Synthesis, crystal structure, and reactions of the 17-valence-electron rhenium methyl complex [(η5-C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3)]*+ B(3,4-C6H3(CF3)2)4-: experimental and computational bonding comparisons with 18-electro

Article abstract of DOI:10.1016/S0022-328X(00)00531-3

Reactions of methyl complex (η⁵-C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3) (2b) and the ferrocenium salt (η⁵-C5H5)2Fe^*+ BAr_F^- (BAr_F^-=B(3,5-C6H3(CF3)2)4^-) or the trityl salt Ph3C

Full text of DOI:10.1016/S0022-328X(00)00531-3

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 616 (2000) 54–66
Synthesis, crystal structure, and reactions of the
17-valence-electron rhenium methyl complex
5
+
−
’
[(h -C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3)] B(3,5-C6H3(CF3)2)4 :
experimental and computational bonding comparisons with
18-electron methyl and methylidene complexes
Jean Le Bras a, Haijun Jiao a, Wayne E. Meyer b, Frank Hampel a, J.A. Gladysz a,*
a Institut fu¨r Organische Chemie, Friedrich-Alexander Uni6ersita¨t Erlangen-Nu¨rnberg, Henkestrasse 42, 91054 Erlangen, Germany
b Department of Chemistry, Uni6ersity of Utah, Salt Lake City, UT 84112, USA
Received 23 June 2000
Abstract
5
5
+
Reactions of methyl complex (h -C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3) (2b) and the ferrocenium salt (h -C5H5)2Fe BArF−
’
(BAr−F =B(3,5-C6H3(CF3)2)−4 ) or the trityl salt Ph3C+ BArF− give the very air sensitive title radical cation 2b BAr−F or the
+
’
robust methylidene complex [(h5-C5Me5)Re(NO)(P(4-C6H4CH3)3)(ꢀCH2)]+ BAr−F (3b+ BAr−F ) as analytically pure powders in
80% yields. The crystal structures of 2b and 2b BAr−F are determined. With the aid of high level density functional calculations
+
’
on model complexes, key structural, bonding, and dynamic properties are compared. Similar quantities are calculated for 3b+
BAr−F , which could not be crystallized, and the ReꢀCH2 rotational barrier is bounded by NMR (DG3‡83 K\17.5 kcal mol−1).
Special attention is given to structural manifestations of backbonding, particularly with the phosphine ligands. Cobaltocene and
+
’
+
2b BAr−F react to give 2b. However, no phosphine exchange or well-defined thermal decomposition products of 2b BAr−F are
’
detected. © 2000 Elsevier Science B.V. All rights reserved.
Keywords: Rhenium; Radical cation; ‘Barf’ anion; Methylidene complex; Density functional calculation
neutral, and anionic radicals have been observed and in
a few cases isolated [4,5].
1. Introduction
As described in the preceding paper [6], we have had
a special interest in complexes where sp carbon chains
bridge two transition metals, and their properties as a
function of redox state [7]. The dirhenium C4 radical
cation[(h5-C5Me5)Re(NO)(P(C6H5)3)(CCCC)((C6H5)3P)-
Organometallic complexes with 17 valence electrons
have numerous important physical and chemical prop-
erties that are not shared by their 18 and 16-valence-
electron counterparts [1]. However, for a variety of
reasons — in some cases their low thermal stability or
high chemical reactivity, and in other cases the re-
stricted utility of NMR for characterization — they
remain far less studied. One specific major class would
be cyclopentadienyl ‘piano stool’ complexes of the for-
mula (h5-C5R5)M(L)(L%)(L%%) [1–3], for which cationic,
5
+
’
+
’
(ON)Re(h -C5Me5)] PF−6 (1 PF6−) is isolable, and
the odd electron (and positive charge) is delocalized
between the two rheniums on the rapid IR and ESR
time scales [7b]. Thus, we have sought to rigorously
characterize similar monorhenium radical cations, in
which the metal would carry higher spin density and a
full formal positive charge. Accordingly, oxidations of
methyl complexes (h5-C5Me5)Re(NO)(PR3)(CH3) (2) to
spectroscopically observable radical cations [(h5-
* Corresponding author. Tel.: +49-9131-8522540; fax: +49-9131-
8526865.
E-mail address: gladysz@organik.uni-erlangen.de (J.A. Gladysz).
C5Me5)Re(NO)(PR3)(CH3)] + X− (2 + X−; X−=PF6−,
’
’
SbF−6 ) that persist for hours in solution at room tem-
0022-328X/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00531-3

J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
55
Scheme 1. Synthesis of new complexes.
perature are detailed in the preceding paper [6]. Phos-
phine ligands that are more electron rich and/or bulkier
than triphenylphosphine were found to be advan-
tageous.
methyl complexes 2 [6]. Accordingly, the previously
reported [9b,11] salt (h -C5H5)2Fe
pared in 80% yield by a standard procedure.
5
+
’
BAr−F was pre-
As shown in Scheme 1, the tri(p-tolyl)phosphine
5
+
methyl complex 2b [6] and (h -C5H5)2Fe BAr−F were
combined in CH2Cl2 at room temperature. Workup
gave a tan powder with properties consistent for the
’
In the course of these efforts, we discovered that
further optimization of the anion allowed one such
radical cation to be isolated in crystalline form. In this
paper, we report the synthesis and structural character-
ization of the ‘barf’ salt of the tri(p-tolyl)phosphine
+
’
radical cation 2b
BAr−F in 80% yield. When the
educts were not of the highest purities, much lower
5
+
’
+
yields were obtained. Complex 2b BAr−F was very air
’
complex [(h -C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3)]
+
’
BAr−F (2b
BAr−F ) [8]. The rationale for a separate
sensitive both in solution and the solid state, and all
measurements utilized freshly prepared samples. Micro-
analyses (C–H–N) were only marginally satisfactory.
However, the IR spectrum showed a strong wNO band
publication is to facilitate comparisons of this scarce
class of compound [4] with 18-valence-electron con-
geners. Towards this end, we also describe: (1) the
crystal structure of the precursor methyl complex 2b;
(2) the isolation and spectroscopic characterization of
the corresponding methylidene complex [(h5-
C5Me5)Re(NO)(P(4-C6H4CH3)3)(ꢀCH2)]+ BArF− (3b+
BAr−F ); and (3) supporting high-level density functional
calculations on models for all of these compounds. The
latter allow a more precise analysis of structural, bond-
ing, and dynamic properties.
at 1714 cm−1, in agreement with the much less stable
+
2b PF−6 [6]. The ESR spectrum was similar to those
’
+
’
of other 2
X− salts [6], as depicted in Fig. 1. As
expected, no NMR signals were detected. A cyclic
voltammogram showed a reversible one electron reduc-
tion, and was equivalent to that of the precursor 2b.
+
Prisms of 2b BAr−F were readily and reproducibly
’
obtained from CH2Cl2–hexane. The crystal structure
was determined as outlined in Table 1 and in Section 4.
Fig. 2 (bottom) confirms the proposed formulation, and
2. Results
2.1. The title radical cation
As exemplified by recent successes involving
organorhenium cations [9], BAr−F is increasingly being
recognized as a ‘stability-conferring’ anion. Possible
favorable factors include its large size, and the presence
of aryl rings that can promote crystallinity. However, it
is not totally innocent [10]. In the previous paper, we
found that the ferrocenium cation readily oxidized
+
Fig. 1. ESR spectrum of 2b BArF−
.
’

