule to be an
alkane complex, with large 1J(C,H) coupling constants of
(96.5 ꢁ 0.5) and (125.0 ꢁ 0.5) Hz for the resonances at d =
ꢀ6.17 and 0.49 ppm, respectively. The latter is close to the
1J(C,H) coupling constants found in free cyclohexane at this
1
temperature ( ꢂ 128 Hz). A H-13C HSQC experiment con-
firms that both protons are attached to the same carbon atom
with a shielded shift (d = ꢀ22.4 ppm) characteristic of a
1
bound alkane. We attribute the two H NMR resonances to
1
center. Similarly, the J(C,H) value decreases as the fraction
of time that a hydrogen atom spends directly complexed with
the metal center increases.
the axial and equatorial protons, respectively, of a bound
methylene unit. The axial hydrogen atom would be antici-
pated to show a large coupling (J > 8 Hz) to three other
hydrogen nuclei (geminal, 2 ” trans), whereas the equatorial
hydrogen atom is expected to have only one large coupling
(geminal). After resolution enhancement (Figure 1b), a
triplet structure is evident for the signal at d = ꢀ6.17 ppm,
indicating two large coupling constants of (9.8 ꢁ 0.5) Hz. No
large coupling constants were resolved for the signal at d =
We have now extended our studies to the binding of
cyclohexane, since, at the temperatures used, it exists in a
chair form and the exchange between the chair forms is
slow.[4] Hence, axial and equatorial protons with discrete
NMR shifts are present and it is possible to observe an alkane
complex that exhibits preferential complexation of one
hydrogen atom within a CH2 unit. The discrete shifts for
axial and equatorial hydrogen atoms will be retained in a
complex-bound cyclohexane even if, as expected, there is
rapid exchange between the two isomeric structures.
When photolysis experiments similar to those described
previously[2,3] [see Eq. (1)] are carried out at 173 K in the
presence of cyclohexane, two new signals of similar intensity
are observed in the 1H NMR spectrum at d = ꢀ6.17 and
2
0.49 ppm. This suggests that the geminal J(H,H) splitting is
unexpectedly small.
In order to confirm assignment of these resonances,
[1,1,2,2,4,4,5,5-D8]cyclohexane was prepared.[6] When this
[D8]cyclohexane was used as alkane ligand under the same
conditions, signals were observed at d = ꢀ6.09 and 0.33 ppm,
the former having a line width of about 7 Hz (Figure 1c) due
to
[CpRe(CO)2([2,2,3,3,5,5,6,6-D8]cyclohexane-h2-C1,H1)]
Angew. Chem. Int. Ed. 2006, 45, 4486 –4490
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4487

Communications
([D8]-1). As all vicinal H,H couplings have been removed
through deuteration, this confirms that the geminal H,H
coupling constant is small (< 6 Hz) and that the signal at d =
ꢀ6.09 ppm corresponds to a bound axial hydrogen atom. This
geminal coupling is weak, possibly because this formally two-
ꢀ ꢀ
bond (H C H) coupling now contains a three-bond compo-
ꢀ
nent (H···Re···C H) of opposite sign to the two-bond
component, and our calculations indicate an increase of the
H-C-H angle to 118.58 (MP2, most stable conformation), both
of which would reduce the coupling strength.[7] The calculated
2
value for this geminal J(H,H) coupling in the most stable
conformation is < 5 Hz (Table 1), and the value for the trans
diaxial 3J(H,H) coupling between the bound axial proton and
vicinal axial protons is calculated to be in the range 7.9–
10.6 Hz, consistent with our experimental observation.[8]
Table 1: NMR parameters of 1 obtained from calculations[a] and from
experiments.
[b]
e
Isomer
d(Ha) [ppm]/ d(He) [ppm]/ 2J(H,H) K1 /1
a
1J(C,H) [Hz]
1J(C,H) [Hz]
[Hz]
Scheme 2. Equilibria between axially and equatorially complexed spe-
cies for various isotopomers with midrange-fitted empirical equilibri-
um constants. All equilibria are in fast exchange on the NMR timescale
at the temperatures employed resulting in only weighted-averaged
shifts being observed. In each case one of the two (equivalent) bonds
antiperiplanar to the complexed bond is shown in bold. Chemical
shifts in brackets are estimated values based on empirical data for
each proton that would be observed if no exchange were occurring.
