New noncentered icosahedral nickel-bismuth carbonyl clusters: Geometric analysis of the homologous [Ni10(EMe)2(CO)18]2- dianions (E = P, As, Sb, Bi) containing noncentered 1,12-Ni10E2 icosahedral cages

Article abstract of DOI:10.1021/om960965k

The designed reaction of [Ni₆(CO)₁₂]^2- (1) with MeBiCl₂ in THF at low temperature (-78 °C) has given rise to the desired [Ni₁₀(BiMe)₂(CO)₁₈]^2- dianion (2) as the major product (ca. 45% yield). This sought-after (methylbismuthinidene)nickel carbonyl cluster containing a noncentered 1,12-Ni₁₀Bi₂ icosahedral cage is the missing member of the known homologous [Ni₁₀(EMe)₂(CO)₁₈]^2- series (E = P, As, Sb) whose members differ from one another only by the substitution of two bridging E atoms of the congeneric main-group V (15) elements into the two centrosymmetrically related (opposite) vertices of the icosahedral cage. The formulation of 2 in the [NMe₄]⁺ salt was established unambiguously from X-ray crystallography and ascertained from a complete elemental analysis. This air-sensitive brown compound was also characterized by IR, cyclic voltammetry, negative-ion ESI mass spectrometry, and low-temperature ¹³C{¹H} and ¹H NMR spectroscopy. The [NMe₄]⁺ salt of the ethylbismuthinidene [Ni₁₀(BiEt)₂(CO)₁₈]^2- (3) was similarly prepared, isolated, and structurally determined via a single-crystal X-ray diffraction study. Solid-state magnetic susceptibility measurements conclusively showed 2 (in the [NMe₄]⁺ salt) to be diamagnetic at room temperature in accordance with it (and the other homologous members) being electronically equivalent (isolobal) with the classic regular icosahedral [B₁₂H₁₂]^2- dianion in likewise possessing 13 skeletal electron pairs. A detailed comparison of the essentially identical molecular parameters of 2 and 3 with those of the three other isostructural [Ni₁₀(EMe)₂(CO)₁₈]^2- clusters (E = P, As, Sb) provides illuminating information regarding particular geometric changes in the empty icosahedral 1,12-Ni₁₀E₂ cages arising from a systematic size variation in the two E atoms of the four congeneric group V (15) elements. Moreover, a comparative analysis of the molecular parameters of 2 with those of the previously reported [Ni₁₁Bi₂(CO)₁₈]^3- trianion (4), which contains a Ni-centered (filled) 1,-12-Ni₁₀Bi₂ icosahedral cage, reveals marked geometric distortions upon the formal insertion of a centered (interior) Ni(i) into the icosahedral cage. Variable-temperature ¹³C{¹H} NMR spectra of 2 in acetone-d₆ show fast terminal-bridge carbonyl exchange of the 18 COs at room temperature but limiting slow exchange at low temperature (-60 °C), as was also found for [Ni₁₀(EMe)₂(CO)₁₈]^2- (E = P, Sb) previously.

Full text of DOI:10.1021/om960965k

Organometallics 1997, 16, 1655-1667
1655
New Non cen ter ed Icosa h ed r a l Nick el-Bism u th Ca r bon yl
Clu ster s: Geom etr ic An a lysis of th e Hom ologou s
[Ni10(EMe)2(CO)18]2- Dia n ion s (E ) P , As, Sb, Bi)
Con ta in in g Non cen ter ed 1,12-Ni10E2 Icosa h ed r a l Ca ges1
Peter D. Mlynek and Lawrence F. Dahl*
Department of Chemistry, University of WisconsinsMadison, Madison, Wisconsin 53706
Received November 15, 1996X
The designed reaction of [Ni6(CO)12]2- (1) with MeBiCl2 in THF at low temperature (-78
°C) has given rise to the desired [Ni10(BiMe)2(CO)18]2- dianion (2) as the major product (ca.
45% yield). This sought-after (methylbismuthinidene)nickel carbonyl cluster containing a
noncentered 1,12-Ni10Bi2 icosahedral cage is the missing member of the known homologous
[Ni10(EMe)2(CO)18]2- series (E ) P, As, Sb) whose members differ from one another only by
the substitution of two bridging E atoms of the congeneric main-group V (15) elements into
the two centrosymmetrically related (opposite) vertices of the icosahedral cage. The
formulation of 2 in the [NMe4]+ salt was established unambiguously from X-ray crystal-
lography and ascertained from a complete elemental analysis. This air-sensitive brown
compound was also characterized by IR, cyclic voltammetry, negative-ion ESI mass
1
spectrometry, and low-temperature 13C{1H} and H NMR spectroscopy. The [NMe4]+ salt
of the ethylbismuthinidene [Ni10(BiEt)2(CO)18]2- (3) was similarly prepared, isolated, and
structurally determined via a single-crystal X-ray diffraction study. Solid-state magnetic
susceptibility measurements conclusively showed 2 (in the [NMe4]+ salt) to be diamagnetic
at room temperature in accordance with it (and the other homologous members) being
electronically equivalent (isolobal) with the classic regular icosahedral [B12H12]2- dianion in
likewise possessing 13 skeletal electron pairs. A detailed comparison of the essentially
identical molecular parameters of 2 and 3 with those of the three other isostructural
[Ni10(EMe)2(CO)18]2- clusters (E ) P, As, Sb) provides illuminating information regarding
particular geometric changes in the empty icosahedral 1,12-Ni10E2 cages arising from a
systematic size variation in the two E atoms of the four congeneric group V (15) elements.
Moreover, a comparative analysis of the molecular parameters of 2 with those of the
previously reported [Ni11Bi2(CO)18]3- trianion (4), which contains a Ni-centered (filled) 1,-
12-Ni10Bi2 icosahedral cage, reveals marked geometric distortions upon the formal insertion
of a centered (interior) Ni(i) into the icosahedral cage. Variable-temperature 13C{1H} NMR
spectra of 2 in acetone-d6 show fast terminal-bridge carbonyl exchange of the 18 COs at
room temperature but limiting slow exchange at low temperature (-60 °C), as was also
found for [Ni10(EMe)2(CO)18]2- (E ) P, Sb) previously.
In tr od u ction
The general synthetic route that has given rise to an
astonishing array of large nickel/main-group carbonyl
clusters has involved reactions of the [Ni6(CO)12]2-
dianion (1)6 with monomeric main-group reagents.
Products isolated and structurally analyzed from X-ray
crystallographic studies include ones with NixEy cores,
where E denotes main-group bridging atoms such as
indium,7 germanium,8,9 tin,10 phosphorus,11-14 arsenic,15
Over the past 10 years there has been much progress
made in transition-metal cluster chemistry, especially
in the preparation and physicochemical characterization
of high-nuclearity metal carbonyl clusters. One par-
ticularly active area in our laboratories has involved the
preparation and physical/chemical characterization of
close-packed nickel carbonyl clusters including ones
with bridging main-group IV (14), V (15), or VI (16)
atoms as well as giant-sized bimetallic nickel clusters
with rhodium,2 palladium,3 gold,4 and copper.5
(3) Kawano, M.; Bacon, J . W.; Campana, C. F.; Dahl, L. F. J . Am.
Chem. Soc. 1996, 7869-7870.
(4) (a) Whoolery, A. J .; Dahl, L. F. J . Am. Chem. Soc. 1991, 113,
6683-6685. (b) Whoolery J ohnson, A. J .; Spencer, B.; Dahl, L. F. Inorg.
Chim. Acta 1994, 227, 269-283.
(5) Mlynek, P. D. Ph.D. Thesis, University of WisconsinsMadison,
1996.
* To whom correspondence should be addressed. E-mail: dahl@
chem.wisc.edu. Fax: +608-262-0381.
(6) (a) Calabrese, J . C.; Dahl, L. F.; Cavalieri, A.; Chini, P.; Longoni,
G.; Martinengo, S. J . Am. Chem. Soc. 1974, 96, 2616-2618. (b)
Longoni, G.; Chini, P.; Cavalieri, A. Inorg. Chem. 1976, 15, 3025-3029.
(7) Demartin, F.; Iapalucci, M. C.; Longoni, G. Inorg. Chem. 1993,
32, 5536-5543.
(8) Ceriotti, A.; Demartin, F.; Heaton, B. T.; Ingallina, P.; Longoni,
G.; Manassero, M.; Marchionna, M.; Masciocchi, N. J . Chem. Soc.,
Chem. Commun. 1989 786-787.
(9) Zebrowski, J . P.; Hayashi, R. K.; Bjarnason, A.; Dahl, L. F. J .
Am. Chem. Soc. 1992, 114, 3121-3123.
X Abstract published in Advance ACS Abstracts, March 15, 1997.
(1) (a) Presented in part at the 208th National Meeting of the
American Chemical Society, Washington, DC, Aug 24, 1994. See:
Mlynek, P. D.; Dahl, L. F. Abstracts of Papers; National Meeting of
the Americal Chemical Society; American Chemical Society: Wash-
ington, DC, 1994; INOR 561. (b) Based in part on the Ph.D. thesis of
P.D.M. at the University of WisconsinsMadison, 1996.
(2) Nagaki, D. A.; Badding, J . V.; Stacy, A. M.; Dahl, L. F. J . Am.
Chem. Soc. 1986, 108, 3825-3827.
S0276-7333(96)00965-X CCC: $14.00 © 1997 American Chemical Society