56
J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
BAr−F and triphenylphosphine, P(C6H5)3 (five equiva-
Table 1
+
Crystallographic data for 2b and 2b BArF− a
’
lents), were combined in CH2Cl2 at 22°C. After 1 h,
cobaltocene was added. Workup gave only 2b. None of
the previously characterized substitution product (h5-
2b BAr−F
+
’
Complex
2b
Empirical formula
Formula weight
Temperature (K)
C32H39NOPRe
670.81
C64H51BF24NOPRe
1534.04
C5Me5)Re(NO)(P(C6H5)3)(CH3) (2a) [6,13] was de-
+
tected. Third, isolated samples of 2b and 2b BArF−
’
173(2)
173(2)
,
Wavelength (A)
0.71073
Monoclinic
P21/c
0.71073
Monoclinic
P21/n
Crystal system
Space group
Cell dimensions
,
a (A)
8.5697(10)
19.787(2)
16.864(3)
93.346(11)
2854.7(6)
4
12.526(3)
14.900(3)
34.288(7)
97.92(3)
6338(2)
4
,
b (A)
,
c (A)
i (°)
V (A )
3
,
Z
Dcalc (g cm−3
Absorption
coefficient
)
1.561
4.337
1.608
2.056
(mm−1
F(000)
)
1344
3044
Crystal size
0.30×0.30×0.20
0.40×0.40×0.40
(mm3)
q limit (°)
Index ranges
2.38–26.28
105h510,
−245k50,
−215l50
5987
2.60–25.01
05h514, 05k517,
−405l540
Reflections
collected
11673
Independent
reflections
Reflections
[I\2|(I)]
5788 [Rint=0.0709]
3918
11124 [Rint=0.0622]
6010
Data/restraints/
parameters
Goodness-of-fit
on F2
5788/0/325
1.031
11124/0/838
1.017
Final R indices
[I\2|(I)]
R indices (all
data)
R1=0.0494,
wR2=0.0949
R1=0.0912,
wR2=0.1115
2.272 and −0.855
R1=0.0700,
wR2=0.1362
R1=0.1508,
wR2=0.1730
1.135 and −0.700
Dp (max),
−3
,
(e A
)
2
2
a R=S(ꢀꢀFoꢀ−ꢀFcꢀꢀ)/S(ꢀFoꢀ); Rw=[S(w(ꢀFoꢀ−ꢀFcꢀ /S(wꢀFoꢀ )]1/2
.
Table 2 lists key bond lengths and angles. Atomic
coordinates and other data have been deposited in the
Cambridge Crystallographic database. In accord with
much precedent, the rhenium is formally octahedral,
with CꢁReꢁP, CꢁReꢁN, and PꢁReꢁN bond angles close
to 90°. One consequence (important below) is that the
methyl ligand cannot adopt a perfectly staggered or
eclipsed conformation with respect to rhenium and the
three other ligands. This requires two tetrahedral
atoms, as in ethane. Additional structural features are
analyzed in conjunction with the computational data.
+
’
Fundamental chemical properties of 2b BAr−F were
explored. First, as shown in Scheme 1, reduction with
5
’
cobaltocene, (h -C5H5)2Co , gave the precursor methyl
+
BAr−F
’
Fig. 2. Molecular structure of complexes 2b (top) and 2b
(bottom).
+
’
complex 2b in 90% yield after workup. Second, 2b

J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
57
Table 2
2.2. Related complexes
+
Key distances (A) and angles (°) in 2b and 2b BArF−
’
,
As previously reported [13], reaction of the
triphenylphosphine methyl complex 2a and trityl hexa-
fluorophosphate, Ph3C+ PF−6 , gives the stable methyli-
dene complex [(h5-C5Me5)Re(NO)(P(C6H5)3)(ꢀCH2)]+
PF−6 (3a+ PF−6 ) [14]. However, the crystal structure
shows severe nitrosyl–methylidene ligand positional
disorder [13]. We sought to synthesize the correspond-
ing tri(p-tolyl)phosphine methylidene complex 3b+ X−
by a similar method, and obtain better structural data if
2b BAr−F
+
’
2b
Bond lengths
Re(1)ꢁC(1)
Re(1)ꢁP(1)
Re(1)ꢁN(1)
N(1)ꢁO(1)
P(1)ꢁC(21)
P(1)ꢁC(31)
P(1)ꢁC(41)
Re(1)ꢁCp* (centroid)
2.123(11)
2.359(2)
1.787(9)
1.175(11)
1.825(8)
1.829(9)
1.839(9)
1.970(9)
2.120(12)
2.455(3)
1.813(12)
1.027(13)
1.827(11)
1.811(11)
1.805(12)
1.985(11)
possible. One of several objectives was to be certain
+
that the crystals represented above as 2b BAr−F were
’
Bond angles
not in fact 3b+ BArF−. These complexes differ by only a
hydrogen atom, and would be very difficult to distin-
guish crystallographically.
C(1)ꢁRe(1)ꢁP(1)
C(1)ꢁRe(1)ꢁN(1)
P(1)ꢁRe(1)ꢁN(1)
O(1)ꢁN(1)ꢁRe(1)
C(21)ꢁP(1)ꢁRe(1)
C(31)ꢁP(1)ꢁRe(1)
C(41)ꢁP(1)ꢁRe(1)
88.1(3)
97.9(5)
92.7(3)
169.4(9)
119.3(3)
116.3(3)
113.2(3)
89.8(3)
99.7(5)
91.8(4)
169.4(14)
113.8(4)
115.3(4)
109.6(4)
As shown in Scheme 1, 2b and the known trityl ‘barf’
salt, Ph3C+ BAr−F [15], were reacted. Workup gave the
methylidene complex 3b+ BAr−F as an analytically pure
1
yellow powder in 80% yield. The H-NMR spectrum
showed two diagnostic low field ꢀCH2 signals (l 15.17,
14.22, CD2Cl2). As expected, the 13C-NMR spectrum
also exhibited a low field ꢀCH2 signal (l 284.7). The IR
spectrum gave an intense wNO band at 1695 cm−1. A
sample was dissolved in C6D5Br, and variable tempera-
were combined in a 95:5 ratio in CH2Cl2. After 10 min,
a more basic triarylphosphine, P(4-C6H4OMe)3 (five
equivalents), was added. The sample was worked up
after 1 h. A 1H-NMR spectrum (C6D6) showed only 2b,
and none of the known complex (h5-C5Me5)Re-
(NO)(P(4-C6H4OMe)3)(CH3) (2e) [6]. Other related at-
tempts to effect phosphine substitution were also un-
1
ture H-NMR spectra were recorded (400 MHz). The
ꢀCH2 protons still gave distinct signals at 110°C (l
15.05, 14.10; 15.06, 14.07 at ambient probe tempera-
ture). Application of the coalescence formula bounds
the DG‡ value for methylidene ligand rotation as \
17.5 kcal mol−1 (110°C) [16].
successful.
Finally, 2b
+
’
BAr−F was dissolved in CD3CN at
Extensive efforts to obtain crystals of 3b+ BArF−
suitable for X-ray analysis were unsuccessful. This in-
cluded many layered solvent combinations (CH2Cl2–
hexane, CH2Cl2–pentane, CHCl2CH3–hexane, CHCl3–
hexane, CCl4–hexane), and vapor diffusion (CHCl3–
pentane) or slow evaporation (CH2Cl2–hexane) condi-
tions. The hexafluoroantimonate salt 3b+ SbF6− was
also prepared, and gave similar results. Nonetheless,
−78°C and slowly warmed to room temperature. Un-
der analogous conditions, the markedly less stable cy-
clopentadienyl radical cation [(h5-C5H5)Re(NO)(P-
+
+
’
’
(C6H5)3)(CH3)]
X− (4a
X−) undergoes clean bi-
molecular decomposition to a 1:1:1 mixture of a meth-
ylidene complex, an acetonitrile complex, and methane,
as shown in Scheme 2 [12]. A 1H-NMR spectrum
showed many new C5Me5 signals, but no low field
resonances. An IR spectrum showed a wNO band at
1633 cm−1, suggestive of neutral 18-valence-electron
+
’
these failures clearly establish that the crystals of 2b
BAr−F described above (which could be reproducibly
grown from CH2Cl2–hexane) are not 3b+ BArF−. An
approximate value for one feature of interest, the
ReꢀCH2 bond length, can be taken from the crystal
structure of the triphenylphosphite complex [(h5-
products. No gas evolution was observed. Hence, the
+
’
bulkier pentamethylcyclopentadienyl ligand in 2b
BAr−F appears to block analogous chemistry.
+
Scheme 2. Thermal disproportionation of the cyclopentadienyl-substituted radical cation 4a X−
.
’

