Different ways to distort a tetracapped tetrahedron on route to forming an E4M4 cubane: The case of [E4(Pd(PPh2Me)2)4] [Ph2EX2]2 (E = Sb, X = Cl; E = Bi, X = Br)

Article abstract of DOI:10.1021/ja982902u

Tetrakis(diphenylmethylphosphine)palladium reacts with diphenylantimony chloride or diphenyl-bismuth bromide to give [E₄(PdL₂)₄][Ph₂EX₂] ₂ (L = PPh₂Me, 1: E = Sb, X = Cl; 2: E = Bi, X = Br) which have been characterized spectroscopically and by single-crystal X-ray diffraction for [Sb₄(PdL₂)₄] [Ph₂SbCl₂]₂· 0.5THF and [Bi₄(PdL₂)₄] [Ph₂BiBr₂]₂. These compounds are electron-rich based on electron counting formalisms. The additional cluster electrons can be rationalized by the ability of group 15 elements to show hypervalency, particularly those elements such as Sb and Bi which show more metal character. The electronic structure of the compounds and of related species has been examined by EHT and DFT calculations. Relationships to other cubane-derived structures are derived, and the stability of structurally related M₄E₄ hypothetical clusters is discussed. Compounds 1 and 2 decompose thermally to give Bi₂Pd and SbPd.

Full text of DOI:10.1021/ja982902u

J. Am. Chem. Soc. 1999, 121, 4409-4418
4409
Different Ways To Distort a Tetracapped Tetrahedron on Route to
Forming an E4M4 Cubane: The Case of [E4(Pd(PPh2Me)2)4][Ph2EX2]2
(E ) Sb, X ) Cl; E ) Bi, X ) Br)
Joseph L. Stark,† Brian Harms,† Ilse Guzman-Jimenez,† Kenton H. Whitmire,*,†
Re´gis Gautier,‡ Jean-Franc¸ois Halet,‡ and Jean-Yves Saillard*,‡
Contribution from the Department of Chemistry, Rice UniVersity, Houston, Texas 77005-1892, and
Laboratoire de Chimie du Solide et Inorganique Mole´culaire, UMR CNRS 6511, UniVersite´ de Rennes 1,
35042 Rennes Cedex, France
ReceiVed August 12, 1998
Abstract: Tetrakis(diphenylmethylphosphine)palladium reacts with diphenylantimony chloride or diphenyl-
bismuth bromide to give [E4(PdL2)4][Ph2EX2]2 (L ) PPh2Me, 1: E ) Sb, X ) Cl; 2: E ) Bi, X ) Br) which
have been characterized spectroscopically and by single-crystal X-ray diffraction for [Sb4(PdL2)4][Ph2SbCl2]2‚
0.5THF and [Bi4(PdL2)4][Ph2BiBr2]2. These compounds are electron-rich based on electron counting formalisms.
The additional cluster electrons can be rationalized by the ability of group 15 elements to show hypervalency,
particularly those elements such as Sb and Bi which show more metal character. The electronic structure of
the compounds and of related species has been examined by EHT and DFT calculations. Relationships to
other cubane-derived structures are derived, and the stability of structurally related M4E4 hypothetical clusters
is discussed. Compounds 1 and 2 decompose thermally to give Bi2Pd and SbPd.
Introduction
structure is that in which A is a transition metal fragment (M)
and B is a main group element functionality (E). These
fragments may be either “naked” or bear a substituent such as
an alkyl or aryl group.10-32 Regular homonuclear cubes of the
The cube, aesthetically satisfying for its high symmetry, is
well-represented in elements in both the p-block and d-block
and in combinations of the two. The prototypical cubane is C8H8,
in which the eight carbon atoms occupy the vertices of a regular
cube. The most widespread type of cubane structure is that in
which two types of vertices are present, A4B4, and the different
elements occupy alternating vertices:
(9) Schulz, S.; Andruh, M.; Pape, T.; Heinze, T.; Roesky, H. W.; Haming,
L.; Kuhn, A.; Herbst-Irmer, R. Organometallics 1994, 13, 4004.
(10) Go´mez-Sal, P.; Martin, A.; Mena, M.; Ye´lamon, C. J. Chem. Soc.,
Chem. Commun. 1995, 2185.
(11) Simon, G. L.; Dahl, L. F. J. Am. Chem. Soc. 1973, 95, 2175.
(12) Fenske, D.; Basoglu, R.; Hachgenei, J.; Rogel, F. Angew. Chem.,
Int. Ed. Engl. 1984, 23, 160.
(13) Scherer, O. J.; Dave, T.; Braun, J.; Wolmersha¨user, J. J. Organomet.
Chem. 1988, 350, C20.
(14) Ro¨ttinger, E.; Vahrenkamp, H. Angew. Chem., Int. Ed. Engl. 1978,
17, 273.
(15) Rheingold, A. L.; Foley, M. J.; Sullivan, P. J., J. Am Chem. Soc.
1982, 104, 4727.
(16) Foust, A. S.; Dahl, L. F. J. Am. Chem. Soc. 1970, 92, 7337.
(17) Ciani, G.; Moret, M.; Fumagalli, A.; Martinengo, S. J. Organomet.
Chem. 1989, 362, 291.
(18) Ro¨ttinger, E.; Vahrenkamp, H. J. Organomet. Chem. 1981, 104,
4727.
(19) Trinh, T.; Teo, B. K.; Ferguson, J. A.; Meyer, T. J.; Dahl, L. F. J.
Am. Chem. Soc. 1977, 99, 408.
There are now numerous examples in which fragments A
and B both are comprised of main group elements. Examples
include [tBuGa(µ3-E)]4 (E ) S, Se, Te),1 [(R)Ga(µ3-E)]4 (R )
Me3C, tBu, EtMe2C, Et2MeC, Et3C and E ) S, Se, Te),2 (C6F5-
NMMe)4 (M ) Ga, In),3 and (LnMIn)4S4 (LnM ) Cp(CO)2Fe,
Cp(CO)3Mo).4 Especially numerous are group 13-16 (III-VI)
cubane molecules.4-9 Perhaps the most famous class of cubane
(20) Simon, G. L.; Dahl, L. F. J. Am. Chem. Soc. 1973, 95, 2164.
(21) Gall, R. S.; Chu, C. T.; Dahl, L. F. J. Am. Chem. Soc. 1974, 96,
4019.
† Rice University.
‡ Universite´ de Rennes 1.
(1) (a) Stoll, S. L.; Bott, S. G.; Barron, A. R. J. Chem. Soc., Dalton
Trans. 1997, 8, 1315 and references therein. (b) Power, M. B.; Barron. A.
R. J. Chem. Soc., Chem. Commun. 1991, 1315. (c) Power, M. B.; Ziller, J.
W.; Barron, A. R. Organometallics 1992, 11, 1055.
(2) Gillan, E. G.; Bott, S. G.; Barron, A. R. Chem. Mater. 1997, 9, 806.
(3) Belgardt, T.; Roesky, H. W.; Noltemeyer, M.; Schmidt, H. G. Angew.
Chem., Int. Ed. Engl. 1993, 32, 1056.
(4) Merzweiler, K.; Rudolph, F.; Brands, L. Naturforsh 1992, 47B, 470.
(5) Uhl, W.; Graupner, R.; Layh, M.; Schu¨tz, U. J. Organomet. Chem.
1995, 493, C1.
(6) Uhl, W.; Graupner, R.; Pohlmann, M.; Pohl, S.; Saak, W. Chem.
Ber. 1996, 129, 143.
(7) Power, M. B.; Barron, A. R.; Hynk, D.; Robertson, H. E.; Rankin,
D. W. H. AdV. Mater Optics. Electron. 1995, 5, 177.
(8) Harlan, C. J.; Gillan, E. G.; Bott, S. G.; Barron, A. R. Organometallics
1996, 15, 5479.
(22) Gall, R. S.; Connelly, N. G.; Dahl, L. F. J. Am. Chem. Soc. 1974,
96, 4017.
(23) Wei, C. H.; Wilker, G. R.; Treichel, P. M.; Dahl, L. F. Inorg. Chem.
1966, 5, 900.
(24) Curtis, M. D.; Williams, P. D. Inorg. Chem. 1983, 22, 2661.
(25) Schunn, R. A.; Fritchie, C. J., Jr.; Prewitt, C. T. Inorg. Chem. 1966,
5, 892.
(26) Teo, B. K. Inorg. Chem. 1984, 23, 1251.
(27) Teo, B. K. Inorg. Chem. 1985, 24, 4209.
(28) Teo, B. K.; Longoni, G.; Chung, F. R. K. Inorg. Chem. 1984×e2
24, 1257.
(29) Harris, S. Inorg. Chem. 1987, 26, 4278.
(30) Harris, S. Polyhedron 1989, 8, 2843.
(31) Bahn, C. S.; Tan; A.; Harris, S. Inorg. Chem. 1998, 37, 2770.
(32) See also: Sung, S.-S.; Glidewell, C.; Butler, A. R.; Hoffmann, R.
Inorg. Chem. 1985, 24, 3856.
