Stable paramagnetic half-sandwich Mo(V) and W(V) polyhydride complexes. Structural, spectroscopic, electrochemical, theoretical, and decomposition mechanism studies of [CP*MH3(dppe)]+ (M = Mo, W)

J. Am. Chem. Soc. 1999, 121, 2209-2225

2209

Stable Paramagnetic Half-Sandwich Mo(V) and W(V) Polyhydride

Complexes. Structural, Spectroscopic, Electrochemical, Theoretical,

and Decomposition Mechanism Studies of [Cp*MH₃(dppe)]⁺(M )

Mo, W)

Brett Pleune,^1aDolores Morales,^1bRita Meunier-Prest,^1bPhilippe Richard,^1b

Edmond Collange,^1bJames C. Fettinger,^1aand Rinaldo Poli*^,1b

Contribution from the Laboratoire de Synthe`se et d’Electrosynthe`se Organome´talliques, Faculte´ des

Sciences “Gabriel”, UniVersite´ de Bourgogne, 6 BouleVard Gabriel, 21000 Dijon, France, and

Department of Chemistry and Biochemistry, UniVersity of Maryland, College Park, Maryland 20742

ReceiVed October 2, 1998

Abstract: Compounds Cp*MH₃(dppe) (M ) Mo, 1; W, 2) are oxidized chemically and electrochemically to

the corresponding 17-electron cations 1⁺and 2⁺. Analogous oxidations of 1-d₃and 2-d₃provide 1⁺-d₃and

2⁺-d₃, respectively. Complex 2⁺is stable in CH₂Cl₂, THF, and MeCN at room temperature. A single-crystal

X-ray analysis of the PF₆^-salt of 2⁺shows a geometry for the cation which is intermediate between octahedral

and trigonal prismatic, which is reproduced by geometry optimization of the [CpWH₃(PH₂CH₂CH₂PH₂)]⁺

model at the B3LYP/LANL2DZ level. Identical calculations on the neutral analogue also reproduce the

previously reported trigonal prismatic structure for 1. A blue shift in the M-H stretching vibrations upon

oxidation for both Mo and W compounds indicates that a M-H bond strengthening accompanies the oxidation

process. The DFT calculations (M-H bond lengths, BDE, and stretching frequencies) are in good agreement

with the experimental results. Complex 1⁺decomposes in solution at room temperature by one or more of

three different mechanisms depending on conditions: H₂reductive elimination, solvent-assisted dispropor-

tionation, or deprotonation. In THF or CH₂Cl₂, a reductive elimination of H₂affords the stable paramagnetic

monohydride Cp*MoH(dppe)PF₆(3), which adds a molecule of solvent in CH₂Cl₂, THF, and MeCN. EPR

studies show that the CH₂Cl₂molecule coordinates in a bidentate mode to afford a 19-electron configuration.

A solvent dependence of the decomposition rate [k(CH₂Cl₂) ≈ 7.8k(THF) at 0 °C] and an inverse isotope

effect [k_H/k_D) 0.50(3) in CH₂Cl₂at 0 °C] indicate the nature of 1⁺as a classical trihydride and suggest a

decomposition mechanism which involves equilibrium conversion to a nonclassical intermediate followed by

a rate-determining associative exchange of H₂with a solvent molecule. In MeCN at 20 °C, a solvent-assisted

disproportionation (rate ) k_disp[1⁺]², k_disp) 3.98(9) × 10³s^-1M^-1) and a deprotonation by residual unoxidized

1 (rate ) k_deprot[1⁺][1], k_deprot) 2.8(2) × 10²s^-1M^-1) take place competitively, as shown by detailed cyclic

voltammetric and thin-layer cyclic voltammetric studies. The stoichiometric chemical oxidation of 1 in MeCN

leads to a mixture of [Cp*MoH₂(dppe)(MeCN)]⁺and [Cp*MoH(dppe)(MeCN)₂]²⁺by the disproportionation

mechanism.

Introduction

ing.^14-19The simplicity of the ligand under scrutiny (the

hydrogen atom) has made the bond between transition metals

and H a favorite subject for a variety of investigations including

Sparked by the potentials of industrial catalysis, reversible

hydrogen storage, and the development of biomimetic hydrogen

activating systems, the interest in transition metal hydride

complexes has led to tremendous advances in the past 15-20

years, ranging from the existence and properties of coordinated

H₂ligands (so-called nonclassical hydride complexes)^2-6to the

existence of stable paramagnetic transition metal mono- and

polyhydride complexes,^7-11and from the mechanism of flux-

ional rearrangements^12-14to M-H^δ-‚‚‚H^δ+hydrogen bond-

(7) Luetkens, M. L., Jr.; Elcesser, W. L.; Huffman, J. C.; Sattelberger,

A. P. Inorg. Chem. 1984, 23, 1718-1726.

(8) Sharp, P. R.; Frank, K. G. Inorg. Chem. 1985, 24, 1808-1813.

(9) Hamon, P.; Toupet, L.; Hamon, J.-R.; Lapinte, C. Organometallics

1992, 11, 1429-1431.

(10) Rømming, C.; Smith, K.-T.; Tilset, M. Inorg. Chim. Acta 1997,

259, 281-290.

(11) Menglet, D.; Bond, A. M.; Coutinho, K.; Dickson, R. S.; Lazarev,

G. G.; Olsen, S. A.; Pilbrow, J. R. J. Am. Chem. Soc. 1998, 120, 2086-

2089.

* To whom correspondence should be addressed. Tel: +33-03.80.39.68.81.

Fax: +33-03.80.39.60.98. E-mail: Rinaldo.Poli@u-bourgogne.fr.

(1) (a) University of Maryland. (b) Universite´ de Bourgogne.

(2) Kubas, G. J.; Ryan, R. R.; Swanson, B. I.; Vergamini, P. J.;

Wassermann, H. J. J. Am. Chem. Soc. 1984, 106, 451-452.

(3) Kubas, G. J. Acc. Chem. Res. 1988, 21, 120-128.

(4) Crabtree, R. H. Acc. Chem. Res. 1990, 23, 95-101.

(5) Jessop, P. G.; Morris, R. H. Coord. Chem. ReV. 1992, 121, 155-

191.

(12) Lee, J. C., Jr.; Yao, W.; Crabtree, R. H.; Ru¨egger, H. Inorg. Chem.

1996, 35, 695-699.

(13) Gusev, D. G.; Berke, H. Chem. Ber. 1996, 129, 1143-1155 and

references therein.

(14) Bosque, R.; Maseras, F.; Eisenstein, O.; Patel, B. P.; Yao, W.;

Crabtree, R. H. Inorg. Chem. 1997, 36, 5505-5511.

(15) Lough, A. J.; Park, S.; Ramachandran, R.; Morris, R. H. J. Am.

Chem. Soc. 1994, 116, 8356-8357.

(16) Crabtree, R. H.; Siegbahn, P. E. M.; Eisenstein, O.; Rheingold, A.

L.; Koetzle, T. F. Acc. Chem. Res. 1996, 29, 348-354 and references therein.

(6) Heinekey, D. M.; Oldham, W. J. Chem. ReV. 1993, 93, 913-944.

Published on Web 02/26/1999

2210 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

structural (X-ray and neutron diffraction and T₁NMR),^4,20-32

theoretical,^33-39and reactivity studies including protonation^40-54

and oxidation.^10,55-76Yet, investigations in this field continue

to produce surprises. In this contribution, we present a com-

prehensive study of the oxidation of Cp*MH₃(dppe) (M ) Mo,

W) systems, which provides relevant new information on the

structure, stability, M-H bond strength, and decomposition

mechanism of paramagnetic polyhydride systems. Steric effects

are shown to protect a paramagnetic polyhydride system, in the

absence of electronically stabilizing π-donating ligands, against

the deprotonation and the disproportionation decomposition

pathways; the M-H interaction is unambiguously shown to

strengthen upon one-electron oxidation; and the H₂reductive

elimination process has been kinetically assessed for the first

time for a paramagnetic polyhydride system. A preliminary

report on some aspects of this work has recently appeared.⁷⁷

(17) Patel, B. P.; Kavallieratos, K.; Crabtree, R. H. J. Organomet. Chem.