58
J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
corresponding vinylidene complex [17]. However, the
+
dirhenium C4 radical cation 1 PF−6 gives a number of
’
visible and near-IR transitions [7b]. Some of these are
diagnostic of its mixed valence redox state. Regardless,
+
none have any counterpart in 2b BAr−F (or 2b).
’
2.3. Computations
In order to further clarify the structural, electronic,
and dynamic properties of the preceding complexes,
computational studies were conducted. These utilized
two series of model compounds. In the first,
+
’
abbreviated {Re(PH3)(CH3)}, {Re(PH3)(CH3)}
and
{Re(PH3)(ꢀCH2)}+, the pentamethylcyclopentadienyl
and tri(p-tolyl)phosphine ligands were replaced by cy-
clopentadienyl and PH3. In the second, cyclopentadi-
enyl and PMe3 ligands were used. Geometries were
optimized at the level of density functional theory
(B3LYP) with the LANL2DZ basis set and an addi-
tional set of polarization functions (LANL2DZp) as
introduced by Hay and Wadt [18] and implemented in
the GAUSSIAN-98 program [19]. For radical cations, the
unrestricted method (UB3LYP) was used. Frequency
calculations were carried out at the same level to fur-
ther characterize the optimized geometries [20].
The computed energies of the model compounds are
listed in Table 3. Natural charges and bond orders,
calculated with the NBO program [21], are given in
Table 4. At the (U)B3LYP/LANL2DZp level, all com-
plexes are energy minima with only real vibrational
modes. Selected bond lengths, and bond and torsion
Fig. 3. Superposition of molecular structure of complexes 2b and
+
’
2b
.
angles, are summarized in Figs. 5–7. The data for
+
’
{Re(PR3)(CH3)} (Fig. 5) and {Re(PR3)(CH3)} (Fig.
6) agree well with the crystal structures, validating the
quality of the computational methodology. Hence, the
metrical parameters of {Re(PR3)(ꢀCH2)}+ (Fig. 7) can
be expected to closely match those of 3b+ BArF−. The
methyl and methylidene ligand conformations agree
with previous studies at the extended Hu¨ckel level
[22,23]. Computed wNO values are also given in Table 4,
and are very close to those observed. Frontier molecu-
lar orbitals are depicted in Fig. 8.
Fig. 4. UV–vis spectra of new complexes.
C5Me5)Re(NO)(P(OC6H5)3)(ꢀCH2)]+ PF6− (5+ PF6−;
,
1.89(2) A) [13].
Rotations about the rheniumꢁCHx and rheniumꢁPH3
bonds were also investigated computationally. Transi-
tion state energies and geometries are summarized in
In contrast to 3b+ BArF−, crystals of the 18-valence-
electron methyl complex 2b could (with some effort) be
obtained. A crystal structure was determined as out-
lined in Table 1 and in Section 4. Fig. 2 (top) gives the
Table
3 and Fig. 9. In the ground state of
{Re(PH3)(CH3)} (Fig. 5), the methyl carbonꢁhydrogen
bonds are roughly staggered with respect to the other
rhenium ligands. However, as observed above, stag-
gered and eclipsed are imprecise conformational de-
scriptors for bonds between formally octahedral and
tetrahedral atoms. Thus, we simply note that one hy-
drogen is anti to the cyclopentadienyl ligand, as
reflected by a HaꢁCꢁReꢁCp torsion angle of −179.8°.
In the transition state for methyl rotation (Fig. 9), a
hydrogen moves into a syn position, as reflected by a
molecular structure, and Table 2 lists key bond lengths
+
and angles side-by-side with those of 2b BAr−F . Fig. 3
’
gives a superposition of the rhenium moieties of 2b and
+
2b BAr−F . The small differences are analyzed below.
’
+
Finally, the UV–vis spectra of 2b, 2b BAr−F , and
’
3b+ BAr−F are illustrated in Fig. 4. Although each
compound is colored, the spectra show only featureless
tails into the visible region. This could have been
anticipated for 3b+ BAr−F based upon data for the

J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
59
Table 3
Energetic data ((U)B3LYP/LANL2DZp) for model complexes [(h5-C5H5)Re(NO)(PR3)(CHx)]n+
Etot (au)
Imaginary frequencies
Ethermal 298.15 K
Entropy 298.15 K
(cm−1
)
(kcal mol−1
)
(cal mol−1
)
Ground states
{Re(PH3)(CH3)}
{Re(PMe3)(CH3)}
{Re(PH3)(CH3)}
−450.86336
−568.84429
−450.62294
−568.62316
−450.02338
−568.02278
None
None
None
105.5
106.0
99.0
111.2
116.5
107.6
+
’
+
’
{Re(PMe3)(CH3)}
{Re(PH3)(ꢀCH2)}+
{Re(PMe3)(ꢀCH2)}+
Transition states
{Re(PH3)(CH3)} (CH3 rotation)
{Re(PH3)(CH3)} (PH3 rotation)
{Re(PH3)(ꢀCH2)}+ (ꢀCH2 rotation)
−450.86122
−450.86142
−499.97685
154i
106i
854i
105.0
104.9
97.6
108.2
108.6
108.9
for phosphine rotation in {Re(PH3)(CH3)} features a
hydrogen anti to the cyclopentadienyl ligand (Fig. 9;
H2ꢁPꢁReꢁCp torsion angle −187.5°), and only slight
changes in bond distances. Importantly, the transition
states for methyl and phosphine rotation exhibit only a
single imaginary frequency (Table 3). Hence, they can
confidently be assigned as real transition states.
Ha%ꢁCꢁReꢁCp torsion angle of −1.3°. Apart from
these rheniumꢁCH3 torsional differences, the ground
and transition state structures are remarkably similar.
The rheniumꢁcarbon bond does lengthen from 2.186 to
,
2.193 A. However, the rheniumꢁphosphorus and rhe-
niumꢁCp bonds, which are not undergoing rotation,
similarly elongate.
The phosphine ligand in {Re(PH3)(CH3)} adopts an
opposite ground state conformation (Fig. 5). Now, one
hydrogen is nearly syn to the cyclopentadienyl ligand,
as indicated by a H3ꢁPꢁReꢁCp torsion angle of −5.7°.
,
The longer rheniumꢁphosphorus bond (2.353 A versus
,
ReꢁCH3 2.186 A) may be a contributing factor. How-
ever, when the hydrogens are replaced by larger methyl
groups in {Re(PMe3)(CH3)} (Fig. 5), the conformation
changes. One methyl group becomes nearly anti to the
cyclopentadienyl ligand, as reflected by a C1ꢁPꢁReꢁCp
torsion angle of 187.1°. Regardless, the transition state
Table 4
Computed natural charges, bond orders, and wNO vibrational modes
(cm−1) for the [(h5-C5H5)Re(NO)(PH3)(CHx)]n+ ground states
+
{Re(PH3)(ꢀCH2}+
’
{Re(PH3)(CH3)} {Re(PH3)(CH3)}
Re
CHx
PH3
0.222
−0.276
0.404
−0.166
−0.182
0.574
−0.155
0.482
0.003
0.096
0.336
0.036
0.539
0.021
0.068
NO
h5-C5H5
ReꢁC
ReꢁN
ReꢁP
PꢁH1
PꢁH2
PꢁH3
0.709
1.793
0.609
0.942
0.958
0.979
0.903
1.489
0.576
0.921
0.918
0.942
1.428
1.491
0.651
0.924
0.929
0.933
a
wNO
1613
(1603)
1714
(1714)
1714
(1695)
a
+
Experimental values for 2b, 2b BAr−F and 3b+ BArF− are given
,
’
5
Fig. 5. Structural parameters (A, °) for [(h -C5H5)Re(NO)(PR3)(Me)]
in parenthesis.
at the B3LYP/LANL2DZp level.