10.1021/ja982902u CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/27/1999

4410 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Stark et al.
transition metals are also known, but they have only been found
to exist with ligands bridging the square faces or edges and
may also possess interstitial atoms.33-44
Scheme 1a
The bonding in a regular cubane is straightforward. Each edge
of the cube can be thought to represent a localized 2-center,
2-electron bond. When one considers cubic structures in which
metal atoms are present, however, the bonding situations rapidly
become more interesting because of the variability in oxidation
states which are generally not possible for cubanes made of the
lighter main group elements. The metals also possess d-orbitals
which allow cross cube bonding interactions. Removal of
electrons from the cubane either via redox processes or by
change in the electron count of the fragments of A or B may
thus be accommodated theoretically by formation of cross cube
bonds. The different possible distortions are shown in Scheme
1.
Taking M-M bond formation to its logical completion, one
arrives at the tetracapped tetrahedral structure (TTd) M4(µ3-E)4.
Alternatively, beginning with the TTd M4(µ3-E)4 structure, one
can derive the cubane structure by successive bond breakage.
A number of the structures illustrated have been reported for
the class of compounds E4M4.19-25 The compounds have been
examined theoretically by Teo and a set of formalisms derived
to help explain the pattern of M-M bond formation.26-28 In
the same context, Harris has also provided a qualitative
molecular orbital rationalization of the clusters having a M4S4
core.29-32
What has been missing from the list of known compounds is
the alternative bonding pattern E4(µ3-M)4, in which the core
structure is a tetrahedron of main group atoms capped by
transition metals. The existence of this inverse bonding scheme
was suggested by preliminary calculations.45 This in an intrigu-
ing possibility because the group 15 elements phosphorus and
arsenic are known to adopt the E4 tetrahedron in one allotropic
form. Direct reaction of E4 (E ) P, As) has not led to formation
of cubane-like derivatives.46 The predominant outcome has been
for the added metal fragments to break E-E bonds and to add
to the resulting fragment in µ-Ex (x ) 2, 4-6) configurations.
Cubane structures are known for group 15 and 16 elements
arising from other types of reactions which do not begin with
an intact E4 unit. Structurally characterized examples include
[Cp′Ni(µ3-P)]4,13 E4{Fe(CO)3}4 (E ) S, Se, Te),16 and E4{Co-
a Each arrow represents the loss of one bond of the original
tetrahedrom An * denotes a chiral structure.
(CO)3}4 (E ) Sb, Bi).17 These molecules possess no M-M or
E-E bonding as anticipated from their electron counts. One
molecule that has been identified in which an E4 core is
recognizable and which retains some E-E bonding is the anion
[Bi4{Fe(CO)3}3{Fe(CO)4}]2- (12). In that molecule, three faces
of a Bi4 tetrahedron are capped by Fe(CO)3 groups but the third
face remains intact with Bi-Bi distances consistent with single-
bond values.47,48 Two other examples that are structually similar
to this are (CpxRu)3Ru(η3-As3)(µ3-η3-As3)(µ3-As)3 (13)49 and
Cp*3Fe3(P3)(µ4-P)3Fe(η3-µ-P3) (14).50 Other examples of this
rare class of molecule are herein reported.
(33) Zebrowski, J. P.; Hayashi, R. K.; Bjarnason, A.; Dahl, L. F. J. Am.
Chem. Soc. 1992, 114, 3121.
(34) Fenske, D.; Persau, C. Z. Anorg. Allg. Chem. 1991, 593, 61.
(35) Fenske, D.; Fleischer, H.; Persau, C. Angew. Chem., Int. Ed. Engl.
1989, 28, 1665.
(36) Fenske, D.; Ohmer, J.; Merzweiler, K. Angew. Chem., Int. Ed. Engl.
1988, 27, 1512.
(37) Brennen, J. G.; Siegrist, T.; Stuczyrski, S. M.; Steigerwald, M. L.
J. Am. Chem. Soc. 1989, 111, 9240.
(38) Whitmire, K. H.; Eveland, J. R. J. Chem. Soc., Chem. Commun.
1994, 1335.
(39) Furet, E.; Le Beuze, A.; Halet, J.-F.; Saillard, J.-Y. J. Am. Chem.
Soc. 1994, 116, 274 and references therein.
(40) Furet, E.; Le Beuze, A.; Halet, J.-F.; Saillard, J.-Y. J. Am. Chem.
Soc. 1995, 117, 4936 and references therein.
(41) Halet, J. F.; Saillard, J.-Y. Struct. Bond. 1997, 87, 81 and references
therein.
(42) Ro¨sch, N.; Ackeman, L.; Pacchioni. G. Inorg. Chem. 1993, 32, 2963.
(43) (a) McCandlish, L. E.; Bissell, E. C.; Coucouvanis, D.; Facker, J.
P.; Knox, K. J. Am. Chem. Soc. 1968, 90, 7357. (b) Hollander, F. J.;
Coucouvanis, D. J. Am. Chem. Soc. 1974, 96, 5646. (c) Avdeef, A.; Fackler,
J. P. Inorg. Chem. 1978, 17, 2182.
(44) (a) Haiduc, I.; Sowerby, D. B.; Lu, S. F. Polyhedron 1995, 14, 3389.
(b) Liu, C. W.; Stubbs, T.; Staples, R. J.; Fackler, J. P. J. Am. Chem. Soc.
1995, 117, 9778.
(45) Albright, T. A.; Yee, K. A.; Saillard, J.-Y.; Kahlal, K.; Halet, J. F.;
Leigh, J. S.; Whitmire, K. H. Inorg. Chem. 1991, 30, 1179.
(46) Whitmire, K. H. AdV. Organomet. Chem. 1998, 42, 1.

Distortion of a Tetracapped Tetrahedron
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4411
Table 1. Crystallographic Data for [Sb4{Pd(Ph2PMe)2)4]-
Experimental Section
[Ph2SbCl2]2‚0.5THF (1) and [Bi4{Pd(Ph2PMe)2)4][Ph2BiBr2]2 (2)a
Unless otherwise specified, all operations were conducted under an
inert atmosphere using standard drybox and Schlenk techniques.51 All
solvents were distilled under nitrogen from appropriate drying agents.52
1H and 31P NMR spectra were recorded on a Bruker AC-250
spectrometer, operating at 250 MHz for 1H and 101 MHz for31P.
Infrared spectra were recorded as Nujol mulls on a Perkin-Elmer 1640
spectrophotometer. The compounds Ph2SbCl53 and Ph2BiBr54 were
prepared by literature methods. The compound Pd[PPh2Me]4 was
purchased from Aldrich and used as received. Powder X-ray diffraction
patterns were recorded on a Siemens 5000 powder diffractometer.
Elemental analyses were performed by Desert Analytics.
Preparation of [Sb4{Pd(PPh2Me)2}4][Ph2SbCl2]2 (1). Solid portions
of PdL4 (0.15 g, 0.16 mmol) and Ph2SbCl (0.051 g, 0.16 mmol) were
weighed into a flask, dissolved in tetrahydrofuran (30 mL), and stirred
at room temperature for 15 h, during which time the solution became
dark red to black and a precipitate developed. The volume of solution
was reduced to ca. 5 mL. The dark red solid was collected by filtration,
washed twice with ether (ca. 30 mL), and then dried under vacuum.
Crystals suitable for X-ray diffraction were obtained by dissolving PdL4
(0.029 g, 0.032 mmol) into 2 mL of THF in a 12 × 75 mm test tube
and then suspending Ph2SbCl (0.01 g, 0.032 mmol) on top of the
solution. Block-shaped crystals deposited over a period of 1 week at
room temperature. Yields were typically 15% yield. Anal. Calcd for
empirical formula C130H128Cl4O0.50P8Pd4Sb6 C128H124Bi6Br4P8Pd4
formula weight
crystal system
space group
a (Å)
b (Å)
c (Å)
3243.98
triclinic
P1h
3909.15
triclinic
P1h
14.176(3)
14.770(3)
32.465(7)
85.08(3)
77.96(3)
85.23(3)
6608.6(23)
2
14.395(3)
14.974(3)
32.363(7)
84.75(3)
77.78(3)
85.59(3)
6777.3(24)
2
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
GOF on F2
final R indices
[I > 2σ(I)]
R indices
(all data)
1.108
R1 ) 0.0386
1.039
R1 ) 0.0550
wR2 ) 0.1334
wR2 ) 0.1658
a w ) I/[σ2(Fo ) + (0.0867P)2], w ) I/[(σ2(Fo ) + (0.0200P)2 +
2
2
35.9373P], P ) (Fo + 2Fc2)/3.
2
give black polycrystalline material. Formation of Bi2Pd was confirmed
by powder X-ray diffraction analysis. Yield: 8 mg (0.15 mmol) of
black polycrystalline material was recovered, making the yield 33%
(based on Bi) for formation of Bi2Pd.
C
130H128O0.5Cl4P8Pd4Sb6 (1‚0.50C4H8O): C, 48.15; H, 3.98; P, 7.64.
1
Found: C, 49.35; H, 3.99; P, 8.12. H NMR (CD2Cl2) 27 °C: δ 7.72
(m, o-H, Ph), 7.17 (m, m,p-H, Ph), 1.98 (d, JP-H ) 1.4 Hz, 24 H, Me)
31P NMR (CD2Cl2): δ -9.06 (s, PPh2Me). NMR analyses of the
reaction solution showed free phosphine and L2PdCl2 as byproducts of
the reaction, but these were not quantitated.
