First paramagnetic zerovalent transition metal isocyanides. Syntheses, structural characterizations, and magnetic properties of novel low-valent isocyanide complexes of vanadium 1

Article abstract of DOI:10.1021/ja000212w

The first homoleptic paramagnetic transition metal isocyanide, V(CNXyl)₆ (2, Xyl = 2,6-dimethylphenyl), can be isolated in high yield by reacting bis(naphthalene)vanadium(0) (1a) or bis(1-methylnaphthalene)vanadium(0) (1b) with 6 equiv of CNXyl

Full text of DOI:10.1021/ja000212w

4678
J. Am. Chem. Soc. 2000, 122, 4678-4691
First Paramagnetic Zerovalent Transition Metal Isocyanides.
Syntheses, Structural Characterizations, and Magnetic Properties of
Novel Low-Valent Isocyanide Complexes of Vanadium1
Mikhail V. Barybin,*,† Victor G. Young, Jr., and John E. Ellis*
Contribution from the Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
ReceiVed January 20, 2000
Abstract: The first homoleptic paramagnetic transition metal isocyanide, V(CNXyl)6 (2, Xyl ) 2,6-
dimethylphenyl), can be isolated in high yield by reacting bis(naphthalene)vanadium(0) (1a) or bis(1-
methylnaphthalene)vanadium(0) (1b) with 6 equiv of CNXyl in tetrahydrofuran/heptane. Reduction of 2 with
excess cesium graphite in THF affords excellent yields of [V(CNXyl)6]- (3) as an unsolvated Cs+ salt, the
first homoleptic octahedral isocyanide metalate. Cs3 reacts with 2 equiv of 18-Crown-6 to give [Cs(18-Crown-
6)2]3. Anion 3 can also be isolated as a practically insoluble [K(Crypt{2.2.2.})]+ salt by reducing 2 with
potassium naphthalenide in the presence of Crypt{2.2.2.}. Complex 3 reduces [Et3NH]Cl to form paramagetic
2. Oxidation of 2 by ferricinium hexafluorophosphate in THF provides nearly quantitative yields of the 16-
electron paramagnetic [V(CNXyl)6][PF6] (4[PF6]), analogous to the exceedingly unstable [V(CO)6]+. Interaction
of V(CO)6 with excess CNXyl in heptane results in the efficient formation of trans-[V(CO)2(CNXyl)4] (5).
Such a substitution reaction is highly unusual for V(CO)6. Oxidation of compound 5 by ferricinium
hexafluorophosphate in THF affords homoleptic 4[PF6]. Complexes 2, 3, 4, and 5 were characterized by a
variety of spectroscopic methods and X-ray crystallography. Spectroscopic, magnetic, and structural features
of these novel electron rich vanadium isocyanides are discussed in detail. The average V-CN bond length
increases in the series [V(CNXyl)6]- < V(CNXyl)6 < trans-V(CO)2(CNXyl)4 < [V(CNXyl)6]+. Well-resolved
1H and 13C NMR spectra were obtained for paramagnetic 2 and its chromium congener, [Cr(CNXyl)6]+. The
importance of back-bonding in the mechanism of unpaired spin delocalization within 2 was demonstrated.
Contrary to the previous prediction for dπ(M)-pπ*(L) unpaired spin delocalization in low-spin d5 octahedral
complexes, negative spin appears to be induced on the CNXyl ligands of 2 by means of dπ(V)-pπ*(CNXyl)
back-bonding.
Introduction
diamagnetic and have dimeric structures4,5 with a Co-Co bond
supported by two bridging CNR ligands in the solid state, similar
to their carbonyl congener Co2(CO)8.6 Until the work presented
herein, no mononuclear (and, hence, paramagnetic) complexes
of this class were known. The elegant study by Cooper and
Warnock, describing isolation of the initial homoleptic isocya-
nide metalate, [Co(CNXyl)4]-, appeared in 1989.5,7
Despite the original claim that group 5 transition metals “did
not show any tendency to form complexes with isocyanides”,2a
a number of such species have been subsequently reported.
These include Lippard’s [V(CNtBu)6]2+,8 which remained the
sole homoleptic group 5 metal isocyanide prior to our recent
contributions.9 Most vanadium isocyanides, isolated to date,
contain the metal in a positive formal oxidation state.10 In these
Similar to metal carbonyls, complexes of transition metals
with isocyanides have been proven to be valuable reagents in
synthetic chemistry and catalysis.2 One of the first recognized
substantial differences between carbon monoxide and isocya-
nides as ligands was the fact that the former seemed to be ideal
for stabilizing zero- and subvalent metals while the latter were
most common in complexes of metals with higher oxidation
states. As it was pointed out in 1998, “the number of neutral
homoleptic metal carbonyls known clearly exceeds that of the
corresponding isocyanide complexes”.2g The lack of zerovalent
isocyanides of the elements with odd atomic numbers and iso-
cyanide analogues of carbonyl metalates was originally em-
phasized in 1959.2a Only 18 years later the first zerovalent
isocyanides of an “odd” metal, Co2(CNR)8 (R ) tBu, Xyl, 4-Br-
Xyl, Mes), were prepared.3 These neutral compounds are
(3) (a) Barker, G. K.; Galas, A. M. R.; Green, M.; Howard, J. A. K.;
Stone, F. G. A.; Turney, T. W.; Welch, A. J.; Woodward, P. J. Chem. Soc.,
Chem. Commun. 1977, 256. (b) Yamamoto, Y.; Yamazaki, H. Inorg. Chem.
1978, 17, 3111.
(4) Carroll, W. E.; Green, M.; Galas, A. M. R.; Murray, M.; Turney, T.
W.; Welch, A. J.; Woodward, P. J. Chem. Soc., Dalton Trans. 1980, 80.
(5) Leach, P. A.; Geib, S. J.; Corella, J. A., II; Warnock, G. F.; Cooper,
N. J. J. Am. Chem. Soc. 1994, 116, 8566.
(6) Sumner, G. G.; Klug, H. P.; Alexander, L. E. Acta Crystallogr. 1964,
17, 732.
(7) Warnock, G. F.; Cooper, N. J. Organometallics 1989, 8, 1826. See
also ref 5.
† Present address: Department of Chemistry, MIT, Cambridge, MA
02139.
(1) Highly Reduced Organometallics. Part 53. Part 52: Chen, Y. S.; Ellis,
J. E. Inorg. Chim. Acta. In press.
(2) For reviews on transition metal complexes with isocyanides see: (a)
Malatesta, L. Progr. Inorg. Chem. 1959, 1, 283. (b) Malatesta, L.; Bonati,
F. Isocyanide Complexes of Transition Metals; Wiley: New York, 1969.
(c) Treichel, P. M. AdV. Organomet. Chem. 1973, 11, 21. (d) Bonati, F.;
Minghetti, G. Inorg. Chim. Acta 1974, 9, 95. (e) Yamamoto, Y. Coord.
Chem. ReV. 1980, 32, 193. (f) Singleton, E.; Oosthuizen, H. E. AdV.
Organomet. Chem. 1983, 22, 209. (g) Weber, L. Angew. Chem., Int. Ed.
1998, 37, 1515.
(8) (a) Silverman, L. D.; Dewan, J. C.; Giandomenico, C. M.; Lippard,
S. J. Inorg. Chem. 1980, 19, 3379. (b) Silverman, L. D.; Corfield, P. W.
R.; Lippard, S. J. Inorg. Chem. 1981, 20, 3106.
10.1021/ja000212w CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/27/2000

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4679
Experimental Section
complexes the CNR ligands function primarily as σ-donors.2g,10e
Given the diversity of group 6 metal(0) isocyanides,2e the lack
of any V(0) compounds possessing at least one isocyanide ligand
is striking. It is worth mentioning that cyclic voltammograms
of [V(CNtBu)6]2+ revealed one reversible (E1/2 ) -0.95 V) and
one quasireversible (E1/2 ≈ -1.5 V) wave assigned to
[V(CNtBu)6]2+/+ and [V(CNtBu)6]+/0 couples, respectively.11 No
other data on the otherwise elusive species [V(CNtBu)6]+/0 are
available. Complexes cis-[V(NO)2(CNR)4]+,12a CpV(NO)2-
(CNR),12b VX(NO)2(CNR)3,12c and [V(CO)5(CNR)]- 13 repre-
sent the only examples of subvalent vanadium isocyanides. On
the basis of usual conventions,2 the above nitrosyl-isocyanide
compounds are formally classified as derivatives of V(1-).
However, as evidenced by IR data in the νCN region, the CNR
ligands in these complexes are not much different electronically
from those in the V(2+) compounds. Indeed, the C-N
stretching bands of, for example, [V(NO)2(CNtBu)4]+ occur at
General Procedures, Starting Materials, and Equipment. All
operations were performed under an atmosphere of 99.5% argon further
purified by passage through columns of activated BASF catalyst and
molecular sieves. All connections involving the gas purification systems
were made of glass, metal, or other materials impermeable to air. Solu-
tions and slurries were transferred via stainless steel needles (cannulas)
whenever possible. Standard Schlenk techniques were employed with
a double manifold vacuum line.16 Solvents were freed of impurities by
standard procedures and stored under argon. Naphthalene was sublimed
under vacuum and 1-methylnaphthalene was distilled under reduced
pressure from sodium metal. Literature procedures were employed to
prepare VCl3(THF)3,17 V(C10H8)2 (1a),18 V(CO)6,19 Cr(CNXyl)6 (6),20
and CsC8.21 Other reagents were obtained from commercial sources
and were freed of oxygen and moisture before use.
Solution infrared spectra were recorded on a Mattson Galaxy 6021
FTIR spectrometer with samples sealed in 0.1 mm gastight CaF2 cells.
Nujol (mineral oil) mulls of all products for IR spectra were prepared
in a Vacuum Atmospheres Corp. drybox. NMR samples were sealed
under argon into 5 mm tubes and were analyzed on a Varian Unity-
300 spectrometer. 1H and 13C chemical shifts are given with reference
to residual 1H and 13C solvent resonances relative to TMS. Such
referencing eliminated bulk susceptibility effects for paramagnetic
samples.22 19F and 31P chemical shifts are reported relative to external
CFCl3 and 85% H3PO4, respectively. In the variable-temperature NMR
studies, samples were allowed to equilibrate for at least 20 min at a
given temperature prior to data acquisition. The temperature indicated
by the console was checked using known temperature dependence of
the peak separation of methanol.23 Solutions for ESR experiments were
prepared in the drybox and placed in quartz tubes sealed with septa
and Parafilm. The data were recorded using an IBM-Bruker ESP-300
spectrometer. A small Dewar flask was inserted into the microwave
cavity to obtain measurements at the liquid N2 temperature. ESR spectra
were referenced to the standard DPPH, 2,2-bis(4-tert-octylphenyl)-1-
picrylhydrazyl (g ) 2.0036 at 25 °C). Melting points are uncorrected
and were determined for samples in sealed under argon capillaries on
a Thomas-Hoover Unimelt apparatus. Microanalyses were carried out
by H. Malissa and G. Reuter Analytische Laboratorien, Lindlar,
Germany.
2199 and 2183 cm-1 12a
cm-1 8a
,
while [V(CNtBu)6]2+ absorbs at 2190
.
The well-established propensity of group 5 transition metals
to induce reductive coupling and polymerization of isocya-
nides10a,14 together with the fact that a highly reduced metal
center should facilitate carbon-carbon bond formation between
CNR molecules14a may explain the scarcity of low-valent
vanadium isocyanides. In this article, we report on the chemistry
of new thermally stable electron rich isocyanide complexes of
vanadium. Of particular significance are the first paramagnetic
zerovalent metal isocyanides, V(CNXyl)6 and trans-V(CO)2-
(CNXyl)4 (Xyl ) 2,6-dimethylphenyl). Remarkably, the former
17-electron homoleptic compound is the only binary octahedral
paramagnetic metal(0) complex for which well-resolved NMR
spectra have been recorded.15 Analysis of the 1H and 13C NMR
patterns, obtained for V(CNXyl)6, in terms of unpaired spin
delocalization is presented. Also described are the first 16- and
18-electron homoleptic isocyanide complexes of vanadium,
namely [V(CNXyl)6]+ and [V(CNXyl)6]-. The latter ion and
its tantalum congener, [Ta(CNXyl)6]-,9b constitute the first
binary octahedral isocyanide metalates.
Magnetic Susceptibility Measurements. Solid-state volume mag-
netic susceptibilities were measured on a Johnson Matthey MSB-AUTO
balance at ambient temperature and converted into the corresponding
molar susceptibilities in the usual way.24 Air-sensitive samples were
packed into stoppered gastight tubes (0.400 cm o.d. × 0.324 cm i.d.)
to a depth of ca. 3.0 cm in the drybox. The air correction of 0.029 ×
10-6 was applied to volume susceptibilities of the samples packed under
argon. Diamagnetic corrections, applied to the molar susceptibilities
of the paramagnetic substances, are reported as ødiam. The latter
corrections were obtained experimentally and/or were calculated using
the standard Pascal’s constants.24 The molar susceptibilities of pure
2,6-dimethylphenyl isocyanide, CNXyl, and Cr(CNXyl)6 were measured
to be -71.6 × 10-6 and -452.3 × 10-6 cm3 mol-1, respectively. Unless
otherwise specified, the effective magnetic moments are reported for
solid samples. Solution magnetic susceptibilities were obtained by the
(9) (a) Barybin, M. V.; Young, V. G., Jr.; Ellis, J. E. J. Am. Chem. Soc.
1998, 120, 429. (b) Barybin, M. V.; Young, V. G., Jr.; Ellis, J. E. J. Am.
Chem. Soc. 1999, 121, 9237.
(10) (a) Rehder, D.; Bo¨ttcher, C.; Collazo, C.; Hedelt, R.; Schmidt, H.
J. Organomet. Chem. 1999, 585, 294. (b) Bo¨ttcher, C.; Schmidt, H.; Rehder,
D. J. Organomet. Chem. 1999, 580, 72. (c) Davies, S. C.; Hughes, D. L.;
Janas, Z.; Jerzykiewicz, L.; Richards, R. L.; Sanders, J. R.; Sobota, P. Chem.
Commun. 1997, 14, 1261. (d) Nomura, K.; Schrock, R. R.; Davis, W. M.
Inorg. Chem. 1996, 35, 3695. (e) Bo¨ttcher, C.; Rodewald, D.; Rehder, D.
J. Organomet. Chem. 1995, 496, 43. (f) Petillon, F. Y.; Schollhammer, P.;
Talarmin, J. J. Organomet. Chem. 1991, 411, 159. (g) Carafiglio, T.;
Floriani, C.; Chiesi-Villa, A.; Guastini, C. Inorg. Chem. 1989, 28, 4417.
(h) Anderson, S. J.; Wells, F. J.; Wilkinson, G.; Hussain, B.; Hursthouse,
M. B. Polyhedron 1988, 7, 2615. (i) Coville, N. J.; Harris, G. W.; Rehder,
D. J. Organomet. Chem. 1985, 293, 365. (j) Fachinetti, G.; Del Nero, S.;
Floriani, C. J. Chem. Soc., Dalton Trans. 1976, 1046. (k) Crociani, B.;
Nicolini, M.; Richards, R. L. J. Organomet. Chem. 1975, 101, C1.
(11) Shah, S. S.; Maverick, A. W. J. Chem. Soc., Dalton Trans. 1987,
2881.
(16) (a) Shriver, D. F.; Drezdon, M. A. The Manipulation of Air-SensitiVe
Compounds, 2nd ed.; Wiley-Interscience: New York, 1986. (b) Ellis, J. E.
ACS Symp. Ser. 1987, 357, 34.
(12) (a) Herberhold, M.; Trampisch, H. Inorg. Chim. Acta 1983, 70, 143.
(b) Herberhold, M.; Trampisch, Z. Naturforsch. 1982, 37b, 614. (c)
Na¨umann, F.; Rehder, D. Z. Naturforsch. 1984, 39b, 1654.
(13) (a) Ellis, J. E.; Fjare, K. L. Organometallics 1982, 1, 898. (b) Ihmels,
K.; Rehder, D. Organometallics 1985, 4, 1340.
(17) Manzer, L. E. Inorg. Synth. 1982, 21, 135.
(18) Pomije, M. K.; Kurth, C. J.; Ellis, J. E.; Barybin, M. V. Organo-
metallics 1997, 16, 3582.
(19) Ellis, J. E.; Faltynek, R. A.; Rochfort, G. L.; Stevens, R. E.; Zank,
G. A. Inorg. Chem. 1980, 19, 1082.
(14) (a) Carnahan, E. M.; Protasiewicz, J. D.; Lippard, S. J. Acc. Chem.
Res. 1993, 26, 90 and references therein. (b) Collazo, C.; Rodewald, D.;
Schmidt, H.; Rehder, D. Organometallics 1996, 15, 4884 and references
therein. (c) Rehder, D.; Gailus, H. Trends Organomet. Chem. 1994, 1, 397
and references therein.
(20) Yamamoto, Y.; Yamazaki, H. J. Organomet. Chem. 1985, 282, 191.
(21) Bergbreiter, D. E.; Killough, J. M. J. Am. Chem. Soc. 1978, 100,
2126.
(22) La Mar, G. N.; Horrocks, W. D., Jr.; Holm, R. H., Eds. NMR of
Paramagnetic Molecules; Academic Press: New York, 1973.
(23) (a) Van Geet, A. Anal. Chem. 1970, 42, 679. (b) Van Geet, A. Anal.
Chem. 1968, 40, 2227.
(24) (a) Earnshaw, A. Introduction to Magnetochemistry; Academic
Press: New York, 1968. (b) Kahn, O. Molecular Magnetism; VCH
Publishers: New York, 1993.
(15) For instance, an 1H NMR spectrum of Co(PPh2Me)4, a homoleptic
paramagnetic metal(0) complex, has been obtained (La Mar, G. N.; Sherman,
E. O.; Fuchs, G. A. J. Coord. Chem. 1971, 1, 289). The tetrahedral geometry
of this molecule has been assumed solely based on the very narrow line
1
width of the H NMR peaks.