60
J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
Activation parameters can be derived from Table 3.
phosphite complexes (although none of the ‘piano
stool’ variety). They found that the metalꢁphosphorus
(MꢁPX3) bonds lengthened upon oxidation. However,
the PꢁX bond distances decreased. This was interpreted
as evidence for the participation of PꢁX s*-orbitals in
phosphine ligand backbonding. A stronger interaction
would be expected in the 18-electron systems, leading to
longer PꢁX bonds. This is readily visualized from the
HOMO of {Re(PH3)(CH3)} shown in Fig. 8(B). There
is a large degree of rhenium d-orbital ‘lone pair’ charac-
ter, resulting in slight PꢁH s* character for the appro-
The DH‡ values for methyl and phosphine rotation in
{Re(PH3)(CH3)} are 0.9 and 0.6 kcal mol−1, respec-
tively (298.15 K). The corresponding DS‡ values are
−3.0 and −2.6 cal mol−1, and DG‡ values are 1.8 and
1.4 kcal mol−1. There is ample precedent for rotational
barriers of this magnitude in transition metal methyl
complexes [24], although much higher barriers are pos-
sible with more congested metal fragments, and alkyl or
phosphine ligands [25]. Attempts to locate the transi-
tion state for methyl rotation in the radical cation
{Re(PH3)(CH3)} + failed. This indicates that the energy
’
priately aligned bond. The singly-occupied HOMO
barrier cannot be higher than that of the neutral
complex.
+
’
(SOMO) of radical cation {Re(PR3)(CH3)}
much different electron distribution.
has a
The geometry of the transition state for methylidene
rotation in {Re(PH3)(ꢀCH2)}+ matches that calculated
earlier by extended Hu¨ckel theory [22a]. As shown in
Fig. 9, the plane of the ligand rotates ca. 90°, as
reflected by the new HaꢁCꢁReꢁP and Ha%ꢁCꢁReꢁN
torsion angles (−82.6° versus −176.5°; 9.3° versus
−78.8°). Unlike methyl rotation in {Re(PH3)(CH3)},
the conformation of the phosphine significantly
changes. The DH‡ and DS‡ values computed from
Table 3 are 27.9 kcal mol−1 and 1.3 cal mol−1 (298.15
The first question is then ‘how different are the
crystal structures of 2b and 2b + BArF−?’ The overlay in
’
Fig. 3 shows that the rhenium moieties exhibit very
similar structures. The eight-atom Re(NO)(CH3)P-
(Cipso 3 segments are virtually superimposable. The
)
greatest difference is in one PꢁCipso conformation, a
slight effect that could easily be due to packing forces.
However, the esd values associated with the crystallo-
graphic data are only of average quality
K). These give a DG3‡83
value of 27.4 kcal mol−1
,
K
consistent with that bounded by NMR for the much
more congested species 3b+ BArF− (\17.5 kcal mol−1),
as well as triphenylphosphine and triphenylphosphite
analogs 3a+ PF−6 and 5+ PF6− (\19 kcal mol−1) [13].
Some cyclopentadienyl analogs [(h5-C5H5)Re(NO)-
(PPh3)(ꢀCHR)]+ X− can be generated as non-equi-
librium mixtures of geometric isomers, and rate studies
give DG‡ values of ca. 20 kcal mol−1 [22a,26].
3. Discussion
+
To our knowledge, 2b and 2b BAr−F represent only
’
the second structurally characterized pair of 18- and
17-valence-electron complexes of the ‘piano stool’ for-
mula (h5-C5R5)M(L)(L%)(L%%). Crystal structures of the
chromium complexes (h5-C5H5)Cr(NO)(P(OMe)3)2 and
5
+
PF−6 were previously
’
[(h -C5H5)Cr(NO)(P(OMe)3)2]
reported by Legzdins, Einstein and coworkers [4f]. The
closely related 17/18-electron species (h5-C5Me5)Fe-
+
’
(dppe)(CH2OMe)]
PF−6 and (h5-C5Me5)Fe(dppe)-
(CꢂCH)
[4a,27],
and
(h5-C5H5)Cr(NO)(PPh3)-
(CH2SiMe3) and (h5-C5H5)Fe(CO)(PPh3)(CH2SiMe3)
[4e,28] also deserve mention. However, more exact
comparisons can be made with otherwise identical com-
+
pounds, such as 2b and 2b BAr−F .
’
In this context, a landmark study of Orpen and
Connelly merits emphasis [29]. These authors compared
nine high-quality crystal structures of homologous neu-
tral 18- and cationic 17-valence-electron phosphine and
5
,
Fig. 6. Structural parameters (A, °) for [(h -C5H5)Re(NO)(PR3)-
+
’
(Me)] at the UB3LYP/LANL2DZp level.