1997, 528, 205-207.

(18) Crabtree, R. H. J. Organomet. Chem. 1998, 577, 111-115.

(19) Chu, H. S.; Lau, C. P.; Wong, K. Y.; Wong, W. T. Organometallics

1998, 17, 2768-2777.

(20) Hamilton, D. G.; Crabtree, R. H. J. Am. Chem. Soc. 1988, 110,

4126-4133.

(21) Bautista, M. T.; Earl, K. A.; Maltby, P. A.; Morris, R. H.;

Schweitzer, C. T.; Sella, A. J. Am. Chem. Soc. 1988, 110, 7031-7036.

(22) Bautista, M. T.; Cappellani, E. P.; Drouin, S. D.; Morris, R. H.;

Schweitzer, C. T.; Sella, A.; Zubkowski, J. J. Am. Chem. Soc. 1991, 113,

4876-4887.

Experimental Section

(23) Gusev, D. G.; Vymenits, A. B.; Bakhmutov, V. I. Inorg. Chim. Acta

1991, 179, 195-201.

(24) Desrosiers, P. J.; Cai, L.; Lin, Z.; Richards, R.; Halpern, J. J. Am.

Chem. Soc. 1991, 113, 4173-4184.

(25) Brammer, L.; Howard, J. A. K.; Johnson, O.; Koetzle, T. F.; Spencer,

J. L.; Stringer, A. M. J. Chem. Soc., Chem. Commun. 1991, 241-243.

(26) Brammer, L.; Zhao, D.; Bullock, R. M.; McMullan, R. K. Inorg.

Chem. 1993, 32, 4819-4824.

(27) Shin, J. H.; Parkin, G. Polyhedron 1994, 13, 1489-1493.

(28) Jime´nez-Tenorio, M.; Puerta, M. C.; Valerga, P. Organometallics

1994, 13, 3330-3337.

(29) Klooster, W. T.; Koetzle, T. F.; Jia, G.; Fong, T. P.; Morris, R. H.;

Albinati, A. J. Am. Chem. Soc. 1994, 116, 7677-7681.

(30) Lemke, F. R.; Brammer, L. Organometallics 1995, 14, 3980-3987.

(31) Luo, X.-L.; Crabtree, R. H. J. Am. Chem. Soc. 1990, 112, 4813-

4821.

General Procedures. Unless otherwise stated, all manipulations were

carried out under an inert atmosphere of dinitrogen or argon by the

use of Schlenk line or glovebox techniques. Methanol was degassed

by three freeze-pump-thaw cycles prior to use. Other solvents were

dried by conventional methods (Et₂O from Na/K/benzophenone, toluene

and heptane from Na, MeCN from CaH₂, and CH₂Cl₂from P₄O₁₀) and

distilled under dinitrogen prior to use. Deuterated solvents were dried

over molecular sieves and degassed by three freeze-pump-thaw cycles

prior to use. ¹H and ³¹P{¹H} NMR measurements were carried out on

Bruker AF200, WP200, or AM400 spectrometers; the peak positions

are reported with positive shifts downfield of TMS as calculated from

the residual solvent peaks (¹H) or downfield of external 85% H₃PO₄

(³¹P). For each ³¹P NMR spectrum, a sealed capillary containing 85%

H₃PO₄was immersed in the same NMR solvent as that used for the

measurement, and this was used as the reference. EPR measurements

were carried out at the X band microwave frequency on a Bruker ER

200 D spectrometer upgraded to ESP 300, equipped with a cylindrical

ER/4103 TM 110 cavity. The spectrometer frequency was calibrated

with DPPH (g ) 2.004). Cyclic voltammograms were obtained at 20

°C in a three-electrode cell with an EG&G 283 potentiostat connected

to a personal computer. The working electrode was a 3-mm-diameter

carbon disk or a 0.5-mm-diameter platinum disk. Bu₄NPF₆was used

as supporting electrolyte at a concentration of 0.1 M. All potentials

are reported vs the Cp₂Fe/Cp₂Fe⁺couple, which has an E_1/2of +0.50

V relative to SCE under conditions identical to those of the other

experiments. The cyclic voltammograms were fitted by simulations

performed with the DIGISIM 2.1 software (BAS Inc.).⁷⁸The solid-

state magnetic susceptibility measurements were carried out with a

Johnson Matthey magnetic susceptibility balance. The solution con-

ductivity measurements were carried out at 25 °C with a Tacussel type

(32) Gusev, D. G.; Nietlispach, D.; Eremenko, I. L.; Berke, H. Inorg.

Chem. 1993, 32, 3628-3636.

(33) Hay, J. P. J. Am. Chem. Soc. 1987, 109, 705-710.

(34) Haynes, G. R.; Martin, R. L.; Hay, P. J. J. Am. Chem. Soc. 1992,

114, 28-36.

(35) Craw, J. S.; Bacskay, G. B.; Hush, N. S. J. Am. Chem. Soc. 1994,

116, 5937-5948.

(36) Lin, Z. Y.; Hall, M. B. Coord. Chem. ReV. 1994, 135, 845-879

and references therein.

(37) Bayse, C. A.; Couty, M.; Hall, M. B. J. Am. Chem. Soc. 1996, 118,

8916-8919.

(38) Dapprich, S.; Frenking, G. Organometallics 1996, 15, 4547-4551.

(39) Gelabert, R.; Moreno, M.; Lluch, J. M.; Lledo´s, A. J. Am. Chem.

Soc. 1997, 119, 9840-9847.

(40) Allison, J. D.; Walton, R. A. J. Chem. Soc., Chem. Commun. 1983,

401-403.

(41) Parkin, G.; Bercaw, J. E. Polyhedron 1988, 7, 2053-2082.

(42) Jia, G.; Morris, R. H. J. Am. Chem. Soc. 1991, 113, 875-883.

(43) Jia, G.; Lough, A. J.; Morris, R. H. Organometallics 1992, 11, 161-

171.

(44) Michos, D.; Luo, X.-L.; Faller, J. W.; Crabtree, R. H. Inorg. Chem.

1993, 32, 1370-1375.

(61) Bruno, J. W.; Caulton, K. G. J. Organomet. Chem. 1986, 315, C13-

C16.

(62) Lemmen, T. H.; Lundquist, L. F.; Sutherland, B. R.; Westerberg,

D. E.; Caulton, K. G. Inorg. Chem. 1986, 25, 3915-3917.

(63) Detty, M. R.; Jones, W. D. J. Am. Chem. Soc. 1987, 109, 5666-

5673.

(45) Chin, B.; Lough, A. J.; Morris, R. H.; Schweitzer, C. T.; D’Agostino,

C. Inorg. Chem. 1994, 33, 6278-6288.

(46) Christ, M. L.; Sabo-Etienne, S.; Chaudret, B. Organometallics 1994,

13, 3800-3804.

(47) Kiss, G.; Nolan, S. P.; Hoff, C. D. Inorg. Chim. Acta 1994, 227,

285-292.

(64) Costello, M. T.; Walton, R. A. Inorg. Chem. 1988, 27, 2563-2564.

(65) Chen, L.; Davies, J. A. Inorg. Chim. Acta 1990, 175, 41-45.

(66) Roullier, L.; Lucas, D.; Mugnier, Y.; Antin˜olo, A.; Fajardo, M.;

Otero, A. J. Organomet. Chem. 1990, 396, C12-C16.

(67) Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990, 112, 2843.