1656 Organometallics, Vol. 16, No. 8, 1997
Mlynek and Dahl
antimony,16-19 bismuth,20 sulfur,21 selenium,22 or tel-
lurium;22 each of the main-group bridging atoms is
either “bare” or ligated by alkyl or aryl substituents.
Although these results indicate a rich diversity of large
nickel carbonyl clusters with different main-group
atoms, much experimental/theoretical work still needs
to be carried out. In addition to the preparation of
particular clusters in order to provide operational tests
of structural/bonding hypotheses, reliable synthetic
methodology must continue to be developed in order to
obtain large metal clusters in sufficient quantities that
their physical properties in both solid state and solution
can be measured and their chemical reactivities ex-
plored.
Initial attempts in our laboratories eight years ago
to obtain the p-tolylbismuthinidene [Ni10(BiR)2(CO)18]2-
dianion (R ) p-MeC6H4) containing a noncentered 1,12-
Ni10Bi2 icosahedral cage by reaction of the [NMe3Ph]+
salt of 1 with (p-tolyl)BiBr2 in THF at -78 °C were
unsuccessful; no carbonyl-containing Ni-Bi products
were isolated.16 On the basis of observed geometric
trends in the known [Ni10(ER)2(CO)18]2- dianions (R )
Me13 for E ) P; R ) Me15 for E ) As; R ) Ph16 for E )
Sb) with empty icosahedral 1,12-Ni10E2 cages, it was
then concluded that the intrapentagonal Ni-Ni′ dis-
tances, whose mean values uniformly increase by 0.1 Å
from E ) P to E ) As and by 0.1 Å from E ) As to E )
Sb, would be too long for E ) Bi to support reasonable
intrapentagonal Ni-Ni′ bonding interactions. However,
Longoni and co-workers20 subsequently reported the
synthesis of the [Ni11Bi2(CO)18]n- anions (n ) 2, 3) as
[NEt4]+ salts by reaction of 1 with BiCl3 in THF at room
temperature; the corresponding tetraanion (n ) 4) was
also obtained in solution by chemical or electrochemical
reduction of the di- or trianion. In addition to CV
measurements that showed reversible electron-transfer
behavior among the three anions (n ) 2-4), an X-ray
crystallographic study of [NEt4]3[Ni11Bi2(CO)18] revealed
that the trianion possesses a Ni-centered 1,12-Ni10Bi2
icosahedral cage that is geometrically analogous to the
Ni-centered 1,12-Ni10Sb2 cage found in the [Ni11-
{SbNi(CO)3}2(CO)18]n- anions (n ) 2-4); these latter
anions were previously reported by Longoni and co-
workers18 at the same time as our synthetic/structural
results16 on the related [Ni10(SbPh)2(CO)18]2- dianion
with the noncentered 1,12-Ni10Sb2 cage. The lack of
residual basicity of the lone pair on each Bi atom in the
[Ni11Bi2(CO)18]n- (n ) 2-4) anions was evident by the
tetraanion being oxidized to the dianion instead of either
being protonated upon treatment with protonic acids or
methylated upon reaction with methyl iodide.20 In
contrast, the closely related Ni-centered [Ni11{SbNi-
(CO)3}2(CO)18]4- tetraanion was found to react with MeI
to give the Ni-centered [Ni11(SbMe)2(CO)18]2-; however,
an attempt to remove the two Sb-attached Ni(CO)3
substituents by reaction with CO was unsuccessful.18
The proven existence of the Ni-centered [Ni11Bi2-
(CO)18]n- anions (n ) 2-4)20 inspired us to design an
appropriate synthetic route to prepare the missing
bismuth member (2) of the aforementioned [Ni10-
(EMe)2(CO)18]2- series (E ) P,13 As,15 Sb19) for a
comparative geometrical analysis with the other three
members. Of particular interest was whether 2 could
be prepared or whether the centered (interior) nickel
atom was a necessary ingredient in order to stabilize
the 1,12-Ni10Bi2 icosahedral cage. If successful, we also
wanted to determine the salient geometrical differences
between the desired [Ni10(BiMe)2(CO)18]2- (2) and the
crystallographically characterized [Ni11Bi2(CO)18]3- tri-
anion20 in order to assess structural/bonding effects due
to the formal insertion of the Ni-centered atom into the
empty icosahedral Ni10Bi2 cage of 2.
On the basis of the previous optimization of the
synthetic procedure used in isolating the methylstib-
inidene Ni10Sb2 cluster,19 a similar reaction led to the
isolation of 2 as the major product. Analogous synthetic
conditions also gave rise to the ethylbismuthinidene 3.
The results presented herein have fulfilled our goals.
However, our repeated attempts to prepare the corre-
sponding arylbismuthinidene Ni10Bi2 cluster were again
unsuccessful.
Resu lts a n d Discu ssion
Syn th eses of Tr ia lk ylbism u th , Alk ylbism u th Di-
h a lid e, a n d Dia lk ylbism u th Ha lid e Mon om er s. The
trimethylbismuth intermediate was prepared via a
modified procedure of that reported by Scherer et al.,23
who in turn had altered the original Scha¨fer-Hein
synthesis24 of this compound. We observed that reac-
tions of the methyl Grignard reagent with bismuth
tribromide instead of bismuth trichloride gave better
results in obtaining trimethylbismuth. In order to
obtain MeBiX2 (X ) Cl, Br), metathesis of trimethyl-
bismuth with 2 equiv of bismuth trihalide was found to
work adequately; this procedure is similar to that used
to obtain the MeSbX2 analogues (X ) Cl, Br).25 When
the reactant ratios were reversed, both dimethylbismuth
chloride and bromide were produced in almost 100%
yields. This synthetic pathway to dimethylbismuth
bromide not only was found to be easier to perform but
also gave a higher yield (54% vs 37%) than that obtained
from an alternate route, developed by Marquardt,26
involving the formation of dimethylbismuth bromide
(10) Zebrowski, J . P.; Hayashi, R. K.; Dahl, L. F. J . Am. Chem. Soc.
1993, 115, 1142-1144.
(11) (a) Lower, L. D.; Dahl, L. F. J . Am. Chem. Soc. 1976, 98, 5046-
5047. (b) Arif, A. M.; J ones, R. A.; Schwab, S. T. J . Coord. Chem. 1987,
16, 51-57. (c) Rieck, D. F.; Rae, A. D.; Dahl, L. F. J . Chem. Soc., Chem.
Commun. 1993, 585-586. (d) Montag, R. A. Ph.D. Thesis, University
of WisconsinsMadison, 1982.
(12) Olmsted, N. M.; Power, P. P. J . Am. Chem. Soc. 1984, 106,
1495-1496.
(13) Rieck, D. F.; Gavney, J . A., J r.; Norman, R. L.; Hayashi, R. K.;
Dahl, L. F. J . Am. Chem. Soc. 1992, 114, 10369-10379.
(14) Gavney, J . A., J r.; Hayashi, R. K.; Underiner, T. L.; Safford, L.
K.; Roth, J . D.; Bjarnason, A.; Weaver, M. J .; Dahl, L. F. To be
published.
(15) Rieck, D. F.; Montag, R. A.; McKechnie, T. S.; Dahl, L. F. J .
Am. Chem. Soc. 1986, 108, 1330-1331.
(16) DesEnfants, R. E., II; Gavney, J . A., J r.; Hayashi, R. K.; Rae,
A. D.; Dahl, L. F.; Bjarnason, A. J . Organomet. Chem. 1990, 383, 543-
572.
(17) Albano, V. G.; Demartin, F.; Iapalucci, M. C.; Longoni, G.;
Sironi, A.; Zanotti, V. J . Chem. Soc., Chem. Commun. 1990, 547-548.
(18) Albano, V. G.; Demartin, F.; Iapalucci, M. C.; Laschi, F.;
Longoni, G.; Sironi, A.; Zanello, P. J . Chem. Soc., Dalton Trans. 1991,
739-748.
(19) Mlynek, P. D.; Dahl, L. F. Organometallics 1997, 16, 1641.
(20) Albano, V. G.; Demartin, F.; Iapalucci, M. C.; Longoni, G.;
Monari, M.; Zanello, P. J . Chem. Soc., Dalton Trans. 1992, 497-502.
(21) Kahaian, A. J . Ph.D. Thesis, University of WisconsinsMadison,
1992.
(23) Scherer, O. J .; Hornig, P.; Schmidt, M. J . Organomet. Chem.
1966, 6, 259-264.
(24) Scha¨fer, K.; Hein, F. Z. Anorg. Allg. Chem. 1917, 100, 297-
302.
(25) Nunn, M.; Sowerby, D. B.; Wesolek, D. M. J . Organomet. Chem.
1983, 251, C45-C46.
(26) Marquardt, A. Ber. Deut. Chem. Ges. 1887, 20, 1516-1523.
(22) Kahaian, A. J .; Thoden, J . B.; Dahl, L. F. J . Chem. Soc., Chem.
Commun. 1992, 353-355.

Nickel-Bismuth Carbonyl Clusters
Organometallics, Vol. 16, No. 8, 1997 1657
from the spontaneous decomposition of trimethylbis-
muth dibromide which was synthesized by the bromi-
nation of trimethylbismuth.
contain a Ni-centered icosahedral Ni10Bi2 cage with a
lone electron-pair on each Bi atom20 instead of a possible
Bi-attached one-electron-donating halogen substituent.
Another indicated requirement of this reaction is that
the bismuth reagent contain a halide ligand. Our
attempted reactions of several trialkylbismuth reagents
with 1 under various conditions did not generate any
nickel-bismuth carbonyl clusters. The apparent driv-
ing force behind the formation of nickel-E(15) carbonyl
clusters from reactions of various monomeric E(15)
reagents with 1 is the formation of nickel(II) halide as
a decomposition byproduct.
Whereas reactions of aryl-containing phosphorus,
arsenic, and antimony halide reagents with 1 have
resulted in the formation of Ni-E carbonyl clusters (E
) P, As, Sb) with E-attached aryl substituents, prior
attempts16 to prepare the analogous Ni10Bi2 carbonyl
cluster with Bi-attached p-tolyl substituents and our
recent attempts to synthesize the same hypothetical
cluster with phenyl substituents both failed. Instead,
reactions of 2 equiv of either (p-MeC6H4)BiCl2 or Ph-
BiCl2 with 1 in THF at low temperature (-78 °C)
produced dark insoluble material and Ni(CO)4. Note-
worthy is that the mean bond dissociation energy of a
Bi-R fragment (R ) Me or Ph) is markedly smaller than
that of the same Sb-R fragment (i.e., in general, the
mean bond dissociation energy, Dh (E-R), decreases upon
going from E ) P to E ) Bi).30 Nonetheless, this
inability to isolate an aryl-substituted Ni10Bi2 carbonyl
cluster is perplexing in light of the fact that the mean
bond dissociation energy for Bi-Me compounds is
significantly less than that for Bi-Ph compounds.30 One
possible rationalization is that even though the Bi-Ph
bond is stronger than a Bi-Me bond, it is still too weak
in the hypothetical [Ni10(BiPh)2(CO)18]2- dianion to
overcome the energetic effects caused by increased steric
crowding of the ligand polyhedron about the icosahedral
1,12-Ni10Bi2 cage.
Ethylbismuth dichloride was prepared in a similar
manner to that of the methylbismuth dichloride; the
only significant difference was that higher temperatures
were needed to distill the triethylbismuth intermediate.
As observed with other trialkylpnicogens, caution must
also be used when handling the intermediate trialkyl-
bismuth reagents.27
Syn th eses of th e [Ni10(BiR)2(CO)18
2-
]
Dia n ion s
(R ) Me (2), Et (3)). The preparations of 2 and 3 were
achieved by reactions of 1 with 2 equiv of the appropri-
ate alkylbismuth dichloride at -78 °C in THF, followed
by solvent removal and subsequent extractions. Similar
preparations of the homologous phosphorus and arsenic
members, [Ni10(ER)2(CO)18]2- (E ) P, As), also gave rise
to other intriguing Ni-E carbonyl byproducts including
the electronically equivalent (isolobal) [Ni9(ER)3(CO)15]2-
and [Ni8(ER)4(CO)12]2- dianions (E ) P,13 As15,28) con-
taining empty icosahedral 1,2,12-Ni9E3 and 1,2,9,12-
Ni8E4 cages, respectively. On the other hand, aside from
decomposition products, reaction of either the methyl-
or ethylbismuthine reagent with 1 only produced the
corresponding [Ni10(BiR)2(CO)18]2- dianion (R ) Me (2),
Et (3)) with the empty 1,12-Ni10Bi2 icosahedral cage.
Similarly, only the homologous [Ni10(SbR)2(CO)18]2-
i
t
dianions (R ) Me, Et, Pr, Bu, Ph, p-FC6H4) were
isolated from reactions of stibine reagents with 1, in
spite of considerable efforts16,19 to detect the possible
existence of the hypothetical [Ni9(SbR)3(CO)15]2- and
[Ni8(SbR)4(CO)12]2- dianions.
In contrast to icosahedral clusters of the other con-
generic group V (15) elements being obtained from
reactions performed at room temperature, low-temper-
ature reactions must be carried out in order to obtain 2
and 3; at room temperature only black insoluble pow-
ders were isolated. Similar reactions attempted in
acetonitrile at -40 °C also did not produce any CO-
containing clusters.
Syntheses of the analogous [Ni10(SbR)2(CO)18]2- di-
anions (R ) alkyl, Ph) were accomplished by reactions
of 1 with either R2SbX or RSbX2.16,19 However, we were
not successful in obtaining any detectable quantities of
2 or 3 or any other Ni-Bi carbonyl cluster under similar
conditions with either monoalkylbismuth dibromide or
dialkylbismuth bromide or chloride reagents. It is
possible that the light-sensitive dialkylbismuth chloride
or bromide29 decomposed preferentially to other unre-
active products instead of reacting under our boundary
conditions with 1. Hence, the only known synthetic
pathway (to date) in isolating 2 or 3 involves the
reaction of the appropriate monoalkylbismuth dichloride
with 1.
St r u ct u r a l/Bon d in g F ea t u r es of [Ni10(BiR )2-
(CO)18]2- Dia n ion s (R ) Me (2), Et (3)). Each of these
isostructural dianions, 2 and 3, can be viewed as a
pentagonal antiprism of 10 surface Ni(s) atoms that is
capped on the upper and lower pentagonal faces by
bismuth atoms (Figure 1b). The resulting Ni10Bi2 icosa-
hedral cage in 2 and 3 is enveloped by essentially
identical arrangements of the 2 Bi-attached alkyl sub-
stituents and 18 carbonyl ligands. Each dianion lies on
a crystallographic inversion center such that one-half
of the icosahedral cage corresponding to a Bi-capped
nickel pentagon constitutes the crystallographically
independent unit. Thus, of the 10 Ni(s)-E, 10 intrap-
entagonal Ni(s)-Ni(s′), and 10 interpentagonal Ni(s)-
Ni(s′) edges in each icosahedral 1,12-Ni10Bi2 cage, only
5 edges of each kind are independent. As expected for
an empty Ni10Bi2 cage, each dianion is greatly deformed
from a regular spherical icosahedron of Ih symmetry (as
It is necessary that the bismuth reagent has one
attached alkyl group in order that each of the two
bismuth atoms in the final Ni10Bi2 carbonyl cluster
possess a one-electron-donating substituent; reactions
by Longoni and co-workers20 of bismuth trihalide with
1 gave rise to [Ni11Bi2(CO)18]n- anions (n ) 2, 3) that
in [B12H12]
2-) by a large oblate compression along the
pseudo-5-fold axis passing through Bi and Bi′. Although
the average overall geometry of the empty icosahedral
Ni10
Bi2 cage conforms to D5d symmetry, inclusion of the
18 carbonyl ligands reduces the symmetry of both 2 and
(27) Me3Bi is a very active poison when contacted by inhalation,
orally, or though the intact skin. For a review of pharmacology of
trimethylbismuth, see: Sollmann, T.; Seifter, J . J . Pharmacol. 1939,
67, 17-49.
(30) (a) Dh (Bi-Me) ) 36 kcal/mol, Dh (Bi-Ph) ) 48 kcal/mol: Skinner,
H. A.; Conner, J . A. Pure Appl. Chem. 1985, 57, 79-88. (b) Dh (Bi-Me)
for BiMe3 ) 34.1 ( 1.1 kcal/mol, Dh (Bi-Ph) for BiPh3 ) 42.1 kcal/mol:
Wieber, M. Bismuth-Organische Verbindungen. In Gmelin Handbuch
der Anorganischen Chemie, 8th ed.; Springer-Verlag: Berlin, 1977; Vol.
47, pp 47, 68.
(28) Whoolery, A. J. Ph.D. Thesis, University of WisconsinsMadison,
1993.
(29) Wieber, M.; Sauer, I. Z. Naturforsch. 1984, 39b, 887-889.

