Synthesis and characterization of a bimetallic boratabenzene cobalt complex

Article abstract of DOI:10.1021/om049750y

Reaction of 1,4-(BBr₂)₂C₆H₄ with cobaltocene gives a bimetallic cobalt complex with cyclopentadienyl and boratabenzene ligands. Oxidation allows the dicationic dicobalt(III) complex to be isolated; the bis(hexafluorophosphate) salt has been structurally characterized. Variable-temperature magnetic susceptibility measurements on the neutral dicobalt(II) complex showed substantial deviation from the Curie -Weiss law; a fit to the Bleaney-Bowers equation for a singlet ground state and a triplet excited state gave an exchange interaction of J = ca. -28 cm ^-1. Electrochemical studies, as well as near-infrared data on the mixed-valence species, suggest that the interaction between the metal centers is weak.

Full text of DOI:10.1021/om049750y

3808
Organometallics 2004, 23, 3808-3813
Syn th esis a n d Ch a r a cter iza tion of a Bim eta llic
Bor a ta ben zen e Coba lt Com p lex
Tony Hascall,† Victoria Beck,† Stephen Barlow,‡ Andrew R. Cowley,† and
Dermot O’Hare*,†
Chemistry Research Laboratory, Department of Chemistry, University of Oxford,
Mansfield Road, Oxford OX1 3TA, U.K., and School of Chemistry and Biochemistry,
Georgia Institute of Technology, Atlanta, Georgia 30332-0400
Received April 6, 2004
Reaction of 1,4-(BBr2)2C6H4 with cobaltocene gives a bimetallic cobalt complex with
cyclopentadienyl and boratabenzene ligands. Oxidation allows the dicationic dicobalt(III)
complex to be isolated; the bis(hexafluorophosphate) salt has been structurally characterized.
Variable-temperature magnetic susceptibility measurements on the neutral dicobalt(II)
complex showed substantial deviation from the Curie-Weiss law; a fit to the Bleaney-
Bowers equation for a singlet ground state and a triplet excited state gave an exchange
interaction of J ) ca. -28 cm-1. Electrochemical studies, as well as near-infrared data on
the mixed-valence species, suggest that the interaction between the metal centers is weak.
In tr od u ction
anion3,5a with an appropriate metal precursor.3a This
approach is limited by the multistep, nontrivial syn-
theses of boratabenzenes.5 The original synthesis of
boratabenzene complexes, however, was via a straight-
forward reaction involving treatment of cobaltocene with
borane reagents RBX2 (X ) halogen), leading to the
insertion of boron into a cyclopentadienyl C-C bond and
the formation of cyclopentadienyl(boratabenzene) cobalt
complexes (eq 1).6 The byproduct of this reaction is
Bridged dinuclear or multinuclear metallocenes may
exhibit interesting properties due to interactions be-
tween the metal centers1a and can serve as models for
organometallic polymers.1b-e The nature and extent of
the metal-metal interactions can depend on the nature
of the coordinating ligand, as well as the bridging group.
We have been interested in studying the properties of
bridged metallocene analogues with noncyclopentadi-
enyl ligands2 and have now prepared a bimetallic cobalt
complex featuring a boratabenzene ligand.3 The bo-
ratabenzene anion, [C5H5BR]-, is isoelectronic with
benzene and, as a ligand, can be viewed as a six-
membered-ring analogue of cyclopentadienyl. Com-
plexes of boratabenzene ligands have been shown to
possess properties similar to those of their Cp ana-
logues.4
3(C5H5)2Co + RBX2 f
(C5H5)Co(C5H5BR) + 2(C5H5)2Co+X- (1)
cobaltocenium halide; thus, cobaltocene also serves as
a reducing agent. A second boron insertion can also
occur, yielding a bis(boratabenzene) complex. While the
reaction of boranes with cobaltocene to give borataben-
zene complexes has been studied extensively,3c there
appears to be only one previous example of the use of
such a reaction to prepare a bimetallic species, namely
the reaction of cobaltocene with FcBBr2 (Fc ) ferrocenyl)
to give the mixed iron-cobalt complex CpCo(C5H5-
BFc).7,8
Boratabenzene complexes have generally been pre-
pared by the reaction of a salt of a boratabenzene
* To whom correspondence should be addressed. E-mail: dermot.
ohare@chem.ox.ac.uk.
Resu lts a n d Discu ssion
† University of Oxford.
‡ Georgia Institute of Technology.
9
Reaction of 1,4-(BBr2)2C6H4 with excess cobaltocene
(1) (a) Barlow, S.; O’Hare, D. Chem. Rev. 1997, 97, 7, 637-669. For
some recent reviews on metal-containing polymers, see: (b) Manners,
I. Macromol. Symp. 2003, 196, 57-62. (c) Abd-El-Aziz, A. S. Macromol.
Rapid Commun. 2002, 23, 995-1031. (d) Manners, I. J . Polym. Sci.