1a
1e
ꢀ9.74/101.3 +2.33/156.4 ꢀ4.73
+2.38/156.6 ꢀ9.78/103.4 ꢀ4.65
2.4
5.6
experimental values ꢀ6.17/96.5
+0.49/125
<6
2.4–3.4
[a] In each case, the values shown are calculated using the lowest-energy
conformation as predicted by the MP2 calculations. Calculations were
performed using B3LYP-GIAO on the HCTH-optimized geometries and
the default SCRF-PCM treatment in cyclohexane solvent.[8] [b] The
equilibrium constant calculated is the value required to give the
observed d at 173 K assuming that the calculated d values are correct.
equatorial protons in the two isomers are not known.
However, further measurements such as equilibrium isotope
effect (isotopic perturbation of resonance) experiments on
[CpRe(CO)2([1-D1]cyclohexane-h2-C1,H1)] ([D1]-1) aid the
prediction of the chemical shifts and coupling constants of
both the complexed and uncomplexed hydrogen atoms in the
two isomers of the complex. Combined with knowledge of the
pentane system, sensible estimates for the equilibrium con-
stants and NMR parameters can be extracted. In the case of
the bound methyl group of [CpRe(CO)2(pentane-h2-C1,H1)],
an analytical solution was possible: the directly complexed
hydrogen atom has a shielded shift of d = ꢀ8.22 ppm and the
uncomplexed hydrogen atoms within the same methyl group
a shift of d = 1.12 ppm, slightly deshielded (by ꢂ 0.20 ppm)
compared to a free methyl group. By making assumptions
about the chemical-shift change upon binding of the more
remote proton in a bound methylene group (i.e., the hydrogen
atoms shown in italics in Scheme 2) it is possible to extract
The observed chemical shift of both the axial and
equatorial protons is the result of an equilibrium between
two isomeric complexes, 1a and 1e which have the axial and
the equatorial hydrogen atom, respectively, directly com-
plexed with the metal center (Scheme 2). Interconversion of
1a and 1e is fast on the NMR timescale at the temperatures
employed. This is confirmed by variable-temperature NMR
experiments on 1. Decreasing the temperature results in a
significant increase in the separation of the axial and
equatorial shifts, indicative of an increasing equilibrium
constant at lower temperatures. The separation of the two
shifts increases from 6.65 ppm at 173 K to 6.91 ppm at 145 K.
The amount of shielding can be used to estimate the
ꢀ
proportion of time that each type of C H bond is complexed
with the rhenium center. Likewise, the reduction of the C H
ꢀ
values for the equilibrium constants K1 /1 and KH/D.
a e
coupling gives an independent measure of the same param-
eters. Given that the axial proton is highly shielded, by about
7.4 ppm compared to free cyclohexane, and 1J(C,H) is greatly
reduced (96.5 Hz compared to 128 Hz in free cyclohexane),
while the shielding (by ꢂ 1.2 ppm) and reduction in coupling
(1J(C,H) = 125 Hz) is much smaller for the equatorial proton,
it is clear that binding of the axial proton is significantly
preferred in this system.
We have chosen to make assumptions about the chemical
shift of this type of proton as it is likely to be less affected by
different binding situations than the strongly complexed
proton. Using a range of values for the change in chemical
shift for this uncomplexed proton that vary from 0 to
+ 2.5 ppm, a range that brackets the observed value for
pentane (deshielded by 0.20 ppm) and the larger deshielding
predicted by calculations (up to 1.24 ppm), a range of values
It is not possible to extract a precise equilibrium constant,
K1 /1 , for this case of axially vs. equatorially complexed as only
for K1 /1 = 3.4–2.4 and
KH/D = 4.7–1.9, respectively, is
a
e
obtained.[9] Using a value for this deshielding of 1 ppm
relative to free cyclohexane leads to the midrange values
shown in Scheme 2, K1 /1 = 2.9 and KH/D = 2.6. Note that in
a
e
the weighted-averaged chemical shifts and 1J(C,H) values are
observed and the values of these parameters for the axial and
a
e
4488 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4486 –4490

Angewandte
Chemie
ꢀ
general, there is no reason to assume that the KH/D value will
be the same for cyclohexane as for pentane, or indeed that the
same value applies to axial and equatorial sites.