1658 Organometallics, Vol. 16, No. 8, 1997
Mlynek and Dahl
F igu r e 2. Configuration of the [Ni10(BiEt)2(CO)18]2- di-
anion (3) in the [NMe4]+ salt.
closed triangulated polyhedron) should possess (n + 1)
electron pairs for skeletal cluster bonding. The 12-
vertex icosahedral Ni10Bi2 cage should therefore have
13 skeletal electron pairs involved in globally delocalized
bonding. The PSEP model assumes that only three
internal valence orbitals at each vertex are involved in
skeletal cluster bonding. An observed electron-pair
count of 13 for each of the icosahedral Ni10Bi2 cages in
2 and 3 results from each Ni(CO), BiR, and CO donating
zero, two, and one skeletal electron pairs, respectively,
with the remaining skeletal electron pair being provided
by the -2 charge of the dianion (i.e., 10 × 0(NiCO) + 2
× 2(BiR) + 8 × 1(CO) + 1(charge) ) 13). The other
electronically equivalent (isolobal) members of the
homologous [Ni10(ER)2(CO)18]2- series (E ) P, As, Sb,
F igu r e 1. (a) Configuration of the [Ni10(BiMe)2(CO)18]2-
dianion (2) in the [NMe4]+ salt. This dianion, which
possesses 10 terminal, 4 doubly bridging, and 4 triply
bridging carbonyl ligands, has crystallographic Ci-1h site
symmetry and closely conforms to C2h-2/m symmetry. The
independent atoms are labeled. (b) Same view of the
noncentered Ni10Bi2 cage of 2. This empty cage may be
envisioned as either a 1,12-disubstituted icosahedron or a
pentagonal antiprism of 10 surface Ni(s) atoms that is
capped on both pentagonal faces by bismuth atoms. Means
of the five independent Ni(s)-Bi edges, five independent
intrapentagonal Ni(s)-Ni(s′), and five independent inter-
pentagonal Ni(s)-Ni(s′) edges are 2.63, 2.88, and 2.50 Å,
respectively.
Bi) and the prototype [B12H12]
2- dianion and its carbo-
rane derivatives likewise possess 13 skeletal electron
pairs. For any of the Ni10(ER)2 members this number
of electron pairs corresponds to the observed number
of cluster valence electrons (CVEs) being 150 electrons
(i.e., 10 × 10(Ni) + 2 × 6(ER) + 18 × 2(CO) + 2(charge)
) 150). This equivalence in electron counting arises
because the skeletal electron-pair model assumes that
each icosahedral vertex utilizes only 3 internal orbitals
for metal cluster bonding; consequently, 12 valence
electrons from the filled 6 external metal orbitals for
each of the 10 Ni(CO) fragments and 2 valence electrons
from the one filled external main-group E orbital for
each of the 2 ER fragments must be subtracted from
150 CVEs. The resulting 26 skeletal valence electrons
occupy 13 internal bonding MOs in accordance with the
skeletal electron-pair count. Application of the different
valence electron-counting models developed by Mingos,32
by Teo and Zhang,33 and by King34 to any of the
3 from D5d to C2h-2/m. Especially noteworthy is that
the virtually identical dispositions of the 10 terminal
and 4 doubly and 4 triply bridging COs under pseudo-
C2h symmetry in both dianions produce analogous
geometrical deformations of the 1,12-Ni10Bi2 cages.
Examination of corresponding independent individual
distances between 2 and 3 in Tables 2 and 3, respec-
tively, shows analogous large variations of one inter-
pentagonal and two intrapentagonal Ni(s)-Ni(s′) dis-
tances from their mean values under assumed D5d
symmetry. The two tables also show that the nonbond-
ing Bi‚‚‚Bi′ distance in 2 is significantly shorter by 0.07
Å than that in 3; this somewhat larger oblate compres-
sion along the pseudo-5-fold Bi‚‚‚Bi′ axis in 2 containing
two smaller Bi-attached methyl substituents compared
to 3 containing two Bi-attached ethyl substituents may
be rationalized from steric effects of the ligand polyhe-
dron involving nonbonding interactions of the two alkyl
substituents with the closest carbonyl groups.
(31) (a) Wade, K. J . Chem. Soc., Chem. Commun. 1971, 792-793.
(b) Wade, K. Electron Deficient Compounds; Thomas Nelson and Sons,
Ltd.: London, 1971. (c) Wade, K. Adv. Inorg. Chem. Radiochem. 1976,
18, 1-66.
(32) Mingos, D. M. P.; Zhenyang, L. J . Chem. Soc., Dalton Trans.
1988, 1657-1664 and references therein. (b) Mingos, D. M. P.; May,
A. P. In The Chemistry of Metal Cluster Complexes; Shriver, D. F.,
Kaesz, H. D., Adams, R. D., Eds.; VCH Publishers: New York, 1990;
Chapter 2, pp 11-119. (c) Mingos, D. M. P.; Wales, D. J . Introduction
to Cluster Chemistry; Prentice Hall: Old Tappan, NJ , 1990.
(33) Teo, B. K.; Zhang, H.; Kean, Y.; Dang, H.; Shi, X. J . Chem. Phys.
1993, 99, 2929-2941. (b) Teo, B. K.; Longoni, G.; Chung, F. R. K.
Inorg. Chem. 1984, 23, 1257-1266. (c) Teo, B. K. Inorg. Chem. 1985,
24, 1627-1638.
According to the polyhedral skeletal electron-pair
(PSEP) model,31 a closo deltahedron (i.e., a completely

Nickel-Bismuth Carbonyl Clusters
Organometallics, Vol. 16, No. 8, 1997 1659
2-
Ta ble 1. Cr ysta l Da ta a n d Str u ctu r a l Refin em en t P a r a m eter s for th e [NMe4]+ Sa lts of [Ni10(BiR)2(CO)18
]
(R ) Me (2), Et (3))
Crystal Data
identification code
formula
2
C
3
28H30Bi2N2Ni10O18
C30H34Bi2N2Ni10O18
dark brown long plate
0.40 × 0.15 × 0.05
triclinic
cryst color, habit
cryst size (mm)
cryst system
space group
unit cell dimens
a (Å)
black opaque prism
0.42 × 0.22 × 0.20
monoclinic
P21/c
P1h
9.9390(2)
11.4488(2)
18.7425(3)
90
94.986(2)
90
2124.63(7)
6484
133(2)
0.710 73
2
9.795(1)
11.295(1)
11.413(2)
117.96(1)
92.29(1)
92.23(1)
1112.0(2)
40
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å)
peaks to determine cell
temp (K)
wavelength (Å)
Z
148(2)
0.710 73
1
fw
1687.60
2.638
12.639
1604
1715.65
2.562
12.076
818
D (calcd, g/cm3)
abs coeff (mm-1
F(000)
)
Data Collection
Siemens P4/CCD
2.06-25.98
-12 e h e 11
-6 e k e 13
-17 e l e 21
φ scan frames
10 s/frame
diffractometer
θ range for data collcn (deg)
index ranges
Siemens P4
2.02-22.50
0 e h e 10
-10 e k e 10
-12 e l e 12
Wyckoff ω
variable (3-30 deg/min)
0.8° in ω
3067
scan type
scan time or speed
scan range
no. of reflcns collcd
no. of indepdt reflcns
0.3° in φ
6981
3630 (Rint ) 0.0235)
2768 (Rint ) 0.0441)
Solution and Refinement
solution
refinement method
H atoms
direct methods
full-matrix least squares on F2
riding
weighting scheme a, ba
data/restraints/params
goodness-of-fit on F2
final Rb indices [I > 2σ(I)]
0.06720, 0
3630/0/272
1.268
R1(F) ) 0.046
wR2(F2) ) 0.12
R1(F) ) 0.050,
wR2(F2) ) 0.12
3268
0.0720, 8.2746
2766/34/281
1.060
R1(F) ) 0.044
wR2(F2) ) 0.11
R1(F) ) 0.055,
wR2(F2) ) 0.12
2379
Rb indices (all data)
obsd data [I > 2σ(I)]
largest diff peak and hole (e Å-3
largest and mean ∆/esd
)
3.80 and -6.17
0.001 and 0.000
2.46 and -2.26
0.000 and 0.000
a
2
b
2
w ) 1/[σ2(Fo2) + (aP)2 + bP]; P ) [Fo + 2Fc2]/3. R1(F) ) |∑||Fo| - |Fc||/∑|Fo|; wR2(F2) ) [∑[wi(Fo - Fc2)2]/∑[wi(Fo2)2]]1/2
.
[Ni10(ER)2(CO)18]2- members (E ) P, As, Sb, Bi) also
gives predicted values in agreement with the observed
electron count.
Noteworthy is that the observed diamagnetism of 2
obtained from solid-state magnetic susceptibility mea-
surements at room temperature is in harmony with the
electron-counting models.
Geom etr ica l An a lysis of th e [Ni10(EMe)2(CO)18]2-
Dia n ion s (E ) Bi, Sb) Con ta in in g Em p ty Icosa h e-
d r a l Ni10E2 Ca ges w ith Cor r esp on d in g An ion s
in the icosahedral Ni10Bi2 architecture resulting from
the formal transformation of 2 into 4; the interconver-
sion involves the addition of the internal d10 Ni(i) into
the empty Ni10Bi2 cage (at the inversion center) with
concomitant removal of the two one-electron Bi-attached
methyl substituents and a negative ionic charge in-
crease by one electron. This conceptual transformation
of 2 into 4 causes a marked 1.1 Å elongation of the
diameter of the highly compressed icosahedral cage of
2 along the principal 5-fold Bi‚‚‚Bi′ axis from 3.9 Å in 2
to 5.0 Å in 4. The resulting shape of the icosahedral
cage thereby changes from an oblate spheroidal one in
2 to a nearly regular spherical one in 4, as evidenced in
4 by the mean cage diameter of 5.2 Å along the five
centrosymmetrically-related Ni(s)‚‚‚Ni(s′) axes being
nearly the same as that of 5.0 Å along the Bi‚‚‚Bi′ axis.
This considerable icosahedral-cage elongation along the
pseudo-5-fold axis in 4 is caused by the formal insertion
of the interior Ni(i) at the inversion center of the
icosahedron, which gives rise to two identical short Ni-
(i)-Bi distances of 2.51 Å; the fact that this mean Ni-
(i)-Bi distance is 0.31 Å shorter than the mean Ni(s)-
Con ta in in g Ni-Cen ter ed (F illed ) Ni10E2 Icosa h e-
2-
d r a l Ca ges. (a ) [Ni10(BiMe)2(CO)18
]
Dia n ion (2)
3-
ver su s th e [Ni11Bi2(CO)18
]
Tr ia n ion (4). A geo-
metrical comparison (Table 4) of the noncentered (empty)
1,12-Ni10Bi2 cage of pseudo D5d symmetry in the 150-
electron [Ni10(BiMe)2(CO)18]2- dianion (2) of the [NMe4]+
salt with the Ni-centered 1,12-Ni10Bi2 cage of pseudo
D5d symmetry in the 159-electron [Ni11Bi2(CO)18]3-
trianion (4)20 of the [NEt4]+ salt reveals major changes
(34) (a) King, R. B. Inorg. Chim. Acta 1992, 198-200, 841-861. (b)
King, R. B. Inorg. Chim. Acta 1995, 235, 111-115. (c) King, R. B.
Polyhedron 1994, 13, 2005-2016.