Single-Crystal X-ray Diffraction Studies of 1 and 2. Data were
collected on a Rigaku AFC5S fully automated, four-circle single-crystal
X-ray diffractometer (Rigaku CONTROL Automatic Data Collection
Series, Molecular Structure Corp., The Woodlands, TX)55 using graphite
monochromated Mo KR radiation (0.7107 Å). Data collection and
refinement parameters are given in Table 1. The crystals were mounted
on a glass fiber with Epoxy cement, and data were collected with ω
scans at 4 deg/min. Three standard reflections were monitored for decay
every 150 reflections throughout data collection. An absorption
correction from azimuthal (ψ) scans was applied to each data set. The
programs used in solving each structure were part of the Siemens
Analytical X-ray Instruments data reduction and refinement package
SHELXTL PC,56 and refinement of each structure was performed using
the data refinement program package SHELXL-93.57 The scattering
factors used were those found in the International Tables for X-ray
Crystallography.58
Preparation of [Bi4{Pd(PPh2Me)2}4][Ph2BiBr2]2 (2). Solid PdL4
(0.15 g, 0.16 mmol) was added to a solution of Ph2BiBr (0.073 g, 0.16
mmol) in THF (30 mL) and stirred for 15 h, forming a black solution
that contained a precipitate. The volume of the solution was reduced
to ca. 5 mL, after which the solid was collected by filtration, washed
twice with ether (ca 30 mL), and dried under vacuum. Crystals suitable
for X-ray diffraction were obtained by dissolving PdL4 (0.0205 g, 0.023
mmol) and Ph2BiBr (0.01 g, 0.023 mmol) into 2 mL of THF in a 12 ×
75 mm test tube and allowing the test tube to stand for 1 week. Yields
were typically 16% yield. Anal. Calcd for C128H124Br4P8Pd4Bi6: C,
1
39.33; H, 3.20; P, 6.34. Found: C, 39.23; H, 3.60; P, 6.70. H NMR
(CD2Cl2, 27° C): δ 7.72 (m, o-H, Ph), 7.17 (m, m,p-H, Ph), 1.95 (d,
JP-H ) 1.4 Hz, 24 H, Me). 31P NMR (CD2Cl2): δ -4.23 (s, PPh2Me).
NMR analyses of the reaction solution showed free phosphine and L2-
PdBr2 as byproducts of the reaction, but these were not quantitated.
Formation of PdSb (3). A portion of 1 (44 mg, 0.014 mmol) was
initially suspended in 50 mL of toluene and refluxed overnight. The
toluene was filtered off, and the residual black solid was placed in a
platinum pan and heated to 600 °C for 1 h under argon, after which
time the sample was cooled to give 10 mg of black polycrystalline
material, which was an 80% yield based upon Pd. Formation of
polycrystalline PdSb was confirmed by powder X-ray crystallography.
Formation of Bi2Pd (4). Compound 2 (60 mg, 0.015 mmol) was
suspended into 50 mL of toluene and refluxed overnight to give a black
precipitate. The toluene was filtered off, and the black residue was
placed in a platinum pan and heated for 1 h at 600 °C under argon to
Computational Details. (a) Extended Hu1ckel Calculations. Cal-
culations have been carried out within the extended Hu¨ckel formalism59
using the weighted Hij formula60 with the program CACAO.61 The
exponents (ú) and the valence shell ionization potentials (Hii in eV)
were (respectively): 1.3, -13.6 for H 1s; 1.625, -24.4 for C 2s; 1.625,
-11.4 for C 2p; 2.275, -32.3 for O 2s; 2.275, -14.8 for O 2p; 1.6,
-18.6 for P 3s; 1.6, -14.0 for P 3p; 2.323, -18.8 for Sb 5s; 1.999,
-11.7 for Sb 5p; 1.9, -9.10 for Fe 4s; 1.9, -5.32 for Fe 4p; 2.19,
-7.32 for Pd 5s; 2.152, -3.75 for Pd 5p. Hii values for Fe 3d and Pd
4d were set equal to -12.6 and -12.02, respectively. A linear
combination of two Slater-type orbitals of exponents ú1 ) 5.35, ú2 )
1.80 and ú1 ) 5.983, ú2 ) 2.613 with the weighting coefficients c1 )
0.5366, c2 ) 0.6678 and c1 ) 0.5264, c2 ) 0.6373 were used to
represent the Fe 3d and Pd 4d atomic orbitals, respectively. The
following bond distances (Å) and angles (deg) were used in the Td
Fe4(CO)12(µ3-Sb)4 model: Fe-Sb ) 2.50, Sb-Sb ) 3.05, Fe-C )
1.85, C-O ) 1.15. The molecular model Sb4[µ3-Pd(PH3)2]4 used was
based on the averaged idealized (D2d) experimental structure of 1. The
(47) Whitmire, K. H.; Albright, T. A.; Kang, A.-K.; Churchill, M. R.
Fettinger, J. C. Inorg. Chem. 1986, 25, 2799.
(48) Whitmire, K. H.; Churchill, M. R.; Fettinger, J. C. J. Am. Chem.
Soc. 1985, 107, 1056.
(49) Scherer, O. J.; Blath, C.; Heckmann, G.; Wolmersha¨user. J.
Organomet. Chem. 1991, 409, C15.
(50) Ahlrichs, R.; Fenske, D.; Fromm, K. Krautscheid, H.; Treutler, O.
Chem. Eur. J. 1996, 2, 238.
(55) TEXSAN: Single-Crystal Structure Analysis Software 5.0, Molecular
Structure Co., The Woodlands, TX, 1990.
(56) Sheldrick, G. M. SHELXTL PLUS PC 5.0, Siemens Crystallographic
Research Systems, Madison, WI, 1995.
(51) Shriver, D. F.; Drezdon, M. A. The Manipulation of Air-SensitiVe
(57) Shedrick, G. M. SHELXL-93, Go¨ttingen, Germany, 1993.
(58) International Tables for Crystallography; Kluwer Academic Pub-
lishers: Dordrecht, The Netherlands, 1992; Vol. C; Tables 4.2.6.8. and
6.1.1.4.
(59) Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.
(60) Ammeter, J. H.; Bu¨rgi, H.-B.; Thibeault, J. C.; Hoffmann, R. J.
Am. Chem. Soc. 1978, 100, 3686.
Compounds; Wiley: New York, 1986.
(52) (a) Gordon, A. J.; Ford, R. A. The Chemist’s Companion; John Wiley
and Sons: New York, 1972. (b) Pangborn, A. B.; Giardello, M. A.; Grubbs,
R. H.; Rosen, R. K.; Timmers, F. J. Organometallics 1996, 15, 1518.
(53) Nunn, M.; Sowerby, D. B.; Wesolek, D. M. J. Organomet. Chem.
1983, 251, C45.
(54) Wilkinson, J. F.; Challenger, F. J. Chem. Soc. 1924, 12.
(61) Mealli, C.; Proserpio, D. J. Chem. Educ. 1990, 67, 399.

4412 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Stark et al.
Table 2. Selected Bond Lengths (Å) and Angles (deg) for
following bond distances (Å) were used in this model: Pd-Sb ) 2.64
and 2.76, Sb-Sb ) 3.05 and 3.39, Pd-P ) 2.36, P-H ) 1.42.
(b) Density Functional Calculations. Density functional calcula-
tions62 were carried out on [Sb4Fe4(CO)12]2+, [Sb4Pd4(PH3)8]2+, E4Co4-
(CO)12 (E ) Bi, Sb), and [E′4{µ3-Fe(CO)3}4]2+ (E ) Ga, In, Tl), using
the Amsterdam Density Functional (ADF) program.63 Electron cor-
relation was treated within the local density approximation (LDA).64
The numerical integration procedure applied for the calculations was
developed by te Velde et al.65 The atom electronic configurations were
described by a triple-ú Slater-type orbital (STO) basis set for H 1s, C
2s and 2p, O 2s and 2p, P 3s and 3p, Ga 3d, 4s, and 4p, In 4d, 5s, and
5p, Sb 4d, 5s, and 5p, Tl 5d, 6s, and 6p, Bi 5d, 6s, and 6p, Fe 4s and
3d, Co 4s and 3d, Pd 5s and 4d and single-ú STO function for H 2p,
C 3d, O 3d, P 3d, Ga 4d, Fe 4p, Co 4p and Pd 5p. The frozen-core
approximation66 was used.