(68) Roullier, L.; Lucas, D.; Mugnier, Y.; Antin˜olo, A.; Fajardo, M.;

Otero, A. J. Organomet. Chem. 1991, 412, 353-362.

(69) Ryan, O. B.; Tilset, M. J. Am. Chem. Soc. 1991, 113, 9554-9561.

(70) Ryan, O. B.; Tilset, M.; Parker, V. D. Organometallics 1991, 10,

298-304.

(48) Feracin, S.; Burgi, T.; Bakhmutov, V. I.; Eremenko, I.; Vorontsov,

E. V.; Vimenits, A. B.; Berke, H. Organometallics 1994, 13, 4194-4202.

(49) Angelici, R. J. Acc. Chem. Res. 1995, 28, 51-60.

(50) Rothfuss, H.; Gusev, D. G.; Caulton, K. G. Inorg. Chem. 1995, 34,

2894-2901.

(51) Shubina, E. S.; Krylov, A. N.; Belkova, N. V.; Epstein, L. M.;

Borisov, A. P.; Mahaev, V. D. J. Organomet. Chem. 1995, 493, 275-277.

(52) Bullock, R. M.; Song, J.-S.; Szalda, D. J. Organometallics 1996,

15, 2504-2516.

(53) Castillo, A.; Esteruelas, M. E.; On˜ate, E.; Ruiz, N. J. Am. Chem.

Soc. 1997, 119, 9691-9698.

(71) Westerberg, D. E.; Rhodes, L. F.; Edwin, J.; Geiger, W. E.; Caulton,

K. G. Inorg. Chem. 1991, 30, 1107-1112.

(54) Luther, T. A.; Heinekey, D. M. Inorg. Chem. 1998, 37, 127-132.

(55) Sanders, J. R. J. Chem. Soc., Dalton Trans. 1973, 748-749.

(56) Sanders, J. R. J. Chem. Soc., Dalton Trans. 1975, 2340-2342.

(57) Pilloni, G.; Schiavon, G.; Zotti, G.; Zecchin, S. J. Organomet. Chem.

1977, 134, 305-318.

(72) Tilset, M. J. Am. Chem. Soc. 1992, 114, 2740-2741.

(73) Amatore, C.; Frau´sto da Silva, J. J. R.; Guedes da Silva, M. F. C.;

Pombeiro, A. J. L.; Verpeaux, J.-N. J. Chem. Soc., Chem. Commun. 1992,

1289-1291.

(74) Smith, K.-T.; Tilset, M. J. Organomet. Chem. 1992, 431, 55-64.

(75) Smith, K.-T.; Rømming, C.; Tilset, M. J. Am. Chem. Soc. 1993,

115, 8681-8689.

(58) Klinger, R. J.; Huffman, J. C.; Kochi, J. K. J. Am. Chem. Soc. 1980,

102, 208-216.

(59) Allison, J. D.; Cameron, C. J.; Wild, R. E.; Walton, R. A. J.

Organomet. Chem. 1981, 218, C62-C66.

(60) Rhodes, L. F.; Zubkowski, J. D.; Folting, K.; Huffman, J. C.;

Caulton, K. G. Inorg. Chem. 1982, 21, 4185-4192.

(76) Zlota, A. A.; Tilset, M.; Caulton, K. G. Inorg. Chem. 1993, 32,

3816-3821.

(77) Pleune, B.; Poli, R.; Fettinger, J. C. J. Am. Chem. Soc. 1998, 120,

3257-3258.

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2211

CD6 N conductimeter equipped with an XE 110 cell which had been

calibrated with a 0.1 M KCl solution. The elemental analyses were

carried out by Desert Analytics (Tucson, AZ), by Atlantic Microlab

(Norcross, GA), or by the analytical service of the Laboratoire de

Synthe`se et d’Electrosynthe`se Organome´talliques (LSEO) of the

Universite´ de Bourgogne. LiAlD₄and CH₃OD (Aldrich) were used as

received, without further purification. [Cp₂Fe]PF₆,⁷⁹CDFCl₂,⁸⁰and

Cp*MH₃(dppe) (M ) Mo, 1; W, 2)⁸¹were prepared according to

literature procedures.

Synthesis of Cp*MoD₃(dppe) (1-d₃). This reaction follows the

protocol reported previously for the preparation of Cp*MoH₃(dppe).⁸¹

Cp*MoCl₄(555 mg, 1.488 mmol), dppe (592 mg, 1.488 mmol), and

LiAlD₄(0.6 g, 14.3 mmol) were slurried in a 70 mL:15 mL toluene/

diethyl ether solvent mixture at room temperature. The mixture was

stirred for 12 h, and CH₃OD (3 mL) as added dropwise to the stirring

mixture at room temperature, causing a vigorous gas evolution. The

solvent mixture was removed under reduced pressure, and the residue

was extracted with heptane (100 mL). The heptane solution was filtered

through Celite and concentrated to ca. 1 mL, precipitating a yellow

powder, which was washed with cold (-80 °C) heptane and dried under

vacuum. Yield: 480 mg (51%). The NMR properties of 1-d₃are

identical with those of 1,⁸¹except for the undetectable hydride

resonances in the ¹H NMR and the P-D coupling in the ³¹P{¹H} NMR

1

spectra. H NMR (C₆D₆, δ): 7.8-7.0 (m, Ph, 20H), 2.05-1.75 (m,

CH₂, 4H), 1.84 (s, Cp*, 15H). ³¹P{¹H} NMR (C₆D₆, δ): 92.1 (1:3:6:

7:6:3:1 septet, J_PD)12.3 Hz).

Synthesis of Cp*WD₃(dppe) (2-d₃). This reaction follows the

protocol reported previously for the preparation of Cp*MoH₃(dppe).⁸¹

Cp*WCl₄(282 mg, 0.612 mmol), dppe (243 mg, 0.612 mmol), and

LiAlD₄(0.3 g, 7.1 mmol) were slurried in a 100 mL:15 mL toluene/

diethyl ether solvent mixture at room temperature. The mixture was

stirred for 12 h, and CH₃OD (0.5 mL) was added to the stirring mixture,

causing gas evolution. The solvent mixture was then removed under

reduced pressure, and the residue was extracted into heptane (100 mL).

The heptane solution was filtered through Celite and concentrated to

ca. 0.5 mL, precipitating an orange powder which was washed with

cold (-80 °C) heptane and dried under vacuum. Yield: 91 mg (21%).

To the remaining residue was added 75 mL of heptane, and 3 mL of

CH₃OD was added dropwise at room temperature. Gas vigorously

evolved, and the solution developed a red-orange color. The reaction

mixture was stirred for an additional 30 min and then filtered through

Celite. The solution was evaporated to ca. 0.5 mL, precipitating a second

crop of orange powder (116 mg). Combined yield: 207 mg (47%).

The NMR properties of 2-d₃are identical to those of 2,⁸¹except for

the undetectable hydride resonances in the ¹H NMR and the P-D

Figure 1. Room-temperature EPR spectra of compounds [Cp*MX₃-

(dppe)]⁺(M ) Mo, X ) H, 1⁺; D, 1⁺-d₃. M ) W, X ) H, 2⁺; D,

2⁺-d₃) in THF. The asterisk in the spectrum of 2⁺-d₃indicates an

unknown impurity.

1

coupling in the ³¹P{¹H} NMR spectra. H NMR (C₆D₆, δ): 7.8-7.0

(m, Ph, 20H), 2.25-2.00 (m, CH₂, 4H), 1.96 (s, 15H, Cp*). ³¹P{¹H}

NMR (C₆D₆, δ): 67.9 (1:3:6:7:6:3:1 septet, J_PD) 8.2 Hz).