4680 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Barybin et al.
Evans method25 on a Varian Unity 300 instrument at 23 °C. The
difference in the solution and solvent densities was disregarded, as
usual.23 Mass susceptibilities of pure solvents were either measured
on the Johnson Matthey MSB-AUTO balance or taken from a literature
source.26
H, 5.43; N, 8.76. IR (THF): νCN 2028 vw br, 1823 vs br, 1772 sh br
cm-1. IR (Nujol mull): νCN 1999 vw, 1856 s sh, 1802 vs br cm-1. 1H
NMR (300 MHz, CD3CN, 25 °C): δ 2.39 (s, 6H, o-CH3), 6.69 (t, 3JH-H
) 7.5 Hz, 1H, p-H), 6.90 (d, 3JH-H ) 7.5 Hz, 2H, m-H) ppm. 13C{1H}
NMR (75.5 MHz, CD3CN, 25 °C): δ 20.3 (o-CH3), 122.9 (p-C), 128.5
(m-C), 133.2 (o-C) ppm.
Synthesis of V(C10H7Me)2 (1b). A deep green solution of NaC10H7-
Me was prepared by transferring a solution of 1-methylnaphthalene
(84 mmol, 12 mL) in 200 mL of THF into a flask containing sodium
metal (65.3 mmol, 1.50 g), cut into ca. 20 small pieces, and a large
glass covered magnetic stir bar. After being stirred for 5 h in the dark
at room temperature, the mixture was cooled to -55 °C. Then a cold
(-55 °C) solution/slurry of VCl3(THF)3 (16.1 mmol, 6.00 g) in 150
mL of THF was added via cannula to the cold NaC10H7Me solution.
The resulting brown mixture was slowly warmed to room temperature
for 15 h with constant stirring. Subsequently, the mixture was cooled
to 0 ˚C. Alumina (45 g) was added by a bent Schlenk tube at 0 ˚C
over a period of 10 min to give a dark red slurry. This slurry was
vigorously stirred for 2 h at 0 °C and then filtered through a 3 cm
column of alumina. The filter cake was washed with additional THF
until the washings were colorless. All THF was removed in vacuo from
the deep maroon-red filtrate, giving a red-brown solution/slurry of
V(C10H7Me)2 in 1-methylnaphthalene. The excess ligand was carefully
removed by short-path distillation (10-2 Torr, 40-50 °C). The brown
residue was recrystallized from toluene/pentane at -78 °C, resulting
in a 43% yield of finely divided dark brown 1b (6.95 mmol, 2.33 g).
Anal. Calcd for C22H20V: C, 78.80; H, 6.01. Found: C, 78.25; H, 5.86.
µeff(24 °C) ) 1.69 µB (ødiam ) -219 × 10-6 cm3 mol-1).
Synthesis of [Cs(18-Crown-6)2]3. A cold (-60 °C) solution of 18-
Crown-6 (3.78 mmol, 1.00 g) in 60 mL of THF was transferred into a
flask containing a cold (-60 °C) solution of Cs3 (1.03 mmol, 1.00 g)
in 100 mL of THF via cannula. The reaction turned deep purple-violet
and was stirred for ca. 15 h while slowly warming to room temperature.
An additional 50 mL of THF was used to completely dissolve a small
amount of green microcrystalline precipitate formed and the mixture
was filtered through a medium porosity frit. Heptane (100 mL) was
added to the filtrate and about 130 mL of the solvent was removed in
vacuo to give a dichroic green-violet precipitate. The solid was filtered
off, washed with pentane (5 × 30 mL), and dried under vacuum.
Recrystallization from THF/pentane followed by drying in vacuo for
18 h (important!) afforded [Cs(18-Crown-6)2]3 (0.90 mmol, 1.350 g)
in an 87% yield as a finely divided dark umber microcrystalline solid.
Mp: 193-195 °C dec. Anal. Calcd for C78H102N6CsO12V: C, 62.48;
H, 6.86; N, 5.60. Found: C, 62.29; H, 6.74; N, 5.71. IR (THF): νCN
2021 vw br, 1872 s sh, 1818 vs br, νCO 1113 s cm-1. IR (Nujol mull):
1
νCN 2021 vw, 1874 s sh, 1801 vs br, νCO 1111 s cm-1. H NMR (300
MHz, CD3CN, 25 °C): δ 2.41 (s, 6H, o-CH3), 3.51 (s, 8H, 18-C-6),
6.66 (t, 3JH-H ) 7.5 Hz, 1H, p-H), 6.92 (d, 3JH-H ) 7.5 Hz, 2H, m-H)
ppm. 13C{1H} NMR (75.5 MHz, CD3CN, 25 °C): δ 20.3 (o-CH3), 71.2
(18-C-6), 122.9 (p-C), 128.4 (m-C), 133.8 (o-C) ppm.
Synthesis of V(CNXyl)6 (2). A clear ice-cold solution of 2,6-
dimethylphenyl isocyanide(CNXyl, 46.6 mmol, 6.11 g) in heptane/THF
(70 mL/30 mL) was added to an ice-cold solution of 1b (6.62 mmol,
2.22 g) in heptane/THF (130 mL/130 mL) via cannula. The reaction
mixture changed from a dark red-brown to a deep purple during the
addition. The mixture was stirred for 15 h while slowly warming to
room temperature. Ca. 150 mL of the solvent was removed in vacuo,
resulting in a dichroic green-purple microcrystalline precipitate. The
solid was filtered off, washed with pentane (3 × 20 mL), and dried in
vacuo. Recrystallization from THF/heptane afforded very air- and
moisture-sensitive microcrystalline golden-purple 2 (5.37 mmol, 4.50
g) in an 81% yield. A 91% yield of 2 was obtained using a smaller
reaction scale (2.62 mmol, 0.88 g of V(C10H7Me)2). Mp: 159-160 °C
dec. Anal. Calcd for C54H54N6V: C, 77.40; H, 6.50; N, 10.03; V, 6.08.
Found: C, 77.23; H, 6.52; N, 9.93; V, 6.06. IR (THF): νCN 2008 w
sh, 1939 vs, 1852 w sh cm-1. IR (Nujol mull): νCN 2006 w sh, 1929
Synthesis of [K(Crypt{2.2.2})]3. A cold solution of 2 (1.19 mmol,
1.000 g) in 50 mL of THF was rapidly added to a cold green solution/
slurry of [K(Crypt{2.2.2})][C10H8] at -70 °C. The latter was prepared
by stirring potassium metal (1.20 mmol, 0.047 g), naphthalene (2.42
mmol, 0.310 g), and Cryptand{2.2.2} in 50 mL of THF for 4 h. The
reaction mixture turned dark violet and was stirred for 15 h while slowly
warming to room temperature. A green microcrystalline precipitate
formed. The mixture was filtered to provide an iridescent green solid.
The solid was washed with THF (150 mL), pentane (5 × 15 mL), and
ether (3 × 30 mL) and dried in vacuo for 4 h to afford free green
microcrystalline [K(Crypt{2.2.2})]3 (1.14 mmol, 1.430 g) in a 96%
yield. Mp: 238-239 °C dec, darkens at 230 °C. IR (HMPA): νCN
2022 vw br, 1998 vw br, 1874 s sh, 1810 vs br cm-1. IR (Nujol mull):
νCN 2022 vw, 1999 vw, 1875 s sh, 1764 vs br cm-1
.
Reaction of Cs3 with [Et3NH]Cl. A cold (-60 °C) dark brown
solution of Cs3 (0.515 mmol, 0.500 g) in 60 mL of THF was rapidly
transferred into a flask containing a cold (-60 °C) suspension of
[Et3NH]Cl (0.515 mmol, 0.071 g) in 20 mL of THF. The mixture turned
deep purple within 2 h of stirring at -60 °C. Then the reaction was
warmed to room temperature and the mixture was filtered. Heptane
(100 mL) was added to the filtrate, which was concentrated to about
20 mL to give a green-purple precipitate. The solid was filtered off,
washed with pentane (2 × 10 mL), and dried under vacuum for 2 h to
give a 75% yield of the product (0.388 mmol, 0.325 g), which was
spectroscopically (IR) and magnetically identical with genuine 2.
Synthesis of [V(CNXyl)6][PF6] (4[PF6]). A cold (-70 °C) solution
of 2 (1.07 mmol, 0.900 g) in 80 mL of THF was transferred to a cold
(-70 °C) suspension of [Cp2Fe][PF6] (1.07 mmol, 0.357 g) in 20 mL
of THF. The reaction mixture turned deep orange-red and was stirred
for 15 h while slowly warming to room temperature. Heptane (70 mL)
was introduced into the reaction flask and ca. 85 mL of the solvent
was removed in vacuo to give a deep red microcrystalline precipitate.
The solid was filtered off, washed with pentane (3 × 20 mL) until the
washings were colorless, and dried in vacuo. Recrystallization from
THF/pentane followed by drying in vacuo for 3 h afforded dark red
microcrystalline 4[PF6] (0.99 mmol, 0.972 g) in a 92% yield. Mp: 179-
181 °C dec, darkens at 150 °C. Anal. Calcd for C54H54N6F6PV: C,
65.98; H, 5.54; N, 8.55. Found: C, 65.71; H, 5.48; N, 8.40. IR (THF):
1
vs br, 1845 m sh cm-1. H NMR (300 MHz, C6D5CD3, 19.9 °C): δ
3
-2.42 (t, JH-H ) 7.2 Hz, 1H, p-H), 9.51 (s, 6H, o-CH3), 13.96 (d,
3JH-H ) 7.2 Hz, 2H, m-H) ppm. 13C{1H} NMR (75.5 MHz, C6D5CD3,
19.8 °C): δ -93.9 (W1/2 ) 112 Hz, i-C), -4.5 (W1/2 ) 8.8 Hz, o-CH3),
101.6 (W1/2 ) 9 Hz, m-C), 178.4 (W1/2 ) 21 Hz, p-C), 339.0 (W1/2
)
103 Hz, o-C) ppm. µeff(25 °C) ) 1.76 µB (ødiam ) -452.3 × 10-6 cm3
mol-1). Substitution of V(C10H8)2 (1a) for 1b in the above synthesis
resulted in practically identical yields of 2.
Synthesis of Cs[V(CNXyl)6] (Cs3). A cold (-50 °C) dark purple
solution of 2 (2.36 mmol, 1.980 g) in 250 mL of THF was added to an
efficiently stirred cold (-50 °C) orange-brown suspension of excess
CsC8 (4.78 mmol, 1.094 g) in 50 mL of THF via cannula. The reaction
turned dark brown and was stirred for ca. 18 h while slowly warming
to room temperature. Then the mixture was filtered to give a greenish-
black filter cake and a dark brown filtrate. The filter cake was washed
with additional THF (ca. 30 mL) until the washings were colorless.
Heptane (200 mL) was added to the filtrate and about 300 mL of the
solvent was removed in vacuo to afford a microcrystalline brown
precipitate, which was filtered off and dried under vacuum. Recrys-
tallization from THF/heptane followed by drying in vacuo for 4 h
resulted in an 84% yield (1.99 mmol, 1.930 g) of Cs3 as an extremely
air- and moisture-sensitive microcrystalline bright brown solid. Anal.
Calcd for C54H54N6CsV: C, 66.80; H, 5.61; N, 8.66. Found: C, 66.47;
ν
CN 2142 w, 2033 vs, 1995 m, νPF 844 vs cm-1. IR (Nujol mull): νCN
(25) (a) Evans, D. F. J. Chem. Soc. 1959, 2003. (b) Crawford, T. H.;
Swanson, J. J. Chem. Educ. 1971, 48, 382. (c) Sur, S. K. J. Magn. Reson.
1989, 82, 169.
(26) Weast, R. C., Ed. Handbook of Chemistry and Physics; CRC
Press: Cleveland, OH, 1976.
1
2143 w, 2107 m sh, 2018 vs br, 1984 s, νPF 840 vs cm-1. H NMR
(300 MHz, THF-d8, 22 °C): δ 2.35 (s, 6H, o-CH3), 7.13 (s br, 3H, m-
and p-H) ppm. 13C{1H} NMR (75.5 MHz, THF-d8, 22 °C): δ 19.0
(o-CH3), 129.1 (p-C), 129.2 (m-C), 129.3 (i-C), 134.2 (o-C) ppm. 31P