J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
61
smaller than the decreases found by Orpen and Con-
nelly [29]. The small anisotropy in the reductions has
prompted us to consider other contributions. For ex-
ample, in work in progress, we compute substantial
PꢁC bond contractions upon protonation of PPh3 to
[HPPh3]+ (1.854 versus 1.801 A) [30]. The examples of
,
Orpen and Connelly similarly feature neutral and
cationic homologs. Table 4 shows that positive charge
on phosphorus increases upon going from
+
’
{Re(PR3)(CH3)} to {Re(PR3)(CH3)}
.
The complexes of Legzdins [4f], (h5-C5H5)Cr(NO)-
5
+
’
(P(OMe)3)2 and [(h -C5H5)Cr(NO)(P(OMe)3)2] PF−6 ,
exhibit similar relationships. The chromiumꢁ
phosphorus bonds lengthen from 2.227(2)–2.240(2) A
(two independent molecules in unit cell) to 2.343(2)–
,
,
2.346(3) A, which the authors attributed to a diminu-
tion in back-bonding. The phosphorusꢁoxygen bond
lengths were not interpreted, perhaps due to disorder of
the phosphite ligand in both species. However, the
refined data indicate a contraction from 1.580(4)–
,
1.623(6) A (average and standard deviation for twelve
,
,
bonds: 1.601(3)90.0038 A) to 1.562(5)–1.578(4) A (av-
erage and standard deviation for six bonds: 1.568(8)9
,
0.0026 A).
+
’
The rheniumꢁmethyl bond lengths in 2b and 2b
5
,
Fig. 7. Structural parameters (A, °) for [(h -C5H5)Re(NO)(PR3)-
(ꢀCH2)]+ at the B3LYP/LANL2DZp level.
BAr−F can also be analyzed. The crystallographic dis-
tances are essentially identical (2.123(11) versus
,
2.120(12) A), a trend we are attempting to further
(Table 2), and it is often difficult to be certain that one
bond length or angle is greater than another. In these
cases, the computational data (Figs. 5 and 6) provide
reliable answers.
support with analogs of 2b. For example, the cyclopen-
tadienyl benzyl complex (R)-(h5-C5H5)Re(NO)(PPh3)-
(CH2C6H5) exhibits a longer rheniumꢁcarbon bond
,
(2.203(8) A) [31], as does a similar secondary alkyl
complex [32]. The computational data for
{Re(PR3)(CH3)} and {Re(PR3)(CH3)} show distinct
contractions (R=H, 2.186 versus 2.143 A; R=Me,
2.191 versus 2.137 A) and increased bond order (Table
4), opposite to the trends with the phosphine and
nitrosyl ligands. This can be rationalized from several
perspectives. First, methyl is the poorest p accepting
ligand, so the p bond order should be least affected by
oxidation. Second, the carbonꢁhydrogen bonding elec-
tron pairs are possible sources of repulsive interactions
with filled metal orbitals in 18-valence-electron systems
[33]. For example, an antibonding interaction is appar-
ent in the HOMO of {Re(PH3)(CH3)} shown in Fig.
8(B). Third, the carbonꢁhydrogen bonding electron
pairs can supply hyperconjugative stabilization after
oxidation.
Regardless, the crystallographic distances between
the heavy rhenium and phosphorus atoms in 2b and
+
’
+
’
2b
BAr−F are very accurately determined. The in-
,
,
crease from 2.359(2) to 2.455(3) A is typical of those
,
found by Orpen and Connelly [29]. The computational
+
’
data for {Re(PR3)(CH3)} and {Re(PR3)(CH3)} show
,
nearly identical increases (R=H, 2.353–2.474 A; R=
,
Me, 2.384–2.469 A), as well as an accompanying de-
crease in bond order (Table 4). The rheniumꢁnitrogen
bonds show a less pronounced lengthening (R=H,
,
,
1.766 versus 1.779 A; R=Me, 1.759 versus 1.772 A),
but a more pronounced bond order decrease. The rhe-
+
niumꢁnitrogen bond lengths of 2b and 2b BAr−F have
esd values too large to allow meaningful comparisons
’
,
(1.787(9) versus 1.813(12) A).
It is similarly difficult to contrast the phospho-
+
’
rusꢁcarbon bond lengths in 2b and 2b
BAr−F
Methylidene ligands are strong p acceptors [34], and
numerous complexes are known with high degrees of
metal-double bond character [14]. Accordingly, the
,
(1.825(8) versus 1.827(11) A; 1.829(9) versus 1.811(11)
A; 1.839(9) versus 1.805(12) A). Two of the three
appear to be shorter. However, the computational data
,
,
+
’
rheniumꢁcarbon
bond
lengths
computed
for
for {Re(PR3)(CH3)}
show unambiguous and rela-
tively uniform PꢁR bond contractions averaging 0.010
{Re(PR3)(ꢀCH2)}+ are much shorter than those in
methyl complexes {Re(PR3)(CH3)} (R=H, 1.923 ver-
,
,
A and 0.013 A (R=H, Me). These are somewhat

62
J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
,
,
,
,
2.430 versus 2.353 A; R=Me, 2.448 versus 2.384 A),
sus 2.186 A; R=Me, 1.918 versus 2.191 A), and in good
agreement with the triphenylphosphite complex 5+ PF6−
although still slightly shorter than in radical cations
+
’
,
,
,
{Re(PR3)(CH3)} (R=H, 2.474 A; R=Me, 2.469 A).
Also, the PꢁR distances are shorter than those in the
neutral complexes, and close to those of the radical
cations. However, the rheniumꢁphosphorus bond order
is higher in {Re(PH3)(ꢀCH2)}+ than {Re(PR3)(CH3)}
(Table 4), suggesting an underlying s-orbital effect.
(1.89(2) A) [13]. As expected, the rheniumꢁcarbon bond
order is much higher (Table 4). In the ground state
conformation, the methylidene ligand competes with the
phosphine for the d-orbital HOMO of the rhenium
fragment (Fig. 8(A)). Thus, the rheniumꢁphosphorus
bonds are longer than those in {Re(PR3)(CH3)} (R=H,
Fig. 8. Frontier molecular orbitals of model complexes.

J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
63
tions of related species failed to give SCF convergence
[7b]. Similarly, the frontier orbitals in Fig. 9 resemble
those obtained with lower levels of theory [22,23d,36],
and are not analyzed further. However, we are conduct-
ing an extensive computational study of dirhenium Cx
complexes of the formula [(h5-C5R5)Re(NO)(PR3)-
(Cx)(R3P)(ON)Re(h5-C5R5)]n+, and the preceding ener-
gies, geometries, and orbitals constitute important
benchmarks for this much more involved undertaking
[37].
Some purely experimental aspects of this study merit
further analysis. First, we are somewhat surprised by
+
the apparent lack of reactivity of 2b BAr−F towards
’
phosphine substitution. There are many well docu-
mented associative substitutions of 17-valence-electron
‘piano stool’ complexes, involving both transient [38]
and isolated [4e,39] species. However, marked steric
effects have been observed [4e]. Except for (h5-
’
C5H5)W(CO)3 [38e] and the chain process in the previ-
ous paper [6], examples involving third-row metals
appear to be rare. Nonetheless, other interesting reac-
tivity modes can be anticipated. For example, in collab-
orative studies with Parker, we showed that reactions of
cyclopentadienyl alkyl complexes (h5-C5H5)Re(NO)-
(PPh3)(CHRCHR%) and trityl cations proceed by initial
electron transfer to give rhenium radical cations and
trityl radicals [40]. Then hydrogen atom transfer occurs.
+
Hence, 2b BAr−F is likely to be an effective hydrogen
’
,
Fig. 9. Structural parameters (A, °) of the transition states for ligand
rotation at the B3LYP/LANL2DZp level.
atom donor, and is a probable intermediate in the
synthesis of 3b+ BAr−F in Scheme 1.
Second, the original motivation of this study was to
The transition state for methylene rotation in
{Re(PH3)(ꢀCH2)}+ (Fig. 9) features a longer rhe-
niumꢁcarbon bond than the ground state (1.983 versus
obtain longer chain homologs of the C4 dirhenium
radical cation 1 PF−6 and dication 12+ 2PF−6 with
+
’
enhanced stabilities [6]. The above data show that an
anion effect can be added to the phosphine effect de-
scribed in the preceding paper [6]. Such anion effects
have precedent [9], but are generally not amenable to
single-parameter explanations. Nonetheless, it is in-
,
1.923 A). However, the bond remains much shorter
than in the ground states of methyl complexes
+
’
{Re(PH3)(CH3)} and {Re(PH3)(CH3)}
(2.186 and
,
2.143 A). This indicates a reduced but still significant p
bond order. In the transition state conformation, the
methylene ligand competes with the nitrosyl for another
of the d donor orbitals on rhenium. Accordingly, the
rheniumꢁnitrogen bond is much longer than in the
structive to note that the formula weight and unit cell
+
’
volume of 2b
BAr−F are factors of 2.28 and 2.22
greater than those of 2b (both crystallize as monoclinic
systems with Z=4). In conclusion, this paper describes
an important refinement in our ‘first generation’ ap-
proach to the Cx target complexes described above. The
next strategic step in this ultimately successful quest will
be the subject of a future communication [41].
,
ground state (1.826 versus 1.780 A). Also, the rhe-
niumꢁphosphorus bond contracts (2.399 versus 2.430
,
A), and one of the three phosphorusꢁhydrogen bonds
,
lengthens (average: 1.424 versus 1.419 A). The Orpen–
Connelly model predicts longer phosphorusꢁhydrogen
bonds.
The preceding calculations also illustrate a number of
conformational features and subtleties that have been
analyzed in previous publications in this series
[22,23,35], and in related work by others [28]. However,
4. Experimental
4.1. General data
the purpose of this discussion is to emphasize the new
Many general procedures were identical with those in
the previous paper [6], but most work was conducted at
a new location. Thus, the following instruments dif-
fered: NMR, Jeol JMN-400CX spectrometer (stan-
+
insights afforded by the radical cations 2b BAr−F and
’
{Re(PR3)(CH3)} +. These represent both experimental
’
and computational breakthroughs, as earlier calcula-