Chemical Oxidation of 1 in THF at -80 °C. Cp₂FePF₆(0.025 g,

0.075 mmol) was added to a stirring solution of 1 (0.047 g, 0.074 mmol)

in THF (1 mL) at -80 °C. An immediate reaction occurred which

quenched the blue color of the Cp₂FePF₆. This solution exhibited a

triplet of quartets in the EPR spectrum at -80 °C (g ) 1.989, a_P(triplet)

) 28.9 G, a_H(quartet) ) 11.8 G), attributable to [Cp*MoH₃(dppe)]-

[PF₆] (1⁺). This spectrum is shown in Figure 1. IR (THF, cm^-1): 1896

(w, sh), 1824 (m, broad). The spectrum was recorded immediately after

charging the cell with the cold solution. By comparison, a solution of

compound 1 showed M-H stretching vibrations at 1775 and 1815 cm^-1

(see Figure 2).

Chemical Oxidation of 1 in THF at Room Temperature. Cp₂-

FePF₆(0.029 g, 0.087 mmol) was added to a stirring solution of 1

(0.055 g, 0.087 mmol) in THF (1 mL) at room temperature. An

immediate reaction occurred which quenched the blue color of the Cp₂-

FePF₆. An aliquot of this solution was immediately transferred into an

Figure 2. Room-temperature IR spectra of (a) 1 (top) and 1⁺(bottom)

in THF and (b) 2 (top) and 2⁺(bottom) in THF. The normal modes

and relative intensities corresponding to the frequencies ν₁, ν₂, and ν₃

in part b are obtained from DFT calculations on geometry-optimized

[CpWH₃(H₂PCH₂CH₂PH₂)]ⁿ⁺(n ) 0, 1; see text).

(78) Rudolph, M.; Reddy, D. P.; Feldberg, S. W. Anal. Chem. 1994, 66,

589A-600A.

(79) Yang, E. S.; Chan, M.; Wahl, A. C. J. Phys. Chem. 1975, 79, 2049-

2051.

EPR tube and frozen at the liquid nitrogen temperature until prior to

the insertion into the EPR probe. The resulting spectrum showed 1⁺as

the only paramagnetic product. The solution gradually developed a red-

(80) Siegel, S. S.; Anet, F. A. J. Org. Chem. 1988, 53, 2629-2630.

(81) Pleune, B.; Poli, R.; Fettinger, J. C. Organometallics 1997, 16,

1581-1594.

2212 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

orange color, which was accompanied by gas evolution. EPR monitoring

showed decomposition to afford a new EPR signal consisting of a

doublet of triplets (g ) 1.950, a_P) 16.5 G, a_H) 24.0 G), which is

attributed to [Cp*MoH(dppe)(THF)]⁺(vide infra).

of this solution did not exhibit any detectable resonances. After

evaporation of the mixture to dryness, an NMR investigation (both ¹H

NMR and ³¹P NMR) of the residue in CD₃CN revealed the resonances

of the previously described⁸¹complexes [Cp*MoH₂(MeCN)(dppe)]⁺

and [Cp*MoH(MeCN)₂(dppe)]²⁺in a ca. 17:83 ratio. No other species

were detected in the NMR spectra.

Chemical Oxidation of 1 in CH₂Cl₂at Room Temperature. Cp₂-

FePF₆(0.024 g, 0.071 mmol) was added to a stirring solution of 1

(0.045 g, 0.071 mmol) in CH₂Cl₂(1 mL) at room temperature. An

immediate reaction occurred which quenched the blue color of the Cp₂-

FePF₆, followed by gas evolution. The solution gradually developed a

red-orange color. EPR monitoring as described in the previous section

showed initially 1⁺as the only paramagnetic product. This was followed

by rapid decomposition, with gas evolution, to afford a new EPR signal

consisting of a doublet of triplets (g ) 1.950, a_P) 16.5 G, a_H) 24.0

G). In a separate experiment, 0.5 mL of CD₂Cl₂was trasferred via

cannula into an NMR tube containing 1 (0.040 g, 0.063 mmol) and

Cp₂FePF₆(0.021 g, 0.063 mmol) at -80 °C, which was then further

cooled to -196 °C, and the tube was flame-sealed under vacuum. The

tube was subsequently warmed to room temperature, resulting in gas

Chemical Oxidation of Compound 2 at Room Temperature.

Synthesis of [Cp*WH₃(dppe)][PF₆] (2⁺PF₆^-). To a yellow-orange

solution of 2 (115 mg, 0.160 mmol) in 4 mL of CH₂Cl₂was added

Cp₂FePF₆(53 mg, 0.160 mmol) at room temperature. The solution

immediately turned red-orange. The solution was filtered and concen-

trated under reduced pressure to ca. 0.5 mL. Orange single crystals of

2⁺PF₆^-were obtained by diffusion of a layer of diethyl ether into this

solution. Yield: 97 mg (70%). A suitable single crystal obtained in

this manner was used for the X-ray analysis. EPR spectrum (CH₂Cl₂):

triplet (br, g ) 2.017), see Figure 1. IR (THF, cm^-1): 1897 (m, broad),

1830 (w, broad), see Figure 2. By comparison, a solution of compound

2 showed M-H stretching vibrations at 1815 and 1885 cm^-1(Figure

2). No diamagnetic impurities were detected in the ¹H NMR spectrum

(CD₃CN). The oxidation of 2 can also be conducted using THF, MeCN,

and acetone as solvents, with identical spectroscopic results.

Chemical Oxidation of Compound 2-d₃at Room Temperature.

To a yellow-orange solution of 2-d₃(0.030 g, 0.042 mmol) in 1 mL of

CH₂Cl₂was added Cp₂FePF₆(0.014 g, 0.042 mmol) at room temper-

ature. The solution immediately turned red-orange. An EPR spectrum

of this solution exhibited a broad triplet resonance with distinguishable

phosphorus coupling (g ) 2.022, a_P) 27.6 G), see Figure 1, attributable

to [Cp*WD₃(dppe)][PF₆] (2⁺-d₃). W-D bands could not be located in

the IR spectra for compounds 2-d₃and 2⁺-d₃(no significant changes

were observed upon oxidation).

X-ray Crystallography. Compound 2⁺PF₆^-. All operations were

routine. Crystal parameters and data collection and refinement details

are summarized in Table 1. Data were corrected for Lorentz and

polarization factors and by absorption (ψ-scan method); a decay

correction was not necessary. All H atoms except the three hydride

ligands were placed in calculated positions and refined with the riding

model. The three hydride ligands were located from the highest peaks

near the W atom with W-H distances ranging from 1.4 to 1.7 Å. They

were initially refined freely but found to possess varying bond lengths

with the W atom, and therefore their refinement was assisted by

restraining the W-H distance to be near 1.70 Å (ADFIX with esd of

0.03 Å). Thermal parameters were allowed to refine freely. Selected

bond distances and angles for complex 2⁺are reported in Table 2.

Compound [Cp*WH₄(dppe)]+PF₆^-([2-H]⁺PF₆^-). Colorless single

crystals of this compound were obtained by slow diffusion of a diethyl

ether layer into a saturated dichloromethane solution. Crystal parameters

and data collection and refinement details are summarized in Table 1.

Data were corrected for Lorentz and polarization factors and by

absorption (ψ-scan method); no decay correction was necessary. The

structure solution and refinement were routine. Handling of the H atoms

attached to carbon followed the protocol described above for compound

2⁺PF₆^-. Four hydride ligands were located at reasonable positions and

W-H distances, but an attempted refinement gave wide variations in

U’s and W-H distances. Attempts at restrained hydride refinement

with d(W-H_initial) ) 1.75 Å in SHELXL93 proved futile. The final

structure was refined to convergence without the presence of the hydride

atoms. Selected bond distances and angles for complex [2-H]⁺are

reported in Table 2.