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4681
1
NMR (121.4 MHz, THF-d8, 21 °C): δ -142.9 (septet, JP-F ) 709
carefully layering cold (0 °C) pentane over a cold (0 °C) THF solution
of the appropriate compound. Crystals of Cs3 were obtained by layering
cold (0 °C) pentane over a cold (0 °C) DME solution of Cs3. All
manipulations with the crystals prior to transfer to the goniometer were
performed in an argon filled glovebag to protect the samples from the
atmosphere. A selected crystal was coated with viscous oil, attached
to a glass fiber, and mounted on the Siemens SMART Platform CCD
diffractometer for a data collection (Mo KR radiation with λ ) 0.71073
Å was employed). An initial set of cell constants was calculated from
reflections harvested from three sets of 20 frames. These initial sets of
frames were oriented such that orthogonal wedges of reciprocal space
were surveyed. Final cell constants were calculated from the actual
data collection. During the data collection a randomly oriented region
of reciprocal space was surveyed to the extent of 1.3 hemispheres to a
resolution of 0.84 Å. Three major swaths of frames were collected with
0.30° steps in ω. Space groups were determined based on systematic
absences and intensity statistics. Successful direct-methods solutions
were calculated, which provided most non-hydrogen atoms from the
E-maps. Several full-matrix least squares/difference Fourier cycles were
performed to locate the remainder of the non-hydrogen atoms. All non-
hydrogen atoms were refined anisotropically. All hydrogen atoms were
placed in ideal positions and refined as riding atoms with individual
or group isotropic displacement parameters. For Cs3, one relatively
large difference peak (2.39 e Å-3) was found near Cs(2). This peak
appears to be spurious. For 5, there were two space groups available,
P4nc and P4/mnc in Laue class 4/mmm. The mean |E2 - 1| ) 0.809
favored the noncentrosymmetric space group P4nc.29 The Flack absolute
structure parameter, x, of 0.01(3) indicated30 that the specimen was
enantiomerically pure and its absolute configuration was chosen
correctly. All calculations were performed on SGI INDY R4400-SC
or Pentium computers using the SHELXTL V5.0 suite of programs
(SHELXTL-Plus V5.0, Siemens Industrial Automation, Inc, Madison,
WI). For all structures, the program SADABS was used to correct the
data for absorption. The program PLATON/SQUEEZE31 was employed
to take into account the effects of disordered solvent in the cases of
Cs3 and 4[PF6]‚THF. Crystal data, data collection, solution, and
refinement information for 2, Cs3, 4[PF6]‚THF, and 5 is provided in
Table 1. Additional crystallographic details can be found in the
Supporting Information.
Hz) ppm. 19F NMR (282.2 MHz, THF-d8, 21 °C): δ -74.35 (d, 1JF-P
) 710 Hz) ppm. µeff(24 °C) ) 2.74 µB (ødiam ) -519.3 × 10-6 cm3
mol-1).
Synthesis of trans-V(CO)2(CNXyl)4 (5). A solution of V(CO)6 (2.28
mmol, 0.500 g) in 150 mL of heptane was added to a solution of 2,6-
dimethylphenyl isocyanide (16.01 mmol, 2.100 g) in 100 mL of heptane
at room temperature in a flask open to a mercury bubbler. The reaction
rapidly changed from a yellow-green color of V(CO)6 to an orange
and then to an orange-brown. An extensive gas evolution was observed.
In about 7 h, a deep red solution/slurry formed. The mixture was
allowed to stir in the dark for a total of 72 h at room temperature. All
but 40 mL of heptane was removed in vacuo and the slurry was filtered
to collect a bright red solid, which was washed thoroughly with pentane
(5 × 15 mL) until the washings were pale orange and dried under
vacuum. This solid was redissolved in 150 mL of toluene to give a
deep orange-red solution and filtered. Removal of all toluene from the
filtrate and recrystallization from THF/heptane afforded lustrous deep
mulberry, microcrystalline 5 (1.82 mmol, 1.150 g) in an 80% yield.
Compound 5 decomposed without melting above ca. 100 °C to a black
material. Anal. Calcd for C38H36N4O2V: C, 72.26; H, 5.74; N, 8.87.
Found: C, 71.93; H, 5.76; N, 8.93. IR (toluene): νCN 2011 m, 1984
vs, νCO 1889 w br, 1840 vs cm-1. IR (Nujol mull): νCN 2011 m, 1986
vs, νCO 1897 vw, 1808 vs cm-1. µeff(22 °C) ) 1.75 µB (ødiam ) -336.2
× 10-6 cm3 mol-1). ESR (5mM solution in toluene glass, -196 °C):
gav ) 2.023 (g ) 2.038, g| ) 1.993); Aav ) 47 G (A ) 31 G, A )
||
80 G).
Oxidation of 5 with [Cp2Fe]+ in the Presence of CNXyl. A solution
of 5 (0.791 mmol, 0.500 g) and CNXyl (2.371 mmol, 0.311 g) in 60
mL of THF was transferred to a suspension of [Cp2Fe][PF6] in 10 mL
of THF at 0 °C. A moderate gas evolution was observed. The mixture
was stirred for 15 h while slowly warming to room temperature. The
resulting deep orange solution was filtered and the filtrate was
concentrated to ca. 20 mL. Addition of heptane (100 mL) produced a
deep red precipitate and an orange solution. The solid was filtered off,
washed with pentane (4 × 15 mL), and dried under vacuum to afford
an 80% yield of free flowing dark red 3[PF6], which was identical with
the bona fide [V(CNXyl)6][PF6].
Synthesis of [Cr(CNXyl)6][BF4] (7[BF4]). Microcrystalline golden-
yellow 7[BF4] was isolated in a 91% yield by oxidizing Cr(CNXyl)6
(6) with 1 equiv of Ag[BF4] in O2-free acetone.27 This procedure is a
variation of the method established by Treichel and Essenmacher to
prepare other Cr(1+) aryl isocyanide complexes.28 Mp: 190-192 °C.
IR (THF): νCN 2055 vs, 1996 w sh cm-1. 1H NMR (300 MHz, CDCl3,
19.9 °C): δ -1.48 (t, 3JH-H ) 7.5 Hz, 1H, p-H), 8.74 (s, 6H, o-CH3),
Results and Discussion
Synthetic Work. Recently, we have established a conven-
tional synthesis of bis(naphthalene)vanadium(0), V(C10H8)2
(1a).18 This undeservedly forgotten32 substance reacted smoothly
with carbon monoxide under very mild conditions to afford
V(CO)6,18 and, therefore, seemed to be a very promising labile
precursor to other compounds of zerovalent vanadium. Because
of the rather modest yields (14-29%) obtained in the prepara-
tion of 1a,18,27 synthesis of its methylated derivative, bis(1-
methylnaphthalene)vanadium(0), V(C10H7Me)2 (1b), was at-
tempted in hopes of a more efficient generation of a synthon
for low-spin atomic vanadium. Thus, “overreduction” of VCl3-
(THF)3 with 4 equiv of NaC10H7Me at -55 °C,18 followed by
addition of alumina, provided a dark red solution from which
pure dark brown 1b was reproducibly isolated in a 43% yield.
The substantially better yield of 1b (43%) compared to that of
1a (14%) under identical experimental conditions (magnetic
stirring) may be attributed to a greater solubility of NaC10H7-
Me vs NaC10H8 in THF at low temperature. The above facile
synthesis of 1b compares favorably with Timms’ original
12.53 (d, JH-H ) 7.5 Hz, 2H, m-H) ppm. 13C{1H} NMR (75.5 MHz,
3
CDCl3, 19.9 °C): δ -32.9 (W1/2 ) 50 Hz, i-C), -0.8 (W1/2 ) 6 Hz,
o-CH3), 90.6 (W1/2 ) 12 Hz, m-C), 179.6 (W1/2 ) 16 Hz, p-C), 224.2
(W1/2 ) 28 Hz, o-C) ppm. µeff(26 °C) ) 1.92 µB (ødiam ) -491.3 ×
10-6 cm3 mol-1). Magnetic and spectroscopic properties of this
substance are fully in accord with the assigned formulation.
Synthesis of [Cr(CNXyl)6][PF6]2 (8[PF6]2). Microcrystalline yellow-
orange 8[PF6]2 was isolated in a 73% yield by oxidizing 6 with 2 equiv
of Ag[PF6] in CH2Cl2.27 This synthesis is based on the procedure
established for the preparation of other Cr(2+) aryl isocyanides.28
Mp: 242-244 °C dec, darkens at 218 °C. IR (CH2Cl2): νCN 2140 vs
cm-1. IR (Nujol mull): νCN 2152 vs sh, 2138 vs, 1962 vw cm-1. H
1
NMR (300 MHz, CD2Cl2, 21.0 °C): δ -15.87 (s, br, 1H, p-H), 19.06
3
1
(d, JH-H ) 6.9 Hz, 2H, m-H), 19.10 (s, 6H, o-CH3) ppm. H NMR
(300 MHz, CD2Cl2, -63.5 °C): δ -26.51 (s, br, 1H, p-H), 23.94 (s,
br, 2H, m-H), 26.60 (s, br, 6H, o-CH3) ppm. 31P NMR (121.4 MHz,
CD2Cl2, 21.0 °C): δ -143.3 (septet, 1JP-F ) 711 Hz) ppm. 19F NMR
1
(282.2 MHz, CD2Cl2, 21.0 °C): δ -73.88 (d, JF-P ) 711 Hz) ppm.
µeff(26 °C) ) 3.02 µB (ødiam ) -586 × 10-6 cm3 mol-1). Magnetic
and spectroscopic properties of this substance are fully consistent with
the assigned formulation.
(29) Stout, G. H.; Jensen, L. H. X-ray Structure Determination; John
Willey & Sons: New York, 1989.
(30) Flack, H. D. Acta Crystallogr. 1983, A39, 876.
(31) Spek, A. L. Acta Crystallogr. 1990, A46, C34.
X-ray Crystallographic Characterization of 2, Cs3, 4[PF6], and
5. X-ray quality crystals of 2, 4[PF6]‚THF, and 5 were grown by
(32) Complex 1a was originally isolated by Ku¨ndig and Timms by
condensing vanadium atoms in a cold methylcyclohexane/THF solution of
naphthalene (Ku¨ndig, E. P.; Timms, P. L. J. Chem. Soc., Chem. Commun.
1977, 912). However, the yield of 1a was not specified and no chemical
properties of this substance were reported prior to our18 recent study.
(27) Barybin, M. V. Ph.D. Dissertation, University of Minnesota, 1999.
(28) Treichel, P. M.; Essenmacher, G. P. Inorg. Chem. 1976, 15, 146.