64
J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
dards: 1H, residual internal C6D5H (l 7.16), CDHCl2 (l
5.32), CD2HCN (l 1.94); 13C, internal CD2Cl2 (l
54.00); 31P, external 85% H3PO4 (l 0.00)); IR, Bruker
IFS 25 or React-IRTM 1000 Reaction Analysis System;
UV–vis, Shimadzu UV-3102 PC; ESR, Bruker ESP-
3000E (ER 4116 DM dual mode X-band cavity, Oxford
Instruments ESR-900 helium flow cryostat; spectra at
modulation amplitude 12.6 G and sweep rate 100 G
140–148°C. IR (CH2Cl2) wNO=1714 cm−1 s.
C64H51BF24NOPRe: Anal. Found: C, 49.56; H, 3.14; N,
0.95. Calc. C, 50.10; H, 3.35; N, 0.91%.
+
4.4. Reduction of 2b BAr−F
’
+
A Schlenk flask was charged with 2b BAr−F (0.045
’
g, 0.029 mmol) and CH2Cl2 (2 ml). Then solid (h5-
s
−1), mass spectrometry, Micromass Zabspec; cyclic
C5H5)2Co (0.005 g, 0.029 mmol) was added with stir-
’
voltammetry, BAS model CV-50W; microanalyses,
Carlo Erba model EA1110 (all measurements in-house).
All manipulations were carried out under N2 that had
been dried with KOH and CaCl2. Hexane (\95% GC
grade, Fluka) was distillated from sodium, CH2Cl2
(99.9% HPLC grade, Sigma-Aldrich) from CaH2, and
benzene from Na–benzophenone. CD3CN was refluxed
over CaH2, vacuum-transferred, and freeze-pump-thaw
ring. After 10 min, solvent was removed by oil pump
vacuum. The residue was extracted with benzene (3×
10 ml). The solution was filtered through a silica gel
pad. The filtrate was taken to dryness by oil pump
vacuum. The residue was recrystallized from CH2Cl2–
hexane to give 2b as an orange powder (0.017 g, 0.026
mmol, 90%). 1H-NMR (C6D6, 400 MHz): l=7.59
(pseudo t, JHP=JHH=8.6, 3oꢁC6H4), 6.93 (d, JHH
=
degassed. Other materials, (h5-C5H5)2Fe (Fluka), and
8.2 Hz, 3m-C6H4), 1.99 (s, 3ArCH3), 1.62 (s, C5(CH3)5),
1.37 (br s, ReCH3).
5
’
(h -C5H5)2Co (Acros), were used as received.
5
+
4.2. (p -C5H5)2Fe BAr−F [9b]
4.5. [(p5-C5Me5)Re(NO)(P(4-C6H4CH3)3)(ꢀCH2)]+
BAr−F (3b+ BAr−F )
’
A Schlenk flask was charged with (h5-C5H5)2Fe
(0.150 g, 0.806 mmol), acetone (1.2 ml), and water (4
ml). The suspension was stirred, and solid FeCl3·6H2O
(0.360 g, 1.331 mmol) was added. After 15 min, the
dark blue solution was filtered. Then Na+ BArF− (0.710
g, 0.801 mmol) [15,42] was added to the filtrate. After
20 min, a dark precipitate was removed by filtration,
washed with water (2×2 ml) and hexane (2×2 ml),
and dried by oil pump vacuum (3 h). The residue was
extracted with CH2Cl2 (10 ml). The mixture was
filtered, and hexane (5 ml) was added to the filtrate.
The sample was concentrated under vacuum. The dark
blue solid was collected by filtration and washed with
A Schlenk flask was charged with 2b (0.061 g, 0.091
mmol) and CH2Cl2 (5 ml) and cooled to −78°C (ace-
tone–CO2). Then Ph3C+ BAr−F (0.101 g, 0.091 mmol)
[15] was added against a N2 flow with stirring. After 3.5
h at −78°C, hexane (10 ml) was added. The sample
was concentrated under vacuum at 0°C, giving a precip-
itate. Then more hexane (15 ml) was added. After 10
min, the yellow precipitate was collected by filtration,
washed with hexane (2×5 ml), and dried by oil pump
vacuum (1 h) to give 3b+ BAr−F (0.112 g, 0.073 mmol;
80%), m.p. (dec.) 160–170°C. IR (solid film) wNO
=
1
1695 cm−1 s. H-NMR (CD2Cl2, 400 MHz) l=15.17
and 14.22 (2dd, JHH%=6 Hz, JHP=1.6 Hz, ꢀCHH%),
7.72 (br s, 4o-BC6H3), 7.55 (br s, 4p-BC6H3), 7.35–
7.32, 7.22–7.17 (m, 3o,m-PC6H4), 2.41 (s, 3ArCH3),
1.88 (s, C5(CH3)5). 13C{1H} (100.4 MHz) l=284.7 (br
s, ꢀCH2), 162.5 (q, JCB=49.8 Hz, i-BC6H3), 144.3 (s,
p-PC6H4), 135.5 (br s, o-BC6H3), 133.4 (d, JCP=12 Hz,
o-PC6H4), 130.8 (d, JCP=10 Hz, m-PC6H4), 130.5–
129.1 (m, i-PC6H4, m-BC6H3), 125.2 (q, JCF=270 Hz,
CF3), 118.1 (br s, p-C6H3), 111.3 (s, C5(CH3)5), 21.6 (s,
ArCH3), 10.3 (s, C5(CH3)5). 31P{1H} (161.7 MHz) l=
22.8 (s). MS (positive FAB, 3-NBA–CH2Cl2): 670
(3b+, 40%), 656 (3b-CH2+, 100%). C64H50BF24NOPRe:
Anal. Found: C, 50.44; H, 3.36; N, 0.89. Calc. C, 50.14;
H, 3.29; N, 0.91%.
5
+
hexane (2×5 ml) to give (h -C5H5)2Fe BAr−F (0.676
g, 0.644 mmol; 80%). C42H22BF24Fe: Anal. Found: C,
47.70; H, 2.07. Calc. C, 48.08; H, 2.11%.
’
5
+
’
4.3. [(p -C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3)]
+
BAr−F (2b BAr−F )
’
A
Schlenk flask was charged with (h5-
C5Me5)Re(NO)(P(4-C6H4CH3)3)(CH3) (2b [6]; 0.030 g,
0.044 mmol) and CH2Cl2 (3 ml). Then solid (h5-
+
’
C5H5)2Fe
BAr−F (0.047 g, 0.044 mmol) was added
with stirring. After 10 min, solvent was removed from
the orange–brown solution by oil pump vacuum. The
residue was washed with hexane (2×3 ml) and dis-
solved in CH2Cl2 (1 ml). The solution was added to
rapidly stirred cold hexane (30 ml, −78°C, acetone–
CO2). After 10 min, solvent was removed by cannula.
The tan powder was dried by oil pump vacuum as the
cold bath was allowed to warm to room temperature
4.6. Crystallography
Slow evaporation of a saturated CH2Cl2–hexane so-
lution of 2b afforded orange prisms. A CH2Cl2 solution
+
of 2b BAr−F was layered with hexane in a dry box.
’
(this period includes 1 h at room temperature) to give
+
’
2b
BAr−F (0.054 g, 0.035 mmol; 80%), m.p. (dec)
Dark brown prisms formed. Data were collected as