[Sb4{Pd(PPh2Me)2}4][Ph2ECl2]2‚0.5THF (1)
Bond Lengths
Sb(1)-Pd(2)
Sb(1)-Pd(1)
Sb(1)-Sb(2)
Sb(2)-Pd(1)
Sb(2)-Sb(4)
Sb(3)-Pd(4)
Sb(3)-Sb(4)
Sb(4)-Pd(2)
Pd(1)-P(2)
Pd(2)-P(3)
Pd(3)-P(6)
Pd(4)-P(7)
2.6380(10)
2.8245(10)
3.0726(10)
2.6501(10)
3.0530(12)
2.6508(9)
3.0920(13)
2.6705(12)
2.355(2)
Sb(1)-Pd(3)
Sb(1)-Sb(3)
Sb(2)-Pd(4)
Sb(2)-Pd(2)
Sb(3)-Pd(1)
Sb(3)-Pd(3)
Sb(4)-Pd(3)
Sb(4)-Pd(4)
Pd(1)-P(1)
Pd(2)-P(4)
Pd(3)-P(5)
Pd(4)-P(8)
2.6662(10)
3.0442(12)
2.6288(11)
2.7734(10)
2.6358(10)
2.7745(10)
2.6273(9)
2.8059(9)
2.375(2)
2.369(2)
2.362(2)
2.374(2)
2.376(2)
2.361(2)
2.359(2)
Bond Angles
89.18(4) Pd(2)-Sb(1)-Pd(1) 110.44(4)
Pd(2)-Sb(1)-Pd(3)
Results
Pd(3)-Sb(1)-Pd(1) 110.82(4) Pd(2)-Sb(1)-Sb(3) 105.24(4)
Formation of 1 and 2 occurs by the direct addition of 1 equiv
of PdL4 to 1 equiv of either Ph2SbCl or Ph2BiBr (eq 1).
Pd(3)-Sb(1)-Sb(3)
Pd(2)-Sb(1)-Sb(2)
Pd(1)-Sb(1)-Sb(2)
Pd(4)-Sb(2)-Pd(1)
57.68(3) Pd(1)-Sb(1)-Sb(3)
57.51(2) Pd(3)-Sb(1)-Sb(2) 103.27(3)
53.21(3) Sb(3)-Sb(1)-Sb(2) 67.72(3)
86.42(4) Pd(4)-Sb(2)-Pd(2) 112.82(4)
53.22(2)
PdL4 + Ph2EX f [E4{PdL2}4][Ph2EX2] +...
(1)
Pd(1)-Sb(2)-Pd(2) 111.67(4) Pd(4)-Sb(2)-Sb(4)
Pd(1)-Sb(2)-Sb(4) 103.82(4) Pd(2)-Sb(2)-Sb(4)
Pd(4)-Sb(2)-Sb(1) 102.90(3) Pd(1)-Sb(2)-Sb(1)
58.62(2)
54.30(3)
58.60(2)
66.62(3)
1: E ) Sb, X ) Cl; 2: E ) Bi, X ) Br; L ) PPh2Me
Pd(2)-Sb(2)-Sb(1)
Pd(1)-Sb(3)-Pd(4)
53.35(3) Sb(4)-Sb(2)-Sb(1)
The reactions were performed in dry THF and led to the
formation of dark red crystals for 1 in 15% yield and black
blocklike crystals for 2 in 16% yield. The compounds are not
soluble in most organic solvents. They are sparingly soluble in
acetonitrile and in methylene chloride; however, dissolving the
compounds in any solvent leads to slow decomposition of the
products which contributes to the low isolated yields. The
decomposition was evident while attempting to obtain 31P NMR
of the products. Upon dissolving 1 or 2 in CD2Cl2 and
immediately collecting the NMR data, one resonance was
observed for 1 at δ ) -9.06 ppm and for 2 at δ ) -4.23 ppm.
If the samples are allowed to set for any period of time other
resonances begin to appear for 1 at δ ) 28.2 and 8.8 ppm and
for 2 at δ ) 28.2 and 7.4 ppm. The signal at δ ) 28.2 ppm is
due to the presence of free PPh2Me, but determination of the
nature of the other compound(s) has not yet been successful. If
excess Ph2EX is used, L2PdX2 is formed.
Taking the lability of the phosphine ligands into consideration,
an attempt was made to utilize the clusters as molecular
precursors for the synthesis of solid-state compounds. Following
the work of Steigerwald and co-workers on metal tellurides,67
1 and 2 were pyrolyzed in refluxing toluene to generate
antimony-palladium and bismuth-palladium solid-state com-
pounds. Upon pyrolysis, black precipitates were obtained but
the materials produced in this fashion are amorphous. Subse-
quent thermolysis of the black precipitates at 600 °C for 1 h
under argon led to the formation of black polycrystalline
material. Powder X-ray diffraction measurements confirmed the
compounds to be polycrystalline SbPd (3) from 1 and Bi2Pd
(4) from 2. These pyrolyses led to the thermodynamically stable
phases and, given the differing compositions of the Sb and Bi
alloys, must be independent of the starting cluster E-M
stoichiometry.
86.27(3) Pd(1)-Sb(3)-Pd(3) 113.34(4)
Pd(4)-Sb(3)-Pd(3) 110.50(4) Pd(1)-Sb(3)-Sb(1)
Pd(4)-Sb(3)-Sb(1) 103.12(3) Pd(3)-Sb(3)-Sb(1)
Pd(1)-Sb(3)-Sb(4) 103.12(4) Pd(4)-Sb(3)-Sb(4)
59.12(3)
54.30(3)
57.89(2)
66.48(3)
Pd(3)-Sb(3)-Sb(4)
Pd(3)-Sb(4)-Pd(2)
52.88(3) Sb(1)-Sb(3)-Sb(4)
89.31(3) Pd(3)-Sb(4)-Pd(4) 110.23(4)
Pd(2)-Sb(4)-Pd(4) 110.52(4) Pd(3)-Sb(4)-Sb(2) 104.76(3)
Pd(2)-Sb(4)-Sb(2)
Pd(3)-Sb(4)-Sb(3)
Pd(4)-Sb(4)-Sb(3)
Sb(3)-Pd(1)-Sb(2)
Sb(2)-Pd(1)-Sb(1)
Sb(1)-Pd(2)-Sb(2)
Sb(4)-Pd(3)-Sb(1)
Sb(1)-Pd(3)-Sb(3)
Sb(2)-Pd(4)-Sb(4)
57.50(2) Pd(4)-Sb(4)-Sb(2)
53.12(3)
57.35(2) Pd(2)-Sb(4)-Sb(3) 103.15(3)
53.15(3) Sb(2)-Sb(4)-Sb(3)
80.30(4) Sb(3)-Pd(1)-Sb(1)
68.20(3) Sb(1)-Pd(2)-Sb(4)
69.14(3) Sb(4)-Pd(2)-Sb(2)
78.90(3) Sb(4)-Pd(3)-Sb(3)
68.01(4) Sb(2)-Pd(4)-Sb(3)
68.26(3) Sb(3)-Pd(4)-Sb(4)
67.37(3)
67.67(4)
78.64(4)
68.19(3)
69.77(3)
80.41(3)
68.96(3)
Structures of 1‚0.5THF and 2. The crystal data for 1‚
0.5THF and 2 (Table 1) show the two structures to be
isomorphous, consisting of 2:1 ratios of [Ph2EX2]- ions to [E4-
{PdL2)4]2+ dications (E ) Sb, X ) Cl; E ) Bi, X ) Br). One
of the anions for 1‚0.5THF and for 2 has disordered aromatic
rings. For 1‚0.5THF, both of the rings of that one anion showed
resolvable disorder components, but for 2, disorder for only one
of the rings could be resolved. The second ring has slightly
larger than normal thermal ellipsoids, suggesting that it also
has disorder, but it was not resolvable. Thermal ellipsoid plots
of these anions showing the disorder components are included
in the Supporting Information. Selected bond distances and
angles are found in Tables 2 and 3 for 1‚0.5THF and 2. Thermal
ellipsoid plots of the metal cores of the cations are provided in
Figures 1 and 2. For 1‚0.5THF there is also one-half of a THF
molecule per cluster unit in the crystalline lattice (located on a
general position but only at half-occupancy). Despite the fact
that the two compounds are isomorphous, no lattice solvent was
found in the structure for 2 and the elemental analyses are
consistent with a solventless formulation.
(62) (a) Baerends, E. J.; Ellis, D. E.; Ros, P. Chem. Phys. 1973, 2, 41.
(b) Baerends, E. J.; Ros, P. Int. J. Quantum. Chem. 1978, S12, 169. (c) te
Velde, G.; Baerends, E. J. J. Comput. Phys. 1992, 99, 84.
(63) Baerends E. J.; et al. Amsterdam Density Functional (ADF) program,
version 2.3, Vrije Universiteit, Amsterdam, Netherlands, 1997.
(64) Vosko, S. D.; Wilk, L.; Nusair, M. Can. J. Chem. 1980, 58, 1200.
(65) Boerrigter, P. M.; te Velde, G.; Baerends, E. J. Int. J. Quantum
Chem. 1988, 33, 87.
The four main group atoms define a pseudo-tetrahedron in
which the triangular faces are capped by µ3-PdL2 fragments.
The Pd atoms are ligated by three E atoms and two phosphine
ligands in a distorted square pyramidal array. The longest Pd-E
distances are those to the E atoms in the apical positions. For
1‚0.5THF the range is Pd-Sbapical ) 2.7734(10)-2.8245(10)
Å vs Pd-Sbbasal ) 2.6273(9)-2.6705(12) Å), while for 2 Pd-
(66) Baerends, E. J. Ph.D. Thesis, Vrije Universiteit, Amsterdam,
Netherlands, 1975.
(67) Steigerwald, M. L.; Rice, C. E. J. Am. Chem. Soc. 1988, 110, 4228.