4682 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Barybin et al.
Table 1. Crystal Data, Data Collection, Solution, and Refinement for 2, Cs3, 4[PF6]‚THF, and 5
2
Cs3
4[PF6]‚THF
5
empirical formula
formula weight
crystal habbit, color
crystal size (mm)
crystal system
space group
a (Å)
C54H54N6V
837.97
plate, deep purple
0.22 × 0.22 × 0.10
triclinic
C54H54CsN6V
970.88
block, green-purple
0.45 × 0.35 × 0.30
triclinic
C58H62N6OPV
1055.05
irregular block, red
0.40 × 0.32 × 0.23
monoclinic
P21/c
15.6942(4)
20.8924(6)
19.6343(5)
90
98.749(1)
90
6363.0(3)
4
1.101
0.237
2208
C38H36N4O2V
631.65
block, red
0.36 × 0.25 × 0.14
tetragonal
P4nc
13.7245(1)
13.7245(1)
9.1904(1)
90
P1h
P1h
10.7588(3)
11.2539(2)
11.8350(3)
117.760(1)
109.184(1)
92.153(2)
1166.05(5)
1
12.0077(2)
14.3698(1)
17.1379(2)
66.534(1)
79.878(1)
83.106(1)
2666.23(5)
2
b (Å)
c (Å)
R (deg)
â (deg)
90
90
γ (deg)
V (Å3)
1731.12(3)
2
1.212
0.324
662
Z
F
calc (Mg m-3
)
1.193
0.255
443
1.209
0.897
996
µ (mm-1
F(000)
)
temp (K)
θ range (deg)
index range
173(2)
173(2)
173(2)
153(2)
2.06-24.97
-12 e h e 11
-13 e k e 11
0 e l e 14
6555
3941
0.0335
1.31-25.00
-13 e h e 14
-15 e k e 17
0 e l e 20
15296
9071
0.0184
1.31-25.03
-18 e h e 18
0 e k e 24
0 e l e 23
30108
11001
0.0360
2.10-24.99
0 e h e 11
0 e k e 16
-10 e l e 8
8539
1379
0.0248
no. of rflns collected
no. of unique rflns
a
Rint
max/min transmission
data/restraints/params
no. of rflns with I > 2σ(I))
R1;b wR2c (I > 2σ(I))
GOF on F2
1.0000/0.6986
3941/0/283
2522
0.0630; 0.1216
1.01
1.000/0.809
9067/0/565
7740
0.0524; 0.1568
1.02
1.000/0.806
10998/35/670
6612
0.0658; 0.1628
1.067
1.000/0.892
1379/1/108
1313
0.0253; 0.0640
1.150
largest diff peak/hole (e Å-3
)
0.263/-0.370
2.394/-0.507
0.554/-0.321
0.140/-0.220
a Rint ) ∑|Fo - Fc2 |/∑|Fo |. b R1 ) ∑||Fo| - |Fc||/∑|Fo|. c wR2 ) [∑(w(Fo - Fc2)2)/∑(w(Fo )2)]1/2
.
2
2
2
2
Scheme 1
Compound 2 is stable at room temperature both in the solid
state and in solution (THF, benzene, toluene) under argon.
However, this complex is very air- and moisture-sensitive.
Indeed, deep purple solutions of 2 decolorized within seconds
upon exposure to atmospheric air to give free CNXyl and
uncharacterized yellow-brown decomposition product(s).
Compound 2 is the first vanadium(0) complex to contain an
isocyanide ligand and may be considered an analogue of the
long-established V(CO)6.35 It is well-known that V(CO)6
undergoes rapid disproportionation in THF.36 On the contrary,
2 is perfectly stable in this solvent for weeks at room
temperature. While V(CO)6 is easily reduced by cobaltocene
and decamethylcobaltocene to give [V(CO)6]-,37 2 is unreactive
toward these substances. The latter fact is in accord with CO
being a better π-acceptor than CNXyl. Solution and Nujol mull
infrared spectra of 2 in the νCN stretching region resemble those
reported for the analogous octahedral Cr(0) complex, Cr-
(CNXyl)6.20,38 The frequency of the most intense peak is
depressed by 178 cm-1 compared to νCN found for free CNXyl
(2117 cm-1 in THF) indicating a significant degree of back-
bonding within 2. Similarly, νCO of V(CO)6 is about 170 cm-1
lower in energy than that of free CO.35
preparation of this compound involving condensation of potas-
sium atoms into a solution of VCl3 and C10H7Me in THF at
-100 °C.33 The latter provided a 25% yield of 1b, which
presumably exists as a mixture of several isomers.33 Both 1a
and 1b proved to be essential for establishing the isocyanide
chemistry of low-valent vanadium summarized in Scheme 1.
V(CNXyl)6 (2). In a typical synthesis of 2, a slight excess (7
equiv) of 2,6-dimethylphenyl isocyanide, was reacted with 1b
in THF/heptane at 0 °C. The red-brown color of 1b rapidly
changed to a deep purple. After the reaction had been completed,
most of the THF was removed in vacuo to give beautiful
dichroic green-purple microcrystals and a deep purple solution.
Workup and recrystallization afforded analytically pure micro-
crystalline golden-purple 2 in 81-91%yields (Scheme 1). The
use of 1a instead of 1b in the above procedure worked equally
well. The remarkably clean substitution of the naphthalene
ligands from 1a,b by CNXyl nicely demonstrates the so-called
“naphthalene effect”34 exhibited by the V(C10H7R)2 complexes.
The bulky nature of the 2,6-xylyl isocyanide (vide infra) was
essential for the stability of 2. Indeed, reactions of 1b with
(34) (a) Ku¨ndig, E. P.; Perret, C.; Spichiger, S.; Bernardinelli, G. J.;
Organomet. Chem. 1985, 286, 183. (b) Basolo, F. New J. Chem. 1994, 18,
19. (c) Albright, T. A.; Hofmann, P.; Hoffmann, R.; Lillya, C. P.; Dobosh,
P. A. J. Am. Chem. Soc. 1983, 105, 3396.
(35) Natta, G.; Ercoli, R.; Calderazzo, F.; Alberola, A.; Corrandini, P.;
Allegra, G. Atti Acad. Naz. Lincei, Classe Sci. Fis. Mater. Nat. 1959, 27,
107.
(36) Hieber, W.; Winter, E.; Schubert, E. Chem. Ber. 1962, 95, 196.
(37) (a) Calderazzo, F.; Pampaloni, G. J. Chem. Soc., Chem. Commun.
1984, 1249. (b) Calderazzo, F.; Pampaloni, G.; Pelizzi, G.; Vitali, F.
Polyhedron 1988, 7, 2039.
(33) (a) Hawker, P. N.; Ku¨ndig, E. P.; Timms, P. L. J. Chem. Soc., Chem.
Commun. 1978, 730. (b) Hawker, P. N.; Timms, P. L. J. Chem. Soc., Dalton
Trans. 1983, 1123.
(38) Bullock, J. P.; Mann, K. R. Inorg. Chem. 1989, 28, 4006.

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4683
excess 4-methylphenyl isocyanide, CNTol, afforded only un-
characterized isocyanide polymerization product(s). The initial
product, presumably V(CNTol)6, was stable for short periods
of time in THF or toluene solutions below -60 °C (IR evi-
dence).27 All attempts to isolate this extremely reactive species
failed. Thus, the methyl substituents in the ortho positions of
the coordinated aryl isocyanide ligands appear to be absolutely
necessary for the stability of V(CNAr)6. Perhaps they effec-
tively shield the V(CN)6 core from an outside attack and/or
prevent the ligating carbon atoms from coming too close to each
other to ultimately form a C-C bond. Notably, treatment of 1a
or 1b with a large excess of CNtBu at low temperature did not
yield V(CNtBu)6 and led only to partial decomposition of
CNtBu. This result is in marked contrast with the fact that
Cr(C10H8)2 reacts smoothly with CNtBu at 0 °C to give
Cr(CNtBu)6.39
[V(CNXyl)6]- (3). Recently, Cooper and co-workers have
isolated the first four- and five-coordinate binary isocyanide
metalates, [Co(CNXyl)4]- 5,7 and [Mn(CNXyl)5]-.40 However,
due to the well-established tendency of isocyanides to couple
in higher coordinate environments, the accessibility of anionic
complexes possessing six discrete isocyanide ligands has
remained in doubt.40 The anion 3, described below, is the first
example of such species.41 Notably, its carbonyl analogue,
[V(CO)6]-, has been known for nearly 40 years.42
One-electron reduction of 2 was achieved by stirring its
solution in THF with a suspension of excess (2 equiv) cesium
graphite, CsC8, at -50 °C. Workup followed by recrystallization
of the obtained brown solid afforded an 84% yield of unsolvated
Cs3 as free flowing bright brown microcrystals (Scheme 1).
Cs3 is thermally stable in the solid state. Solutions of the product
in THF, DME, and CH3CN did not show signs of decomposition
for days when kept under argon at room temperature. However,
unlike its isoelectronic analogues Cr(CNXyl)620 and [V(CO)6]-,42
Cs3 is extremely air- and moisture-sensitive both in solution
and in the solid state and requires exceptional care to be isolated
as an analytically pure substance.
The vanadate 3 can also be isolated as a [K(Crypt{2.2.2})]+
salt by reacting 2 with 1 equiv of [K(Crypt{2.2.2}][C10H8] in
THF at low temperature. Unfortunately, this iridescent green
microcrystalline substance is virtually insoluble in THF, DME,
CH3CN, and many other solvents. [K(Crypt{2.2.2})]3 can be
dissolved in hexamethylphosphoramide (HMPA) to give dark
violet solutions. However, IR spectra of these solutions showed
signs of decomposition within several hours of standing at room
temperature. Attempts to prepare the unsolvated potassium salt
of 3, analogous to K[Co(CNXyl)4],5 were unsuccessful.27 It is
worth mentioning that neither 2 nor 3 could be accessed by
reducing VCl3(THF)3 with excess sodium amalgam in the
presence of xylyl isocyanide,27 even though reactions, similar
to this, worked well to prepare related neutral group 6 metal
isocyanides, M(CNXyl)6 (M ) Cr, Mo).20 Reductions of VCl3-
(THF)3 with 4 equiv of naphthalene or anthracene radical anions
followed by treatment with CNXyl led to polymerization of the
isocyanide.9a,27 The latter procedures, however, were quite
efficient in the syntheses of other [ML6]- (M ) V, Nb, Ta; L
) CO, PF3)43 and [Co(CNXyl)4]-.5
with [Et3NH]Cl. However, this resulted in oxidation of the
vanadate 3 to neutral 2 (Scheme 1), presumably accompanied
by the formation of H2, Et3N, and CsCl. Complex Cs3 com-
bined with two molecules of 18-Crown-6 to give dark umber
microcrystalline [Cs(18-Crown-6)2]3 (Scheme 1). The pres-
ence of two 18-Crown-6 units per cesium cation in this
compound is not unusual45 and was confirmed by elemental
1
analyses and integration of the H NMR resonances for [Cs-
(18-Crown-6)2]3. This substance is reasonably soluble in THF
and presently constitutes the only convenient alternative to
the parent Cs3 complex. The IR bands in the νCN region,
observed for 3, are quite broad and similar to those pre-
viously reported for [Co(CNXyl)4]-, which is believed to be
appreciably distorted from its expected tetrahedral symmetry
in solution.5 The qualitative similarities in the IR spectral
signatures of the cobaltate and the vanadate 3 suggest that the
geometry of the latter species is not strictly octahedral. Ap-
preciable bending of the isocyanide ligands in this very electron
rich molecule (vide infra) is among the factors reducing the
ideal octahedral symmetry of 3.46 Using notation for the Oh
group,47 the bands at 2021, 1872, and 1818 cm-1 in the IR
spectrum (THF) of [Cs(18-Crown-6)2]3 should correspond to
A1, Eg, and T1u νCN modes, respectively. Indeed, (20212 -
18722)/(18722 - 18182) ) 2.91, a value very close to the
required 3.0.48 The most intense νCN band for 3 is ca. 120 cm-1
lower in energy than that for neutral 2, indicating increase in
the extent of back-donation upon reduction of 2. Employing
the Cotton-Kraihanzel approximation,48 one can estimate the
C-N force constant, kCN, for the isocyanide ligands in [Cs(18-
Crown-6)2]3 to be about 13.3 mdyn Å-1, while kCN for free
CNXyl is ca. 17.1 mdyn Å-1. Similarly, kCO for [V(CO)6]- is
only 14.9 mdyn Å-1, as opposed to kCO ) 18.7 mdyn Å-1 for
free carbon monoxide.49
[V(CNXyl)6]+ (4). Oxidation of 2 with 1 equiv of [Cp2Fe]-
[PF6] in THF produced an orange-red solution from which a
deep red solid was obtained by adding excess heptane. This
precipitate was washed with pentane to remove the byproduct,
Cp2Fe, and recrystallized to afford an almost quantitative yield
of analytically pure 4[PF6], a dark red microcrystalline substance
(Scheme 1). Solutions of 4[PF6] in THF or CH2Cl2 were stable
for months at 4 °C under argon but deteriorated within minutes
upon exposure to air. This compound proved to be moderately
air-sensitive in the solid state as well. Infrared data indicated
that one of the decomposition pathways of the anion 3 on
exposure to air involved sequential oxidation of 3 to 4 (3 f 2
f 4 f decomposition products). It is worth noting that cation
4 was also isolated in ca. 70% yields as [BF4]- and [BPh4]-
(42) Ercoli, R.; Calderazzo, F.; Alberola, A. J. Am. Chem. Soc. 1960,
82, 2966.
(43) (a) Barybin, M. V.; Ellis, J. E.; Pomije, M. K.; Tinkham, M. L.;
Warnock, G. F. Inorg. Chem. 1998, 37, 6518. (b) Barybin, M. V.; Pomije,
M. K.; Ellis, J. E. Inorg. Chim. Acta 1998, 269, 58. (c) Ellis, J. E.; Warnock,
G. F.; Barybin, M. V.; Pomije, M. K. Chem. Eur. J. 1995, 1, 521.
(44) (a) Calderazzo, F.; Pampaloni, G.; Vitali, D. Gazz. Chim. Ital. 1981,
111, 455. (b) Calderazzo, F.; Pampaloni, G.; Lanfranchi, M.; Pelizzi, G. J.
Organomet. Chem. 1985, 296, 1.
(45) Selected examples: (a) Vidal, J. L.; Troup, J. M. J. Organomet.
Chem. 1981, 213, 351. (b) Tinkham, M. L.; Dye, J. L. J. Am. Chem. Soc.
1985, 107, 6129.
(46) (a) Cotton, F. A.; Zingales, F. J. Am. Chem. Soc. 1961, 83, 351. (b)
Mann, K. R.; Cimolino, M.; Geoffroy, G. L.; Hammond, G. S.; Orio, A.
A.; Albertin, G.; Gray, H. B. Inorg. Chim. Acta 1976, 16, 97.
(47) Strictly speaking, the highest possible symmetry for [V(CNXyl)6]-
is Th, which is a subgroup of Oh. See ref 46b.
In an attempt to generate the hydride HV(CNXyl)6, analogous
to the extremely unstable possible HV(CO)6,44 Cs3 was treated
(39) Ku¨ndig, E. P.; Timms, P. L. J. Chem. Soc., Dalton Trans. 1980,
991.
(48) Cotton, F. A.; Kraihanzel, C. S. J. Am. Chem. Soc. 1962, 84,
4432.
(49) (a) Abel, E. W.; McLean, R. A. N.; Tyfield, S. P.; Braterman, P.
S.; Walker, A. P.; Hendra, P. J. J. Mol. Spectrosc. 1969, 30, 29. (b) Dobson,
G. R. Inorg. Chem. 1965, 4, 1673.
(40) Utz, T. L.; Leach, P. A.; Geib, S. J.; Cooper, N. J. Chem. Commun.
1997, 847.
(41) Very recently, we have also isolated the tantalum analogue of 3,
[Ta(CNXyl)6]-. See ref 9b.