Polym. Chem. 2002, 40, 179-191. (e) Manners, I. Science 2001, 294,
1664-1666.
gives a bimetallic cyclopentadienyl-boratabenzene com-
plex, [CpCo(C5H5B)]2C6H4 (Co2; Scheme 1). In theory
6 equiv of cobaltocene is required per molecule of 1,4-
(BBr2)2C6H4, as 2 equiv of cobaltocenium halide is also
formed per boron insertion; however, less than the
(2) Hascall, T.; Drewitt, M. J .; O’Hare, D. Chem. Commun. 2001,
2600-2601.
(3) Reviews of boratabenzene (also known as borabenzene) com-
plexes: (a) Fu, G. C. Adv. Organomet. Chem. 2001, 47, 101-119. (b)
Ashe, A. J .; Al-Ahmad, S.; Fang, X. G. J . Organomet. Chem. 1999, 581,
92-97. (c) Herberich, G. E.; Ohst, H. Adv. Organomet. Chem. 1986,
25, 199-236. Huttner, G.; Krieg, B.; Gartzke, W. Chem. Ber. 1972,
105, 3424-3436.
(4) (a) Bazan, G. C.; Rodriguez, G.; Ashe, A. J .; Al-Ahmad, S.; Muller,
C. J . Am. Chem. Soc. 1996, 118, 2291-2292. (b) Ashe, A. J .; Kampf,
J . W.; Muller, C.; Schneider, M. Organometallics 1996, 15, 387-393.
(5) (a) Ashe, A. J .; Shu, P. J . Am. Chem. Soc. 1971, 93, 1804-1805.
(b) Herberich, G. E.; Schmidt, B.; Englert, U.; Wagner, T. Organome-
tallics 1993, 12, 2891-2893. (c) Herberich, G. E.; Schmidt, B.; Englert,
U. Organometallics 1995, 14, 471-480.
(6) (a) Herberich, G. E.; Greiss, G. Chem. Ber. 1972, 105, 3413-
3423. (b) Herberich, G. E.; Greiss, G.; Heil, H. F.; Muller, J . Chem.
Commun. 1971, 1328-1329. (c) Herberich, G. E.; Greiss, G.; Heil, H.
F. Angew. Chem., Int. Ed. Engl. 1970, 9, 805-806.
10.1021/om049750y CCC: $27.50 © 2004 American Chemical Society
Publication on Web 06/29/2004

A Bimetallic Boratabenzene Cobalt Complex
Organometallics, Vol. 23, No. 16, 2004 3809
Sch em e 1
F igu r e 1. Molecular structure of Co2(PF6)2, showing the
40% thermal probability ellipsoids. Hydrogen atoms have
been omitted; only the cation is shown.
rings and the bridging phenylene ring are all essentially
coplanar (interplanar angle 6.8°). The boratabenzene
rings are slightly distorted from planarity, with the
boron atom displaced away from the metal, leading to
a fold angle of ca. 7°. The boratabenzene moiety can be
considered to be bound in an η6 manner, although, as
has been observed previously for the C5H5BPh ligand,3
the coordination of the metal to the ring is slipped away
from the boron such that the shortest boratabenzene
Co-C bond length (2.056(3) Å) is to the carbon para to
the boron. The Co-B distance is 2.277(3) Å, which is
shorter than in the related cation CpCo(C5H5BFc)+
(2.307(4) Å)7 but is almost identical with the corre-
sponding Co-B bond lengths for the neutral complexes
(CO)2Co(C5H5BPh) (2.274(2) Å)3c and Co(C5H5BMe)2
(2.283(5) Å).3c
stoichiometric amount was used, as previous related
work had found that the best yields were obtained with
a slight deficiency of cobaltocene, presumably because
the formation of side products in which both Cp rings
of cobaltocene undergo boron insertion is minimized.
Attempts to recrystallize the dicobalt species Co2
from solution led to decomposition; therefore, the crude
Co2 was purified by oxidation in situ to give the air-
stable dication, a bimetallic analogue of cobaltocenium,
which could be isolated and purified as its PF6 salt
Co2(PF6)2. The dicationic complex is, as expected,
diamagnetic and exhibits three resonances in the 1H
NMR spectrum due to the boratabenzene ligand in the
region of δ 6.4-7.2, as well as singlets due to the Cp
protons at δ 6.06 and the four equivalent protons of the
arene bridge at δ 8.22. A signal is observed in the 11B
NMR spectrum at δ 25.0, consistent with related bo-
ratabenzene complexes.