1

evolution over a 30-45-min period. H NMR spectroscopy exhibited

a resonance (δ 4.61) due to the formation of H₂gas as a product of

this reaction. A control experiment in which H₂gas was bubbled directly

into CD₂Cl₂exhibited a resonance at δ 4.60 in the ¹H NMR spectrum.

Chemical Oxidation of 1-d₃in THF at -80 °C. Cp₂FePF₆(0.016

g, 0.049 mmol) was added to a stirring solution of 1-d₃(0.031 g, 0.049

mmol) in THF (1 mL) at -80 °C. An immediate reaction occurred

which quenched the blue color of the Cp₂FePF₆. An aliquot of this

solution was transferred into an EPR tube which was maintained at

-80 °C. The EPR spectrum at -80 °C consisted of a broad triplet (g

) 1.991, a_P) 28.9 G), attributable to [Cp*MoD₃(dppe)][PF₆] (1⁺-d₃)

(Figure 1). IR (THF, cm^-1): 1318 (m, broad). The spectrum was

recorded immediately after the cell was charged with the cold solution.

By comparison, a solution of compound 1-d₃showed a M-H stretching

vibration centered at 1307 cm^-1. Upon the solution warming to room

temperature, the EPR signal due to 1⁺-d₃disappeared rapidly and was

replaced by a binomial triplet of 1:1:1 triplets (g ) 1.954, a_P) 18.4

G, a_D) 4.0 G), which is assigned to [Cp*MoD(dppe)(THF)]⁺, followed

by further decomposition (vide infra).

Synthesis of {Cp*MoH(dppe)PF₆} (3). Cp₂FePF₆(0.135 g, 0.407

mmol) was added to a stirring solution of 1 (0.257 g, 0.407 mmol)

dissolved in 5 mL of THF at -80 °C. An immediate reaction occurred

which quenched the blue color of the Cp₂FePF₆. The solution was

allowed to stir for 45 min at -80 °C. The Schlenk flash was then

removed from the acetone-dry ice bath, and the solution was allowed

to warm to room temperature with stirring for an additional 30 min.

The solution gradually evolved gas and developed a red-orange color.

After filtration, the solution was concentrated to ca. 2 mL. Heptane

(20 mL) was added, precipitating a light purple solid, which was filtered,

washed with heptane (3 × 15 mL), and dried in vacuo. Yield: 221 mg

(64%). Anal. Calcd for C₃₆H₄₀F₆MoP₃: C, 55.8; H, 5.2. Found: C,

55.1; H, 5.3. EPR spectrum (THF, room temperature): dt (g ) 1.950,

a_P) 16.5 G, a_H) 24.0 G). EPR spectrum (CH₂Cl₂, room tempera-

ture): dt (g ) 1.950, a_P) 16.5 G, a_H) 24.0 G). EPR spectrum (CH₂-

Cl₂, 0 °C): dt of 1:2:3:4:3:2:1 septets (g ) 1.950, a_P) 16.5 G, a_H

)

24.0 G, a_Cl) 1.0 G). Next, 135 mg (0.174 mmol) of this solid was

dissolved in 0.5 mL of MeCN at room temperature, forming a red

solution. The EPR spectrum of this solution was identical, within

experimental error, to the spectra of the THF and CH₂Cl₂solutions.

Et₂O (15 mL) was added to the stirring solution, precipitating a red-

brown solid, which was washed with Et₂O (2 × 5 mL). Yield: 49 mg

(36%). The solids crystallized from THF and from MeCN have identical

IR spectra in Nujol mull. µ_eff) 1.50 µ_B. No diamagnetic impurities

were detected in the ¹H NMR spectrum (CD₃CN). Molar conductivity

(Λ, S cm²mol^-1, CH₂Cl₂): 15.5 (7.1 × 10^-3M); 29.1 (7.1 × 10^-4

Computational Details. All the calculations were performed using

GAUSSIAN 94⁸²on an Origin200 SGI workstation. The LanL2DZ set

(basis I) was employed to perform complete geometry optimization

with a density functional theory (DFT) approach. The three-parameter

(82) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;

Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson,

G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,

V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;

Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;

Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;

Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-

Gordon, M.; Gonzales, C.; Pople, J. A. Gaussian 94 (ReVision E.1); Gaussian

Inc.: Pittsburgh, PA, 1995.

M). Λ_∞) 35.4 S cm²mol^-1

.

Chemical Oxidation of 1 in MeCN at Room Temperature. Cp₂-

FePF₆(0.027 g, 0.081 mmol) was added to a stirring slurry of 1 (0.051

g, 0.081 mmol) in MeCN (1 mL) at room temperature. An immediate

reaction occurred which dissolved the insoluble yellow precipitate; the

solution turned orange. Gas evolution was noted during this period,

which ceased almost immediately (a few seconds). An EPR spectrum

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2213

Table 1. Crystal Data and Structure Refinement for 2⁺PF₆and [2-H]⁺PF₆

-

2⁺PF₆

[2-H]⁺PF₆

-

empirical formula

formula weight

temperature

C₃₆H₄₂F₆P₃W

865.46

153(2) K

C₃₆H₄₃F₆P₃W

866.46

153(2) K

wavelength

crystal system, space group

unit cell dimensions

0.710 73 Å

monoclinic, P2(1)/n

a ) 14.3763(8) Å

b ) 16.684(2) Å

c ) 15.0123(12) Å

â ) 100.424(6)°

3541.4(5) Å³, 4

1.623 Mg m^-3

3.454 mm^-1

monoclinic, P2(1)/n

a ) 14.4683(15) Å

b ) 16.4788(15) Å

c ) 15.0809(15) Å

â ) 100.140(8)°

3539.4(6) Å³, 4

1.626 Mg m^-3

3.456 mm^-1

volume, Z

density (calcd)

absorption coeff

F(000)

θ range for data colln

reflns collected

1724

1.80-24.99°

19 459

1728

1.80-22.50°

9640

independent reflns

max. and min. transmission

data/restraints/paramrs

goodness-of-fit on F²

final R indices [I > 2σ(I)]

6226 [R(int) ) 0.0729]

0.4216 and 0.3384

6226/4/432

4617 [R(int) ) 0.0826]

0.3590 and 0.2640

4617/3/423

1.076

R1 ) 0.0412

1.056

R1 ) 0.0558

wR2 ) 0.0767 [4577 data]

R1 ) 0.0713

wR2 ) 0.1101 [3033 data]

R1 ) 0.1035

R indices (all data)

wR2 ) 0.0861

wR2 ) 0.1265

largest diff peak and hole

1.351 and -0.945 e Å^-3

1.380 and -1.002 e Å^-3

Table 2. Bond Lengths (Å) and Angles (deg) for 2⁺PF₆and

[2-H]⁺PF₆^-, and Comparison with the Previously Reported Values

-

for [2-H]⁺BF₄

- a

-

- b

2⁺PF₆

[2-H]⁺PF₆

[2-H]⁺BF₄

W1-CNT

W(1)-P(1)

W(1)-P(2)

W(1)-H(1)

W(1)-H(2)

W(1)-H(3)

1.999(15)

2.474(2)

2.506(2)

1.71(2)

1.67(3)

1.69(3)

1.996(5)

2.484(3)

2.516(3)

1.990(9)

2.477(2)

2.508(2)

P(1)-W(1)-P(2)

CNT -W1-P1

CNT -W1-P2

CNT-W1-H1

CNT -W1-H2

CNT -W1-H3

78.85(5)

162.0(2)

119.1(2)

109(2)

103(2)

113(2)

78.79(11)

163.6(2)

117.7(2)

78.72(7)

162.8(2)

118.4(2)

Figure 3. Peak potentials for the oxidation of 1 and reduction of 1⁺

in MeCN as a function of scan rate (V in V s^-1) at [1] ) 8.9 × 10^-4

M. Circles refer to CV data, and squares refer to TLCV data. The lines

fitting the data are obtained from simulations as described in the text.