axial and equatorial C H bonds would be the same, although
long-range steric interactions with the ring would favor the
equatorially complexed isomer. This is supported by the
crystal structure of a complex of uranium with a very bulky
Ab initio and DFT calculations[8] on 1 give structures in
which the alkane is complexed through one hydrogen atom at
any time. The PW91 and HCTH DFT calculations prefer the
more stable axially complexed isomer 1a by 1.3 and
0.4 kJmolꢀ1, respectively, and ab initio MP2 calculations
prefer it by 4.2 kJmolꢀ1 relative to the equatorially com-
plexed isomer 1e; in comparison a value of 1.5 kJmolꢀ1 is
estimated from the empirical data. In contrast, the B3LYP
hybrid density functional prefers 1e by 1.2 kJmolꢀ1. As seen in
calculated structures of smaller alkanes,[10] the Re-H-C angle
ꢀ
hexadentate ligand with the equatorial C H bond of meth-
ylcyclohexane being complexed,[11] though it is noted that this
is in the solid state rather than in solution as observed here
and stabilizing interactions between the cyclohexane ring and
the co-ligand may be significant. Below, we propose an
electronic argument to explain the observed preference for
binding.
In 1981, Cieplak proposed a radical argument to account
for the selectivity observed in nucleophilic additions to
cyclohexanones.[12a] The Cieplak model relies on the assump-
ꢀ
for the complexed C H moiety is large: the lowest-energy
ꢀ
ꢀ
conformations of 1a and 1e have values of 114.48 and 112.68,
respectively (MP2), although these values can be as large as
1288 in the other conformations.
tion that a C H bond is a better electron donor than a C C
bond when both are available for hyperconjugation, for
example, when they have an antiperiplanar relationship to an
electron-depleted bond. This assumption sparked some con-
troversy and has stimulated many groups to examine its
validity.[12–17] The earliest experimental evidence supporting
Cieplakꢀs hypothesis used the 19F NMR chemical shifts of
bicycloalkyl fluorides as probes to the extent of delocalization
of electrons from C X bonds into the s* orbital of the C F
bond.[13a,b] This indicated that alkyl groups are electron-
withdrawing when compared to hydrogen. Subsequent exper-
imental studies, including those examining the hydrochlori-
nation and fluorination of 5-substituted adamant-2-ols,[13e] the
reduction of 5-substituted adamantan-2-ones,[14a,b] and the
conformations of a,b-unsaturated esters and aldehydes,[15a]
further support this model.
The barrier to exchange between the two isomers is
calculated to be 10.0 and 14.6 kJmolꢀ1 at the HCTH and MP2
levels, respectively. This energy barrier is sufficiently low for
the two isomers to remain in fast exchange on the NMR
timescale even at the lowest temperatures employed in this
study, consistent with our observations. The calculated
barriers for 1,2-(axial-equatorial) and 1,3-(axial-axial) hydro-
gen shifts are 21 and 29 kJmolꢀ1, respectively, at HCTH and
52 and 13 kJmolꢀ1, respectively, at MP2. Experimental data
indicate that these 1,2 and 1,3 shifts are not occurring rapidly
on the NMR timescale and that these barriers are under-
estimated. The calculated binding energies at the different
levels vary widely but in the expected fashion. PW91 and MP2
apparently overestimate the binding, whereas B3LYP and
HCTH agree moderately well. Given the disagreement
between B3LYP and the other methods as to the relative
stabilities of the isomers, the HCTH binding energies are
probably most reliable.