The bimetallic dicationic cobalt complex Co2(PF6)2
was also characterized by X-ray crystallography, with
the cation shown in Figure 1. The compound crystallizes
in the space group P1h, with the cation located on the
crystallographic inversion center. Thus, the two CpCo-
(C5H5B) units are crystallographically equivalent, and
the metals are located in a trans configuration on
opposite sides of the bridging ligand. The Co‚‚‚Co
distance within the molecule is 9.879 Å; inspection of
the crystal packing reveals intermolecular Co‚‚‚Co
separations as short as 6.683 Å. The two boratabenzene
Complexes of bis(cyclopentadienyl)benzene and -bi-
phenylene ligands, arene-bridged dinuclear metal-
locenes related to Co2(PF6)2, have been synthesized
previously10-13 and have been suggested as possible
building blocks for organometallic polymers. Recently
a hexylated version of a bis(cyclopentadienyl)benzene
ligand was in fact used to prepare soluble arene-bridged
poly(ferrocenes).14 These ligands, however, suffer the
disadvantage that multistep organic synthesis is re-
quired for their preparation; thus, the reaction pre-
sented here represents a significantly more facile syn-
thesis of this type of compound.15
(10) (a) Patoux, C.; Coudret, C.; Launay, J -.P.; J oachim, C.; Gourdon,
A. Inorg. Chem. 1997, 36, 5037-5049. (b) Manriquez, J . M.; Ward, M.
D.; Calabrese, J . C.; Fagan, P. J .; Epstein, A. J .; Miller, J . S. Mol. Cryst.
Liq. Cryst. 1989, 176, 527-534. (c) Bunel, E. E.; Campos, P.; Ruz, J .;
Valle, L.; Chadwick, I.; Ana, M. S.; Gonzalez, G.; Manriquez, J . M.
Organometallics 1988, 7, 474-476.
(11) For fluorinated analogues see: (a) Deck, P. A.; Lane, M. J .;
Montgomery, J . L.; Slebodnick, C.; Fronczek, F. R. Organometallics
2000, 19, 1013-1024. (b) Hollandsworth, C. B.; Hollis, W. G.; Slebod-
nick, C.; Deck, P. A. Organometallics 1999, 18, 3610-3614.
(12) For examples containing carborane ligands see: (a) Meng, X.
S.; Sabat, M.; Grimes, R. N. J . Am. Chem. Soc. 1993, 115, 6143-6151.
(b) Stephan, M.; Davis, J . H.; Meng, X. S.; Chase, K. J .; Hauss, J .;
Zenneck, U.; Pritzkow, H.; Siebert, W.; Grimes, R. N. J . Am. Chem.
Soc. 1992, 114, 5214-5221. (c) Davis, J . H.; Sinn, E.; Grimes, R. N. J .
Am. Chem. Soc. 1989, 111, 4784-4790.
(13) For biferrocenes bridged by 1,4-diboracyclohexadiene see: (a)
Herberich, G. E.; Kucharska-J ansen, M. M. J . Organomet. Chem. 1983,
243, 45-49. (b) Herberich, G. E.; Hessner, B. Chem. Ber. 1982, 115,
3115-3127.
(14) Southard, G. E.; Curtis, M. D. Organometallics 2001, 20, 508-
522.
(7) Hagenau, U.; Heck, J .; Hendrickx, E.; Persoons, A.; Schuld, T.;
Wong, H. Inorg. Chem. 1996, 35, 7863-7866.
(8) Acetylene-bridged heterobimetallic complexes with borataben-
zene and Cp ligands were reported recently: Behrens, U.; Meyer-
Friedrichsen, T.; Heck, J . Z. Anorg. Allg. Chem. 2003, 629, 1421-1430.
(9) Haubold, W.; Herdtle, J .; Gollinger, W.; Einholz, W. J . Orga-
nomet. Chem. 1986, 315, 1-8.
(15) In a reaction related to the synthesis of Co2(PF6)2, Grimes et
al. treated a cyclopentadienyl(nido-carborane)cobalt dianion with 1,4-
(BBr2)2C6H4 to give a phenylene-bridged apically substituted complex.
See: Curtis, M. A.; Sabat, M.; Grimes, R. N. Inorg. Chem. 1996, 35,
6703-6707.

3810 Organometallics, Vol. 23, No. 16, 2004
Hascall et al.
F igu r e 2. Variable-temperature magnetic susceptibility data (circles) and the derived effective magnetic moment (crosses)
for Co2. The solid lines are fits to the Bleaney-Bowers equation, with additional terms added to account for a Curie
paramagnet impurity and for temperature-independent paramagnetism. The broken lines show the Bleaney-Bowers and
Curie components of the susceptibility fit separately.
The neutral dicobalt(II) species Co2 could be isolated
most conveniently by reduction of the dicobalt(III) salt
Co2(PF6)2 (Scheme 1). Electrochemical data (vide supra)
suggested that cobaltocene would be a suitable reducing
agent, and this was indeed found to be the case. Co2
was isolated from solution after all the cobaltocenium
hexafluorophosphate was removed by multiple precipi-
tations; the identity of the compound was confirmed by
dication salt).10b This similarity, when compared with
data for 1,3-bis(3,4,1′,2′,3′,4′,5′-heptamethylcobaltocen-
1-yl)benzene,10b which shows much weaker magnetic
interactions than its para-bridged analogue, suggests
that the principal interactions in Co2 are likely to be
intramolecular in nature (unfortunately the magnetic
couplings, J , were not reported for these dicobaltocenes).