The insets show representative thin-layer cyclic voltammograms [V )

4 mV s^-1(a)] and cyclic voltammograms [V ) 0.02 (b) and 1 (c) V

s^-1].

^aCNT is the centroid of atoms C(1)-C(5). ^bFrom ref 81.

form of the Becke, Lee, Yang, and Parr functional (B3LYP)⁸³was

employed. The LanL2DZ basis set includes both Dunning and Hay’s

D95 sets for H and C⁸⁴and the relativistic electron core potential (ECP)

sets of Hay and Wadt for the heavy atoms.^85-87Electrons outside the

core were all those of H and C atoms, the 3s and 3p electrons of the

P atoms, and the 5s, 5p, 5d and 6s electrons of the W atom. Single-

point calculations on the B3LYP/LANL2DZ-optimized geometries were

also carried out with an extended basis set (basis II) in which the hydride

ligands are represented in a triple-ê basis set. All calculations were

carried out without imposed symmetry.

The mean value of the spin of the first-order electronic wave function,

which is not an exact eigenstate of S²for unrestricted calculations on

open-shell systems, was considered suitable for the unambiguous

identification of the spin state. Spin contamination was carefully

monitored, and the energies shown in the Results section correspond

to unrestricted B3LYP (UB3LYP) calculations. The value of S²for

the UB3LYP calculations was 0.7546 for [CpWH₃(H₂PCH₂CH₂PH₂)]⁺,

0.7545 for CpWH₂(H₂PCH₂CH₂PH₂), and 2.0117 and 2.0126 for cis

and trans triplet [CpWH₂(H₂PCH₂CH₂PH₂)]⁺, respectively, indicating

minor spin contamination.

Results

(a) Cyclic Voltammetric Studies. Both compounds 1 and 2

exhibit reversible cyclic voltammograms at fairly negative

potentials (E_1/2) -0.73 V for 1 and -0.88 V for 2 in THF),

indicating electron richness. As expected, the W compound (2)

is easier to oxidize than the Mo compound. Changing the solvent

to CH₂Cl₂, MeCN, or acetone does not affect the reversibility

nor the E_1/2value for 2, nor is any change observed upon

changing the solvent to CH₂Cl₂for 1. On the other hand, 1

exhibits a semireversible system (-0.85 V) in MeCN, indicating

a faster decomposition of 1⁺in this solvent (vide infra). The

return scan shows a smaller intensity for the reduction wave of

1⁺and the appearance of a new reduction peak (-1.58 V).

Faster scan rates resulted in an increased reversibility for the

oxidation of 1. Representative cyclic voltammograms at different

scan rates are shown in the insets of Figure 3. The oxidation of

1 as a function of scan rate and concentration was investigated

in both bulk (CV) and thin-layer (TLCV) cyclic voltammetry

at 20 °C. The results yield useful information on the decomposi-

(83) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.

(84) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry;

Schaefer, H. F., III, Ed.; Plenum Press: New York, 1976; pp 1-28.

(85) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283.

(86) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284-298.

(87) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310.

2214 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

Table 3. Slope and Intercept with E ) E° of the Straight Lines Fitting the Anodic Peak Potentials at Low Scan Speeds (V in V s^-1) for the

Oxidation of 1 in MeCN

CV

TLCV

slope of E_p,a

(mV) vs log V

log V at E ) E°

log

k_disp

slope of E_p,a

(mV) vs log V

log V at E ) E°

log

k_deprot

b

log c^a

(log V_{CV )}

(log V_{TLCV )}

i

-3.30

-3.20

-3.19

-3.08

-3.05

-3.03

-3.20

-3.14

-3.11

-3.06

-3.05

-2.96

3.65

3.61

3.63

3.57

3.55

3.62

52(4)

53(1)

56(5)

53(4)

49(1)

54(1)

-2.1

-1.95

-1.7

-1.9

-2.18

2.49

2.39

2.53

2.67

2.44

2.14

9(1)

14(5)

17(7)

22(5)

14(1)

^ac ) [1] in mol L^{-1 b}See text (k in s^-1M^-1).

.

tion mechanism of the primary oxidation product 1⁺(see

Discussion). In the region of scan rates where a distinctive return

wave is observable (scan rate > 100 mV s^-1), both anodic and

cathodic peak potentials are independent of scan speed and

concentration. At lower scan rates, the anodic peak shifts to

more negative potentials while the cathodic peak disappears.

At constant concentration, the shift of the anodic peak potential

is linear with respect to the logarithm of the scan rate in both

CV and TLCV experiments. Figure 3 shows a representative

study at [1] ) 8.9 × 10^-4M. Table 3 reports the slopes of

these lines for all concentrations used in this study, as well as

the value of log V at the intercept with the E ) E° horizontal

line. The large errors on some of the slopes obtained in the CV

study are the consequence of the limited range of scan rates

(e0.5 log unit) in the kinetic region.

1-d₃to 1⁺-d₃similar to that observed for the Mo-H vibrations

of 1⁺relative to those of 1. In addition, the expected isotope

shift is observed upon H/D exchange on going from 1 to 1-d₃

and from 1⁺to 1⁺-d₃.

Compound 2⁺can be generated in THF, CH₂Cl₂, MeCN, or

acetone at room temperature. It exhibits a broad triplet resonance

in the EPR spectrum (g ) 2.017; see Figure 1). Unlike complex

1⁺, complex 2⁺is rather stable at room temperature in any of

the aforementioned solvents, permitting its isolation in the solid

state and the growth of single crystals suitable for X-ray analysis

(vide infra). Cooling of the solution to -80 °C does not affect

the line shape of the EPR signal. IR investigations for 2 and 2⁺

also reveal a blue shift upon oxidation, as observed for the Mo

analogue (see Figure 2). Oxidation of 2-d₃leads to 2⁺-d₃, which

better reveals the phosphorus coupling (g ) 2.022, a_P) 27.6

G; see Figure 1). An IR investigation of these two complexes

did not allow the identification of the W-D stretching vibrations

(no visible change is detected in the IR spectrum upon

oxidation), probably because these weak and broad bands are

overshadowed by the other stronger bands.

At constant scan rate, the shift of the anodic peak potential

is linear with respect to the logarithm of the concentration. The

data were fit to straight lines, giving average slopes of -15 (

3 mV per log unit for the CV data and -48 ( 7 mV per log

unit for the TLCV data.

(b) Chemical Generation and Spectroscopic Characteriza-

tions of 1⁺and 2⁺. The interaction of 1 and 2 with 1 equiv of

Cp₂FePF₆affords the corresponding cations, 1⁺and 2⁺,

respectively (see eq 1). Complex 1⁺can be generated in THF

(c) Chemical Oxidation of 1 in CH₃CN. The chemical

oxidation of 1 with 1 equiv of Cp₂FePF₆in MeCN provides

very different results (see eq 2) relative to those of the same

reaction in THF or CH₂Cl₂. An immediate and short-lasting (a

+

Cp₂Fe

8 {[Cp*MoH₃(dppe)]⁺} ^MeCN8

Cp*MH₃(dppe) + Cp₂FePF₆f

Cp*MoH₃(dppe)

[Cp*MoH₂(dppe)(MeCN)]⁺+

[Cp*MoH(dppe)(MeCN)₂]²⁺(2)

fast addition: ca. 17:83

[Cp*MH₃(dppe)][PF₆] + Cp₂Fe (1)

(M ) Mo, 1⁺PF₆; W, 2⁺PF₆)

-

or CH₂Cl₂at -80 °C and is indefinitely stable at this

temperature. When the same reaction is carried out at room

temperature, the formation of 1⁺still takes place, but this is

followed by a relatively rapid decomposition (t_1/2of a few

minutes in both solvents). We shall examine this phenomenon

and the nature of the decomposition product in more detail in

section d. The EPR spectrum of 1⁺exhibits a triplet of quartets

(g ) 1.989, a_P) 28.9 G, a_H) 11.8 G; see Figure 1), consistent

with coupling to three equivalent H and two equivalent P nuclei.