Distortion of a Tetracapped Tetrahedron
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4413
Table 3. Selected Bond Lengths (Å) and Angles (deg) for
[Bi4{Pd(Ph2PMe)2}4][Ph2BiBr2]2 (2)
Bond Lengths
Bi(1)-Pd(2)
Bi(1)-Pd(1)
Bi(1)-Bi(2)
Bi(2)-Pd(4)
Bi(2)-Pd(2)
Bi(2)-Bi(3)
Bi(3)-Pd(4)
Bi(3)-Bi(4)
Bi(4)-Pd(2)
Pd(1)-P(2)
Pd(2)-P(4)
Pd(3)-P(6)
Pd(4)-P(7)
2.730(2)
2.957(2)
3.2194(13)
2.715(2)
2.887(2)
3.5093(14)
2.748(2)
3.239(2)
2.765(2)
2.343(5)
2.346(6)
2.350(6)
2.344(5)
Bi(1)-Pd(3)
Bi(1)-Bi(3)
Bi(1)-Bi(4)
Bi(2)-Pd(1)
Bi(2)-Bi(4)
Bi(3)-Pd(1)
Bi(3)-Pd(3)
Bi(4)-Pd(3)
Bi(4)-Pd(4)
Pd(1)-P(1)
Pd(2)-P(3)
Pd(3)-P(5)
Pd(4)-P(8)
2.763(2)
3.201(2)
3.4518(13)
2.745(2)
3.2177(14)
2.718(2)
2.892(2)
2.716(2)
2.912(2)
2.352(5)
2.355(6)
2.353(5)
2.347(5)
Bond Angles
91.93(6) Pd(2)-Bi(1)-Pd(1) 109.56(6)
Pd(2)-Bi(1)-Pd(3)
Pd(3)-Bi(1)-Pd(1) 109.46(6) Pd(2)-Bi(1)-Bi(3) 105.09(5)
Figure 1. Thermal ellipsoid plot (50% probability level) of the cationic
portion of the cluster [Sb4(PdL2)4][Ph2SbCl2]2‚0.5THF. The carbon
atoms have been omitted for clarity. Full thermal ellipsoid plots and
labeling scheme are provided in the Supporting Information.
Pd(3)-Bi(1)-Bi(3)
Pd(2)-Bi(1)-Bi(2)
Pd(1)-Bi(1)-Bi(2)
Pd(2)-Bi(1)-Bi(4)
Pd(1)-Bi(1)-Bi(4)
Bi(2)-Bi(1)-Bi(4)
57.46(4) Pd(1)-Bi(1)-Bi(3)
52.21(4)
57.36(4) Pd(3)-Bi(1)-Bi(2) 103.54(5)
52.57(4) Bi(3)-Bi(1)-Bi(2)
51.53(4) Pd(3)-Bi(1)-Bi(4)
93.75(5) Bi(3)-Bi(1)-Bi(4)
57.55(3) Pd(4)-Bi(2)-Pd(1)
66.27(3)
50.35(4)
58.13(3)
88.88(5)
Pd(4)-Bi(2)-Pd(2) 111.29(6) Pd(1)-Bi(2)-Pd(2) 111.18(5)
Pd(4)-Bi(2)-Bi(4)
Pd(2)-Bi(2)-Bi(4)
Pd(1)-Bi(2)-Bi(1)
Bi(4)-Bi(2)-Bi(1)
Pd(1)-Bi(2)-Bi(3)
Bi(4)-Bi(2)-Bi(3)
Pd(1)-Bi(3)-Pd(4)
58.05(4) Pd(1)-Bi(2)-Bi(4) 103.43(5)
53.52(4) Pd(4)-Bi(2)-Bi(1) 102.84(4)
58.80(4) Pd(2)-Bi(2)-Bi(1)
64.86(3) Pd(4)-Bi(2)-Bi(3)
49.70(4) Pd(2)-Bi(2)-Bi(3)
57.37(3) Bi(1)-Bi(2)-Bi(3)
52.77(4)
50.44(4)
94.51(5)
56.61(3)
88.77(5) Pd(1)-Bi(3)-Pd(3) 112.71(6)
Pd(4)-Bi(3)-Pd(3) 109.32(5) Pd(1)-Bi(3)-Bi(1)
Pd(4)-Bi(3)-Bi(1) 102.57(4) Pd(3)-Bi(3)-Bi(1)
Pd(1)-Bi(3)-Bi(4) 103.50(5) Pd(4)-Bi(3)-Bi(4)
59.28(4)
53.64(4)
57.50(4)
64.82(3)
49.62(4)
57.12(3)
92.19(6)
Pd(3)-Bi(3)-Bi(4)
Pd(1)-Bi(3)-Bi(2)
Pd(3)-Bi(3)-Bi(2)
Bi(4)-Bi(3)-Bi(2)
52.22(4) Bi(1)-Bi(3)-Bi(4)
50.38(4) Pd(4)-Bi(3)-Bi(2)
94.19(5) Bi(1)-Bi(3)-Bi(2)
56.78(3) Pd(3)-Bi(4)-Pd(2)
Pd(3)-Bi(4)-Pd(4) 109.63(5) Pd(2)-Bi(4)-Pd(4) 109.14(6)
Pd(3)-Bi(4)-Bi(2) 104.68(5) Pd(2)-Bi(4)-Bi(2)
Pd(4)-Bi(4)-Bi(2) 52.29(4) Pd(3)-Bi(4)-Bi(3)
Pd(2)-Bi(4)-Bi(3) 103.26(5) Pd(4)-Bi(4)-Bi(3)
57.12(4)
57.30(4)
52.74(4)
51.55(4)
93.50(5)
57.05(3)
68.51(5)
77.83(5)
69.36(5)
70.48(5)
79.94(5)
69.76(5)
Figure 2. Thermal ellipsoid plot (50% probability level) of the cationic
portion of the cluster [Bi4(PdL2)4][Ph2BiBr2]2. The carbon atoms have
been omitted for clarity. Full thermal ellipsoid plots and labeling scheme
are available in the Supporting Information.
Bi(2)-Bi(4)-Bi(3)
Pd(2)-Bi(4)-Bi(1)
Bi(2)-Bi(4)-Bi(1)
Bi(3)-Pd(1)-Bi(2)
Bi(2)-Pd(1)-Bi(1)
Bi(1)-Pd(2)-Bi(2)
Bi(4)-Pd(3)-Bi(1)
Bi(1)-Pd(3)-Bi(3)
Bi(2)-Pd(4)-Bi(4)
65.84(3) Pd(3)-Bi(4)-Bi(1)
50.64(4) Pd(4)-Bi(4)-Bi(1)
57.60(3) Bi(3)-Bi(4)-Bi(1)
79.93(5) Bi(3)-Pd(1)-Bi(1)
68.63(5) Bi(1)-Pd(2)-Bi(4)
69.87(5) Bi(4)-Pd(2)-Bi(2)
78.10(5) Bi(4)-Pd(3)-Bi(3)
68.90(5) Bi(2)-Pd(4)-Bi(3)
69.66(5) Bi(3)-Pd(4)-Bi(4)
more extensively examined in the theoretical discussion, gives
the [E4(PdL2)4]2+ clusters idealized D2d symmetry.
The Pd-Sb distances in 1‚0.5THF may also be compared to
those observed in [Sb6Pd9(PPh3)8], which is essentially a Pd8
cube with a Pd atom residing in the center of the cube and µ5-
Sb atoms capping each face of the cube. Here there is also one
long Pd-Sb distance, that of the center Pd to the bridging Sb
atoms (2.868(1) Å), and two short Pd-Sb distances (2.621(1)
and 2.601(1) Å). The alloy PdSb, which is formed by pyrolysis
of compound 1, shows a Pd-Sb distance (2.737 Å) intermediate
to the bonding and nonbonding distances of compound 1.70
Compound 2 may be compared to the [Bi4Fe4(CO)13]2-
dianion.71 In this molecule the four bismuth atoms define a
tetrahedron in which three triangular faces are capped by µ3-
Fe(CO)3 fragments and a fourth face is bare. There are two
unique Bi-Bi distances (ca. 3.16 and 3.46 Å) that compare
Biapical ) 2.887(2)-2.957(2) Å vs Pd-Bibasal ) 2.715(2)-2.765-
(2) Å). If one neglects the apical E atoms, the Pd atoms can be
considered to have the common cis-L2PdX2 geometry, but the
calculations indicate that the bonding to the apical E atom is
significant (vide infra). Only four E-E distances of the six
present in the tetrahedral array may be considered bonding, even
though they are somewhat long. For 1‚0.5THF these four
bonding distances are 3.0442(12)-3.0920(12) Å versus two
nonbonding distances of 3.36-3.41 Å. They may be compared
to the Sb-Sb interactions in the pure crystalline element (2.908
and 3.355 Å).68 Similarly, the Bi-Bi bonding distances observed
for 2 (3.201(2)-3.239(2) Å) and nonbonding interactions (3.45-
3.51 Å) may be compared to the Bi-Bi contacts observed in
the pure crystalline element (3.07 and 3.53 Å).69 The presence
of only four E-E bonding interactions, whose nature will be
2-
favorably to those observed in 2. The Bi4 anion has been
crystallographically characterized and can be viewed as a square-
planar array of Bi atoms with two unique Bi-Bi distances of
(70) Pratt, J. N.; Myles, K. M.; Darby, J. B., Jr.; Mueller, M. H. J. Less
Common Met. 1968, 14, 427.