The magnetic coupling in Co2 is somewhat stronger
than in bis(1′,2′,3′,4′,5′-pentamethylcobaltocen-1-yl)di-
methylsilane (J ) -15 cm-1) and bis(1′,2′,3′,4′,5′-pen-
1
high-resolution mass spectrometry. The H NMR spec-
trum of Co2 features broad, shifted peaks, as would be
expected due to the paramagnetic Co(II) centers.
The solid-state magnetic susceptibility of Co2 was
studied using SQUID magnetometry. The derived mag-
netic moment of Co2 approaches ca. 2.4 µB at room
temperature. This value is close both to that expected
tamethylcobaltocen-1-yl)dimethylgermane (-12 cm-1 19
)
but considerably weaker than in bis(fulvalene)dicobalt20
and bis(pentalene)dicobalt,21 which are both diamag-
netic.
To further assess the nature and extent of metal-
metal interactions in this system, cyclic voltammetric
experiments were performed. A reversible wave was
observed at -0.86 V relative to the ferrocene/ferro-
cenium couple in THF, corresponding to the reduction
of both cobalt centers of the dication to the neutral
dicobalt(II) species. This represents a positive shift of
ca. 0.5 V of the potential relative to cobaltocene,
demonstrating the electron-withdrawing nature of the
boratabenzene ligand compared to Cp. A second revers-
ible wave was seen at -2.01 V in THF, which can be
assigned to the overlapping independent reductions of
the two Co centers to CoI: i.e. overlapping Co20/- and
Co2-/2- couples. For comparison, a similar reduction to
the 20-electron species was observed for CoCp2 at -2.60
1
for two independent S ) /2 (µeff ) 60.5) cobalt centers
and to that expected for a high-spin S ) 1 configuration
(µeff ) 80.5); however, the temperature-dependent data
reveal significant deviations from Curie-Weiss behavior
indicative of spin-spin exchange interactions (Figure
2). An excellent fit to the magnetic susceptibility data
was obtained by considering the dinuclear species to
adopt an S ) 0 ground state and an S ) 1 excited state,
using the Bleaney-Bowers equation (eq 2) (with an
additional correction for a Curie impurity)16 with pa-
2NAg2 µB
2
1
ømol
)
(2)
(
)(
)
1 - (1/3)e-2J /kT
3k(T - Θ)
V vs FeCp2 in THF22 and for CoCpCp* at -2.67 V vs
+/0
rameters J ) -28 cm-1, Θ ) -9.6 K, and g ) 1.73. The
g value is somewhat dependent upon how the correction
is made for the Curie impurity but is of similar
magnitude to the isotropic g values of 2.03, 1.77, and
1.8 obtained by EPR for CpCo(C5H5BFc),7 CoCp2,17 and
CoCp*2,18 respectively. The temperature dependence of
the magnetic data for Co2 is broadly similar to that
previously shown for 1,4-bis(3,4,1′,2′,3′,4′,5′-heptameth-
ylcobaltocen-1-yl)benzene (and an isoelectronic dinickel
FeCp2+/0 in EtCN,23 which again demonstrates that Co2
is more readily reduced than the purely cyclopentadi-
enyl systems. The related mononuclear boratabenzene
complex CpCo(C5H5BPh) showed reductions at -0.84
(19) J ones, S. C.; Barlow, S.; O’Hare, D. Manuscript in preparation.
(20) Smart, J . C.; Pinsky, B. L. J . Am. Chem. Soc. 1977, 99, 956-
957.
(21) Katz, T. J .; Acton, N.; McGinnis, J . J . Am. Chem. Soc. 1972,
94, 6205-6206.
(16) Carlin, R. L. Magnetochemistry; Springer-Verlag: Berlin, 1986.
(17) Ammeter, J . H. J . Magn. Reson. 1978, 30, 299-325.
(18) Robbins, J . L.; Edelstein, N.; Spencer, B.; Smart, J . C. J . Am.
Chem. Soc. 1982, 104, 1882-1893.
(22) Bard, A. J .; Garcia, E.; Kukharenko, S.; Strelets, V. V. Inorg.
Chem. 1993, 32, 3528-3531.
(23) Hudeczek, P.; Ko¨hler, F.; Bergerat, P.; Kahn, O. Chem. Eur. J .
1999, 5, 70-78.

A Bimetallic Boratabenzene Cobalt Complex
Organometallics, Vol. 23, No. 16, 2004 3811
and -1.86 V vs FeCp2 in CH3CN.24,25 Thus, introduc-
agent. Addition of 1 equiv of a CoCp2 solution to
Co2(BAr′4)2 gave a highly air-sensitive dark orange-
brown solution (solids were rapidly precipitated on
exposure to air).30 In the visible region, intense absorp-
tions were seen with maxima at 426 and 375 nm. The
reduced solution showed a very weak band in the near-
IR region; accurate measurement of this band was
complicated by its weakness, by solubility consider-
ations, and by overlap with several vibrational features.