The equivalence of the P and H hyperfine couplings in the EPR

spectrum of 1⁺clearly indicates that fluxional processes are

occurring. IR investigations show broad bands attributable to

M-H stretching vibrations, which blue-shift upon oxidation (see

Figure 2).

The chemical oxidation of 1-d₃in THF leads to the formation

of 1⁺-d₃, which is characterized by an EPR broad triplet at room

temperature (g ) 1.991, a_P) 28.9 G; see Figure 1), confirming

the assignments made for the spectrum of 1⁺. The triplet

coupling is essentially identical with that observed for the

corresponding trihydride complex, whereas the quartet coupling

in 1⁺disappears on going to complex 1⁺-d₃. IR investigations

show a blue shift for the Mo-D vibrations upon oxidation of

slow addition (40 min): ca. 36:64

few seconds) H₂evolution occurs, and no observable amount

of 1⁺or any other paramagnetic species is shown by EPR

investigation. The ¹H NMR analysis of the final solution

revealed that two hydridic products, [Cp*MoH₂(MeCN)(dppe)]⁺

and trans-[Cp*MoH(dppe)(MeCN)₂]²⁺, are formed in a 17:83

ratio. Both products have previously been obtained by carrying

out the protonation of 1 with HBF₄in MeCN.⁸¹It is interesting

to note, however, that the double protonation of 1 (or the

protonation of [Cp*MoH₂(MeCN)(dppe)]⁺) gives rise to two

different isomers of the dicationic product in a 3:1 ratio, the

major product having a pseudo-trigonal prismatic structure and

the minor, thermodynamically more stable isomer having a

pseudo-octahedral structure.⁸¹The oxidation process reported

here, on the other hand, gives the thermodynamically more

stable isomer selectiVely. This demonstrates that the dicationic

product cannot result from a proton transfer from 1⁺to either

1 or [Cp*MoH₂(MeCN)(dppe)]⁺(see Discussion).

An additional oxidation experiment carried out under the same

conditions, except that the 1 equiv of ferrocenium was admin-

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2215

crystals for this compound have been unsuccessful. The

measured magnetic moment for solid 3 is 1.50 µ_B, close to the

expected value for one unpaired electron. A spin doublet

configuration is systematically observed for 17-electron half-

sandwich complexes of Mo(III).⁸⁹

Solutions of 3 in CH₂Cl₂give rise to electrical conductivity

(Λ_∞) 35.4 S cm²mol^-1), in agreement with its formulation as

a 1:1 salt;⁹⁰thus, the PF₆^-anion is not coordinated to the metal

in this solvent. The nonexistence of 15-electron half-sandwich

Mo(III) complexes leads to the hypothesis that one loosely

bound molecule of CH₂Cl₂completes the coordination sphere

in solution. This is confirmed by the EPR spectrum at 0 °C

(Figure 4b). Each of the lines observed at room temperature

splits, upon cooling of the solution to 0 °C, into a 1:2:3:4:3:2:1

septet, unambiguously indicating coupling to two equivalent Cl

3

nuclei (³⁵Cl and ³⁷Cl, both I ) /₂, similar gyromagnetic ratio,

100% total abundance). It is reasonable to propose that a single

molecule of CH₂Cl₂is coordinated to the metal in a chelating

mode, as observed, for instance, in Ag₂(CH₂Cl₂)₄Pd(OTeF₅)₄

and in [RuH(CH₂Cl₂)(CO)(PBu^t₂Me)₂][B(C₆H₃(CF₃)₂-3,5)₄],^91,92

rather than two separate molecules in a monodentate mode.

Thus, all evidence points to the formulation of 3 in CH₂Cl₂as

[Cp*MoH(dppe)(η²-Cl₂CH₂)]⁺PF₆^-, having a 19-electron con-

figuration for the metal valence shell. The inability to observe

the Cl coupling at room temperature must be ascribed to a fast

exchange between the loosely bound CH₂Cl₂ligand and the free

solvent. Cooling the THF and MeCN solutions did not result

in any change in the EPR spectrum. In particular, no coupling

to the ¹⁴N (I ) 1, 100%) nucleus/nuclei was observable in

MeCN down to -45 °C. On the basis of the results obtained in

CH₂Cl₂, it is reasonable to assume that the complex also

coordinates a solvent molecule when dissolved in THF or

MeCN, since the latter Lewis bases have stronger coordinating

properties than CH₂Cl₂. The most likely structure in these cases

is a four-legged piano stool for [Cp*MoH(dppe)(S)]⁺(S ) THF

or MeCN) with a 17-electron configuration at the metal center.

This formulation is in agreement with the electrochemical

studies. Compound 3 shows an irreversible oxidation wave at

1 V s^-1, whose peak potential depends on the nature of the

solvent: 0.719 V in THF, 0.597 V in CH₂Cl₂, and 0.504 V in

MeCN. A reverse reduction wave is not observed, even when

carrying out the experiments at 100 V s^-1. Although these

potential shifts should be taken with some caution, they are too

large to be attributed solely to differences (if any) in follow-up

chemical processes and to a solvent dependence of the ferrocene

redox process. Therefore, the anodic peak potential can be taken

as a qualitative measure of the electron richness at the metal.

The donor power of the three solvents in the order MeCN >

THF > CH₂Cl₂should lead to anodic peak potentials in the

reverse order. However, the observed order is different (MeCN

< CH₂Cl₂< THF). The unexpected CH₂Cl₂/THF reversal may

be reconciled with the proposed chelating coordination mode

of dichloromethane and with monocoordination for THF. In

other words, dichloromethane is effectively a better electron

donor, by virtue of its bidentate coordination mode, than a single

molecule of THF.

Figure 4. EPR spectrum of compound Cp*MoH(dppe)(PF₆). Solvent

) CH₂Cl₂. (a) T ) room temperature. (b) T ) 0 °C. The peaks marked

with an asterisk are due to an unknown impurity.

istered slowly over 40 min instead of all at once, gives the same

two products of eq 2, but the dihydride/monohydride ratio is

now 36:64. Once again, the monohydride complex is present

in a single isomeric form, corresponding to the thermodynami-

cally stable isomer.

(d) Decomposition of 1⁺in THF and CH₂Cl₂. Warming a

THF or CH₂Cl₂solution of 1⁺PF₆^-(generated at -80 °C, vide

supra) to room temperature, or direct oxidation of 1 with Cp₂-

FePF₆at room temperature in THF or CH₂Cl₂, causes decom-

position of 1⁺with gas evolution. The decomposition was

complete within a few minutes at room temperature, as shown

by EPR monitoring. The nature of the gas as H₂is confirmed

by its resonance at δ 4.61 in the ¹H NMR spectrum, in

comparison with an authentic sample. The reaction affords a

new, highly air-sensitive compound (3), which is characterized

by a doublet of triplets EPR resonance (g ) 1.950, a_P) 16.5

G, a_H) 24.0 G; see Figure 4a), consistent with coupling to

two equivalent P nuclei and one proton. The spectral parameters

are identical whether the spectrum is taken in a THF, CH₂Cl₂,

or MeCN solution. Notice that this product cannot be obtained

directly by oxidation of 1 in MeCN (vide supra) but is rather

stable in this solvent after it is formed. Solutions of 3 in MeCN

decompose quite slowly (t_1/2≈ 4 h) to afford a mixture of the

monohydride (thermodynamic isomer only) and dihydride

products of eq 2, plus other noncharacterized, nonhydridic

products.