ꢀ
ꢀ
ꢀ
While these experimental results indicate that C H bonds
are better hyperconjugative donors than C C bonds, the
ꢀ
magnitude of this difference has not been derived exper-
imentally. Computational studies of 2-substituted 1-fluoro-
ethanes[18a] and 2,5-dimethyl-2-adamantyl cations[18b] report
the difference as about 4 and 1.6 kJmolꢀ1, respectively.
The preference for axial coordination in 1 is an electronic,
In agreement with experiment, calculations predict very
different NMR parameters (d and J) for the complexed and
uncomplexed hydrogen atoms within the bound methylene
group (Table 1).
ꢀ
not a vibrational effect. The slightly weaker axial C H bonds
are better donors than their equatorial counterparts. There is
a stereoelectronic preference for conformations in which the
best donor lone pair or bond is antiperiplanar to the best
Our empirical data suggest that the calculated shifts of the
complexed hydrogen atoms are slightly more extreme than
the experimental ones, but the basic principle is well
acceptor bond.[19] A C H bond complexed with a rhenium
ꢀ
1
ꢀ
demonstrated. The J(C,H) values are clearly overestimated
center is more electron-deficient than a regular C H bond (as
as they cannot give a weighted average of 96.5 Hz as was
there is insignificant electron back donation from the metal
ꢀ
observed in the case of the axial C H coupling. However,
they do predict the reduction of J(C,H) upon complexation
center[20]) and thus is a better electron acceptor. Since axial
1
ꢀ
C H bonds in chair conformations of cyclohexanes are
antiperiplanar to some of the other C H bonds, while the
equatorial C H bonds are antiperiplanar to certain C C
bonds (Scheme 2), the selectivity described in this paper
suggests that a C H bond is a better hyperconjugative donor
ꢀ
and an increase of it for the uncomplexed hydrogen atom
compared to the free alkane. This behavior was observed
empirically in the case of [CpRe(CO)2(pentane-h2-C1,H1)]
(1J(C,H) of the complexed H: 85 Hz; of the uncomplexed:
132 Hz).
ꢀ
ꢀ
ꢀ
ꢀ
than a C C bond. Taking into account that two such effects
The expected mode of complexation of cyclohexane to the
rhenium center, confirmed by our calculations, is shown in
are operating in each case, this corresponds to an energy
difference of about 0.75 kJmolꢀ1, which is in reasonable
agreement with the computed differences, given the small
magnitude of the effect.[8]
It is worth noting that preferential reaction of equatorial
C H bonds has been found in the monooxygenation of
Scheme 2, with the metal center positioned between the axial
[1]
ꢀ
and equatorial C H bonds. Steric interactions with the
cyclohexane ligand would be expected to disfavor any other
options. From this, it is not apparent that a steric argument
could account for the selective formation of the complex of an
ꢀ
[21]
ꢀ
methylene C H bonds in substituted cyclohexanes. This
was explained qualitatively using long-range hyperconjuga-
ꢀ
axial C H bond of cyclohexane. At best, complexation of the
Angew. Chem. Int. Ed. 2006, 45, 4486 –4490
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4489

Communications
[5] See the Supporting Information for details of experimental
NMR procedures.
[6] See the Supporting Information for the preparation of labeled
alkanes.
tive stabilization, though this difference could not be quanti-
fied.
The equilibrium constant K1 /1 is reduced somewhat by
a
e
the introduction of deuterium into vicinal sites (using
[D8]cyclohexane). Assuming that the “normal” deuterium
isotope shifts are similar for axial and equatorial protons,
which is the case in free cyclohexane, it is apparent
(Figure 1c) that the shifts of the two protons are closer
together in this isotopomer and that the equilibrium constant
is therefore reduced (from ꢂ 2.9 to ꢂ 2.7). Ascribing this
[7] J. Geertsen, J. Oddershede, W. T. Raynes, Magn. Reson. Chem.
1993, 31, 722.
[8] See the Supporting Information for details of the quantum
mechanical calculations.
[9] Details of the derivation of these values are given in the
Supporting Information.
[10] For calculated structures of [CpRe(CO)2(ethane)] and
[CpRe(CO)2(methane)], respectively, see: a) H. Yang, M. C.