Nevertheless, the band maximum, νmax, was estimated
+/0
tion of a p-phenylene bridge between two cyclopenta-
dienyl(boratabenzene) cobalt units appears to have little
effect on the electrochemistry of the species.
The separation, ∆E1/2, between the individual waves
for the Co22+/+ and Co2+/0 couples could not be resolved
using either cyclic voltammetry or differential pulse
voltammetry. The magnitude of ∆E1/2 in such systems
is often indicative of the extent of communication
between the metal centers. On this basis it would
appear, therefore, that metal-metal interactions in Co2
are limited. Such a separation of waves was also not
observed in the purely cyclopentadienyl analogue, p-bis-
(3,4,1′,2′,3′,4′,5′-heptamethylcobaltocen-1-yl)benzene, al-
though two oxidation waves were observed for the
as ca. 6000 cm-1, and the peak molar absorptivity, ꢀmax
,
was determined to be no greater than 25 M-1 cm-1
(correction for disproportionation was made assuming
Kcomp, defined as the equilibrium constant for the
reaction
related iron and nickel derivatives.10b Furthermore, in
the related ferrocene system p-Fc2C6H4
Co22+ + Co2 h 2Co2+
2+/+/0
“the begin-
ning of a splitting” was observed, and a ∆E1/2 value of
0.10 V was reported in chloroform.10a
to have a statistical value of 4, since Kcomp could not be
determined electrochemically due to the unresolved
separation of the Co22+/+ and Co2+/0 couples). The
maximum possible value of the electronic coupling, V,
was estimated to be < ca. 7 meV (60 cm-1) using Hush’s
equation:27
In view of the fact that it has recently been shown
that ∆E1/2 can be highly solvent dependent,26 electro-
chemistry experiments were performed in additional
solvents. CV in CH3CN revealed the CoIII/CoII couples
+/0
as a reversible feature at E1/2 ) -0.82 V vs FeCp2
,
while in CH2Cl2 the corresponding value was -0.80 V.
V ) ((4.5 × 10-4)ꢀmax∆ν1/2νmax
)
1/2/r
The reversible CoII/CoI redox process was also observed
+/0
in CH3CN at -1.87 V vs FeCp2
and at -1.96 V in
where ∆ν1/2 (cm-1) is the width at half-height of the
intervalence band and r (Å) is the intermetallic distance.
∆ν1/2 for Co2+ was estimated by assuming a typical
value for a class II mixed-valence species of ca. 1.3 times
CH2Cl2. Once again the separation between the redox
processes of the individual metal centers was not
observed in either solvent. There was, however, evidence
of imperfect reversibility (Iox < Ired) and additional
features possibly due to decomposition, particularly in
CH3CN, suggesting that the reduced species may be less
stable in these more polar solvents.
the Hush limit of ∆ν1/2 ) (2310νmax)
1/2; r was assumed
to be equal to the Co‚‚‚Co distance in the crystal
structure of Co2(PF6)2: i.e., 9.879 Å.
While several studies of CoII/CoIII and FeII/CoIII met-
allocene systems indicate increased M-M interactions
relative to those of their FeII/FeIII analogues,31 the
We were also interested in probing the electronic
coupling between the two cobalt centers in the mixed-
valence CoII/CoIII derivative Co2+; this can be estimated
by applying Hush theory27 to the near-infrared inter-
valence charge transfer expected in such a mixed-
valence system.28 The CV results described above
suggested that THF might be a suitable solvent for this
experiment, as there was evidence of decomposition in
more polar solvents. Co22+(PF6)2 was insufficiently
soluble in THF, however; thus, the BAr′4 salt Co2(BAr′4)2
(Ar′ ) 3,5-(CF3)2C6H3) was synthesized by metathesis
electronic coupling in Co2+ is clearly rather weak in
+
comparison to the ferrocene analogue, p-Fc2C6H4
,
which shows a coupling of 43 meV,10a at least 6 times
greater than that estimated for Co2+. In cobaltocenium/
cobaltocene systems, the increased coupling has been
attributed to increased ligand character of the frontier
orbitals relative to those of their iron analogues; this
also explains the higher self-exchange rates found for
-
of PF6 with NaBAr′4.29
+/0
+/0
CoCp2
and CoCp*2
relative to those of the analo-
gous iron systems.32 Although the redox centers in Co2
are isoelectronic with those of cobaltocene, perhaps the
inclusion of the boron atoms in the coordinated rings
significantly reduces the metal-metal coupling. On the
other hand, electrochemical data appear to suggest
weaker interactions in the closely analogous 1,4-bis-
(3,4,1′,2′,3′,4′,5′-heptamethylcobaltocen-1-yl)benzene sys-
tem than in 1,4-Fc2C6H4; unfortunately, near-IR data
for the 1,4-bis(3,4,1′,2′,3′,4′,5′-heptamethylcobaltocen-
The mixed-valence species was generated from the
dication by again using cobaltocene as the reducing
(24) Koelle, U. J . Organomet. Chem. 1978, 152, 225-228.