1660 Organometallics, Vol. 16, No. 8, 1997
Mlynek and Dahl
Ta ble 2. Com p a r a tive Geom etr ica l An a lysis of th e
Ta ble 3. Selected In ter a tom ic Dista n ces for th e
closo-1,12-Disu bstitu ted Icosa h ed r a l Ni10Bi2 Ca ge
2-
Cen tr osym m etr ic [Ni10(EMe)2(CO)18
]
Hom ologu es
(E ) P ,a As,b Sb,c Bid ) u n d er Assu m ed D5d
in th e Cen tr osym m etr ic
[NMe4]2[Ni10(BiEt)2(CO)18]
Sym m etr y w ith All Clu ster s a s [NMe4]+ Sa lts
2-
1,12-Ni10E2 cage
Ni10P2
Ni10As2
Ni10Sb2
Ni10Bi2
dist between
centrosymmetrically
related core atoms
intrapentagonal
Ni-Ni dists
A. Nonbonding Distances between
Centrosymmetrically Related Cage Atoms (Å)
Ni(1)-Ni(1a)
Ni(2)-Ni(2a)
Ni(3)-Ni(3a)
Ni(4)-Ni(4a)
Ni(5)-Ni(5a)
4.948(2)
4.866(2)
4.942(2)
4.956(2)
4.838(2)
5.043(3)
4.974(3)
5.034(3)
5.037(3)
4.957(3)
5.235(4)
5.142(3)
5.194(5)
5.184(3)
5.153(3)
5.357(2)
5.275(2)
5.263(2)
5.248(2)
5.260(2)
Ni(1)-Ni(1a)
5.307(3)
5.334(3)
5.232(3)
5.195(3)
5.272(3)
Ni(1)-Ni(2)
3.019(2)
2.745(2)
2.874(2)
2.721(2)
2.973(2)
Ni(2)-Ni(2a)
Ni(3)-Ni(3a)
Ni(4)-Ni(4a)
Ni(5)-Ni(5a)
Ni(2)-Ni(3)
Ni(3)-Ni(4)
Ni(4)-Ni(5)
Ni(5)-Ni(1)
Ni-Ni(av)
E(1)-E(1a)
4.91(av)
3.586(2)
5.01(av)
3.689(2)
5.18(av)
3.848(3)
5.28(av)
3.881(1)
Ni-Ni(av)
5.27(av)
3.951(1)
Ni-Ni(av)
2.87(av)
B. Nickel-Pnicogen (E) Distances (Å)
Bi(1)-Bi(1a)
E-Ni(1)
E-Ni(2)
E-Ni(3)
E-Ni(4)
E-Ni(5)
2.375(2)
2.329(2)
2.366(2)
2.355(2)
2.324(2)
2.450(2)
2.412(2)
2.431(2)
2.440(2)
2.405(2)
2.576(2)
2.538(2)
2.575(2)
2.563(2)
2.538(2)
2.638(1)
2.603(1)
2.635(1)
2.647(1)
2.610(1)
interpentagonal
Ni-Ni dists
Ni-Bi dists
Bi-Ni(1)
Bi-Ni(2)
Bi-Ni(3)
Bi-Ni(4)
Bi-Ni(5)
2.609(2)
Ni(1)-Ni(3a)
2.520(2)
2.523(2)
2.549(2)
2.343(2)
2.553(2)
E-Ni(av)
2.35(av)
2.43(av)
2.56(av)
2.63(av)
2.652(2)
2.621(2)
2.625(2)
2.640(2)
Ni(1)-Ni(4a)
Ni(2)-Ni(4a)
Ni(2)-Ni(5a)
Ni(3)-Ni(5a)
C. Intrapentagonal Nickel-Nickel Distances (Å)
Ni(1)-Ni(2)
Ni(2)-Ni(3)
Ni(3)-Ni(4)
Ni(4)-Ni(5)
Ni(5)-Ni(1)
2.636(1)
2.575(1)
2.622(1)
2.584(1)
2.660(1)
2.731(2)
2.635(2)
2.712(2)
2.621(2)
2.731(2)
2.873(2)
2.709(2)
2.856(2)
2.731(2)
2.863(2)
2.985(2)
2.757(2)
2.932(1)
2.731(1)
2.977(2)
Bi-Ni(av)
2.63(av)
Ni-Ni(av)
2.50(av)
Ni-Ni(av)
2.62(av)
2.69(av)
2.81(av)
2.88(av)
Bi distance of 2.82 Å in 4 implies the existence of strong
radial Ni(i)-Bi bonding in 4. One net consequence of
the two Bi atoms strongly interacting with the interior
Ni(i) in 4 is a substantial weakening in their surface-
cage bonding interactions with the 10 surface Ni(s)
atoms, as indicated by a 0.19 Å increase of the mean
Ni(s)-Bi distance from 2.63 Å in 2 to 2.82 Å in 4.
Other salient structural consequences due to the
formal insertion of the Ni(i) into the empty icosahedral
Ni10Bi2 cage of 2 at the inversion center are a small
decrease of 0.05 Å in the mean inversion center (IC)-
to-Ni(s) distance and a corresponding small decrease of
0.07 Å in the mean intrapentagonal Ni(s)-Ni(s′) dis-
tance. These geometrical changes are consistent with
the view that the bonding interactions between the Ni-
(i) and the 10 Ni(s) atoms in 4 are relatively weak
compared to the strong bonding Ni(i)-Bi interactions.
If the radial Ni(i)-Ni(s) bonding in 4 were also strong,
one would expect weaker surface-cage Ni(s)-Ni(s′)
bonding interactions and hence longer (instead of the
observed shorter) Ni(s)-Ni(s′) distances in 4 versus
those in 2. The corresponding change in the mean
interpentagonal Ni(s)-Ni(s′) is only 0.01 Å; this invari-
ance of the mean value for the 10 interpentagonal Ni-
(s)-Ni(s′) distances, whether the icosahedral cage is
empty or filled, may be attributed to the dispositions of
the 8 bridging carbonyl ligands in both 2 and 4.
D. Interpentagonal Nickel-Nickel Distances (Å)
Ni(1)-Ni(3a)
Ni(1)-Ni(4a)
Ni(2)-Ni(4a)
Ni(2)-Ni(5a)
Ni(3)-Ni(5a)
2.505(2)
2.493(2)
2.538(2)
2.421(2)
2.497(2)
2.502(2)
2.500(2)
2.510(2)
2.417(1)
2.522(2)
2.519(2)
2.516(2)
2.524(2)
2.414(2)
2.507(2)
2.516(2)
2.520(1)
2.515(2)
2.387(2)
2.539(2)
Ni-Ni(av)
2.49(av)
2.49(av)
2.50(av)
2.50(av)
E. Ni-CO Bridging Distances (Å)
Ni(1)-C(13)
Ni(1)-C(14)
1.943(7)
1.920(6)
1.897(7)
1.888(7)
2.123(7)
1.921(6)
1.991(6)
1.967(6)
1.944(6)
2.262(6)
1.93(1)
1.90(1)
1.86(1)
1.90(1)
2.26(1)
1.90(1)
1.95(1)
1.96(1)
1.91(1)
2.19(1)
1.934(7)
1.901(6)
1.883(6)
1.878(6)
2.221(6)
1.892(7)
2.008(7)
1.966(7)
1.887(6)
2.339(6)
1.89(1)
1.94(1)
1.88(1)
1.88(1)
2.32(1)
1.87(1)
1.95(1)
1.98(1)
1.88(1)
2.15(1)
Ni(3)-C(13a)
Ni(4)-C(14a)
Ni(2)-C(235a)
Ni(2)-C(245a)
Ni(3)-C(235a)
Ni(4)-C(245)
Ni(5)-C(235)
Ni(5)-C(245)
F. Ni-CO Terminal Distances (Å)
Ni(1)-C(1)
Ni(2)-C(2)
Ni(3)-C(3)
Ni(4)-C(4)
Ni(5)-C(5)
1.775(7)
1.769(6)
1.796(7)
1.792(7)
1.753(8)
1.81(1)
1.76(1)
1.78(1)
1.77(1)
1.76(1)
1.783(6)
1.749(6)
1.789(7)
1.766(7)
1.758(6)
1.80(1)
1.78(1)
1.82(1)
1.80(1)
1.77(1)
Ni-C(av)
1.78(av)
1.78(av)
1.77(av)
1.79(av)
G. C-O Bridging Distances (Å)
C(13)-O(13)
C(14)-O(14)
C(235)-O(235)
C(245)-O(245)
1.16(1)
1.17(1)
1.17(1)
1.17(1)
1.17(1)
1.15(1)
1.17(1)
1.20(1)
1.14(1)
1.16(1)
1.16(1)
1.17(1)
1.19(1)
1.16(1)
1.19(1)
1.19(1)
H. C-O Terminal Distances (Å)
C(1)-O(1)
C(2)-O(2)
C(3)-O(3)
C(4)-O(4)
C(5)-O(5)
1.14(1)
1.12(1)
1.13(1)
1.12(1)
1.16(1)
1.13(av)
1.11(1)
1.12(1)
1.13(1)
1.15(1)
1.12(1)
1.13(av)
1.12(1)
1.14(1)
1.12(1)
1.15(1)
1.12(1)
1.13(av)
1.12(1)
1.13(1)
1.12(1)
1.13(1)
1.15(1)
1.13(av)
In the Ci-1h [Ni11Bi2(CO)18]3- trianion20 the 8 bridging
COs of the 18 carbonyl ligands adopt a different cen-
trosymmetric mode of coordination to the 10 Ni(s) atoms
from that found for all of the other [Ni10(ER)2(CO)18]2-
I. E-CH3 Distance (Å)
1.868(6) 1.99(1)
dianions (E ) P, As, Sb, Bi) under the same pseudo C2h
-
E-C(Me)
2.174(6)
2.28(1)
2/m symmetry. Again, there are 10 terminal COs, one
per surface Ni(s), and 8 bridging COs in the equatorial
region of the cluster. However, the bridging COs now
consist of 6 edge-bridging COs and 2 face-capping COs.
The 2 triply bridging COs cap the opposite (centrosym-
metrically related) triangular Ni3 faces, each of which
consists of 1 intrapentagonal and 2 interpentagonal
edges; the remaining 6 doubly bridging COs span the
other 6 interpentagonal edges (i.e., those not belonging
to the two OC-capped trinickel faces). The pseudo
horizontal mirror plane passes through the 2 triply
bridging COs, the Ni(i), the 2 Bi, and 2 Ni(s) atoms,
J . Ni-E-Ni Angles (deg)
Ni(1)-E-Ni(2)
Ni(2)-E-Ni(3)
Ni(3)-E-Ni(4)
Ni(4)-E-Ni(5)
Ni(5)-E-Ni(1)
68.5(1)
66.1(1)
67.8(1)
66.4(1)
68.6(1)
68.3(1)
65.9(1)
67.7(1)
65.5(1)
68.5(1)
68.36(4)
63.98(5)
67.56(5)
64.73(4)
68.09(6)
69.42(4)
63.50(3)
67.44(3)
62.59(3)
69.12(3)
Ni-E-Ni(av)
67.5(av)
67.2(av)
66.5(av)
66.4(av)
K. Oblate Spheroidal Approximation of the Ni10E2 Core
minor: major
semiaxes ratio
0.657(11)
0.671(7)
0.687(6)
0.682(10)
a
22 °C/ref 41 (for the interatomic distances of the Ni10P2 cluster
b
in the [PPN]+ salt,13 see ref 16). 22 °C/ref 15. c -150 °C/ref 19.
d
-140 °C/this work.