(71) Whitmire, K. H.; Albright, T. A.; Kang, S.-K.; Churchill, M. R.;
Fettinger, J. C. Inorg. Chem. 1986, 25, 2799.
(68) Barrett, C. S.; Cucka, P.; Haefner, K. Acta Crystallogr. 1963, 16,
451.
(69) Cucka, P.; Barrett, C. S. Acta Crystallogr. 1962, 15, 865.

4414 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Stark et al.
2.936(2) and 2.941(2) Å.72 Those distances are considerably
shorter and are thought to arise due to multiple-bonding
interactions in the square-planar molecule which is an aromatic,
6π-electron system. The tetragonal phase of the â-Bi2Pd alloy
has a known crystal stucture, and was the phase isolated from
the thermolysis of cluster 2. This Bi2Pd alloy has sheets of cubic
arrays of Bi atoms with the cubes centered by Pd atoms.73 The
Bi-Pd distances of the alloy (2.969 Å) are slightly longer than
the long Bi-Pd distances of 2. The two Bi-Bi distances (3.362
and 3.557 Å) observed in the alloy are comparable to the
nonbonding Bi-Bi contacts in 2.
ometries are
4PdL4 + 6Ph2BiBr f [Bi4Pd4L8][Ph2BiBr2]2 + 8L +
3Ph-Ph + 2Ph-Br
5PdL4 + 6Ph2BiBr f [Bi4Pd4L8][Ph2BiBr2]2 + 10L +
4Ph-Ph + L2PdBr2
4PdL4 + 7Ph2BiBr f [Bi4Pd4L8][Ph2BiBr2]2 + 8L +
5Ph-Ph + BiBr3
Discussion
Mass spectral results of the byproducts indicate the formation
of Ph-Ph for E ) Sb and Bi, and for E ) Bi, a very small
amount of PhX was also observed. This suggests that the first
possibility does not represent the reaction stoichiometry, but
given the low yield of the reaction and the instability of the
product in solution, it is not possible to conclude more than
this. When greater than a 1:1 ratio of Ph2EX to PdL4 was used
in the starting reaction all that was isolated was L2PdX2.76 Free
phosphine was always found in the reaction solutions. Since
an excess of Ph2EX led to the formation of L2PdX2, it is not
clear if that compound appears as a byproduct of the main
reaction or results from subsequent reaction of the product
cluster with Ph2EX.
Electron Counting Considerations. The stable electron
count for an E4M4 cubane is 80 electrons, which is the count
predicted for conventional localized 2-center, 2-electron bond-
ing. For the cases where bond formation occurs between
transition metals, most of the compounds continue to obey the
EAN formalism, i.e., M generally follows the 18-electron rule77
and E obeys the octet rule consistent with Teo’s analysis.27 Each
successive loss of two electrons is generally accompanied by
the formation of one M-M bond, intermediate diamagnetic
situations being sometimes observed.30 Thus the TTd M4(µ3-
E)4 structure has 18 localized 2-electron, 2-center bonds (6
M-M + 12 M-E) and a resulting favored closed-shell cluster
electron (CE) count of (18 × 4) + (8 × 4) - (2 × 18) ) 68.
Conversely, starting with the 68-CE TTd structure and breaking
one M-M bond increases by two electrons the CE count until
one arrives again at the 80-CE cubane structure.
A number of synthetic methodologies reviewed elsewhere
have been developed for preparing E-M cluster compounds.46
Our interest in compounds with Pd arose from the possibility
of using the well-established ability of the noble metals to
undergo oxidative-addition reactions into E-X bonds. We
anticipated that the first step of a reaction between organo main
group halides where the main group element was a heavy atom
such as Bi or Sb and a Pd0 complex would be insertion of the
Pd into the E-X bond as illustrated in eq 2.
Ph2EX + PdL4 f L2Pd(EPh2)(X) + 2L
Ph2EX + X- f [Ph2EX2]-
(2)
(3)
We have observed empirically that organic groups attached to
metalated heavy main group elements are extremely easily lost
and intended to use such compounds as intermediates in the
formation of cluster compounds. In this case, however, this
hypothesis evidently works so well that an intermediate PdEPh2
complex is not observed and a cluster compound is the product
isolated directly from the reaction. These compounds have the
formulation [E4(PdL2)4][Ph2EX2]2 (E ) Sb, X ) Cl; E ) Bi, X
) Br) and are obtained directly from the reaction of PdL4 and
the appropriate Ph2EX. One of the most interesting features of
the product is the cationic nature of the cluster core. Most cluster
compounds are either negatively charged or neutral with
positively charged species being extremely rare. In many cases,
those that are known possess hydride ligands and could be
consider as protonated compounds, which is not the case here.
The reaction mechanism is obviously complex, but it is clear
from the product formulation that Ph2EX has been used to
scavenge halide ions, a well-precedented process for organo-
bismuth and antimony halide compounds (eq 3) in their direct
reaction with free X-.75,75 The dibromodiphenylbismuthide and
dichlorodiphenylantimonide ions are readily formed by adding
the appropriate tetraphenylphosphonium halide to either Ph2-
BiBr(thf) or in the case of antimony to either PhSbX2 or Ph2-
SbX. It is possible that Ph2EX may abstract a halide ion directly
from an intermediate L2Pd(EPh2)(X) to give L2Pd(EPh2)+.
There is a significant change in oxidation state between the
starting materials and the final cluster product. Considering Pd
to begin in the 0 oxidation state and E to be trivalent, the final
cluster stoichiometry requires the addition of 10 electrons.
Elimination of Ph (either as Ph-H or Ph-Ph) could provide at
least some of those electrons. Three possible overall stoichi-
The cluster core [E4(PdL2)4]2+ possesses a total of 74
electrons for which one would expect to obtain either one of
the C3V or C2 structures (5, 6, and 7, respectively, in Scheme
1). The fact that it possesses a D2d structure instead (3 in Scheme
1) automatically indicates that a more delocalized bonding
picture is required. Noting that 7 is closely related to the
observed D2d symmetry, one could analyze the compound as
being a resonance hybrid of four structures of type 7 in which
the missing E-E bond moves around the cluster framework.
This would predict E-E bond orders of 3/4, consistent with the
long distances observed here compared to Sb-Sb and Bi-Bi
single bonds.
In the TTd M4(µ3-E)4 structure, localized 2-electron, 2-center
bonding is possible because each transition metal atom has a
set of nine valence orbitals, which is large enough to build six
bonds within the cluster in addition to the metal-ligand bonds.
This is not the case for the hypothetical TTd E4(µ3-M)4 structure
in which each main group atom has only four valence orbitals
to ensure six cluster bonds and still retain a nonbonding lone
pair. This is the typical situation for hypercoordination where
(72) Cisar, A.; Corbett, J. D. Inorg. Chem. 1977, 16, 2482.
(73) Zhuravlev, N. N. SoViet Phys. JETP 1957, 5, 1064.
(74) (a) Clegg, W.; Errington, J.; Fisher, G. A.; Hockless, D. C. R.;
Norman, N. C.; Orpen, A. G.; Stratford, S. E. J. Chem. Soc., Dalton Trans.
1992, 1967. (b) Clegg, W. R.; Errington, R. J.; Fisher, G. A.; Flynn, R. J.;
Norman, N. C. J. Chem. Soc., Dalton Trans. 1993, 637.
(76) Stark, J. L.; Whitmire, K. H. Acta Crystallogr. 1997, C53 (6),
IUC9700007/1-2.
(77) Clusters with electron-deficient early-transition-metal centers have
been reported. Their bonding is analyzed in ref 30.
(75) Sheldrick, W. S.; Martin, C. Z. Naturforsch 1992, 47b, 919.

Distortion of a Tetracapped Tetrahedron
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4415
Figure 3. Qualitative MO interaction diagrams of a 74-CE E4(µ3-MLn)4 cluster (left) and a 66-CE E′4(µ3-MLn)4 cluster (right) (E′ ) early main-
group element).
each atomic orbital (or combination thereof) is participating in
more than one bond or lone pair.78 In such a situation the number
of bonding MOs is lower than the number of bonding contacts.
Therefore the CE count of the TTd M4(µ3-E)4 structure is
expected to be different from the one of its E4(µ3-M)4 homo-
logue. Owing to the tendency of group 15 elements to show
hypervalency, the CE count of the E4(µ3-M)4 (E ) group 15
element) is expected to be larger than 68.
interaction diagram expected for a hypothetical TTd E4(µ3-MLn)4
cluster is sketched in the left half of Figure 3. Only two E4
levels are left nonbonding with respect to interaction with the
metal fragments. One is one of the two E4 a1 orbitals (or a
combination of these), the other one being the antibonding t2
level, which one can anticipate to be too high in energy to
interact significantly with the metal frontier orbitals. Closed-
shell stability requires the filling of all the bonding and
nonbonding levels, leaving the antibonding ones vacant. This
corresponds to the occupation of the 13 lowest skeletal orbitals
of the E4(µ3-MLn)4 cluster. To these 26 electrons one has to
add 12 electrons per metal center (vide infra), which leads to
the favored 74-CE count for the TTd E4(µ3-MLn)4 cluster. This
counts exceeds by six the corresponding value of the localized
M4(µ3-E)4 TTd species.