(25) Corrected by assuming that the potential of the Cp2Fe+/0 couple
is +0.40 V vs SCE in CH3CN/0.1 M Bu4N+PF6-. See: Connelly, N. G.;
Geiger, W. E. Chem. Rev. 1996, 96, 877-910.
(26) Barrie`re, F.; Camire, N.; Geiger, W. E.; Mueller-Westerhoff, U.
T.; Sanders, R. J . Am. Chem. Soc. 2002, 124, 7262-7263.
(27) Hush, N. S. Prog. Inorg. Chem. 1967, 8, 391-444.
(28) It should be noted that although the individual redox potentials
for each metal center, ∆E1/2, were not observed in the electrochemical
experiments, this does not necessarily mean that there is no electronic
coupling between the metal centers. See, for example, footnote 8 in
ref 10a and: Barlow, S.; Marder, S. R. Chem. Commun. 2000, 1555-
1562.
(30) The solution was reinvestigated by cyclic voltammetry after
spectroscopic analysis; features due to Co22+/+/0/-/2- and CoCp2+/0 were
observed in the expected ratios, and no additional features were seen,
indicating that the reduced neutral dicobalt species had not signifi-
cantly decomposed during the spectroscopic measurements.
(31) (a) Barlow, S. Inorg. Chem. 2001, 40, 7047-7053. (b) McManis,
G. E.; Nielson, R. M.; Weaver, M. J . Inorg. Chem. 1988, 27, 1827-
1829.
(32) (a) Nielson, R. M.; McManis, G. E.; Golovin, M. N.; Weaver, M.
J . J . Phys. Chem. 1988, 92, 3441-3450. (b) Nielson, R. M.; Govin, M.
N.; McManis, G. E.; Weaver, M. J . J . Am. Chem. Soc. 1988, 110, 1745-
1749.
-
(29) BAr′4 salts of related species have been prepared previously
-
-
by similar metatheses of BF4 or PF6 salts with NaBAr′4. See for
example: (a) Cha´vez, I.; Alvarez-Carena, A.; Molins, E.; Roig, A.;
Maniukiewicz, W.; Arancibia, A.; Arancibia, V.; Brand, H.; Manriquez,
J . M. J . Organomet. Chem. 2000, 601, 126-132. (b) Barlow, S.; Cowley,
A. R.; Green, J . C.; Brunker, T. J .; Hascall, T. Organometallics 2001,
20, 5351-5359. (c) Barlow, S.; Henling, L. M.; Day, M. W.; Schaefer,
W. P.; Green, J . C.; Hascall, T.; Marder, S. R. J . Am. Chem. Soc. 2002,
124, 6285-6296.

3812 Organometallics, Vol. 23, No. 16, 2004
Hascall et al.
which was removed. To the aqueous layer was added a solution
of 1.13 g of NH4PF6 (6.93 mmol) in 20 mL of H2O, giving an
orange precipitate of Co2(PF6)2 which was isolated by filtra-
tion. The product was purified by dissolving in acetone (150
mL), filtering the solution, and adding diethyl ether to give
an orange precipitate, which was dried under vacuum. Yield:
0.87 g (46% based on CoCp2). X-ray quality crystals were grown
by diffusion of Et2O into a saturated CH3CN solution. NMR
1-yl)benzene monocation, which might help resolve the
reasons behind the weak coupling in Co2+, have not
been reported.
In summary, a bimetallic cobalt complex with bo-
ratabenzene ligands linked by a p-phenylene bridge has
been synthesized. Magnetic data for the neutral species,
showing antiferromagnetic interactions between the
metal centers, and electrochemical data, showing no
resolvable separation between the two CoIII/CoII couples,
are similar to those reported for a structurally similar
dicobaltocene system. The near-IR spectrum of the
mixed-valence derivative shows only rather weak metal-
metal interactions.
1
data are as follows. H (acetone-d6): δ 6.06 (s, 10 H [C5H5Co-
(C5H5B)]2C6H4); 6.37 (d, J H-H ) 9 Hz, 4 H, C5H5B); 7.06 (dd,
J H-H ) 6, 9 Hz, 4 H, C5H5B); 7.21 (t, J H-H ) 6 Hz, 2 H, C5H5B);
8.22 (s, 4 H, [C5H5Co(C5H5B)]2C6H4). 13C{1H} (CD3CN): δ 88.3
([C5H5Co(C5H5B)]2C6H4); 90.7 ([C5H5Co(C5H5B)]2C6H4); 95.6
([C5H5Co(C5H5B)]2C6H4); 108.2 ([C5H5Co(C5H5B)]2C6H4); 133.7
([C5H5Co(C5H5B)]2C6H4); 139.2 (B-C of [C5H5Co(C5H5B)]2C6H4).