Asplund, K. T. Kotz, M. J. Wilkens, H. Frei, C. B. Harris, J. Am.
Chem. Soc. 1998, 120, 10154; b) R. G. Bergman, T. R. Cundari,
A. M. Gillespie, T. B. Gunnoe, W. D. Harman, T. R. Klinckman,
M. D. Temple, D. P. White, Organometallics 2003, 22, 2331.
[11] I. Castro-Rodriguez, H. Nakai, P. Gantzel, L. N. Zakharov, A. L.
Rheingold, K. Meyer, J. Am. Chem. Soc. 2003, 125, 15734.
[12] a) A. S. Cieplak, J. Am. Chem. Soc. 1981, 103, 4540; b) C. T.
Johnson, B. D. Tait, A. S. Cieplak, J. Am. Chem. Soc. 1987, 109,
5875; c) A. S. Cieplak, J. Org. Chem. 1998, 63, 521; d) A. S.
Cieplak, Chem. Rev. 1999, 99, 1265.
[13] a) W. Adcock, A. N. Abeyrickrema, J. Org. Chem. 1982, 47,
2957; b) W. Adcock, A. N. Abeyrickrema, G. B. Kok, J. Org.
Chem. 1984, 49, 1387; c) W. Adcock, V. S. Iyer, J. Org. Chem.
1988, 53, 5259; d) W. Adcock, N. A. Trout, J. Org. Chem. 1991,
56, 3229; e) W. Adcock, J. Cotton, N. A. Trout, J. Org. Chem.
1994, 59, 1867; f) W. Adcock, N. J. Head, N. R. Lokan, N. A.
Trout, J. Org. Chem. 1997, 62, 6177; g) W. Adcock, N. A. Trout,
Magn. Reson. Chem. 1998, 36, 181; h) W. Adcock, N. A. Trout,
Chem. Rev. 1999, 99, 1415.
[14] a) C. K. Cheung, L. T. Tseng, M. H. Lin, S. Srivastava, W. J.
le Noble, J. Am. Chem. Soc. 1986, 108, 1598; b) M. Xie, W. J.
le Noble, J. Org. Chem. 1989, 54, 3836; c) V. R. Bodepudi, W. J.
le Noble, J. Org. Chem. 1991, 56, 2001; d) W.-S. Chung, N. J.
Turro, S. Srivastava, W. J. le Noble, J. Org. Chem. 1991, 56, 5020;
e) H. Li, J. E. Silver, W. H. Watson, R. P. Kashyap, W. J.
le Noble, J. Org. Chem. 1991, 56, 5932; f) H. Li, W. J. le Noble,
Recl. Trav. Chim. Pays-Bas 1992, 111, 199; g) V. R. Bodepudi,
W. J. le Noble, J. Org. Chem. 1994, 59, 3265; h) M. Kaselj, W.-S.
Chung, W. J. le Noble, Chem. Rev. 1999, 99, 1287.
[15] a) B. W. Gung, M. M. Yanik, J. Org. Chem. 1996, 61, 947;
b) B. W. Gung, Tetrahedron 1996, 52, 5263; c) B. W. Gung, Chem.
Rev. 1999, 99, 1377.
[16] a) M. D. Rozeboom, K. N. Houk, J. Am. Chem. Soc. 1982, 104,
1189; b) Y.-D. Wu, K. N. Houk, J. Am. Chem. Soc. 1987, 109, 908;
c) Y.-D. Wu, J. A. Tucker, K. N. Houk, J. Am. Chem. Soc. 1991,
113, 5018.
[17] a) T. Laube, H. U. Stilz, J. Am. Chem. Soc. 1987, 109, 5876; b) T.
Laube, T.-K. Ha, J. Am. Chem. Soc. 1988, 110, 5511; c) T. Laube,
S. Hollenstein, J. Am. Chem. Soc. 1992, 114, 8812.
[18] a) P. R. Rablen, R. W. Hoffmann, D. A. Hrovat, W. T. Boron, J.