For this electron count, the interaction of the E-E bonding
orbitals of the E4 fragment with vacant counterparts on the
(MLn)4 unit is expected to depopulate the former in the final
cluster. On the other hand the vacant t1 antibonding E-E orbitals
of the E4 fragment should be significantly populated after
interaction through their participation in occupied skeletal cluster
orbitals. Both effects are expected to weaken the E-E bonding
as compared to E-E single bonds in an isolated 20-electron E4
tetrahedron. This expected weakening is the result of hyperco-
ordination.
Electronic Structure of the Hypothetical 74-CE TTd [Sb4-
{µ3-Fe(CO)3}4]2+ Compound. To check the above qualitative
analysis in a more quantitative way, we have performed
extended Hu¨ckel theory (EHT) and density functional theory
(DFT) calculations on the hypothetical 74-CE [Sb4{µ3-Fe-
(CO)3}4]2+ cluster, assuming first Td molecular symmetry.
Details of the calculations are given in the Experimental Section.
The molecular geometry was fully optimized at the DFT level.
It is shown in Figure 5a. The major optimized metrical data
are reported in Table 4, together with the computed HOMO-
LUMO gap. The EHT and DFT MO diagrams of [Sb4{µ3-Fe-
(CO)3}4]2+ are given on the left and right sides of Figure 4,
respectively. The occupied levels that are shown are the Fe
nonbonding and the highest Fe-Sb bonding combinations. The
t2 LUMO is the lowest skeletal Fe-Sb antibonding level. The
other vacant levels shown have a dominant π*(CO) character.
It is noteworthy that both types of calculation are in close
agreement, leading to similar electronic configurations, level
Qualitative Approach in Determining the Electronic
Structure of a TTd E4(µ3-M)4 Cluster. The electronic structure
of an hypothetical E4(µ3-MLn)4 cluster of Td symmetry can be
predicted as resulting from the interaction between two tetra-
hedra: an E4 tetrahedron, in which the atoms are bonded, and
an (MLn)4 tetrahedron made of four nonbonded metal atoms in
their local environment of terminal ligands. In the first fragment,
the four s and p AOs on each of the four E atoms combine to
give rise to a set of six bonding combinations (a1 + e + t2), a
set of four nonbonding combinations (a1 + t2), and a set of six
antibonding levels (t1 + t2).79,80 This well-known level ordering
is drawn on the far left side of Figure 3. To cap symmetrically
a triangular face of the E4 tetrahedron, a transition metal atom
needs three frontier orbitals, one of local σ-type and two of
π-type symmetry.80 This means that the six remaining metal
AOs are involved in metal-terminal ligand bonding or hold
the metal nonbonding lone pairs. Six electron pairs must then
be associated with each MLn group in addition to those lying
in its three frontier orbitals. Because there is no interaction
between the M atoms, these frontier orbitals give rise to 12
nonbonding combinations in the tetrahedral (MLn)4 fragment
(a1 + e + t1 + 2t2).79 They correspond to the levels shown in
the middle of Figure 3. To ensure strong M-E bonding, these
metal levels must interact significantly with their E4 counterparts.
In a localized bonding scheme, they would interact only with
the nonbonding levels of the E4 fragment. Since in the present
case there are only four (a1 + t2) E4 nonbonding orbitals, eight
other E4 levels of e, t1, and t2 symmetries are also needed for
the M-L interactions. They can be found in the bonding set (e
+ t2) and in the antibonding t1 set. The resulting skeletal MO
(78) Albright, T. A.; Burdett, J. K.; Whangbo, M.-H. Orbital Interactions
in Chemistry; John Wiley and Sons: New York, 1985.
(79) Cotton, F. A. Chemical Applications of Group Theory, 3rd ed.; John
Wiley and Sons: New York, 1990.
(80) Mingos, D. M. P.; Wales, D. J. Introduction to Cluster Chemistry;
Prenctice Hall International Editions: Engelwood Cliffs, NJ, 1990.

4416 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Stark et al.
Figure 4. EHT and DFT MO level ordering of [Sb4{µ3-Fe(CO)3}4]2+
.
Figure 6. EHT and DFT MO level ordering of [Sb4{µ3-Pd(PPH3)2}4]2+
.
strongly that the hypothetical tetrahedral [Sb4{µ3-Fe(CO)3}4]2+
cluster may be synthesizable.
Electronic Structure of the 74-CE [Sb4(µ3-Pd(PH3)2)4]2+
Model. EHT and DFT calculations have been carried out on
[Sb4(µ3-Pd(PH3)2)4]2+, as a model for compound 1. Ideal D2d
symmetry was assumed. Details of the calculations are given
in the Experimental Section. Important structural data obtained
through the DFT geometry optimization are given in Table 4.
The optimized structure is shown in Figure 5b. The optimized
Pd-Sb distances are in a fairly good agreement with the
experimental distances (Tables 2 and 4), the optimized Sb-Sb
separations (3.15 and 3.54 Å) being slightly longer than the
experimental ones. The EHT and DFT MO diagrams of [Sb4-
(µ3-Pd(PH3)2)4]2+ are given in Figure 6 in a similar frontier
region as for [Sb4{µ3-Fe(CO)3}4]2+ (see above). As for this latter
model, both types of calculations give consistent results. In
particular, both methods give large HOMO-LUMO gaps, in
agreement with the stability of compounds 1 and 2. The relative
strength of the two types of Sb-Sb contacts can be evaluated
by the corresponding computed EHT overlap populations which
are 0.411 and 0.080 for the short (3.05 Å) and long (3.39 Å)
separations, respectively. Clearly, the bonding along the two
long edges of the Sb4 tetrahedron is weak. The average value
is close to the one computed for [Sb4{µ3-Fe(CO)3}4]2+. The
EHT Pd-Sb overlap populations are 0.283 and 0.156 for the
short (2.64 Å) and long (2.76 Å) separations, respectively. These
values indicate that significant bonding is still present along
the long Pd-Sb vectors.
Figure 5. DFT-optimized molecular structures of (a) [Sb4Fe4(CO)12]2+
and (b) [Sb4Pd4(PH3)8]2+
.
ordering, and energy gaps. They are also in full agreement with
the qualitative MO diagram of the E4(µ3-MLn)4 cluster (left side
of Figure 3). Both EHT- and DFT-computed HOMO-LUMO
gaps are large, as expected from the qualitative predictions of
the thermodynamical stability of 74-CE E4(µ3-MLn)4 TTd clusters
and from earlier EHT calculations on the related [Bi4{µ3-Fe-
(CO)3}4]2+ model.45 Nevertheless, the Jahn-Teller stability of
the Td architecture was tested by performing DFT geometry
optimization assuming D2d symmetry. The calculations provided
the same Td molecular geometry.
In agreement with its hypervalent coordination mode, the
DFT-optimized Sb-Sb bond lengths (3.08 Å) are longer than
what one would expect for a standard single bond. A significant
bonding interaction is still present, however, as exemplified by
the corresponding EHT overlap population (+0.253). The DFT-
optimized Sb-Fe distances (2.61 Å) are in the range of those
for normal single bonds. Clearly the calculations suggest
One may wonder why the 74-CE clusters 1 and 2 adopt a
D2d geometry (3 in Scheme 1), while theory predicts that an
isoelectronic species such as [Sb4{µ3-Fe(CO)3}4]2+ should have

Distortion of a Tetracapped Tetrahedron
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4417
Table 4. Selected Structural Data and HOMO-LUMO Gaps Corresponding to the DFT-Optimized Geometries of Various Closed-Shell
E4(MLn)4 Model Clusters
[Sb4Pd4(PH3)8]2+ [Sb4Fe4(CO)12]2+ Sb4Co4(CO)12 Bi4Co4(CO)12 [Ga4Fe4(CO)12]2+ [In4Fe4(CO)12]2+ [Tl4Fe4(CO)12]2+
74 CE, D2d
74 CE, Td
80 CE, Td
80 CE, Td
66 CE, Td
66 CE, Td
66 CE, Td
M-E (Å)
2.69-2.82
3.15-3.54
84.6-110.3
69.7-82.3
2.37
1.43
93.7
2.61
3.08
105.1
72.4
1.78
1.14
95.3
1.42
2.64
3.23
102.9
75.3
1.78
1.15
102.9
2.10
2.74
3.42
101.3
77.4
1.78
1.15
103.1
2.23
2.43
3.33
93.4
86.5
1.81
1.14
99.6
0.88
2.59
3.56
93.1
86.8
1.79
1.15
102.2
0.89
2.69
3.70
93.2
86.7
1.80
1.15
103.3
0.84
E-E (Å)
M-E-M (deg)
E-M-E (deg)
M-L (Å)
P-H/C-O (Å)
L-M-L (deg)
HOMO/LUMO
gap (eV)
1.34
the TTd structure (1 in Scheme 1). Examination of the skeletal
orbitals of [Sb4{µ3-Pd(PH3)2}4]2+ indicates that it has the same
electron configuration than [Sb4{µ3-Fe(CO)3}4]2+, i.e. when
going from the latter TTd cluster to the former D2d compound
there is no HOMO-LUMO level crossing. The Td to D2d
lowering in symmetry perturbs somewhat the level ordering
within the occupied MOs as well as within the vacant MOs,
but maintains a large gap between these two blocks. The reason
for the significant D2d distortion of the E4 tetrahedron present
in the core of 1 and 2 originates from the nonconical nature of
the PdL2 fragments.77 As said above, the capping of a triangular
face of an E4 tetrahedron by a metal fragment requires, inter
alia, the use of two metal π-type frontier orbitals. A symmetrical
capping requires degeneracy of these two orbitals. In the case
of the PdL2 fragment, one is far from this situation. One of its
π frontier orbitals is a low-lying d-type hybrid, which lies in
the PdL2 plane and whose main lobes are pointing in the
direction of the long Sb‚‚‚Sb edges. The other π frontier orbital
is the high-lying Pd 5pz orbital which is perpendicular to the
PdL2 plane. Because of a better energy match and of better
directional properties, the d-type orbital interacts more strongly
with the triangular face than the p-type orbital. As a result, an
unsymmetrical capping and a distortion of the E4 tetrahedron
occur. The two Sb-Sb edges which are in a privileged position
to interact with the four d-type metal orbitals are much more
perturbed. As a result, the Sb-Sb σ-donation and σ*-retrodo-
nation (see qualitative approach) affect preferencially these two
Sb-Sb edges. The effect is strong enough to almost cancel the
corresponding Sb-Sb bonding. On the other hand, the four other
edges which are involved with the weaker interaction with the
metal p-type orbital, are less perturbed. Thus, clusters 1 and 2
are better described as strongly distorted relatives of the TTd
structure 1 of Scheme 1, rather than typical representatives of
a new structural type of D2d symmetry (3 in Scheme 1).