J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
65
Neil, R.J. Batchelor, F.W.B. Einstein, J. Am. Chem. Soc. 117
(1995) 10521.
outlined in Table 1 using a Nonius MACH3 diffrac-
tometer. Cell parameters were obtained from 15 reflec-
tions with 5.0B2qB50.0°. The space groups were
determined from systematic absences (h01, l=2n; 0k0,
k=2n) and subsequent least-squares refinement. Em-
pirical absorption corrections were applied (c-scans).
The structures were solved by direct methods using
[5] Other types of structurally characterized 17-valence-electron rhe-
5
+
nium complexes: (a) [(h -C5H5)Re(NCCH3)(PPh3)2(H)] PF6−
:
’
M.R. Detty, W.D. Jones, J. Am. Chem. Soc. 109 (1987) 5666. (b)
Re(CO)3(P(c-C6H11)3)2 : L.S. Crocker, D.M. Heinekey, G.K.
’
Schulte, J. Am. Chem. Soc. 111 (1989) 405. (c) There is also an
extensive literature involving paramagnetic rhenium complexes
in which the odd electron is ligand-centered: A. Klein, C. Vogler,
W. Kaim, Organometallics 15 (1996) 236.
SHELXS-86. Then SHELXS-93 [43] was used to refine 325
+
(2b) or 838 (2b BAr−F ) parameters with all data by
’
[6] W.E. Meyer, A.J. Amoroso, M. Jaeger, J. Le Bras, W.-T. Wong,
J.A. Gladysz, J. Organomet. Chem. 616 (2000) 44.
full-matrix least-squares on F2. Non-hydrogen atoms
were refined anisotropically. The hydrogen atoms were
fixed in idealized positions using a riding model.
[7] Full papers with extensive literature background: (a) W. Weng,
T. Bartik, M. Brady, B. Bartik, J.A. Ramsden, A.M. Arif, J.A.
Gladysz, J. Am. Chem. Soc. 117 (1995) 11922. (b) M. Brady, W.
Weng, Y. Zhou, J.W. Seyler, A.J. Amoroso, A.M. Arif, M.
Bo¨hme, G. Frenking, J.A. Gladysz, J. Am. Chem. Soc. 119
(1997) 775. (c) S.B. Falloon, S. Szafert, A.M. Arif, J.A. Gladysz,
Chem. Eur. J. 4 (1998) 1033. (d) T. Bartik, W. Weng, J.A.
Ramsden, S. Szafert, S.B. Falloon, A.M. Arif, J.A. Gladysz, J.
Am. Chem. Soc. 120 (1998) 11071. (e) R. Dembinski, T. Bartik,
B. Bartik, M. Jaeger, J.A. Gladysz, J. Am. Chem. Soc. 122
(2000) 810.
5. Supplementary material
All data (excluding structure factors) have been de-
posited with the Cambridge Crystallographic Data
+
’
Centre, CCDC nos. 140674 (2b) and 140675 (2b
[8] BAr−F =B(3,5-C6H3(CF3)2)−4
.
BAr−F ). Copies of this information may be obtained
free of charge from The Director, CCDC, 12 Union
Road, Cambridge CB2 1EZ, UK (Fax: +44-1223-
336033; e-mail: deposit@ccdc.cam.ac.uk or www: http:
//www.ccdc.cam.ac.uk).
[9] (a) J. Huhmann-Vincent, B.L. Scott, G.J. Kubas, Inorg. Chem.
38 (1999) 115 and references therein. (b) D.M. Heinekey, C.E.
Radzewich, Organometallics 17 (1998) 51.
[10] W.V. Konze, B.L. Scott, G.L. Kubas, Chem. Commun. (1999)
1807.
[11] (a) F. Calderazzo, G. Pampaloni, L. Rocchi, U. Englert,
Organometallics 13 (1994) 2592. (b) I. Cha´vez, A. Alvarez-
Carena, E. Molins, A. Roig, W. Maniukiewicz, A. Arancibia, H.
Brand, J.M. Manr´ıquez, J. Organomet. Chem. 601 (2000) 126.
[12] M. Tilset, G.S. Bodner, D.R. Senn, J.A. Gladysz, V.D. Parker,
J. Am. Chem. Soc. 109 (1987) 7551.
Acknowledgements
[13] A.T. Patton, C.E. Strouse, C.B. Knobler, J.A. Gladysz, J. Am.
Chem. Soc. 105 (1983) 5804.
We thank the Deutsche Forschungsgemeinschaft
(DFG, GL 300/2-1) for financial support, and Professor
Dr Ulrich Zenneck and Matthias Zeller of the Institut
fu¨r Anorganische Chemie for assistance with the ESR
spectra.
[14] Leading references to isolable methylidene complexes: J.L. Bru-
maghim, G.S. Girolami, Chem. Commun. (1999) 953.
[15] S.R. Bahr, P. Boudjouk, J. Org. Chem. 57 (1992) 5545.
[16] J. Sandstro¨m, Dynamic NMR Spectrometry, London, Academic
Press, 1982 Chapters 6–7.
[17] W. Weng, T. Bartik, M.T. Johnson, A.M. Arif, J.A. Gladysz,
Organometallics 14 (1995) 889.
[18] (a) P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299. (b) T.H.
Dunning Jr., P.J. Hay, Modern Theoretical Chemistry, in: H.F.
Scheafer III (Ed.), Plenum Press, New York, 1976. The double
zeta basis set and a set of polarization function (DZp) for carbon
(10s5p[3s/2p]721/41) with a polarization function (n=0.6) and
for hydrogen (4s[2s]/31). For phosphorus the pseudopotential
(3s/3p[2s/2p]21/21) with a set of polarization function (n=0.6).
The basis set for rhenium required has the DZ quality {8s/6p/
3d[3s/3p/2d]341/321/21) with (n=0.073), and the (n-1)s and
(n-1)p electrons were considered as part of the valence space.
[19] GAUSSIAN-98, Revision A.5, M.J. Frisch, G.W. Trucks, H.B.
Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G.
Zakrzewski, J.A. Montgomery Jr., R.E. Stratmann, J.C. Burant,
S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C.
Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi,
B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski,
G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Mal-
ick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.
Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T.
Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonza-
lez, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W.
Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S. Re-
plogle, J.A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.
References
[1] D. Astruc, Electron Transfer and Radical Processes in Transi-
tion-Metal Chemistry, VCH, New York, 1995.
[2] M.C. Baird, Chem. Rev. 88 (1988) 1217.
[3] A recent leading reference: K. Costuas, J.-Y. Saillard,
Organometallics 18 (1999) 2505.
[4] Structurally characterized 17-valence-electron (h5-C5R5)M(L)-
(L%)(L%%)
systems:
(a)
[(h5-C5Me5)Fe(Ph2PCH2CH2PPh2)-
(CH2OMe)] PF−6 : C. Roger, L. Toupet, C. Lapinte, J. Chem.
+
’
5
’
Soc. Chem. Commun. (1988) 713. (b) (h -C5H5)Cr(CO)2(PR3)
(R=Ph, Me): S. Fortier, M.C. Baird, K.F. Preston, J.R. Mor-
ton, T. Ziegler, T.J. Jaeger, W.C. Watkins, J.H. MacNeil, K.A.
Watson, K. Hensel, Y. Le Page, J.P. Charland, A.J. Williams, J.
Am. Chem. Soc. 113 (1991) 542. (c) [(h5-C5Me5)Fe(Ph2PCH2-
CH2PPh2)(D)] PF−6 : P. Hamon, L. Loupet, J.-R. Hamon, C.
+
’
Lapinte, Organometallics 11 (1992) 1429. (d) (h5-
’
C5Ph5)Cr(CO)3 : R.J. Hoobler, M.A. Hutton, M.M. Dillard,
M.P. Castellani, A.L. Rheingold, A.L. Rieger, P.H. Rieger, T.C.
Richards, W.E. Geiger, Organometallics 12 (1993) 116. (e) (h5-
’
C5H5)Cr(NO)(PPh3)(CH2SiMe3) : P. Legzdins, M.J. Shaw, R.J.
Batchelor, F.W.B. Einstein, Organometallics 14 (1995) 4721. (f)
5
+
’
[(h -C5H5)Cr(NO)(P(OMe)3)2]
PF−6 : P. Legzdins, W.S. Mc-