11B (CD3CN): δ 25.0 (br s). Anal. Calcd for C26H24B2Co2P2F12
:
Exp er im en ta l Deta ils
C, 40.8; H, 3.2. Found: C, 41.0; H, 3.2.
Syn t h esis of {[Cp Co(C5H 5B)]2C6H 4}2+(BAr ′4
-
)
(Co2-
2
Gen er a l Con sid er a tion s. All manipulations were per-
formed under an inert atmosphere of N2 using a combination
of drybox and Schlenk techniques, except for the oxidation of
Co2 and workup of its PF6- salt, which could be performed in
(BAr ′4)2). CH2Cl2 (25 mL) was added to a mixture of 0.14 g of
Co2(PF6)2 (0.18 mmol) and 0.33 g of NaBAr′4 (0.37 mmol). The
mixture was gently warmed (35 °C) and stirred for 90 min,
before filtering through Celite to remove the fine white
precipitate that had formed. The CH2Cl2 was removed in
vacuo, the orange solid was extracted with ether (2 × 15 mL),
and the extracts were filtered through Celite. The solvent was
removed from the resulting clear orange solution under
reduced pressure to give an orange solid. Yield: 0.34 g (84%
based on Co2(PF6)2). NMR data are as follows. 1H (CD2Cl2):
δ 5.61 (s, 10 H, [C5H5Co(C5H5B)]2C6H4), 6.14 (d, J H-H ) 9 Hz,
4 H, C5H5B), 6.72 (dd, J H-H ) 6, 9 Hz, 4 H, C5H5B), 6.85 (t,
J H-H ) 6 Hz, 2 H, C5H5B), 7.56 (s, 8 H, B[C6H3(CF3)2]4), 7.71
(m, 16 H, B[C6H3(CF3)2]4), 8.09 (s, 4 H, [C5H5Co(C5H5B)]2C6H4).
Anal. Calcd for C90H48B4Co2F48: C, 49.1; H, 2.2. Found: C,
49.0; H, 2.1.
9
33
air. 1,4-(BBr2)2C6H4 and NaBAr′4 were synthesized by
following literature procedures. NMR spectra were recorded
using a Varian Unity Plus 500 MHz spectrometer or a Varian
Mercury VX-Works 300 MHz spectrometer. 1H and 13C chemi-
cal shifts (δ) are quoted in ppm relative to tetramethylsilane
and were referenced via the residual protio solvent peak (1H)
or a 13C resonance of the solvent. 11B spectra were referenced
using BF3‚OEt2 as an external standard.
Variable-temperature magnetic susceptibility data were
acquired in a field of 0.1 T using a MPMS-5 SQUID magne-
tometer, with the sample loaded in a gelatin capsule mounted
between other gelatin capsules in a nonmagnetic plastic straw.
The data were corrected for sample diamagnetism using
Pascal’s constants.34 The molar susceptibility data were fitted
to the Bleaney-Bowers equation16 with additional terms to
account for a paramagnetic impurity (ø ) C/(T - Θ)) and for
temperature-independent paramagnetism. Without knowledge
of the molecular weight and spin state of the Curie paramag-
netic impurity, it is difficult to quantify the exact level present;
however, if for example the impurity is assumed to have S )
3/2 (typical for inorganic Co(II) species) and to have a molecular
weight similar to that of Co2, it would correspond to ca. 6%
of the sample.
Electrochemical measurements were carried out under
argon, with solutions in deoxygenated, dry solvents ca. 10-4
M in analyte (Co2(BAr′4)2 for experiments in THF and CH2Cl2;
Co2(PF6)2 for experiments in CH3CN) and 0.1 M in [nBu4N]-
[PF6], using a BAS or CH Instruments CHI600A potentiostat,
a glassy-carbon working electrode, a platinum auxiliary elec-
trode, and a Pt wire, or a AgCl-coated Ag wire, as a pseudo-
reference electrode. Potentials were referenced to Cp2Fe+/0 by
adding ferrocene to the cell. The reversibility of a redox couple
was judged by comparison with the behavior of the Cp2Fe+/0
couple under the same conditions. Near-IR measurements
were conducted using a Varian Cary 5 spectrometer.
Syn th esis of {[CpCo(C5H5B)]2C6H4}2+(P F6-)2 (Co2(P F6)2).