4684 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Barybin et al.
salts by oxidizing 2 with 1 equiv of Ag[BF4] or Ag[BPh4],
respectively.27 Spectroscopic (IR and NMR) characteristics of
4 were shown to be independent of the counterion.27
Complex 4 is the first homoleptic V(1+) isocyanide. This
thermally robust compound may be regarded as an isocyanide
analogue of exceedingly unstable [V(CO)6]+.50 The most
prominent IR νCN band for cationic 4 in THF occurs 94 cm-1
higher in energy than that for neutral 2. Nevertheless, this peak
is 84 cm-1 below νCN found for free xylyl isocyanide in the
same solvent. Therefore, substantial back-bonding is still present
in 4. Importantly, complex 4 is paramagnetic in the solid state
(vide infra) and does not add a seventh CNXyl ligand. In
contrast, the Nb and Ta congeners of 4, [M(CNXyl)7]+, are both
diamagnetic and contain seven coordinate metals.9b Most likely,
such a difference is a reflection of the smaller size of vanadium
compared to those of niobium and tantalum.
trans-V(CO)2(CNXyl)4 (5). To this day, direct substitution
reactions of V(CO)6, proceeding without change in the metal
oxidation state, have been limited almost exclusively to the
syntheses of a few mixed carbonyl-phosphine19,51 and carbo-
nyl-phosphite52 vanadium(0) derivatives. The potentially di-
verse carbonyl displacement chemistry of V(CO)6 has undoubt-
edly been hampered by its well-documented tendency to
disproportionate in the presence of many ligands, including
donor solvents.53,54
Given that both 2 and V(CO)6 are stable under appropriate
conditions, a reaction between hexacarbonylvanadium(0) and
xylyl isocyanide was investigated. Solutions of V(CO)6 and
excess CNXyl (7 equiv) in heptane were combined at room
temperature (Scheme 1). An extensive gas evolution and a series
of color changes were observed. After 3 days of stirring, the
reaction mixture was concentrated and filtered. The isolated
bright red solid was purified to provide an 80% yield of
analytically pure lustrous dark red 5. Unlike V(CO)6 (but similar
to 2), compound 5 is stable in toluene and THF under argon at
room temperature. Complex 4 is very air- and moisture-sensitive
both in solution and in the solid state.
Surprisingly, no disproportionation of V(CO)6 occurred in
the above reaction and complex 5 was the only isolated product.
Infrared data indicated that shorter reaction times (<72 h) led
to incomplete substitution of four carbonyls from V(CO)6.
Stirring the reaction mixture for 4 additional days (for a total
of 7 days) at room temperature did not produce any further
changes. An attempt to achieve a higher degree of substitution
by slowly raising the temperature of the reaction to 80 °C
ultimately led to decomposition of 5. The assignments of the
IR spectra of 5 were confirmed by preparing 13CO-labeled trans-
V(13CO)2(CNXyl)4, 5*. Indeed, the νCN frequencies for 5 and
5* are practically identical, whereas the ν13CO peaks for 5* are
Figure 1. Molecular structure of 2 (50% thermal ellipsoids); hydrogens
are omitted for clarity. Selected bond lengths (Å) and angles (deg):
V-C1 2.034(3), V-C2 2.022(4), V-C3 2.023(4), C1-N1 1.184(4),
C2-N2 1.183(4), C3-N3 1.191(4), C1-N1-C7 168.1(3), C2-N2-
C15 158.8(3), C3-N3-C23 161.5(3), av cis C-V-C 90(3), av trans
C-V-C 180.0.
shifted as predicted55 toward lower energies relative to the
corresponding ν12CO bands observed for 5. The presence of two
ν
CN and two νCO peaks in the IR spectra of 5 is consistent with
the C4 symmetry of the complex revealed by the X-ray study
(vide infra). The higher energy νCO band is very weak because
the corresponding C-O stretching mode is nearly IR-forbidden
under the C4 symmetry. Interestingly, while we observed two
relatively sharp νCN peaks for 5, Rehder et al. reported a single
but broad νCN band for trans-VI2(CNXyl)4.10a
After homoleptic 2, compound 5 is only the second example
of a group 5 metal(0) complex to contain an isocyanide ligand.
Displacement of four carbonyls from V(CO)6 by CNXyl ligands
under mild conditions is quite remarkable. The only other known
reaction of this type involved interaction of V(CO)6 with dppe
at 120 °C to afford trans-V(CO)2(dppe)2.51b Although complete
carbonyl substitution of V(CO)6 to give 2 could not be
accomplished, oxidation of 5 with 1 equiv of [Cp2Fe]+ in the
presence of free CNXyl afforded the binary cationic vanadium
isocyanide 4 (Scheme 1).
Structural Characterization of 2-5: V(CNXyl)6 (2). Com-
plex 2 crystallizes in the space group Ph1 with the vanadium
atom lying on an inversion center (Figure 1). This results in
three independent vanadium-isocyanide units. The molecule
is essentially octahedral. The average cis C-V-C angle for 2
is 90(3)° while all trans C-V-C angles are required to be 180°
because of the crystallographically imposed symmetry. The
structure of 2 exhibits a possible small tetragonal distortion from
a regular octahedron, very similar in magnitude to that observed
for V(CO)6.56 In contrast, the M-C distances in the related
neutral low-spin d6 complexes, Cr(CNPh)6,57 Mo(CNXyl)6,20
W(CNXyl)6,58 and [Mn(CNPh)6]+,59 are identical within their
estimated standard deviations. The average V-C bond length
in 2 is only 0.025 Å longer than the corresponding distance in
V(CO)6.56 At the same time, it is 0.07 Å shorter compared to
the mean V-C bond in [V(CNtBu)6]2+ 8b despite a significant
(50) (a) Darensbourg, M. Y. Progr. Inorg. Chem. 1985, 33, 221 and
references therein. (b) Kochi, J. K.; Bockman, T. M. AdV. Organomet. Chem.
1991, 33, 51. (c) Horowitz, C. P.; Shriver, D. F. AdV. Organomet. Chem.
1984, 23, 219.
(51) (a) Interrante, L. V.; Nelson, G. V. J. Organomet. Chem. 1971, 25,
153. (b) Behrens, H.; Lutz, K. Anorg. Allg. Chem. 1968, 356, 225. (c)
Hieber, W.; Winter, E. Chem. Ber. 1964, 97, 1037. (d) Werner, R. P. M.
Naturforsch. 1961, 16b, 477. (e) Hieber, W.; Peterhans, J.; Winter, E. Chem.
Ber. 1961, 94, 4, 2572.
(55) IR for 5*: (toluene) νCN 2012 s, 1985 vs, ν CO 1844 w br, 1799 vs
13
(52) (a) Shi, Q. Z.; Richmond, T. G.; Trogler, W. C.; Basolo, F. J. Am.
Chem. Soc. 1984, 106, 71. (b) Shi, Q. Z.; Richmond, T. G.; Trogler, W.
C.; Basolo, F. J. Am. Chem. Soc. 1982, 104, 4032.
cm-1; (Nujol mull) νCN 2015 s, 1988 vs, ν
1855 vw, 1768 vs cm-1
.
13CO
Calcd IR ν CO for 5*: (toluene) 1847, 1799 cm-1; (Nujol mull) 1855, 1768
13
cm-1
.
(53) (a) Berno, P.; Gambarotta, S.; Richeson, D. In ComprehensiVe
Organometallic Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G.,
Eds.; Pergamon: New York, 1995; Vol. 4, pp 2-9 and references therein.
(b) Connelly, N. G. In ComprehensiVe Organometallic Chemistry; Wilkin-
son, G., Stone, F. G. A., Abel, E. W., Eds.; Pergamon: New York, 1983;
Vol. 3, pp 647-654 and references therein.
(56) Bellard, S.; Rubinson, K. A.; Sheldrick, G. M. Acta Crystallogr.
1979, B35, 271.
(57) Ljungstro¨m, E. Acta Chem. Scand. 1978, A32, 47.
(58) See Supporting Information for: Lockwood, M. A.; Fanwick, P.
E.; Rothwell, I. P. Organometallics 1997, 16, 3574 (at http://pubs.acs.org).
(59) Ericsson, M.-S.; Jagner, S.; Ljungstro¨m, E. Acta Chem. Scand. 1980,
A34, 535.
(54) Ellis, J. E.; Faltynek, R. A. J. Organomet. Chem. 1975, 93, 205.