Nickel-Bismuth Carbonyl Clusters
Organometallics, Vol. 16, No. 8, 1997 1661
2-
Ta ble 4. An a lysis of Selected Mea n Dista n ces (Å) of th e Hom ologou s [Ni10(EMe)2(CO)18
]
Dia n ion s (E ) Bi
3-
(2), Sb) Con ta in in g Em p ty Icosa h ed r a l 1,12-Ni10E2 Ca ges w ith th e [Ni11Bi2(CO)18
]
Tr ia n ion (4) a n d th e
n -
[Ni11(SbNi(CO)18
]
Dia n ion (n ) 2) a n d Tr ia n ion (n ) 3), Con ta in in g Ni-Cen ter ed (F illed ) Icosa h ed r a l
1,12-Ni10E2 Ca ges
Ni10Bi2 dianiona
Ni11Bi2 trianionb
Ni10Sb2 dianionc
Ni11Sb2 dianiond
Ni11Sb2 trianione
empty/filled
empty
filled
empty
filled
filled
1,12-Ni10E2 cage
Ni10Bi2
Ni10Bi2
Ni10Sb2
Ni10Sb2
Ni10Sb2
IC-Ef [2]g
1.95
2.64
2.63
2.88
2.50
2.51
2.59
2.82
2.81
2.51
1.92
2.60
2.56
2.81
2.50
2.40
2.57
2.73
2.77
2.52
2.40
2.58
2.74
2.78
2.52
IC-Ni(s)f [10]
Ni(s)-E [10]
intrapentagonal Ni(s)-Ni(s′) [10]
interpentagonal Ni(s)-Ni(s′) [10]
a
b
d
[NMe4]+/this work. [NEt4]+/ref 20. c [NMe4]+/ref 19. [N(PPh3)2]+/ref 18. e [NMe3(CH2Ph)]+/ref 18. f IC denotes the inversion center
g
of the icosahedral Ni10E2 cage, which for a Ni-centered (filled) icosahedron is occupied by the interior Ni(i). Values in brackets denote
the number of equivalent distances under assumed D5d symmetry having each mean value in the right row.
while the pseudo-2-fold axis passes through 2 doubly
bridging COs and the Ni(i). The pseudo-C2h-2/m sym-
metry exhibited by the 18-vertex carbonyl polyhedron
gives rise to three independent Ni(s) atoms, correspond-
ing to 3 independent Ni(s)-Bi edges and 3 independent
intrapentagonal and 3 interpentagonal Ni(s)-Ni(s′)
edges. The other 8 intrapentagonal Ni(s)-Ni(s′) edges
are not connected by either edge- or face-bridging COs.
The marked influence of the bridging COs on the bond-
length shortening of certain Ni(s)-Ni(s′) edges is readily
apparent; the only independent intrapentagonal Ni(s)-
Ni(s′) edge involved with a bridging CO (viz., the face-
capping CO) is 0.11 Å shorter than the mean of 2.81 Å
for the 5 independent intrapentagonal Ni(s)-Ni(s′)
edges; the 5 independent interpentagonal Ni(s)-Ni(s′)
edges, each of which is carbonyl-linked, are all within
0.02 Å of the 0.3 Å-shorter mean of 2.51 Å. Albano, et
al.20 pointed out that this different carbonyl arrange-
ment in the [Ni11Bi2(CO)18]3- trianion gives rise to 3
COs (one terminal and two bridging) being coordinated
to 8 Ni(s) atoms but only 2 COs (one terminal and one
triply bridging) being linked to the other 2 centrosym-
metrically related Ni(s) atoms on the pseudo mirror
plane; from the observation that a face-capping CO is
0.3 Å nearer to each of these mirror-containing Ni(s)
atoms and thereby forms a stronger Ni-(µ3-CO) bonding
interaction, they concluded that each of the 10 surface
Ni(s) atoms has an approximately equivalent charge
density.
and 4 triply bridging) as found in the [Ni10(SbMe)2-
(CO)18]2- dianion and its homologues including 2.
First, it should be noted from Table 4 that the
corresponding mean distances are experimentally equiva-
lent (within 0.01 Å) for the 158-electron dianion (n ) 2)
and 159-electron trianion (n ) 3) of the [Ni11(SbNi-
(CO)3)2(CO)18]n- series (n ) 2-4). This indicated in-
variance of geometry to the negative ionic charge of the
anions in this series suggests that the structurally
determined Ni-centered Ni10Bi2 cage geometry for the
159-electron trianion (n ) 3) in the corresponding
[Ni11Bi2(CO)18]n- series (n ) 2-4) is essentially un-
changed for the other two members. Second, the
conceptual interconversion of the empty icosahedral
Ni10Sb2 cage in the [Ni10(SbMe)2(CO)18]2- dianion into
the Ni-filled one in the [Ni11(SbNi(CO)3)2(CO)18]2- di-
anion similarly produces a large 0.96 Å elongation of
the diameter of the highly compressed icosahedral cage
along the principal 5-fold Sb‚‚‚Sb′ axis. However, the
resulting Ni-filled cage diameter of 4.80 Å along the
Sb‚‚‚Sb′ axis is still 0.34 Å smaller than the mean cage
diameter of 5.14 Å along the five centrosymmetrically
related Ni(s)‚‚‚Ni(s′) axes. In this case, the Ni-filled
icosahedral cage in the Ni11Sb2 dianion is still somewhat
oblate.
Otherwise, Table 4 shows that the observed trends
resulting from the formal transformation of the Ni10-
Sb2 dianion into the Ni11Sb2 dianion (which includes the
placement of the internal Ni(i) at the inversion center
of the icosahedral cage) parallel those resulting from
the similar conversion of the Ni10Bi2 dianion (2) into the
Ni11Bi2 dianion (4). These analogous trends are as
follows: (1) two short radial Ni(i)-Sb distances of 2.40
Å indicative of strong radial Ni(i)-Sb bonding; (2) the
10 Ni(s)-Sb distances being 0.17 Åslonger than those
in the empty icosahedral cage, corresponding to much
weaker surface-cage Ni(s)-Sb bonding due to the
additional involvement of the Sb atoms in strong radial
bonding with the interior Ni(i); (3) small decreases of
0.02 Å in the IC-to-Ni(s) distances and 0.04 Å in the
intrapentagonal Ni(s)-Ni(s′) distances, indicating that
the interior Ni(i) forms relatively weak bonding interac-
tions with the 10 surface Ni(s) atoms and therefore has
little effect on surface-cage intrapentagonal Ni(s)-Ni-
(s′) bonding; (4) the analogous mean interpentagonal Ni-
(s)-Ni(s′) distances of 2.50 Å in the Ni10Sb2 dianion and
2.52 Å in the Ni11Sb2 dianions comparing favorably
(within 0.02 Å) with those in the corresponding Ni10Bi2
and Ni11Bi2 dianions. In fact, the mean interpentagonal
Ni(s)-Ni(s′) distances in these and the two other
homologous [Ni10(EMe)2(CO)18]2- dianions (E ) P, As)
2-
(b) [Ni10(SbMe)2(CO)18
]
Dia n ion ver su s th e
Dia n ion (n ) 2) a n d
n -
[Ni11{SbNi(CO)3}2(CO)18
]
Tr ia n ion (n ) 3). Table 4 also allows a geometrical
comparison of selected mean distances (under pseudo-
D5d symmetry) between the noncentered (empty) icosa-
hedral 1,12-Ni10Sb2 cage in the 150-electron [Ni10-
(SbMe)2(CO)18]2- homologue19 of 2 and the correspond-
ing Ni-centered (filled) icosahedral 1,12-Ni10Sb2 cage in
the crystallographically determined 158-electron dian-
ion (n ) 2) and 159-electron trianion (n ) 3) of the
[Ni11(SbNi(CO)3)2(CO)18]n- series (n ) 2-4).18 In each
of these anions the Sb-attached Ni(CO)3 substituent is
coordinated as a Lewis-acid adduct to the otherwise lone
electron pair of each Sb (corresponding to the formation
of a dative electron-pair bond). Hence, the same charged
anions in the [Ni11Bi2(CO)18]n- series (n ) 2-4)20 and
[Ni11(SbNi(CO)3)2(CO)18]n- series (n ) 2-4)18 have the
same observed number of metal cluster valence elec-
trons (CVEs). Especially noteworthy is that the 18
carbonyl ligands in both the [Ni11{SbNi(CO)3}2(CO)18]n-
dianion (n ) 2) and trianion (n ) 3) possess the same
pseudo-C2h distribution of bridging COs (viz., 4 doubly

1662 Organometallics, Vol. 16, No. 8, 1997
Mlynek and Dahl
metry. The pseudo horizontal mirror plane passes
through the 2 E and 2 Ni(s) atoms; no atoms lie on the
perpendicular twofold principal axis. Both Table 4 and
Figure 3 show the following trends under assumed D5d
symmetry: (1) the empty 1,12-Ni10E2 icosahedral cage
is similarly deformed for all four members by a large
oblate compression along the pseudo-5-fold E‚‚‚E′ axis.
The nonbonding E‚‚‚E′ distances among the empty
Ni10P2 (3.59 Å), Ni10As2 (3.69 Å), Ni10Sb2 (3.85 Å), and
Ni10Bi2 (3.88 Å) cages are much less than the five
centrosymmetrically related, nonbonding mean Ni(s)-
Ni(s′) distances of 4.91, 5.01, 5.18, and 5.28 Å, respec-
tively. (2) The means of the five independent Ni(s)-E
edges and five independent intrapentagonal Ni(s)-Ni-
(s′) edges in the four Ni10E2 icosahedral cages increase
in an analogous fashion with the increased size of the
main-group V (15) E atoms. Subtraction of the nonpolar
covalent radii35 for P (1.10 Å), As (1.19 Å), Sb (1.38 Å),
and Bi (1.46 Å) from the corresponding mean Ni(s)-E
distances gives decreasing values of 1.25, 1.24, 1.18, and
1.17 Å, respectively, for the covalent nickel radius. (3)
The means of the five independent interpentagonal Ni-
(s)-Ni(s′) edges are invariant to the increased sizes of
the different E atoms capping each pentagonal face; the
observed means of 2.49-2.50 Å are in agreement with
that found36 in ccp nickel metal (2.492 Å).
The 18 carbonyl ligands in each dianion are distrib-
uted in an identical fashion about the 1,12-Ni10E2 cage.
The 10 terminal COs are each coordinated to one surface
Ni(s) atom. The 8 bridging COs, consisting of 4 doubly
and 4 triply bridging COs, are arranged under pseudo-
C2h symmetry in alternating pairs of two doubly (d) and
two triply (t) bridging COs (i.e., corresponding to a cyclic
dttddttd CO sequence) on adjacent interpentagonal
edges and triangular faces formed by one intrapentago-
nal and two interpentagonal edges. This particular
distribution of 18 terminal and bridging COs gives rise
to 30 carbonyl links (i.e., one for each terminal CO, two
for each edge-bridging CO, and three for each face-
capping CO), such that each of the 10 surface Ni(s)
atoms is similarly connected to 3 COs (one terminal and
two bridging). This centrosymmetric carbonyl arrange-
ment results in each of the 10 interpentagonal Ni(s)-
Ni(s′) edges being linked by a doubly bridging CO or by
one or two triply bridging COs; 4 of the 10 intrapen-
tagonal Ni(s)-Ni(s′) edges are linked by one triply
bridging CO. Table 4 reveals similar effects of the
bridging COs in markedly shortening certain individual
intrapentagonal and interpentagonal Ni(s)-Ni(s′) edges
in the four homologues from the mean values under
assumed D5d symmetry for the Ni10E2 cage.
F igu r e 3. Plot showing the variation of the mean dis-
tances for each of the three kinds of five independent
icosahedral edges in the centrosymmetric 1,12-Ni10E2
icosahedral cage as a function of the congeneric E element
for the four members of the homologous [Ni10(EMe)2(CO)18]2-
series (E ) P, As, Sb, Bi). The means of the Ni-E edges
(∆) and intrapentagonal Ni(s)-Ni(s′) edges (]) increase in
a similar fashion with the increased sizes of the main-group
V (15) E atoms (i.e., nonpolar covalent radii35 are 1.10 Å
for P, 1.19 Å for As, 1.38 Å for Sb, and 1.46 Å for Bi). In
sharp contrast, the means of the interpentagonal Ni(s)-
Ni(s′) edges (O) are essentially invariant to the increased
sizes of the four different pairs of E atoms that cap the
two pentagonal Ni(s) faces; the uniformity of these consid-
erably shorter means is attributed to the analogous ar-
rangements in the four members of the 8 bridging COs that
span all of the interpentagonal Ni(s)-Ni(s′) edges.
are alike (within 0.03 Å) as well as much shorter than
the intrapentagonal Ni(s)-Ni(s′) distances (which, in
contrast, uniformly increase from 2.62 Å for E ) P to
2.88 Å for E ) Bi); these structural observations are
readily ascribed to the analogous polyhedral arrange-
ments of the 18 COs and in particular to the bond-length
shortening effects of the 8 bridging carbonyls. The
considerable variations that occur in individual dis-
tances are due primarily to the carbonyl ligands in a
given cluster conforming, in general, to only pseudo-C2h
symmetry instead of the pseudo D5d symmetry assumed
in Table 4 for each icosahedral 1,12-Ni10E2 cage (E )
Bi, Sb). Nevertheless, the observed systematic trends
in mean values based upon averaging equivalent dis-
tances under D5d symmetry provide a self-consistent
representation of the structural/bonding features of this
important class of NixEy icosahedral clusters.
Com p a r ison of t h e F ou r H om ologou s [Ni10
-
2-
(EMe)2(CO)18
]
Mem ber s (E ) P , As, Sb, Bi) a n d
Resu ltin g Im p lica tion s. (a ) Geom etr ica l F ea tu r es.
Because each member of the [Ni10(ER)2(CO)18]2- series
(E ) P,13 As,15 Sb,19 Bi) in all structurally determined
salts invariably possesses a crystallographic inversion
center, only one-half of the 30 icosahedral edges in each
Ni10E2 cage are independent. Table 4 presents a com-
parison of selected distances under assumed D5d sym-
metry for the [NMe4]+ salts of the four centrosymmetric
[Ni10(EMe)2(CO)18]2- members (E ) P, As, Sb, Bi) which
differ from one another only by the group V (15) E
element. Figure 3 gives a plot showing the variation of
the mean distances for each of the five independent Ni-
(s)-E, intrapentagonal Ni(s)-Ni(s′), and interpentago-
nal Ni(s)-Ni(s′) edges in a given 1,12-Ni10E2 cage as a
function of the congeneric E element.
The shape of the icosahedral framework of the 12
Ni10E2 cage atoms was found for each member of the
[Ni10(EMe)2(CO)18]2- series (E ) P, As, Sb, Bi) from use
of the crystallographic Ni-E and centrosymmetric
E‚‚‚E′ and Ni‚‚‚Ni′ distances to approximate an oblate
spheroid with the smaller minor axis (2c) coinciding
with the E‚‚‚E′ axis and the other two larger orthogonal
major axes (2a ) 2b) lying in the equatorial plane of
the Ni10E2 cage. The shape of the oblate spheroid is
defined in terms of the minor (c)-to-major (a) semiaxial
ratio, where the length c of the minor semiaxis corre-
sponds to one-half the E‚‚‚E′ distance; the length a for
(35) Purcell, K. F.; Kotz, J . C. Inorganic Chemistry; W. B. Saun-
ders: Philadelphia, 1977.
(36) Donohue, J . The Structures of the Elements; J ohn Wiley and
Sons: New York, 1974; p 213.
The geometries of these dianions are very similar;
each dianion conforms closely to pseudo C2h-2/m sym-