Another Possible Closed-Shell CE Count for the TTd E4-
(µ3-M)4 Geometry. In a search for alternative stable electron
counts for a E4(µ3-M)4 cluster having structure 1 of Scheme 1,
EHT calculations on a series of E4(µ3-M(CO)3)4 models, with
M varying from V to Ni and E varying from group 13 to group
15 have been carried out. These exploratory calculations
suggested that another CE count, namely 66, should be favored
for this TTd arrangement. This situation is expected to occur
with early main-group elements. The qualitative MO interaction
diagram of such a E′4(µ3-MLn)4 species is sketched in the right
half of Figure 3. In the case of an early main-group E′4
tetrahedron, the mixing between the s-type combinations and
the high-lying p-type combinations is weak. As a result, the 10
lowest orbitals of E′4 are split into two groups, well-separated
in energy. The lowest set is significantly bonding and can be
described as made of the a1 + t2 combinations of the s atomic
orbitals. Although weakly bonding or nonbonding, the other set
(a1 + e + t2) lies at a rather high energy since it derives
primarily from the high-lying p atomic orbitals. This level
ordering is depicted on the right side of Figure 3. The EHT
calculations indicate that the a1 + e + t1 + 2t2 set of frontier
orbitals of the tetrahedral (MLn)4 fragment matches with the
lowest a1 + t2 E′4 orbitals and with the unique e and t1 sets of
the E′4 fragment. The E′4 high-lying t2 level is only weakly
involved in the interaction. As a consequence, one of the t2 sets
of the (MLn)4 fragment remains nonbonding. The resulting MO
level ordering exhibits a group of nine M-E′ bonding orbitals
(a1 + e + t1 + t2) well-separated from the nonbonding t2 level.
A large HOMO-LUMO gap is found when the bonding block
is filled and the nonbonding t2 set remains vacant. This 66-CE
situation corresponds to an electron-deficient system since there
are only nine bonding electron pairs associated with 12 M-E′
bonding contacts. DFT calculations on the hypothetical TTd [E′4-
(µ3-Fe(CO)3)4]2+ (E′ ) Ga, In, Tl) series confirm the EHT
expectations. The computed HOMO-LUMO gaps and major
optimized metrical data are reported in Table 4. In the case of
E′ ) In and Tl, the optimized E′-E′ distances are slightly longer
than in the metal (3.56 and 3.70 Å vs 3.25 and 3.40 Å,
respectively). In the case of E′ ) Ga, the corresponding value
is significantly larger that in the metal (3.33 vs 2.45 Å). In the
three compounds, the M-E′ distances are in the range of
expected values for regular bonds. Significant HOMO-LUMO
gaps are computed, indicating closed-shell stability. These results
strongly suggest that 66-CE TTd E4(µ3-M)4 clusters should be
accessible.
Adding Electrons to the 74-CE TTd E4(µ3-M)4 Arrange-
ment. As said above, in the case of the various structures of
Scheme 1 deriving from the M4(µ3-E)4 TTd geometry, one goes
from the favored closed-shell count of 68 CE (1) to 80 CE (11)
by adding two electrons to the CE count as one sucessively
breaks the M-M bonds. In the case of the possible structures
deriving from the E4(µ3-M)4 TTd geometry, the largest electron
count still corresponds to the 80-CE cubane species. One may
ask if, between the CE counts of 74 (1) and 80 (11), other stable
intermediate arrangements (such as 2-10 in Scheme 1) could
exist for particular closed-shell CE counts. Exploratory EHT
calculations on various E4M4(CO)12 models were not able to
reproduce particular compounds with structures 2-10 for which
the existence of a large HOMO-LUMO gap would clearly
indicate fulfillment of the closed-shell requirement. Some DFT
calculations performed on 76- and 78-CE systems confirmed
the EHT results. Stable closed-shell situations were only found
for the known 80-CE cubane-type structure 11. The major
structural data of the DFT optimized structures and the corre-
sponding HOMO-LUMO gaps of the 80-CE cubane-type
compounds [Co4(CO)12(µ3-E)4] (E ) Bi, Sb) are given in Table
4. The optimized geometries are in good agreement with the
experimental ones.16,17 These results support the suitability of
using DFT calculations for predicting molecular structures in
this class of compounds and suggest that stable closed-shell

4418 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Stark et al.
E4M4 clusters presenting E-E bonding with structures 2-10
are unlikely to exist for CE counts intermediate between 74
and 80. This does not rule out, however, the possibility of
existence of paramagnetic species.
74 and 80 which fulfills the closed-shell requirement for
structures 2-10, assuming conical MLn units.
Acknowledgment. The Centre National de la Recherche
Scientifique (J.-Y.S. and J.-F.H.) and the National Science
Foundation (K.H.W.) are gratefully acknowledged for financial
support of this work as well as for the cooperative international
travel grant that has greatly facilitated these studies. The Robert
A. Welch Foundation (K.H.W.) also provided financial support
for carrying out these studies. J.-F.H. and J.-Y.S. thank Pr.
Baerends and Dr. te Velde (Vrije Universiteit, Amsterdam) for
introducing them to the ADF program. Exchanges with Baer-
ends’ group have been made possible through a European
Human Capital and Mobility Network Grant. J.-F.H. and J.-
Y.S. thank the Centre de Ressources Informatiques of Rennes
and the Institut de Developpement et de Ressources en Infor-
matique Scientifique of Orsay for computing facilities.
Conclusions. The 74-CE clusters [E4(PdL2)4]2+ (E ) Sb, Bi)
have a core arrangement corresponding to structure 3 of Scheme
1 and deriving from the E4(µ3-M)4 TTd structure 1 by the almost
complete breaking of two E-E bonds. Calculations on the model
[Sb4(Pd(PH3)2)4]2+ indicate that these compounds are electroni-
cally strongly related to the isoelectronic hypothetical [Sb4(µ3-
Fe(CO)3)4]2+ TTd species. The lower D2d symmetry of the
[E4(PdL2)4]2+ (E ) Sb, Bi) systems is the consequence of the
nonconical nature of the PdL2 constituents of the clusters. In
particular, the existence in these clusters of two long E‚‚‚E
separations is principally due to the strong d-π donation of the
PdL2 units into the σ*(E‚‚‚E) orbitals. 74-CE TTd E4(µ3-MLn)4
clusters are predicted to be stable diamagnetic compounds in
which the presence of both E-E and M-E bonding is the
consequence of the tendency to hypervalency of heavy main-
group elements. The E4(µ3-M)4 TTd arrangement (structure 1
in Scheme 1) should also be stable for the closed-shell 66-CE
count. Such an electron-poor system is favored with early main-
group elements. The formal adding of electrons to a 74-CE E4-
(µ3-MLn)4 TTd species is expected to induce E-E bond breaking
and successively generates structures 2-11 (Scheme 1). Cal-
culations indicate that there is no CE count intermediate between
Supporting Information Available: Tables of crystal data
and structural refinements, anisotropic and isotropic thermal
parameters, full bond distances and angles, and atomic coor-
dinates, labeling scheme diagrams, and full thermal ellipsoid
plots for the cations and anions of [Sb4(PdL2)4][Ph2SbCl2]2‚
0.5THF and [Bi4(PdL2)4][Ph2BiBr2]2 (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA982902U