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4685
Table 2. Close Cs-C and Cs-N Contacts (Å) within Cs3
Cs(1)-C(2)
Cs(1)-N(2)
Cs(1)-C(3)
Cs(1)-N(3)
Cs(1)-C(4)
Cs(1)-N(4)
3.392(3)
3.417(4)
3.349(3)
3.369(3)
3.506(3)
3.629(3)
Cs(2)-C(1)
Cs(2)-N(1)
Cs(2)-C(2)
Cs(2)-N(2)
Cs(2)-C(5)
Cs(2)-N(5)
3.580(4)
3.798(4)
3.426(3)
3.552(4)
3.327(3)
3.375(3)
Figure 3. Cation-anion interactions within the unit cell of Cs3; xylyl
substituents are omitted for clarity.
Figure 2. Molecular structure of 3 (50% thermal ellipsoids); hydrogens
are omitted for clarity. Selected bond lengths (Å) and angles (deg):
V-C1 2.008(4), V-C2 1.939(4), V-C3 1.985(3), V-C4 1.972(4),
V-C5 1.986(3), V-C6 2.010(4), C1-N1 1.182(6), C2-N2 1.228(5),
C3-N3 1.196(5), C4-N4 1.200(5), C5-N5 1.202(5), C6-N6 1.190-
(5), C1-N1-C7 166.4(8), C2-N2-C15 145.5(4), C3-N3-C23
157.6(4), C4-N4-C31 148.5(4), C5-N5-C39 158.3(4), C6-N6-
C47 170.5(4), av cis C-V-C 90(3), av trans C-V-C 176.0(7).
[Co(CNXyl)4]-,5 respectively. The structure of Cs3 features
rather close contacts between the cesium cations and CN groups
of the isocyanide ligands. These are listed in Table 2 and are
depicted in Figure 3. Cs(1) is approximately octahedrally
coordinated by six CN groups (three from each anion) and may
be regarded as a bridge between two vanadate anions. Cs(2)
also interacts with three CN fragments of the vanadate.61 Crystal
structures of alkali-metal salts of [Co(CNXyl)4]- 5 and certain
carbonyl metalates62 show similar cation-anion interactions.
The large spread in C-N-C angles (and, to a lesser extent, in
the structural parameters of the V(CN)6 core) is undoubtedly a
consequence of the described above perturbation of 2 by the
cesium counterions. The widest C-N-C angle, C6-N6-C47,
of 170.5(4)° corresponds to the isocyanide ligand least affected
by the presence of Cs (Figure 3). As in the case of 2, the bulky
nature of CNXyl is likely to be important for the stability of
even more electron rich 3. There are six contacts between methyl
substituents within 3 that are shorter than 4.0 Å.
[V(CNXyl)6][PF6] (4[PF6]). Compound 4[PF6] crystallizes
in the space group P21/c with one ordered and, possibly, one
disordered THF molecule in the asymmetric unit. The ap-
proximately octahedral geometry of 4 (Figure 4) is relatively
unperturbed by the counterion [PF6]- and the solvent of
crystallization. The shortest interionic contact of 3.536 Å is
between one of the fluorine atoms of [PF6]- and C(30). The
mutually trans V-C(3) and V-C(4) bonds are somewhat longer
than the other four V-C distances. This may be due to a
tetragonal distortion of the low-spin d4 complex 4. In 4, the
average V-C bond is 0.04 Å longer while the average C-NXyl
bond is 0.02 Å shorter compared to corresponding distances in
neutral 2. Also, the mean C-N-C angle in 4 is 10° greater
than that in 2. All of these facts are consistent with reduction
in the extent of back-donation upon proceeding from 2 to 4.
While Me‚‚‚Me contacts within 4 are longer than those within
1 and 2, six of them are still under 4.0 Å.
difference in size of the V(0) and V(2+) ions. In addition,
compound 2 possesses relatively long C-NXyl distances (av
C-NXyl ) 1.186(5) Å) and somewhat bent C-N-Xyl angles
(av C-N-Xyl ) 163(4)°). All of the above features are
consistent with CNXyl functioning as a good π acceptor ligand
in 2. The most noticeable difference between the structures of
2 and M(CNXyl)6 (M ) Mo,20 W58) is the fact that the trans
xylyl rings in 2 are mutually parallel (Figure 1) while these
groups in the analogous Mo(0) and W(0) complexes are
approximately perpendicular to each other. As in the case of
Mo(CNXyl)6,20 the xylyl substituents in 2 are torsional with
respect to the appropriate V(CN)4 planes. This causes reduction
in the overlap between the relevant CN and Xyl π systems. On
the contrary, orientation of the phenyl rings in Cr(CNPh)6
indicates that the CN and phenyl π systems are practically
coplanar.57 Undoubtedly, the arrangement of the xylyl fragments
in 2 is a consequence of balancing electronic and steric factors.
The latter constitutes repulsion between methyl substituents of
the mutually cis isocyanide ligands. Remarkably, four of the
Me‚‚‚Me contacts within 2 are significantly shorter than the
doubled van der Waals radius of a methyl group (4.0 Å).
Cs[V(CNXyl)6] (Cs3). Compound Cs3 (Figure 2) crystallizes
in the space group P1h with two anions 3 in the unit cell. One
of the cesium counterions, Cs(1), is located on the inversion
1
1
center (1/2, /2, /2), while the other, Cs(2), is disordered (50:
50) over two symmetry related general positions. The V(CN)6
core is nearly octahedral. All V-C bonds in 3 are statistically
shorter than the V-C distances in 2. This is consistent with
the expected increase in back-bonding upon reducing 2 to 3.
Also, as anticipated, the mean V-C distance of 1.98(3) Å in 3
is somewhat longer than that in [V(CO)6]- (1.931(9) Å).60 The
average C-NXyl bond length of 1.20(2) Å for the vanadate 3
is indistinguishable from analogous values of 1.20(2) and 1.20-
(3) Å, reported for the only other crystallographically character-
ized homoleptic isocyanide metalates, [Mn(CNXyl)5]- 40 and
(60) Wilson, R. D.; Bau, R. J. Am. Chem. Soc. 1974, 96, 7601.
(61) Disordered solvent, found in the vicinity of Cs(2), may interact with
the Cs(2) cation as well.
(62) (a) Darensbourg, M. Y. Progr. Inorg. Chem. 1985, 33, 221 and
references therein. (b) Kochi, J. K.; Bockman, T. M. AdV. Organomet. Chem.
1991, 33, 51. (c) Horowitz, C. P.; Shriver, D. F. AdV. Organomet. Chem.
1984, 23, 219.

4686 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Barybin et al.
Figure 5. Molecular structure of 5 (50% thermal ellipsoids); hydrogens
are omitted for clarity. Selected bond lengths (Å) and angles (deg):
V-C1 2.044(2), V-C10 1.965(5), V-C11 1.944(6), C1-N1 1.170-
(2), C10-O10 1.154(6), C11-O11 1.150(7), C1-N1-C2 172.2(2),
C10-V-C11 180.0, C1-V-C1A 176.2(2), C1-V-C1B 89.936(5),
C1-V-C10 91.91(7), C1-V-C11 88.09(7).
Table 4. Comparison of V-CN and C-NXyl Distances (Å) and
C-N-C Angles (deg) for 2, 4, 5, and trans-VI2(CNXyl)4
Figure 4. Molecular structure of 4 (50% thermal ellipsoids); hydrogens
are omitted for clarity. Selected bond lengths (Å) and angles (deg):
V-C1 2.048(4), V-C2 2.062(4), V-C3 2.078(4), V-C4 2.100(4),
V-C5 2.061(4), V-C6 2.062(4), C1-N1 1.174(4), C2-N2 1.177(4),
C3-N3 1.165(4), C4-N4 1.162(4), C5-N5 1.168(4), C6-N6 1.166-
(4), C1-N1-C7 170.4(3), C2-N2-C15 176.2(3), C3-N3-C23
173.0(3), C4-N4-C31 173.0(3), C5-N5-C39 171.9(3), C6-N6-
C47 170.8(4), av cis C-V-C 90(2), av trans C-V-C 177(2).
10a
2
5
4
VI2(CNXyl)4
av V-CN
av C-NXyl
av C-N-C
2.026(7)
1.186(5)
163(4)
2.044(2)
1.170(2)
172.2(2)
2.07(2)
1.169(6)
173(2)
2.143(21)
1.145(8)
176(4)
Table 5. Comparison of V-CO and C-O Distances (Å) for
V(CO)6,56 5, and trans-V(CO)2(dmpe)2
64
Table 3. Comparison of Structural and Infrared Data for
V(CO)6
5
V(CO)2(dmpe)2
[V(CNXyl)6]z
av V-CO
av C-O
2.001(6)
1.128(8)
1.955(15)
1.152(7)
1.891(11)
1.184(10)
Z ) 1+
Z ) 0
Z ) 1-
av V-C (Å)
2.07(2)
1.169(6)
173(2)
2033
2.026(7)
1.186(5)
163(4)
1939
1.98(3)
1.20(2)
158(10)
1817
av C-NXyl (Å)
av C-N-C (deg)
As expected, the V-CO bonds are shorter than the V-CN
distances within 5. Judging by the V-CN bond lengths in 2, 4,
5, and trans-VI2(CNXyl)410a (Table 4), the amount of electron
density at the vanadium center, available for π bonding with
CNXyl ligands, decreases in the row 2 > 5 > 4 > trans-VI2-
(CNXyl)4. The shortening of the V-CO bonds and the
concomitant lengthening of the C-O distances in proceeding
from V(CO)6 through 5 to trans-V(CO)2(dmpe)2 (Table 5)
nicely reflect the decrease in electron accepting capability in
the series of ligands CO > CNXyl > dmpe.
Both steric and electronic factors favor the trans arrangement
of the carbonyl ligands in 5. While the steric argument is
straightforward, the electronic reason is less obvious. The latter
is based on the angular overlap considerations originally
proposed by Bursten65 and Mingos66 for complexes [ML′2L4]z,
where L is a poorer π acceptor compared to L′. Figure 6 shows
stabilization energies67 of dπ orbitals for hypothetical cis-
V(CO)2(CNXyl)4 and observed trans-V(CO)2(CNXyl)4, 5, with
respect to the t2g-like set in 2. It turns out that for the 17-electron
complex V(CO)2(CNXyl)4, trans configuration leads to greater
stabilization of five d electrons by means of back-donation to
ν
CN in THF (cm-1
)
back-donation increases f
Table 3 summarizes structural and IR data for all three
homoleptic vanadium isocyanides, 2, 3, and 4. The trends
presented in this table nicely parallel those obtained for the
isoelectronic series [Cr(CNPh)6]z (z ) 0, 1+, 2+)57,63 and are
consistent with the aryl isocyanides functioning as good π
acceptor ligands. Among three structural parameters considered,
the V-C bond length seems to be the most sensitive to changes
in the vanadium oxidation state. Bending of the isocyanide
ligands is the least reliable factor reflecting relative electron
richness of the complexes. The latter is not surprising since, in
addition to electronic effects, magnitudes of the C-N-C angles
are greatly influenced by steric and interionic (e.g., as in 3)
interactions as well as crystal packing forces. Nevertheless, the
C-N-C angles observed for 2 and 4 do correlate well with
the corresponding V-C and C-NXyl distances and infrared
νCN stretching frequencies of these species in the expected
manner.
trans-V(CO)2(CNXyl)4 (5). Complex 5 crystallizes in the
rarely encountered tetragonal space group P4nc. The structure
of 5 is shown in Figure 5. It exhibits mutually trans V-CO
units. The crystallographic 4-fold axis passes through the OC-
V-CO fragment requiring molecular symmetry of 5 to be
strictly C4. The xylyl substituents of four equivalent isocyanide
ligands feature a propeller-like arrangement with no unusually
short Me‚‚‚Me contacts. Compound 5 is practically octahedral.
56
64
2
2
the CO ligands (Estab(trans) ) 8âπSπ vs Estab(cis) ) 7âπSπ ).
On the other hand, a similar approach suggests that for an 18-
(64) Wells, F. J.; Wilkinson, G.; Motevalli, M.; Hursthouse, M. B.
Polyhedron 1987, 6, 1351.
(65) Bursten, B. E. J. Am. Chem. Soc. 1982, 104, 1299.
(66) Mingos, D. M. P. J. Organomet. Chem. 1979, 179, C29.
(67) Stabilization energy, Estab, is defined as (∑Nini) × âπSπ2, where N
is the number of carbonyls interacting with a given d orbital, n is the number
of electrons in that orbital, âπ is a constant, and Sπ is the extent of overlap
between the appropriate CO and metal π orbitals. See the following for
details: Burdett, J. K. Inorg. Chem. 1975, 14, 375.
(63) Bohling, D. A.; Mann, K. R. Inorg. Chem. 1984, 23, 1426.