Nickel-Bismuth Carbonyl Clusters
Organometallics, Vol. 16, No. 8, 1997 1663
the two equivalent major semiaxes was calculated from
the crystallographic distances.37 The differences in
oblate compression along the four icosahedral Ni10E2
frameworks determined from the calculated mean c/a
shape ratios are surprisingly small; although slight
increases are observed on going from E ) P (0.657) to
E ) As (0.671) and then to E ) Sb (0.687), a slight
decrease then occurs for E ) Bi (0.682).
ions15 and the [Ni8(AstBu)4(CO)12]2- dianion.28 It is
likely that the presumed nonexistence of the corre-
sponding antimony or bismuth clusters with either
empty 1,2,12-Ni9E3 or 1,2,9,12-Ni8E4 cages is a conse-
quence of the empty icosahedral cage being thermody-
namically unstable due largely to the resulting effects
of the longer and weaker Sb-Sb and Bi-Bi bonds on
the icosahedral cage architectures.
If the primary cause of a cage deformation involved
only size effects, then the Ni10E2 cages would be
expected to become more prolate spheroids as E becomes
larger. Thus, combined electronic-size effects need to
be considered in order to account for the observation of
the relatively invariant oblate compressions for the four
cages. The aforementioned lengthening of the Ni-E
bonds with larger E atoms, which is due primarily to
the increase of the E covalent radii, is also paralleled
by a similar increase of the intrapentagonal Ni-Ni bond
lengths, leading to relatively constant Ni-E-Ni angles
of ca. 67° for the four Ni10E2 icosahedra. In a M2E2-
bridged system the M-E-M angles are governed not
only by the minimization of all nonbonded repulsions
between atoms which depend on their sizes but also by
interorbital electron-pair interactions, which for a more
electronegative bridging atom favor a wider bridging
angle.38 Thus, the observed sizes and shapes of the
oblate spheroidal Ni10E2 frameworks are dictated by
combined electronic/steric effects including the steric
effects of the ligand polyhedra. This analysis empha-
sizes that the Ni-E connectivities are much stronger
than the intrapentagonal Ni-Ni connectivities in the
Ni10E2 cages. This is evidenced by concomitant length-
ening of the intrapentagonal Ni-Ni distances upon the
increase in Ni-E distances with a larger E atom in
order to maintain relatively constant Ni-E-Ni angles.
If these Ni-Ni distances did not increase, then the Ni10-
Sb2 and Ni10Bi2 cages would have considerably more
prolatelike ellipsoidal shapes along with implausibly
acute Ni-E-Ni angles of 55.0 and 53.6°, respectively.
(b) Ster eoch em ica l Isom er s. A prime trend is that
In opposition to the two main-group V (15) atoms in
the empty or Ni-centered Ni10E2 icosahedral cages being
invariably located at opposite 1,12-vertices, the two E
atoms in disubstituted isolobal main-group V (15)
derivatives of the regular icosahedral [B12H12]
2- dianion
occupy adjacent 1,2-vertices. Specifically, Todd, Little,
and their co-workers39 have systematically prepared and
characterized 1,2-E2B10H10 derivatives (E ) P, As, Sb,
Bi) as well as mixed 1,2-EBiB10H10 derivatives (E ) P,
As, Sb). An X-ray diffraction analysis of Bi2B10H10
revealed a highly distorted C2v icosahedral cage with 1
Bi-Bi edge of 2.96 Å, 8 Bi-B edges of 2.48 Å (av), and
21 B-B edges of 1.80 Å (av).39d
Of the three possible disubstituted A10E2 icosahedral
isomers (viz., the C2v 1,2-, the C2v 1,7-, and the D5d 1,-
12-), it is apparent from stereochemical considerations
that the symmetrical D5d 1,12-isomer would best mini-
mize cage-strain and thereby would be thermodynami-
cally the most stable isomer. Furthermore, only the D5d
1,12-isomer for both empty or Ni-centered Ni10E2 icosa-
hedral cages would give rise to suitable polyhedral
ligand arrangements (i.e., the observed ones) for the 10
terminal and 8 bridging COs to coordinate in a favorable
steric/electronic fashion to the 10 surface Ni(s) atoms
with a resulting equivalent localized charge density on
each Ni(s). In the case of the E2B10H10 derivatives,39
the synthetic methodology allows formation of only the
known 1,2-isomers. In light of its highly deformed
icosahedral framework, the observed behavior of the 1,2-
Bi2B10H10, in being thermally stable in the solid state
up to 350 °C (before darkening) and in giving rise to an
ion-parent peak in an EI mass spectrum, is especially
surprising.39d One can conclude that its existence is due
to the B-B and B-Bi bonding interactions being
sufficiently strong and its 10 hydrogen atoms not
sterically demanding to offset the presumed large icosa-
hedral cage-strain created by the much longer Bi-Bi
bonding edge.
for E ) P and As three different types of electronically
2-
equivalent (isolobal) [Ni12-x(ER)x(CO)24-3x
]
members
(x ) 2-4) containing noncentered 1,12-Ni10E2, 1,2,12-
Ni9E3, and 1,2,9,12-Ni8E4 icosahedral cages were iso-
lated and structurally characterized; in contrast, for E
) Sb and Bi only the [Ni10(ER)2(CO)18]2- members are
known despite extensive unsuccessful attempts to
detect the other two types. For example, reactions
of [Ni6(CO)12]2- (1) with MePCl2 have produced
the [Ni10(PMe)2(CO)18]2-, [Ni9(PMe)3(CO)15]2-, and
[Ni8(PMe)4(CO)12]2- dianions;13 similarly, reactions of
E xp er im en t a l An a lyses of t h e [Ni10(E Me)2-
2-
(CO)18
]
Dia n ion s (E ) P , As, Sb, Bi) a n d Resu lt-
in g Im p lica tion s. (a ) In fr a r ed Sp ectr a An a lyses.
The virtually identical IR spectra of both the 150-
electron 2 and 3 in THF exhibited a strong terminal
carbonyl absorption at 2010 (2005) cm-1 and a broad
medium carbonyl absorption at 1829 (1830) cm-1 that
can be ascribed to edge-bridging COs; in contrast to the
three-band carbonyl pattern observed for the phospho-
rus homologue (but not for the arsenic and antimony
homologues), no distinct maximum characteristic of
face-bridging COs was observed in either 2 or 3.
1
with RAsCl2 reagents have given rise to the
[Ni10(AsMe)2(CO)18]2- and [Ni9(AsPh)3(CO)15]2- dian-
(37) The length a of the two major semiaxes (for which a ) b) of the
spheroid was calculated from the equation a ) [(c2 sin2 θ)/((c/r) - cos2
θ)]1/2, where c is the length of the minor semiaxis corresponding to
one-half of the E‚‚‚E′ distance, r is a vector from the centroid of the
spheroid to each Ni atom corresponding to one-half of the centrosym-
metric Ni‚‚‚Ni′ distance, and θ is the dihedral angle between the
centroid-to-E vector and the centroid-to-Ni vector. The dihedral θ angle
was calculated from trigonometric relationship h2 ) r2 + c2 - 2cr cos
θ, where h denotes the Ni-E bonding distance. The small deviations
((0.04 Å) obtained for the five individual crystallographically inde-
pendent Ni atoms in each of the four Ni10E2 clusters provides the basis
for our premise that the ellipsoidal shape (a * b * c) for each
icosahedral Ni10E2 framework is closely approximated to an oblate
spheroid (a ) b > c). The mean c/a shape ratios are 1.793 Å/2.728 Å
) 0.657 for E ) P, 1.845 Å/2.749 Å ) 0.671 for E ) As, 1.924 Å/2.801
Å ) 0.687 for E ) Sb, and 1.941 Å/2.845 Å ) 0.682 for E ) Bi.
(38) Dahl, L. F.; Rodulfo de Gil, E.; Feltham, R. D. J . Am. Chem.
Soc. 1969, 91, 1653-1664.
A comparison of the infrared solution spectra of the
[Ni10(EMe)2(CO)18]2- dianions (E ) P, As, Sb, Bi) reveals
(39) (a) 1,2-P2B10H10
Todd, L. J . Inorg. Chem. 1989, 28, 1087-1091. (b) 1,2-As2B10H10
Hanusa, T. P.; Roig de Parisi, N.; Kester, J . G.; Arafat, A.; Todd, L. J .
Little, J . L.
: Little, J . L.; Kester, J . G.; Huffman, J . C.;
:
Inorg. Chem. 1987, 26, 4100-4102. (c) 1,2-Sb2B10H10
:
Inorg. Chem. 1979, 18, 1598-1600. (d) 1,2-Bi2B10H10 and 1,2-
EBiB10H10: Little, J . L.; Whitesell, M. A.; Kester, J . G.; Folting, K.;
Todd, L. J . Inorg. Chem. 1990, 29, 804-808.