A solution of 1.13 g of p-(BBr2)2C6H4 (2.71 mmol) in ca. 40 mL
of toluene was added dropwise to a solution of 2.80 g of
cobaltocene (14.8 mmol) in ca. 60 mL of toluene at -78 °C. A
light brown precipitate began to form immediately. The
reaction mixture was allowed to return to room temperature
slowly and then stirred for 1 h. The mixture was filtered and
the solvent removed from the red filtrate under vacuum. To
the residue was added 50 mL of ether, followed by a solution
of 1.69 g of FeCl3‚6H2O (6.25 mmol) in 50 mL of H2O. This
yielded an orange aqueous layer and a yellow organic layer,
Syn th esis of [Cp Co(C5H5B)]2C6H4 (Co2). A solution of
0.12 g of CoCp2 (0.63 mmol) in 10 mL of CH2Cl2 was added
dropwise at room temperature to a slurry of 0.25 g of Co2(PF6)2
(0.33 mmol) in 15 mL of CH2Cl2. After the mixture was stirred
overnight, the Co2(PF6)2 had completely dissolved and a pale
yellow precipitate ([CoCp2]PF6) had formed. The precipitate
was removed by filtration and washed with CH2Cl2, and the
washings were combined with the filtrate. Ether was added
to the solution, and the remaining [CoCp2]PF6 that precipitated
was removed by filtration and washed with ether. The solvent
was removed from the red-orange solution to give a brown
powder, which was dried under vacuum. Yield: 0.11 g (73%
based on CoCp2). Because of its highly air-sensitive nature,
we were unable to obtain a satisfactory elemental analysis for
1
this complex. NMR data are as follows. H (CD2Cl2): δ -61.4
(10 H), -47.7 (4 H), 12.7 (4 H), 15.8 (4 H), 30.2 (2 H). UV-vis
(CH2Cl2): λmax (ꢀmax) 485 (1300), 358 (5000), 318 (8500), 274
(18 000) nm (M-1 cm-1). MS (EI): m/z 476 (M+, 100%), 411
(M+ - Cp, 5%), 238 (M2+, 4%), 189 ([Cp2Co]+, 10%). MS (ES):
m/z 476 (M+, 3%), 238 (M2+, 100%). High-resolution MS (EI):
calcd for C26H24B2Co2, 476.0728; found, 476.0746. Anal. Calcd
for C26H24B2Co2: C, 65.6; H, 5.1; Co, 24.8. Found: C, 62.3; H,
4.7; Co, 20.8.
X-r a y Cr ysta llogr a p h y Stu d y on Co2(P F 6)2. A single
crystal (red-brown block, ca. 0.08 × 0.18 × 0.18 mm) was
mounted on a glass fiber using perfluoropolyether oil and
cooled from ambient temperature to 150 K at a rate of 120
K/h in a stream of cold N2 using an Oxford Cryosystems
CRYOSTREAM unit. Diffraction data were measured using
an Enraf-Nonius KappaCCD diffractometer (graphite-mono-
chromated Mo KR radiation, λ ) 0.710 73 Å). Intensity data
were processed using the DENZO-SMN package.35 A total of
9742 reflections were measured (3192 unique). Crystal data:
(33) Brookhart, M.; Grant, B.; Volpe, A. F. Organometallics 1992,
11, 3920-3922.
(34) O’Conner, C. J . Prog. Inorg. Chem. 1985, 29, 203-283.
(35) Otwinowski, Z.; Minor, W. Processing of X-ray Diffraction Data
Collected in Oscillation Mode, In Methods in Enzymology; Carter, C.
W., Sweet, R. M., Eds.; Academic Press: New York, 1997; Vol. 276.

A Bimetallic Boratabenzene Cobalt Complex
Organometallics, Vol. 23, No. 16, 2004 3813
ment converged satisfactorily to give R ) 0.0397 and Rw
triclinic, a ) 7.0697(2) Å, b ) 9.9528(3) Å, c ) 10.4818(3) Å, R
) 88.1016(12)°, â ) 78.4576(12)°, γ ) 77.1226(14)°, V ) 704.4
Å3.
)
0.0401 for 2481 observed reflections (I > 3σ(I)) and 247 refined
parameters: C26H24B2Co2F12P2, Z ) 1, Fcalcd ) 1.805 Mg/m3, µ
) 1.389 mm-1
.
The structure was solved in the space group P1h (No. 2) using
the direct-methods program SIR92,36 which located all non-
hydrogen atoms. Subsequent full-matrix least-squares refine-
ment was carried out using the CRYSTALS program suite.37
Coordinates and anisotropic thermal parameters of all non-
hydrogen atoms were refined. Hydrogen atoms were positioned
geometrically after each cycle of refinement. A three-term
Chebychev polynomial weighting scheme was applied. Refine-
Ack n ow led gm en t. We thank Dr. Simon J ones for
helpful discussions and Balliol College, Oxford (V.B.)
and the EPSRC for financial support.
Su p p or tin g In for m a tion Ava ila ble: Additional data
from the X-ray crystal structure determination of Co2(PF6)2.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(36) Altomare, A.; Cascarano, G.; Giacovazzo, G.; Guagliardi, A.;
Burla, M. C.; Polidori, G.; Camalli, M. J . Appl. Crystallogr. 1994, 27,
435.
(37) Watkin, D. J .; Prout, C. K.; Carruthers, J . R.; Betteridge, P.
W.; Cooper, R. I. CRYSTALS issue 11; Chemical Crystallography
Laboratory, Oxford, U.K., 2001.
OM049750Y