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4687
Table 6. 1H and 13C Paramagnetic Shifts of 2 and 7[BF4] at 20 °C
vs Cr(CNXyl)6
∆δ, (ppm)
[Cr(CNXyl)6]+ b
a
nucleus
V(CNXyl)6
o-CH3
m-H
p-H
i-C
o-C
o-CH3
m-C
p-C
+7.04
+7.18
-9.20
-225.7
+205.8
-23.6
-25.8
+54.1
+6.27
+5.53
-8.31
-162.8
+90.2
-19.9
-36.4
+55.0
a In C6D5CD3. b In CDCl3.
Figure 6. Stabilization energies (expressed as angular orbital func-
tions)67 of dπ orbitals for isomers V(CO)2L4 relative to the t2g set of
ML6 (L ) CNXyl). N is the number of carbonyls interacting with a
given d orbital. (Based on diagrams from refs 65 and 66.)
back-bonding and its µeff is practically identical with the value
anticipated by the theory for a low-spin Mn(2+) complex.75
It is well-known that V(CO)6 is ESR silent at ambient
temperature.79 No ESR signal could be detected for a toluene
solution of 2 at 20 °C either. At the same time, relatively sharp
1H NMR resonances, showing strong paramagnetic shifts, were
observed for 2 in C6D6 and C6D5CD3. Both of these features
are consistent with the very short electron spin relaxation time
expected for a homoleptic octahedral low-spin d5 complex
[ML6]z.80 The hydrogen NMR resonances for 2 were easily
assigned based on their relative intensities and splitting due to
H-H coupling and confirmed by an 1H COSY picture.
Assignment of the 13C NMR peaks for 2 was facilitated by an
1H-13C{1H} HETCOR experiment. For the sake of comparison,
electron low-spin d6 species, such as Mo(CO)2(CNCy)4,68 cis
and trans isomers should be equally stable (Estab(cis) ) Estab
-
2
(trans) ) 8âπSπ ). However, ligand-ligand overlaps, neglected
in the angular overlap model,66 make cis configuration more
thermodynamically favorable in the case of Mo(CO)2(CNCy)4.
The above arguments were previously used by Walton and
Conner to explain why oxidation of cis-Mo(CO)2(CNCy)4
resulted in trans-[Mo(CO)2(CNCy)4]+.68 Exclusive formations
69
of trans-V(CO)2(dmpe)264 (eq 3.28) and cis-Cr(CO)2(dmpe)2
were rationalized in a similar manner as well.
Magnetic Properties of 2-5: V(CNXyl)6 (2). Complex 2
is the first paramagnetic homoleptic metal(0) isocyanide. Unlike
[M(CNXyl)4]2 (M ) Co,3b,5 Rh70), 2 is monomeric both in the
solid state and in solution. Its effective magnetic moment of
1.76 µB at 25 °C implies a low-spin d5 configuration of 2 and
is virtually identical with those reported for V(CO)6 (1.73-
1.78 µB)71 and V(PF3)6 (1.76 µB).72 The diamagnetic correction,
applied to molar susceptibility of 2, was determined by
measuring molar susceptibility of Cr(CNXyl)6,20 a low-spin d6
complex. Such correction should be very accurate since the
difference between diamagnetic susceptibilities of V(0) and
Cr(0) is negligible.73 The free ion spin-orbit coupling constant,
1
[Cr(CNXyl)6][BF4], 7[BF4], was synthesized and its H, 13C,
and two-dimensional NMR spectra were recorded. To our great
surprise, even though many homoleptic aryl isocyanides of
Cr(1+) have been known for a long time,28 NMR data on such
compounds have not been reported prior to this study.
1H and 13C paramagnetic (isotropic) shifts of 2 and 7[BF4],
∆δ, at 20 °C were determined relative to diamagnetic Cr-
(CNXyl)6 (6) and are collated in Table 6. The NMR spectra of
the reference compound 6 were obtained27 in the appropriate
media (C6D5CD3 or CDCl3) to minimize any solvent effects on
the isotropic shifts of 2 and 7.81 The hydrogen paramagnetic
shifts of 2 and 7 are essentially contact in origin due to the
approximately octahedral symmetry of these species.22,80 Indeed,
as in the case of V(CO)6,71b the average g tensors for 2 and 7
are expected to be isotropic in solution (dynamic Jahn-Teller
effect), so the dipolar field should be averaged to near zero by
rapid tumbling of the complexes in solution. Also, since ∆δ
values, obtained for 7[BF4], are practically insensitive to
variations in solvent (CDCl3 or THF-d8) and concentration, any
significant pseudocontact contributions to the 1H isotropic shifts
of 7 due to ion pairing can be ruled out.81 As can be seen from
λ, for V(0) is about -95 cm-1 74
In the case of an octahedral
.
low-spin d5 species, λ ) -95 cm-1 is expected to lead to a
magnetic moment of ca. 2.3 µB at 25 °C.75 However, µeff of 2
as well as those of V(CO)6, and V(PF3)6 are essentially “spin
only”. Quenching of the orbital angular momentum in these
compounds may be attributed, in part, to significant delocal-
ization of unpaired electron density onto the ligands via back-
donation.76 Incidentally, µeff of V(Me2PCH2CH2PMe2)3 (where
the organophosphine ligand is a poor acceptor, at best) at 25
°C is 2.10 µB.77 Also, the room-temperature effective magnetic
moment increases within the isoelectronic series 2 (1.76 µB) <
[Cr(CNXyl)6][BF4] (1.92 µB) < and [Mn(CNPh)6][BF4]278 (2.32
µB). The latter compound exhibits only a very small degree of
1
Table 6, the H and 13C isotropic shifts of 2 and 7 occur in
both directions and do not attenuate with the increasing number
of bonds from the metal center. This is consistent with unpaired
spin being present in the π systems of the xylyl substituents.
The p-C and p-H resonances are shifted in opposite directions
(atomic exchange coupling mechanism of spin delocalization82).
The same applies to m-C and m-H signals. On the other hand,
the carbon atoms in the ortho positions and the hydrogens of
(68) Conner, K. A.; Walton, R. A. Inorg. Chem. 1986, 25, 4422.
(69) Salt, J. E.; Girolami, G. S.; Wilkinson, G.; Motevalli, M.; Thornton-
Pett, M.; Hursthouse, M. B. J. Chem. Soc., Dalton Trans. 1985, 685.
(70) Yamamoto, Y.; Yamazaki, H. Bull. Chem. Soc. Jpn. 1984, 57, 297.
(71) (a) Calderazzo, F.; Cini, R.; Corrandini, P.; Ercoli, R.; Natta, G.
Chem. Ind. (London) 1960, 79, 500. (b) Bernier, J. C.; Kahn, O. Chem.
Phys. Lett. 1973, 19, 414.
(72) Collong, W.; Kruck, T. Chem. Ber. 1990, 123, 1655.
(73) E. Ko¨nig, E. In Zahlenwerte und Funktionen aus Naturwissen-
schaften und Technik, 6th ed.; Landolt-Bo¨rnstein; Springer-Verlag: Berlin,
1966; Vol. 2, p 16.
(78) Nielson, R. M.; Wherland, S. Inorg. Chem. 1985, 24, 3458.
(79) (a) Pratt, D. W.; Myers, R. J. J. Am. Chem. Soc. 1967, 89, 6740.
(b) Rubinson, K. A. J. Am. Chem. Soc. 1976, 98, 5188. (c) Boyer, M. P.;
Le Page, Y.; Morton, J. R.; Preston, K. F.; Vuolle, M. J. Can. J. Spectrosc.
1981, 26, 181.
(74) Dunn, T. M. Trans. Faraday Soc. 1961, 57, 1441.
(75) Figgis, B. N.; Lewis, J. Progr. Inorg. Chem. 1964, 6, 37.
(76) Gerloch, M.; Miller, J. R. Progr. Inorg. Chem. 1968, 10, 1.
(77) Chatt, J.; Watson, H. R. J. Chem. Soc. 1962, 2545.
(80) Drago, R. S. Physical Methods in Chemistry; W. B. Saunders:
Philadelphia, PA, 1977; Chapter 12.
(81) Wicholas, M.; Drago, R. S. J. Am. Chem. Soc. 1969, 91, 5963.
(82) McConnell, H. M.; Chesnut, D. B. J. Chem. Phys., 1958, 28, 107.

4688 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Barybin et al.
Figure 8. Signs of unpaired spin densities for the CNXyl ligands of
2 (v, positive; V, negative).
Scheme 2. Schematic Representation of Unpaired Electron
Density Delocalization in a d1 ML6 Complex via
pπ(L)-dπ(M) Bonding (a) and in a Low-Spin d5 ML6
Complex via dπ(M)-pπ*(L) Back-Bonding (b) According to
Ref 87
carbon was not experimentally observed for 2, one can easily
deduce86 that this carbon must bear excess negative spin.
Surprisingly, La Mar predicted the spin, induced on the ligating
atom by means of π back-bonding, to be positive for a
homoleptic octahedral low-spin d5 complex.87 To sort out this
discrepancy, we considered ligand-to-metal pπ(L)-dπ(M)
bonding in an octahedral d1 complex and metal-to-ligand
dπ(M)-pπ*(L) back-bonding in a low-spin d5 complex. Be-
cause of the electron-hole formalism, these two processes are
topologically identical. According to La Mar,87 ligand-to-metal
π bonding in a d1 octahedral complex should result in negative
spin density on the ligand. Therefore, electron density will be
transferred from filled π orbitals of L to t2g orbitals of M slightly
more with spin parallel to that on the metal (Scheme 2a). This
should leave a net negative spin on the ligand. Such a
delocalization mode is governed by correlation effects,88 which,
in simple terms, require maintaining the Hund’s rule at the metal
center. Of course, the total spin of the complex remains
unaffected by the pπ(L)-dπ(M) interaction. Now, if back-
donation in the low-spin d5 octahedral species produces excess
positive spin on the ligand, delocalization of the electron density
with excess spin parallel to that on the metal ion is implied
(Scheme 2b). Such a process, however, violates the above-
mentioned correlation requirements. On the other hand, if back-
bonding in a low-spin d5 octahedral complex involves trans-
ferring electrons possessing excess antiparallel spin, these
requirements will be met. The latter mechanism is, in fact,
consistent with the experimental results obtained for 2.
Figure 7. (a) Plot of 13C line width vs magnitude of paramagnetic
shift for 2 in C6D5CD3 at 20 °C. (b) Plot of the 13C line width vs the
magnitude of paramagnetic shift for 7[BF4] in CDCl3 at 20 °C.
the methyl groups exhibit paramagnetic shifts of the same sign,
which is a consequence of hyperconjugative (direct) spin
delocalization.83 The directions of the 13C paramagnetic shifts
for 2 and 7 clearly indicate their predominantly contact nature
and relatively minor importance of the so-called ligand-centered
pseudocontact contribution.22 Interestingly, the line width at half-
height (W1/2) of the 13C signals is nearly proportional to the
magnitude of the paramagnetic shift. While the linearity of the
relationships in Figure 7 has little physical significance, the
general trends indicate the importance of contact relaxation for
the carbon nuclei of 2 and 7.84 The slope in Figure 7a is greater
than that in Figure 7b, which is consistent with more effective
electron relaxation in the complex of higher charge.22
The presence of unpaired spin in the π system of the aryl
substituent of a coordinated ligand is not necessarily evidence
for metal-ligand π-covalency.22,80,85 Nevertheless, since sig-
nificant back-donation within compound 2 was unambiguously
established by means of IR spectroscopy and X-ray crystal-
lography (vide supra), dπ-pπ* back-bonding may be expected
to be an important contributor to the mechanism of unpaired
spin delocalization in 2. Signs of spin densities on the nuclei
of the xylyl isocyanide ligands in 2 can be determined
straightforwardly22 from the directions of the corresponding
paramagnetic shifts (Figure 8). By definition, the spin on the
metal center is aligned parallel to the field, i.e., it is positive or
“v”.22 Even though the 13C resonance of the terminal isocyanide
1
Figure 9 shows Curie-type plots for H paramagnetic shifts
of 2 and 7. In both cases, the contact shifts of the meta
hydrogens follow the Curie law82 exactly within experimental
(83) Levy, D. H. Mol. Phys. 1966, 10, 233.
(84) (a) Bertini, I.; Luchinat, C. In Physical Methods for Chemists; Drago,
R. S., Ed.; Saunders College Publishing: Philadelphia, PA, 1992. (b)
Bloembergen, N. J. Chem. Phys., 1957, 27, 575.
(85) (a) Fitzgerald, R. J.; Drago, R. S. J. Am. Chem. Soc. 1967, 89, 2879.
(b) La Mar, G. N.; Sherman, E. O.; Fuchs, G. A. J. Coord. Chem. 1971, 1,
289.
(86) Horrocks, W. D., Jr.; Taylor, R. C.; La Mar, G. N. J. Am. Chem.
Soc. 1964, 86, 3031.
(87) See ref 22, Table 3-VI on p 107.
(88) (a) Owen, J.; Thornley, J. H. M. Rep. Prog. Phys. 1966, 29, 675.
(b) Orgel, L. J. Chem. Phys. 1959, 30, 1617. (c) Orgel, L. E. Discuss.
Faraday Soc. 1958, 26, 92.