1664 Organometallics, Vol. 16, No. 8, 1997
Mlynek and Dahl
insignificant variations of the two higher-frequency
absorption maxima in the carbonyl region. In THF
solution, these four homologues with E ) P, As, Sb, and
Bi exhibited one strong terminal carbonyl absorption
at 2012, 2005, 2007, and 2010 cm-1, respectively, and
one broad medium edge-bridging carbonyl absorption
at 1830, 1840, 1829, and 1829 cm-1, respectively. It is
apparent that the nature of the group V (15) E atoms
does not significantly alter the average negative charge
density on the surface Ni(s) atoms of the empty icosa-
hedral 1,12-Ni10E2 cage; a significant increase in the
relative negative charge density on the Ni(s) atoms
would be expected to increase dπ(Ni(s))-π*(CO) back-
bonding and thereby lower the terminal carbonyl fre-
quency of the cluster.
[NMe4]+ salts); it is likewise close to that at δ 2.53 ppm
for the methyl substituent in the 36-electron Fe2(CO)8-
(µ2-BiMe)2.40
A room-temperature 13C{1H} NMR spectrum of
[NMe4]2[2] in acetone-d6 displayed two resonances as-
signed to the [NMe4]+ counterions and the Bi-attached
methyl substituents. Signals due to the carbonyl ligands
were not detected at room temperature on account of
fast terminal-bridge carbonyl exchange. However, a
limiting slow-exchange 13C{1H} NMR spectrum was
obtained at -60 °C; two sharp resonances at 193.8 ppm
(assigned to the terminal carbonyls) and 249.4 ppm
(assigned to the unresolved doubly and triply bridging
carbonyls) were observed along with the two resonances
at -8.7 and 54.3 ppm due to the Bi-attached methyl
substituents and methyl groups of the [NMe4]+ coun-
terions, respectively.
A comparison of the IR spectra of 2 and 3 with the
reported IR spectra of the [Ni11Bi2(CO)18]n- anions (n
) 2-4) is also of interest in assessing the combined
electronic effects of the aforementioned structural dif-
ferences and the varying anionic charge of the Ni-
centered anions. For the [Ni11Bi2(CO)18]n- anions (n )
2-4), Longoni and co-workers20 reported that the 158-
electron dianion (n ) 2) in acetonitrile displayed infra-
red carbonyl absorptions at 1995 (s), 1840 (m), and 1780
(sh, br) cm-1 which were attributed to terminal edge-
bridging, and face-bridging COs, respectively; the 159-
electron trianion (n ) 3) in acetonitrile expectedly
showed much lower frequencies at 1963 (s), 1807 (s),
Similar limiting slow-exchange 13C{1H} NMR spectra
(with respect to terminal-bridge carbonyl exchange)
were previously observed for two [Ni10(EMe)2(CO)18]2-
homologues (E ) P,41 Sb19) as [NMe4]+ salts. The
spectrum at -80 °C for E ) P consisted of two sharp
resonances at 190.2 ppm (terminal COs) and 241.4 ppm
(bridging COs) along with a resonance at 25.8 ppm (P-
attached methyl substituents);41 a corresponding spec-
trum at -60 °C for E ) Sb also gave two sharp
resonances at 191.8 ppm (terminal COs) and 240.8 ppm
(bridging COs) together with a resonance at -4.2 ppm
(Sb-attached CH3 substituents).19
and 1740 (sh, br) cm-1
. The terminal carbonyl fre-
These limiting slow-exchange 13C{1H} spectra are
completely consistent with those obtained at -80 °C
and interpreted by Longoni, Heaton, and Chini42
for [Ni5(CO)12]2-, [Ni6(CO)12]2-, [Ni9(CO)18]2-, and
quency for the 150-electron noncentered Ni10Bi2 carbo-
nyl dianion (2) is 15 cm-1 higher than that for the 158-
electron Ni-centered Ni10Bi2 carbonyl dianion (n ) 2),
but the bridging frequency is 12 cm-1 lower. Because
the overall negative ionic charge (-2) is the same, we
conclude that the net electronic effect resulting from the
formal transformation of 2 into the [Ni11Bi2(CO)18]2-
dianion (via the formal insertion of the d10 Ni(i) into
the icosahedral Ni10Bi2 cavity and the concomitant
removal of the two Bi-attached alkyl substituents) does
not markedly change the average localized negative
charge density on the Ni(s) atoms.
[Ni12(CO)21H4-n]
n- (n ) 2,3); resonances for the terminal
COs ranged from 192 to 203 ppm, while those for the
doubly bridging COs ranged from 230 to 257 ppm.
(d ) Ma ss Sp ectr a l An a lysis. A negative-ion mass
spectrum of [NMe4]2[2] was obtained via the electro-
spray ionization (ESI) method.43 Signals corresponding
to 8 relatively abundant doubly charged ion-fragments
were observed. The assignment of each ion-fragment
formula was based upon the resulting simulated isotopic
pattern matching the observed one for each signal.44 No
signal corresponding to the doubly charged ion-parent,
[M]2-, was detected. Nevertheless, the isotopic patterns
for the 8 observed signals correspond to the doubly
charged [M - 2CH3 - nCO]2- ions (n ) 2-9).
(b) Electr och em ica l An a lyses. Cyclic voltammo-
grams of three homologous [Ni10(EMe)2(CO)18]2- mem-
bers (E ) P, Sb, Bi) in THF (for E ) P) or acetonitrile
(for E ) Sb, Bi) did not exhibit any reversible reduction
out to -1.5 V. However, these CVs each displayed an
irreversible oxidation wave at ca. +0.75 V for E ) P,13
+0.60 V for E ) Sb,19 and +0.70 V for E ) Bi, which
indicates that each cluster undergoes a significant
architectural rearrangement or decomposition upon
oxidation. The observed insignificant variations in the
irreversible oxidation waves are also consistent with the
IR data in indicating that no marked changes in the
surface charge density occur on the icosahedral 1,12-
Ni10E2 cage upon formal substitution of the group V (15)
E atoms.
Several important pieces of structural information
emerge from this observed ion-fragmentation pattern,
which illustrates the mildness of the ionization process.
First, the ion fragments correspond to the sequential
loss of COs, which is frequently found from mass
spectral analyses of fragmentation of neutral metal
carbonyl clusters. Second, the presence of only Ni10Bi2-
containing ion fragments points to the icosahedral Ni10-
Bi2 cage being intact and thereby not being inherently
fragile. Third, the absence of both methyl groups in
these ion fragments may be readily attributed to the
1
(c) NMR Sp ectr a l An a lysis. A H NMR spectrum
of [NMe4]2[2] in acetone-d6 at room temperature showed
signals corresponding to the methyl groups attached to
the bismuth atoms and those of the [NMe4]+ counteri-
ons; the signals integrated closely to the theoretical 6:24
ratio. The [NMe4]+ proton resonance at δ 3.46 ppm is
near that at δ 3.21 ppm for [NMe4]Br. The methyl
substituent signal at 2.37 ppm compares favorably with
that of δ 2.40 ppm for E ) P and that of 1.10 ppm for E
) Sb in the [Ni10(EMe)2(CO)18]2- homologues (also as
(40) Whitmire, K. H.; Shieh, M.; Cassidy, J . Inorg. Chem. 1989, 28,
3164-3170.
(41) Gavney, J . A., J r. Ph.D. Thesis, University of Wisconsins
Madison, 1992.
(42) Longoni, G.; Heaton, B.; Chini, P. J . Chem. Soc., Dalton Trans.
1980, 1537-1541.
(43) (a) Hop, C. E. C. A.; Bakhliar, R. J . Chem. Educ. 1996, 73,
A162-169 and references therein. (b) Hop, C. E. C. A.; Saulys, D. A.;
Gaines, D. F. Inorg. Chem. 1995, 34, 1977-1978.
(44) Arnold, L. Isotope Pattern Calculator, Ver. 1.6, 1990.

Nickel-Bismuth Carbonyl Clusters
Organometallics, Vol. 16, No. 8, 1997 1665
low pressure was found to be unnecessary; an unidentified
orange-yellow residue sublimes at these latter conditions, but
no additional trimethylbismuth was collected. The colorless
distillate was fractionally redistilled (43 °C, 30 min) to remove
the diethyl ether. The second fraction was a clear, colorless
liquid (11.1 g) that was assumed to be a mixture of the desired
product, Me3Bi, and the solvent; on the basis of its measured
density of 1.8 g/mL, this liquid was estimated to be about 67%
pure trimethylbismuth. Yield: 56% based on BiBr3.
Trimethylbismuth was slowly dripped into a solution con-
taining 19.7 g (62.5 mmol) of bismuth trichloride dissolved in
100 mL acetone. After being stirred for 2 h, the solvent was
evaporated and the residue washed several times with diethyl
ether and then dried under vacuum. This procedure gave rise
to 26.5 g (89.8 mmol) of methylbismuth dichloride (96% yield
based on Me3Bi). Methylbismuth dichloride is soluble in
acetone, slightly soluble in THF, and insoluble in lower polarity
solvents.
The identity of MeBiCl2 was established from a mass
spectrum (Kratos MS-25; resolution [5000]; electron-impact
ionization, 40 eV). The sample was introduced into the ion
source via a heated probe. The fragment pattern observed in
the positive-ion mass spectrum includes the ion-parent peak
at m/z 294 ([M]+, 21%) and prominant ion peaks at 279
([BiCl2]+, 70%), 259 ([MeBiCl]+, 80%), 244 ([BiCl]+, 45%), and
209 ([Bi]+, 100%).
relative low binding energy of the two Bi-attached
methyl substituents. In contrast, a negative-ion ESI
mass spectrum of an antimony analogue, [Ni10(Sbi-
Pr)2(CO)18]2-, displayed prominent signals that were
assigned on the basis of their observed and calculated
isotopic patterns to ion-fragment formulas containing
Sb-attached isopropyl substituents.19 These mass spec-
tral differences are completely consistent with calori-
metric results which showed that the bond dissociation
energy of a Bi-alkyl bond is considerably smaller than
that of an Sb-alkyl bond.45 The implications of these
mass spectral results are also in harmony with the
unsuccessful attempts by Longoni and co-workers20 to
either protonate or methylate the Bi atoms in the
[Ni11Bi2(CO)18]n- trianion (n ) 3) or tetraanion (n ) 4),
whereas their effort to methylate the closely related
[Ni11{SbNi(CO)3}2(CO)18]4- by reaction with methyl
iodide was successful in producing the [Ni11(SbMe2)-
(CO)18]2- dianion containing the Ni-centered 1,12-Ni10-
Sb2 cage (i.e., the Ni-centered homologue (E ) Sb)
of 2).
Exp er im en ta l Section
Meth od s a n d Ma ter ia ls. All reactions including sample
transfers and manipulations were carried out with standard
Schlenk techniques either on a preparative vacuum line under
nitrogen or within a Vacuum Atmospheres drybox under
nitrogen atmosphere. The following solvents were freshly
distilled under nitrogen from the indicated appropriate drying
agents immediately prior to use: hexane (CaH2); toluene (Na/
benzophenone); tetrahydrofuran (K/benzophenone); acetone
(CaSO4); acetonitrile (Na2CO3); diisopropyl ether (K/benzophe-
none); diethyl ether (K/benzophenone).
The [NMe4]+ salt of the [Ni6(CO)12]2- dianion (1)6 was
prepared by a modification46 of the general method of Longoni,
Chini, and Cavalieri.6b The following materials were used
without further purification: magnesium metal turnings
(Fisher), carbon tetrachloride (Mallinckrodt), bismuth trichlo-
ride (Aldrich), iodomethane (Aldrich), bromine (Mallinckrodt),
iodoethane (Aldrich).
(b) Eth ylbism u th Dich lor id e. A similar preparation of
ethylbismuth dichloride was carried out. Magnesium turnings
(4.5 g, 180 mmol), a stirbar, and 30 mL of ethyl ether were
added to a two-neck 500 mL Schlenk flask. A reflux condenser
and rubber septum were attached, and after the system was
flushed with nitrogen, 14.7 mL (28.7 g, 180 mmol) of iodoet-
hane dissolved in 30 mL of diethyl ether was added over 20
min. Bismuth tribromide (25.0 g, 55.7 mmol) dissolved in 90
mL of diethyl ether was then added over another 20 min. After
the solution was refluxed for 1 h, the condenser was replaced
with a distillation apparatus, and the diethyl ether was
distilled off. The temperature was then raised to 140 °C, and
under low pressure the reaction vessel was heated to dryness
in order to distill a clear liquid into a receiving flask cooled to
-78 °C. This liquid (14.5 g, 16.7 mL) was assumed to be a
mixture of Et3Bi (density: 1.82 g/mL) and Et2O (0.71 g/mL);
the resulting estimated yield of Et3Bi is 12.4% (2.05 g, 6.92
mmol).
Infrared spectra were obtained on a Mattson Polaris FT-IR
or a Nicolet 740 FTIR spectrophotometer. Solution spectra
were obtained by use of nitrogen-purged CaF2 cells.
This intermediate, Et3Bi, was dissolved in 50 mL of acetone
and slowly added to 4.4 g (13.8 mmol) of BiCl3. The milky
white solution quickly turned yellow; after being stirred for 2
h, the solvent was removed under a nitrogen stream. The
residue was washed with diethyl ether and dried under
vacuum; the resulting solid was identified as EtBiCl2 (5.6 g,
18 mmol; 87% yield based on Et3Bi).
Electrochemical measurements were performed with
a
Bioanalytical Systems electrochemical analyzer in concert with
a Princeton Applied Research electrochemical cell operating
inside of a Vacuum Atmosphere drybox. The cell consisted of
a glassy carbon working electrode, a platinum wire auxiliary
electrode, and a standard calomel electrode (SCE). The voltage
potentials were measured relative to the SCE.
P r ep a r a t ion a n d Ch a r a ct er iza t ion of Icosa h ed r a l
[Ni10(BiR)2(CO)18] -
2- Clu ster s a s [NMe4]+ Sa lts. (a ) [Ni10
2-
(BiMe)2(CO)18
]
Dia n ion (2). Methylbismuth dichloride
P r ep a r a t ion a n d Ch a r a ct er iza t ion of Alk ylb ism u t h
Dich lor id es. (a ) Meth ylbism u th Dich lor id e. Magnesium
turnings (4.5 g, 180 mmol), a stirbar, and 30 mL of diethyl
ether were added to a two-neck 500 mL Schlenk flask equipped
with a reflux condenser. Iodomethane (11.5 mL; 26.1 g, 180
mmol) dissolved in 50 mL of diethyl ether was added over 30
min to create a methyl Grignard reagent. Bismuth tribromide
(25.0 g, 55.7 mmol) dissolved in 70 mL of diethyl ether was
added dropwise to the methyl Grignard reagent at room
temperature over another 30 min. An excess of Grignard
reagent was utilized to prevent the development of partially
methylated bismuth bromides. Completion of reaction was
ensured by a 1-h-long reflux. By use of a distillation ap-
paratus, both diethyl ether and trimethylbismuth were readily
distilled with the receiving flask kept at -78 °C. Although
suggested by Scherer,23 distillation at high temperature and
(145 mg, 0.49 mmol) and 359 mg (0.42 mmol) of [NMe4]2[Ni6-
(CO)12] were placed into a 250 mL Schlenk flask as a solid
mixture. After this mixture was stirred and sonicated to
ensure homogeneity of mixing and then cooled to -78 °C, ca.
50 mL of similarly cooled THF was added via hypodermic
tubing. After the mixture was stirred for 3 h, the temperature
was increased to 0 °C, and the solvent and the Ni(CO)4
byproduct formed were removed. Addition of hexane and then
toluene to the residue gave no soluble species. Most of the
residue was found to be soluble in THF; this brown extract
(160 mg) primarily contained [NMe4]2[Ni10(BiMe)2(CO)18]. In
addition to some insoluble powder, the only other substance
detected to be present was a small amount of acetone-soluble
solid, identified as the unreacted nickel reagent. An estimated
yield of ca. 45% is based upon the following assumed reaction:
2[1] + 2MeBiCl2 f [2] + Ni(CO)4 + NiCl2 + 2Cl- + 2CO
(45) Doak, G. O.; Freeman, L. D. Organometallics Compounds of
Arsenic, Antimony, and Bismuth; Wiley-Interscience: New York, 1970;
pp 1-16.
The desired dianion (2) as the [NMe4]+ salt was crystallized
from acetone/Et2O or THF/Et2O layering mixtures.
(46) Ceriotti, A.; Longoni, G.; Piva, G. Inorg. Synth. 1989, 26, 312-
315.