Chem. Soc. Perkin Trans. 2 1999, 1719; b) A. Filippi, N. A. Trout,
P. Brunelle, W. Adcock, T. Sorenson, M. Speranza, J. Org. Chem.
2004, 69, 5537.
ꢀ
effect entirely to the replacement of two trans-diaxial C H
ꢀ
bonds by two C D bonds leads to a reduction in energy of
about 46 Jmolꢀ1 per bond. This is consistent with the fact that
ꢀ
the stronger C D bond is less likely to donate its electrons.
This observation adds further weight to the hypothesis that
the observed phenomenon is indeed caused by an electronic
effect since the steric effect of replacing hydrogen with
deuterium in vicinal sites is expected to be negligible.
Qualitative MO arguments also suggest that the axial
hydrogen atoms are better electron donors than the equato-
rial ones. The degenerate HOMOs of cyclohexane are
ꢀ
ꢀ
delocalized over the C C and equatorial C H bonds. How-
ever, the contribution on each equatorial hydrogen atom is
only 5.6% (sum of the squared coefficients of the two H basis
functions in the Eigenvector). Three occupied combinations
of essentially pure axial sCꢀH orbitals are only roughly 0.5 eV
lower in energy than the HOMO and provide more effective
electron donation than the HOMOs.[22] These orbitals reflect
exactly the hyperconjugation arguments given above.
In conclusion, we have identified an alkane complex,
[CpRe(CO)2(cyclohexane)], in which binding occurs prefer-
ꢀ
entially through the axial C H bond of a methylene group.
ꢀ
The ratio of the complexation of each type of C H bond in
cyclohexane (K1 /1 ꢂ 2.9) suggests the difference in the
a
e
ꢀ
ꢀ
ꢀ
electron-donating abilities of C H, C C, and even C D
bonds may be determined, assuming that there is no steric
preference for either mode of complexation.
Received: January 24, 2006
Published online: June 13, 2006
Keywords: ab initio calculations · alkane ligands ·
.
ꢀ
C H activation · hyperconjugation · NMR spectroscopy
´
[1] a) S. Zaric, M. B. Hall, J. Phys. Chem. A 1997, 101, 4646, and
references therein; b) C. Hall, R. N. Perutz, Chem. Rev. 1996, 96,
3125.
[19] P. Deslongchamps, Stereoelectronic Effects in Organic Chemistry,
Pergamon, Oxford, 1983.
[20] G. J. Kubas, J. Organomet. Chem. 2001, 635, 37.
[21] M. E. Gonzµlez-Nfflꢁez, G. Castellano, C. Andrew, J. Royo, M.
Bµguena, R. Mello, G. Asensio, J. Am. Chem. Soc. 2001, 123,
7487.
[2] Reviews: a) J. A. Labinger, J. E. Bercaw, Nature 2002, 417, 507;
b) R. H. Crabtree, J. Chem. Soc. Dalton Trans. 2001, 17, 2437;
c) W. D. Jones, Acc. Chem. Res. 2003, 36, 140; d) M. Lersch, M.
Tilset, Chem. Rev. 2005, 105, 2471; e) S. R. Klei, J. T. Golden, P.
Burger, R. G. Bergman, J. Mol. Catal. A 2002, 189, 79.
[3] a) S. Geftakis, G. E. Ball, J. Am. Chem. Soc. 1998, 120, 9953; b) S.
Geftakis, G. E. Ball, J. Am. Chem. Soc. 1999, 121, 6336; c) D. J.
Lawes, S. Geftakis, G. E. Ball, J. Am. Chem. Soc. 2005, 127, 4134.
[4] The barrier to ring inversion of cyclohexane is 43 kJmolꢀ1 (F. R.
Jensen, D. S. Noyce, C. H. Sederholm, A. J. Berlin, J. Am. Chem.
Soc. 1960, 82, 1256); so the rate of interconversion of the chair
forms of cyclohexane is ca. 1 sꢀ1 at 180 K.
[22] See: R. Koch, T. Clark, The Chemistꢀs Electronic Book of
Orbitals, Springer, Berlin, 1999.
4490 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 4486 –4490