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4689
Figure 10. 1H isotropic shift ratios for 1 (a) and 6 (b) vs temperature.
group around the N-Xyl bond and/or bending of the isocyanide
ligand at nitrogen. This is a consequence of direct interaction
of the aryl σ system with the π system of the CN group.91 At
the same time, the ortho Me and para hydrogen shifts will be
proportional to the overlap between the π systems of the CN
and Xyl fragments. Should the analogy hold, for a [M(CNXyl)6]z
complex the ratio ∆δ(o-Me)/∆δ(p-H) has to be nearly temper-
ature independent while the ratios ∆δ(m-H)/∆δ(p-H) and ∆δ-
(m-H)/∆δ(p-H) may vary with temperature. Figure 10 clearly
illustrates that this, indeed, is the case for complexes 2 and 7.
Importantly, not only do the ratios ∆δ(o-Me)/∆δ(p-H) remain
practically constant over the studied temperature intervals
(circles in Figure 10), they are also of nearly the same magnitude
for 2 and 7. Thus, temperature dependence of the average
orientation of the xylyl group with respect to the CN π system
can reasonably account for the nonzero intercepts obtained for
the para and methyl hydrogens in Figure 9. The paramagnetic
shifts of the meta H resonances for 2 and 7 follow the Curie
law rather accurately since, as in the case of benzyl radical,91
the magnitudes of meta H isotropic hyperfine coupling constants
may not vary significantly with the orientation of the aryl
substituents.
Finally, it is important to compare parameters ∆δH/µeff92 for
2 and 7. The ratios in eqs 1-3 are all greater than 1. This can
be explained by greater contribution of dπ-pπ* back-bonding
to the mechanism of unpaired spin delocalization for 2 than for
7. The paramagnetic shifts of ortho methyl and para hydrogens
reflect primarily the extent of unpaired spin delocalization into
the π systems of the aromatic rings and, therefore, depend on
the overlap between the relevant CN and Xyl π systems in 2
and 7. This is why the magnitudes of the ratios in eqs 2 and 3
1
Figure 9. Curie plots for H isotropic shifts of 2 (a) and 7[BF4] (b).
errors in the studied temperature ranges. On the other hand,
intercepts obtained for the shifts of the para and methyl hy-
drogens are somewhat large to be attributed to experimental
uncertainties. Complexes 2 and 7 possess degenerate T2g ground
states (assuming ideal octahedral symmetries), so their magnetic
moments (and, hence, average g values) are not necessarily
temperature independent.24a,80,84 This, in principle, may lead to
deviations from the Curie law, which presupposes a constant
isotropic g value.89 Notably, however, the magnetic moments
of both V(CO)6 and V(PF3)6 are nearly constant in the
temperature range considered for 2 and 7 in Figure 9. The use
of equations,90 derived for the contact isotropic shifts in an
octahedral (t2g)5 field, could not explain the temperature behavior
of the 1H paramagnetic shifts observed for 2 and 7. This fact is
not surprising, in particular, because the free ion spin-orbit
coupling constants, λ,74 employed in these equations are
significantly altered by back-bonding, especially in the case of
zerovalent 2.
Taking into account the rather crowded nature of the CNXyl
ligands (vide infra), it is reasonable to suggest that the above
deviations from the Curie behavior may be caused, at least in
part, by temperature dependence of the average orientation of
the xylyl rings relative to the CN π systems.22 By analogy with
the previous study on the benzyl radical,22,91 one might anticipate
that at a given temperature the magnitude of the meta H contact
shift will not change significantly upon rotation of the xylyl
71b
72
(89) For recent discussions of theory of the paramagnetic shift for strong
field d5 complexes, see: McGarvey, B. R.; Batista, N. C.; Bezerra, C. W.
B.; Schultz, M. S.; Franco, D. W. Inorg. Chem. 1998, 37, 2865 and
references therein.
(90) Davis, D. G.; Kurland, R. J. J. Chem. Phys. 1967, 46, 388.
(91) Pople, J. A.; Beveridge, D. L. J. Chem. Phys. 1968, 49, 4725.
(92) To a first approximation, magnitude of a contact shift is proportional
to the isotropic g-value and, therefore, to the magnetic moment of the
complex.22,80,84

4690 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Barybin et al.
are practically identical and remain the same over the studied
temperature interval. Direct interaction of the meta hydrogens
of the aryl substituents with unpaired spin in the CN π sys-
tems91 (vide infra) may account for the higher value of the ratio
in eq 1.
of two vanadium(1+) centers seems to be unlikely because this
does not happen with the isoelectronic low-spin d4 ion 8 and
the related low-spin d5 complexes 2, 7, and [Mn(CNAr)6]2+ 78
.
A reasonable hypothesis, which may account for the unusual
magnetic behavior of 4, would be severe distortion of the
octahedral geometry of 4 in solution. In principle, the distortion
may be caused by an interaction of 4 with the solvent. Such
interaction, however, cannot be very strong since removal of
the solvent (THF or CH2Cl2) from solutions of 4[PF6] quanti-
tatively regenerates the unsolvated and fully paramagnetic
compound.
∆δm-H(2) ∆δm-H(7)
/
) 1.42
) 1.21
(T ) 293 K)
(T ) 293 K)
(T ) 293 K)
(1)
(2)
(3)
µeff(2)
µeff(7)
∆δp-H(2) ∆δp-H(7)
/
µeff(2)
µeff(7)
trans-V(CO)2(CNXyl)4 (5). Compound 5 possesses the
effective magnetic moment of 1.75 µB at 22 °C in the solid
state. This value is consistent with the low-spin d5 configuration
of 5 and is virtually identical with µeff ) 1.76 and 1.73-1.78
µB obtained for 2 and V(CO)6,71 respectively. The trans
geometry of 5, unambiguously established by X-ray crystal-
lography (vide supra), implies a tetragonal distortion from
octahedral symmetry.94 The t2g-like d orbitals of vanadium
should be split into b (dxy) and e (dxz, dyz) sets under C4
symmetry. Since CO is a better π acceptor compared to CNXyl,
the e orbitals must be lower in energy than the b orbital, leading
to a nondegenerate 2B ground state of 5. Therefore, one might
anticipate the electron-spin relaxation time for 5 to be rather
long.80,84a In accord with this expectation, no 1H NMR signals
were detected for C6D6 solutions of 5 at room temperature. On
the other hand, a reasonably resolved ESR spectrum of 5 was
obtained in toluene glass. The latter appears to consist of two
partially overlapping sets of eight equally spaced lines with gav
) 2.023 (g ) 2.038, g| ) 1.993). These g values are similar
to those reported for the analogous low-spin d5 complexes trans-
V(CO)2(dmpe)264 and trans-[Mo(CO)2(CNCy)4]+.95 Hyperfine
∆δo-Me(2) ∆δo-Me(7)
/
) 1.22
µeff(2)
µeff(7)
[V(CNXyl)6]- (3). Pure Cs3 and [Cs(18-Crown-6)2]3 are
diamagnetic in solution and in the solid state. They both exhibit
1H and 13C NMR patterns consistent with equivalently bound
2,6-dimethylphenyl groups. These spectra closely resemble those
obtained for the isoelectronic low-spin d6 Cr(CNXyl)6 com-
20
plex. As neutral 2, anion 3 does not undergo rapid ligand
exchange with free CNXyl. On the other hand, intermolecular
electron transfer between 2 and 3 is very fast on the NMR time
1
scale. Indeed, the room temperature H and 13C NMR spectra
of mixtures, containing 2 and 3, constitute weighted averages
of the patterns observed for the individual components.27
[V(CNXyl)6]+ (4). Complex 4[PF6] is paramagnetic in the
solid state with µeff ) 2.74 µB at room temperature, which is
consistent with low-spin d4 configuration and octahedral
geometry of 4. Surprisingly, solutions of 4[PF6] in THF were
found to be only weakly paramagnetic as determined by the
Evans method.25 1H and 13C NMR spectra of 4[PF6] in THF-d8
also indicated loss of the expected paramagnetism in solution.27
Indeed, these spectra are almost identical with those of
splitting (A ) 31 G, A ) 80 G) in the ESR spectrum of 5 is
||
due to interaction of the single electron spin with 51V nucleus
7
(I ) /2, 99.8%). The largest hyperfine coupling component
corresponds to the smallest g component, which is consistent
1
diamagnetic [M(CNXyl)7]+ (M ) Nb, Ta).9b The H NMR
2
with a B ground state of 5 and, hence, trans geometry of this
pattern of 4[PF6] at room temperature consists of two somewhat
broadened singlets at 2.25 (6H) and 7.13 ppm (3H) correspond-
ing to methyl and aromatic hydrogens of the CNXyl ligands,
respectively. Cooling the sample to -60 °C resulted in only
very small (<0.1 ppm) downfield shifts of both resonances.
Thus, the isocyanides in the complex are equivalent on the NMR
time scale and did not undergo fast exchange with added free
isocyanide. Room temperature 31P and 19F NMR spectra of
3[PF6] were unexceptional and remained unchanged upon
lowering the temperature to -100 °C. Moreover, substitution
of the [PF6]- anion by either [BF4]- or [BPh4]- did not affect
the 1H NMR characteristics of the cation.27 Changing the solvent
from THF-d8 to CD2Cl2 resulted in practically the same 1H and
13C NMR patterns for 4[PF6]. To check whether [Cr(C-
NXyl)6]2+, isoelectronic with 4, would exhibit similar behavior
in solution, complex [Cr(CNXyl)6][PF6]2 (8[PF6]2) was prepared.
compound.94 The average hyperfine splitting of 47 G, observed
for 5, is somewhat greater than Aav ) 36 G, found for V(CO)6
in toluene glass,79b indicating that the vanadium center in the
former species is more electron rich. This is, of course, due to
CO being a better π acceptor compared to CNXyl. Of note, the
average hyperfine coupling constants for bis(η6-naphthalene)-
vanadium(0) and tris(2,2′-dipyridyl)vanadium(0) were reported
to be 6618 and 83.5 G,96 respectively.
Concluding Remarks
Syntheses of the first 16-, 17-, and 18-electron homoleptic
isocyanides of vanadium, [V(CNXyl)6]z (z ) 1+, 0, 1-), were
successfully accomplished. These thermally stable but air-
sensitive complexes were isolated in high yields and were
carefully characterized by a variety of spectroscopic methods
and X-ray crystallography. The bulky nature of xylyl isocyanide
appears to be of critical importance for the stability of zerovalent
V(CNXyl)6, oxidation and reduction of which provided novel
[V(CNXyl)6]+ and [V(CNXyl)6]-, respectively. Synthesis of
trans-V(CO)2(CNXyl)4 under mild conditions is very remarkable
in that no disproportionation of V(CO)6 in the presence of xylyl
isocyanide occurred. Both V(CNXyl)6 and trans-V(CO)2-
(CNXyl)4 constitute the first examples of paramagnetic metal-
(0) compounds possessing at least one isocyanide ligand. The
1
However, the H NMR spectrum of this compound in CD2Cl2
3
was fully in accord with the anticipated ideal degenerate T1g
ground state of the octahedral low-spin d4 dication 8.27 As in
1
the case of low-spin d5 2 and 7, strong H paramagnetic shifts
in both directions were observed for complex 7. Their magni-
tudes increased with decreasing temperature and the m-H
paramagnetic shift followed the Curie law exactly.
On the basis of the above facts, interaction of 4 with
counterions can be safely ruled out as the cause of quenching
of its paramagnetism in solution. Dimerization of 4, either direct
or through the aryl rings,93 leading to antiferromagnetic coupling
(94) Rieger, P. H. Coord. Chem. ReV. 1994, 135, 203.
(95) Conner, K. A.; Walton, R. A. Organometallics 1983, 2, 169.
(96) Ko¨nig, E. Z. Naturforsch. 1964, 19a, 1139.
(93) Bond, A. M.; Colton, R. Inorg. Chem. 1976, 15, 2036.

Paramagnetic ZeroValent Transition Metal Isocyanides
J. Am. Chem. Soc., Vol. 122, No. 19, 2000 4691
average V-CN bond length in the described vanadium isocya-
nides is a sensitive function of the extent of back-bonding to
the isocyanide ligands. This distance increases in the series
isocyanide ligand in stabilizing electron-rich group 5 transition
metals. In particular, the existence of [V(CNXyl)6]-, the first
homoleptic octahedral isocyanide metalate,41 confirms the recent
suggestion that the ability of isocyanides to form subvalent metal
complexes has, indeed, been underestimated for decades.2g The
reaction chemistry of the electron rich vanadium isocyanides,
reported herein, promises to be fruitful and is currently under
examination. Also, extension of this research to group 4
transition elements is underway in this laboratory.
[V(CNXyl)6]- < V(CNXyl)6 < trans-V(CO)2(CNXyl)4
<
[V(CNXyl)6]+. The infrared cyanide νCN stretching frequencies
also increase in the same order.
The vast majority of paramagnetic transition metal com-
pounds, for which NMR spectra have been obtained, contain
metals in positive oxidation states. Comparison of the NMR
patterns for zeroValent 2 and monoValent 7 indicates the
importance97 of back-bonding in the mechanism of unpaired
electron spin delocalization at least in the case of 2. Contrary
to the previous prediction for dπ(M)-pπ*(L) unpaired spin
delocalization in low-spin d5 octahedral complexes,87 it appears
that dπ(V)-pπ*(CNXyl) back-donation within 2 induces excess
negative spin density (spin polarization) on the isocyanide
ligands. This fact, however, is consistent with the correlation
requirements at the vanadium(0) center.
Acknowledgment. We dedicate this article to Professor M.
L. H. Green on the occasion of his 65th birthday. The National
Science Foundation and donors of the Petroleum Research Fund,
administered by the American Chemical Society, are gratefully
acknowledged for financial support. M.V.B. thanks the Uni-
versity of Minnesota for the 1998-1999 Doctoral Dissertation
Fellowship. We are also indebted to Professor Kent R. Mann
for valuable discussions.
Together with our other recent contribution,9b this study
clearly demonstrates the exceptional versatility of the 2,6-xylyl
Supporting Information Available: Details of crystal-
lographic analyses for 2, Cs3, 4[PF6], and 5; complete tables
of atomic coordinates, bond distances, angles, and anisotropic
displacement parameters for the above structures (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(97) Strictly speaking, we can only conclude that back-bonding is an
important but not necessarily dominant contributor22 to the mechanism of
unpaired electron density delocalization within 2. This caution is based on
Drago’s doubt85a that unpaired electron density in the π systems of the aryl
substituents of [Ni(NCPh)6]2+ (Kluiber, R. W.; Horrocks, W. D., Jr. Inorg.
Chem. 1966, 5, 152) is a result of nickel-benzonitrile π-bonding. Similarly,
one has to be cautious invoking metal-ligand π-covalency to explain the
NMR patterns obtained for [Mn(CNAr)6]2+ 78
.
JA000212W