1666 Organometallics, Vol. 16, No. 8, 1997
Mlynek and Dahl
The stoichiometry of this compound, established unambigu-
ously from the X-ray diffraction study, was confirmed from an
elemental analysis (Desert Analytics, Tucson, AZ). Calcd for
C28H30Bi2N2Ni10O18: C, 19.93; H, 1.79; Bi, 24.77; N, 1.66; Ni,
34.79. Found: C, 20.18; H, 1.61; Bi, 24.51; N, 1.76; Ni, 34.53.
An infrared spectrum of 2 in THF exhibited bands at 2009 (s)
and 1828 (m) cm-1. A magnetic susceptibility measurement
(TME balance) at room temperature unequivocally showed a
powdered sample of 2 as the [NMe4]+ salt to be diamagnetic.
CV measurements of recrystallized samples of [NM4]2[2] in
acetonitrile solution showed an irreversible oxidation at ca.
+0.70 V.
pure fragments and not a mixture, and because in each case
the signal matches that for the pure nonmethylated [Ni10
-
Bi2(CO)x]2- ion fragment and not that for the corresponding
Bi-methylated one, we conclude that it is highly unlikely that
any of the 8 centered signals results from inclusion of the
2-
corresponding Bi-methylated [Ni10(BiMe)2(CO)x-1
ment.
]
ion frag-
(b) [Ni10(BiEt)2(CO)18]2- Dian ion (3). Ethylbismuth dichlo-
ride (140 mg, 0.49 mmol) and 360 mg (0.43 mmol) of [NMe4]2-
[Ni6(CO)12] were placed as a solid mixture into a 250 mL
Schlenk flask. After this mixture was briefly sonicated to
ensure homogeneity of mixing and then cooled to -78 °C,
approximately 50 mL of similarly cooled THF was added via
hypodermic tubing. After the solution was stirred for 3 h, the
temperature was increased to 0 °C, and the solvent was
removed. No soluble hexane or toluene species were obtained
via sequential extractions of the residue with these solvents.
The brown THF extract gave the major product of this reaction
which was identified as [NMe4]2[Ni10(BiEt)2(CO)18]2-. Beside
a copious amount of insoluble material, the only other isolated
material consisted of a small amount of acetone-soluble solid,
which was identified as the unreacted [NMe4]+ salt of 1. The
[NMe4]+ salt of 3 did not crystallize as easily as that of the
methylbismuthinidene 2.
About 0.7 mL of purified acetone-d6 was saturated with
selected crystals of the freshly recrystallized [NMe4]2[2] and
diluted to 1.0 mL in a 5 mm NMR tube, which was then flame-
sealed under vacuum. 13C{1H} NMR spectra were obtained
under the following conditions: the Brucker AM-500 instru-
ment was operated at 125.7 MHz synthesizer frequency, 5.0
µs pulse width, repetition time of 0.42 s, for tens of thousands
pulses at 223 K. The downfield acetone peak at 206.00 ppm
was used as an internal standard.42 The spectrum showed
peaks at δ -8.67 (Bi-CH3), 54.31 (J ) 7.0 Hz, N(CH3)4+),
193.83 (terminal CO), 249.38 ppm (bridging CO).
A
1H
spectrum of [NMe4]2[2] at room temperature showed signals
at δ 2.37 (s, 6H, Bi-CH3) and 3.46 ppm (t, 24H, N(CH3)4,
2J (1H-15N) ) 0.61 Hz).
An infrared spectrum of 3 in THF displayed maxima at 2005
(s) and 1830 (m) cm-1
.
X-r a y Cr ysta llogr a p h ic Deter m in a tion s a n d Refin e-
m en ts. (a ) Gen er a l P r oced u r es. Each crystal was mounted
under an argon atmosphere either onto a thin glass fiber and
coated with epoxy or inside of a Lindemann glass capillary
which was then hermetically sealed. Intensity data were
obtained with graphite-monochromated Mo KR radiation on
a Siemens P4 diffractometer47 equipped with a sealed tube
generator. Intensities of standard reflections showed no
significant variations during the entire collection of data.
Empirical absorption corrections based upon ψ-scan measure-
ments at different azimuthal angles were applied to each data
set. Crystallographic computations were carried out with
SHELXTL PLUS run on a Silicon Graphics Indy work sta-
tion.48 Preliminary positions of the heavy atoms were found
by direct methods, while positions of the other non-hydrogen
atoms were determined from successsive Fourier difference
maps coupled with initial isotropic least squares.49 Final
Fourier difference maps exhibited no unusual features for each
structure. Crystal data, data-collection, and refinement pa-
rameters are presented in Table 1.
Mass spectra of [NMe4]2[2] were obtained with
a VG
AutoSpecE triple-sector (EBE) mass spectrometer (mass range
200-4000) equipped with the Mark II electrospray ionization
(ESI) source and operated in the negative-ion mode with an
accelerating voltage of 4000 eV. The ESI source has larger
orifices for the skimmer cones, increased pumping capacity,
and a hexapole focusing unit to improve sensitivity and ease
of use. Experiments were carried out with a resolution of 2000
and a scan speed of 4.0 s per decade. Nitrogen was used as
both the nebulizer gas (16 L/h) and the bath gas. [NMe4]2[2]
was dissolved in acetonitrile (ca. 10-5 M), and the flow rate
was approximately 15 µL/min. The following prominent
centered doubly charged ion-fragment signals were obtained:
m/z 725.9 ([Ni10Bi2(CO)16]2-, 3%), 711.9 ([Ni10Bi2(CO)15]2-, 7%),
697.8 ([Ni10Bi2(CO)14]2-, 5%), 683.8 ([Ni10Bi2(CO)13]2-, 42%),
669.9 ([Ni10Bi2(CO)12]2-, 100%), 655.9 ([Ni10Bi2(CO)11]2-, 37%),
639.9 ([Ni10Bi2(CO)10]2-, 31%), 627.8 ([Ni10Bi2(CO)9]2-, 15%).
Each of these centered ion signals consists of an observed
isotopic pattern that is consistent with the calculated pattern
based on the assigned ion-fragment formula. For example, the
observed relative values of the middle 11 isotopic mass
components in the isotopic pattern for the signal centered at
m/z 670 are 42:11:76:19:100:25:96:25:72:20:42; these values are
completely consistent with those simulated44 for the isotopic
mass distribution in the [Ni10Bi2(CO)12]2- ion fragment, viz.,
43:8:81:17:100:24:92:24:68:19:42. The other observed and
calculated isotopic patterns of each signal also have a similar
alternation of strong and weak isotopic mass components due
(b) [NMe4]+2[Ni10(BiMe)2(CO)18 2- Dia n ion (2). Crystals
]
of this compound were obtained by various methods, including
the layering of low-polarity solvents such as hexane, diethyl
ether, or diisopropyl ether over high-polarity solutions (such
as THF or acetone) containing the dissolved product. X-ray
data were obtained at -140 °C from a block-shaped crystal,
isolated from an acetone/diisopropyl ether solution, which was
found to be monoclinic. X-ray data were initially collected with
a Siemens P4 diffractometer, from which the crystal structure
was determined and refined (R1(F) ) 0.11). Subsequently,
better crystals were grown, and X-ray data were obtained with
a Siemens P4/CCD diffractometer; the resulting refinement
gave much better crystallographic results. Figure 1 shows the
configuration of the dianion and its Ni10Bi2 icosahedral cage
in the [NMe4]+ salt.
to each ion fragment containing the intact icosahedral Ni10
-
Bi2 cage and to nickel having three abundant even mass
number isotopes (i.e., the five naturally occurring nickel
isotopes are as follows: 58Ni, 68.1%; 60Ni, 26.2%; 61Ni, 1.1%;
62Ni, 3.6%; 64Ni, 0.9%) and bismuth consisting of only one
nuclide (viz., 209Bi, 100%).
The possibility that the observed isotopic pattern of each
(c) [NMe4]+2[Ni10(BiEt)2(CO)18 2- Dia n ion (3). Crystals
]
of this compound were obtained by the layering of hexane over
doubly charged ion-fragment signal may be due to a mixture
of corresponding [Ni10Bi2(CO)x]2- and [Ni10(BiMe)2(CO)x-1
]
2-
ion fragments (x ) 9-16) was investigated. Because the two
masses of each pair of the corresponding two ion fragments
(i.e., with the same x value) are only 2 Da apart, and thus the
two ion fragments would display overlapping isotopic patterns,
a detailed analysis of the observed isotopic pattern for each
signal with simulated isotopic patterns for each of the two
pure ion fragments and for mixtures of the corresponding
(47) Data Collection: SMART Software Reference Manual (1994).
Siemens Analytical X-ray Instruments, 6300 Enterprise Dr., Madison,
WI 53719-1173.
(48) Data Reduction: SAINT Version 4 Software Reference Manual
(1995). Siemens Analytical X-ray Instruments, 6300 Enterprise Dr.,
Madison, WI 53719-1173.
(49) Structure Solution, Refinement and Graphics: Sheldrick, G.
M. SHELXTL Version 5 Reference Manual (1994). Siemens Analytical
X-ray Instruments, 6300 Enterprise Dr., Madison, WI 53719-1173.
Neutral atom scattering factors were taken from: International Tables
for Crystallography; Kluwer: Boston, MA, 1995; Vol. C, Tables 6.1.1.4
and 4.2.6.8.
[Ni10Bi2(CO)x]2- and [Ni10(BiMe)2(CO)x-1
]
2- ion fragments was
carried out. Because in each case the observed isotopic pattern
closely matches the calculated pattern for either of the two

Nickel-Bismuth Carbonyl Clusters
Organometallics, Vol. 16, No. 8, 1997 1667
a THF solution containing the dissolved product to give block-
shaped crystals or by vapor diffusion of acetone solution with
diisopropyl ether solvent to give needle-shaped crystals. X-ray
data were collected at -125 °C from a parallelepiped crystal,
which was found to be triclinic. A statistical analysis of the
data indicated that the probable space group to be Ph1h. Figure
2 shows the configuration of the dianion in the [NMe4]+ salt.
Dept., UWsMadison) for the ESI mass spectrometric
measurements, to David Snyder (Chem. Dept.,
UWsMadison) for the EI mass spectrometric measure-
ments, to Professor Charles P. Gibson (Chem. Dept.,
UWsOshkosh) for the magnetic susceptibility measure-
ments, and to Dr. Randy K. Hayashi and Dr. Douglas
R. Powell (Chem. Dept., UWsMadison) for crystal-
lographic assistance.
Ack n ow led gm en t. The authors thank the National
Science Foundation (Grants CHE-9013059 and CHE-
9310428) for support of this research. Departmental
purchase of the Siemens P4/CCD area detector diffrac-
tometry system and associated computer equipment
utilized in this research was made possible by funds
from the NSF (Grants CHE-9105497 and CHE-9310428)
and the Graduate School of UWsMadison. Addition-
ally, we are greatly indebted to Dr. Marcel Hop (Chem.
Su p p or tin g In for m a tion Ava ila ble: Crystallographic
tables for the [NMe4]+ salts of [Ni10(BiMe)2(CO)18]2- (2) and
[Ni10(BiEt)2(CO)18]2- (3) listing atomic coordinates, anisotropic
thermal displacement parameters, interatomic distances, bond
angles, and idealized hydrogen coordinates (12 pages). Order-
ing information is given on any current masthead page.
OM960965K