66
J. Le Bras et al. / Journal of Organometallic Chemistry 616 (2000) 54–66
[29] A.G. Orpen, N.G. Connelly, J. Chem. Soc. Chem. Commun.
(1985) 1310.
[30] H. Jiao, Unpublished computations, University Erlangen-Nu¨rn-
berg.
[20] For theory and applications: (a) W.J. Hehre, L. Radom, P.v.R.
Schleyer, J.A. Pople, Ab Initio Molecular Orbital Theory, New
York, 1986. (b) J.B. Foresman, Æ. Frisch, Exploring Chemistry
with Electronic Structure Methods: A Guide to Using GAUSSIAN
,
[31] J.H. Merrifield, C.E. Strouse, J.A. Gladysz, Organometallics 1
(1982) 1204.
[32] T.-S. Peng, A.M. Arif, J.A. Gladysz, J. Chem. Soc. Dalton Trans.
(1995) 1857.
Second Edition; Gaussian, Inc.: Pittsburgh PA, 1996.
[21] (a) NBO Program, 3.1, E.D. Glendening, A.E. Reed, J.E. Carpen-
ter, F. Weinhold. (b) A.E. Reed, L.A. Curtiss, F. Weinhold, Chem.
Rev. 88 (1988) 899.
[33] B.E.R. Schilling, R. Hoffmann, D.L. Lichtenberger, J. Am. Chem.
Soc. 101 (1979) 585.
[34] A.K. Rappe´, W.A. Goddard, III, J. Am. Chem. Soc. 99 (1977)
3966.
[35] G.L. Crocco, K.E. Lee, J.A. Gladysz, Organometallics 9 (1990)
2819.
[36] B.E.R. Schilling, R. Hoffmann, J.W. Faller, J. Am. Chem. Soc.
101 (1979) 592.
[37] H. Jiao, J.A. Gladysz, (in preparation).
[38] (a) J.W. Hershberger, R.J. Klingler, J.K. Kochi, J. Am. Chem.
Soc. 105 (1983) 61. (b) N.N. Turaki, J.M. Huggins Organometal-
lics 5 (1986) 1703. (c) Y. Huang, G.B. Carpenter, D.A. Sweigart,
Y.K Chung, B.Y. Lee, Organometallics 14 (1995) 1423, and
references therein.
[39] (a) D. Astruc, Acc. Chem. Res. 24 (1991) 36. (b) P. Legzdins, W.S.
McNeil, M.J. Shaw, Organometallics 13 (1994) 562.
[40] G.S. Bodner, J.A. Gladysz, M.F. Nielsen, V.D. Parker, J. Am.
Chem. Soc. 109 (1987) 1757.
[22] (a) W.A. Kiel, G.-Y. Lin, A.G. Constable, F.B. McCormick, C.E.
Strouse, O. Eisenstein, J.A. Gladysz, J. Am. Chem. Soc. 104 (1982)
4865. (b) S. Georgiou, J.A. Gladysz, Tetrahedron 42 (1986) 1109.
[23] For other computational studies involving chiral rhenium com-
plexes, see Ref. 7b and the following: (a) G.S. Bodner, A.T. Patton,
D.E. Smith, S. Georgiou, W. Tam, W.-K. Wong, C.E. Strouse,
J.A. Gladysz, Organometallics 6 (1987) 1954. (b) G.S. Bodner,
D.E. Smith, W.G. Hatton, P.C. Heah, S. Georgiou, A.L. Rhein-
gold, S.J. Geib, J.P. Hutchinson, J.A. Gladysz, J. Am. Chem. Soc.
109 (1987) 7688. (c) W.E. Buhro, B.D. Zwick, S. Georgiou, J.P.
Hutchinson, J.A. Gladysz, J. Am. Chem. Soc. 110 (1988) 2427. (d)
P.T. Czech, J.A. Gladysz, R.F. Fenske, Organometallics 8 (1989)
1806.
[24] R.F. Jordan, J.R. Norton, J. Am. Chem. Soc. 101 (1979) 4853.
[25] (a) F.B. McCormick, R.J. Angelici, R.A. Pickering, R.E. Wagner,
R.A. Jacobson, Inorg. Chem. 20 (1981) 4108. (b) For metal–PPh3
rotational barriers, see S.G. Davies, A.E. Derome, J.P. McNally,
J. Am. Chem. Soc. 113 (1991) 2854, and references therein.
[26] W.A. Kiel, G.-Y. Lin, G.S. Bodner, J.A. Gladysz, J. Am. Chem.
Soc. 105 (1983) 4958.
[41] W.E. Meyer, A.J. Amoroso, C.R. Horn, M. Jaeger, J.A. Gladysz,
submitted to Organometallics.
[27] L. Dahlenburg, A. Weiß, M. Bock, A. Zahl, J. Organomet. Chem.
541 (1997) 465.
[42] M. Brookhart, B. Grant, A.F. Volpe, Jr., Organometallics 11
(1992) 3920.
[28] S.G. Davies, I.M. Dordor-Hedgecock, K.H. Sutton, M. Whit-
taker, J. Am. Chem. Soc. 109 (1987) 5711.
[43] G.M. Sheldrick, SHELXL-93, University of Go¨ttingen, Go¨ttingen,
Germany, 1993.
.