Compound 3 has been isolated as a microcrystalline solid

and is assigned the formula Cp*MoH(dppe)PF₆on the basis of

the EPR spectrum and the elemental composition (see eq 3).

-H₂

[Cp*MoH₃(dppe)][PF₆]

8 Cp*MoH(dppe)PF₆(3)

1⁺PF₆

-

3

(88) Abugideiri, F.; Keogh, D. W.; Kraatz, H.-B.; Poli, R.; Pearson, W.

J. Organomet. Chem. 1995, 488, 29-38.

(89) Poli, R. J. Coord. Chem. B 1993, 29, 121-173.

Elemental analysis indicates no solvent incorporation, and the

IR spectra of batches of the compound recrystallized from THF

or from MeCN are identical. On the basis of the nonexistence

of half-sandwich compounds of Mo(III) with 15 valence

(90) Burgess, J. Ions in Solution: Basic Principles of Chemical Interac-

tions; Wiley: New York, 1988; p 187.

(91) Colsman, M. R.; Newbound, T. D.; Marschall, L. J.; Noirot, M.

D.; Miller, M. M.; Wulfsberg, G. P.; Frye, J. S.; Anderson, O. P.; Strauss,

S. H. J. Am. Chem. Soc. 1990, 112, 2349-2362.

(92) Huang, D.; Huffman, J. C.; Bollinger, J. C.; Eisenstein, O.; Caulton,

K. G. J. Am. Chem. Soc. 1997, 119, 7398-7399.

electrons,⁸⁸we propose that the PF₆anion is coordinated to

-

the metal center in the solid state. Attempts to grow single

2216 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

Figure 6. ORTEP view of the molecular geometry for 2⁺. Ellipsoids

are drawn at the 30% probability level.

W-P1 and CNT-W-P2 angles for 2⁺and [2-H]⁺(while

CNT-W-P1 is larger by 1.6(2)° for [2-H]⁺, CNT-W-P2 is

larger by 1.4(2)° for 2⁺). The W-P and W-Cp* bond distances

(see Table 2) are very similar for 2⁺and [2-H]⁺, suggesting

that the replacement of a W-H bond with a singly occupied

W orbital does not greatly affect the effective metal charge (i.e.,

the W-H bond is highly covalent). The geometry of 2⁺is

intermediate between I (a pseudo-trigonal prism) and II (a

pseudo-octahedron), i.e., the ideal geometries adopted by the

isoelectronic d²complexes 1 and trans-octahedral [Cp*MoH-

(dppe)(MeCN)₂]²⁺(MeCN occupies the two trans H positions),

respectively.⁸¹

Figure 5. Decay from the EPR monitoring experiments of the relative

concentration of (a) 1⁺(b) and 1⁺-d₃(O) in CH₂Cl₂at 0 °C and (b) 1⁺

in THF at 0 °C (b) and 10 °C (O).

The decomposition of 1⁺was investigated kinetically by EPR

monitorning in both THF and CH₂Cl₂below room temperature.

The extensive overlap of the EPR signals required the deter-

mination of the relative concentrations of residual 1⁺and

product by signal subtraction. The quality of the data was not

sufficient to obtain accurate double integrals, but an estimate

of the relative concentration at each time was obtained by

comparison of the subtracted spectrum with the initial spectrum

multiplied by a coefficient. The resulting data show the expected

first-order decay (see Figure 5) in all cases. The decay is much

faster in CH₂Cl₂(k ) 0.125(6) min^-1at 0 °C) relative to that

in THF (k ) 0.016(1) min^-1at 0 °C and 0.100(3) min^-1at 10

°C).

As expected from the behavior of the hydride analogue, 1⁺-

PF₆^--d₃decomposes in THF at room temperature to afford

Cp*MoD(dppe)PF₆(3-d). The EPR spectrum of 3-d at room

temperature allows the observation of the deuterium coupling

as 1:1:1 triplets (a_D) 4.0 G), which is related to the H coupling

of 3 by a measured a_H/a_Dratio of 6.0, close to the theoretical

value of 6.5. The EPR monitoring of this transformation in CH₂-

Cl₂at 0 °C and the subsequent data analysis as outlined above

for the trihydride analogue yielded the data illustrated in Figure

5a and a first-order rate constant of 0.25(1) min^-1. Therefore,

the decay of the oxidized trihydride complex shows an inVerse

isotope effect, k_H/k_D) 0.50(3).

There was some concern, given the remarkable structural

similarities between the [Cp*W(dppe)] moieties in complexes

2⁺and [2-H]⁺, that the crystal examined could correspond to

compound [Cp*WH₄(dppe)]PF₆. This possibility seems excluded

by the simple observation that the crystal of 2⁺PF₆used for

-

the X-ray experiment is orange, while that of [Cp*MoH₃-

-

(dppe)[PF₆]⁺BF₄is colorless. Nevertheless, in the search for

additional evidence, the tetrahydride cation has also been

-

crystallized with the PF₆counterion, and an X-ray structure

has been determined on a genuine (colorless) crystal of

[2-H]⁺PF₆^-. The variations of the metric parameters of the

[Cp*W(dppe)] moiety are rather small (see Table 2), but the

differences in unit cell parameters (Table 1) are sufficient to

establish that the two crystals relate to two similar but different

compounds. Furthermore, dissolution of crystals from the same

batch used for the X-ray investigation showed no NMR signal

attributable to [2-H]⁺for the sample of 2⁺and no EPR signal

attributable to 2⁺for the sample of [2-H]⁺.

(f) DFT Calculations. Calculations on the [CpWH_x(H₂PCH₂-

CH₂PH₂)]ⁿ⁺(x ) 3, 2; n ) 0, 1) model systems were carried

out with the following objectives in mind. A full geometry

optimization starting from different ideal configurations could

reproduce the somewhat bizarre structural preference for both

(e) Structural Characterizations. A single-crystal X-ray

investigation of 2⁺PF₆permitted the location and refinement

-

of the three hydride ligands (see Figure 6). Neglecting the

hydrogen positions, the geometry of 2⁺is almost identical with

that of complex [Cp*WH₄(dppe)]⁺([2-H]⁺), namely the pro-

tonation product of 2. This tetrahydrido cation had been

previously characterized by X-ray diffraction as the BF₄^-salt.⁸¹

The major difference consists of a small displacement (0.025-

(8) Å, as opposed to 0.002(9) Å in the tetrahydrido complex)

of the W atom from the plane defined by CNT, P(1) and P(2),

toward the side of the molecule occupied by H2 and H3 (see

Figure 6). There are also small differences between the CNT-

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2217

latter are subject to large uncertainties. Thus, the results of our

calculations provide support to the geometry proposed on the

basis of X-ray data for 1 and, at the same time, suggest that 2

should adopt a structure identical with that of 1. A calculation

based on a cis-mer pseudo-octahedral starting geometry, as

experimentally observed for the trichloride complexes CpMoCl₃-

(dppe) and CpMoCl₃(dmpe),^94,95led again to IV, showing that

there is an electronic preference for the pseudo-trigonal

prismatic geometry. This preference must be related to the

geometrical constraint of the chelating ligand, since the

CpMoH₃(PMe₂Ph)₂complex with two monodentate phosphorus

ligands was experimentally found to adopt a mer-trans pseudo-

octahedral structure.⁹⁶

The global energy minimum for the cationic system is

obtained for a geometry (VI in Figure 8) which corresponds

remarkably well to that experimentally determined for 2⁺

(compare Figure 8 with Figure 6). Once again, the remarkable

closeness of the hydride positions calculated by theory (see

Supporting Information) to those obtained from the X-ray data

gives stronger confidence in the correctness of the experimental

result. Starting from the geometry of the neutral global

minimum, an eq,eq octahedral local minimum (structure V in

Figure 8) which is only 0.8 kcal/mol higher in energy than the

global minimum (0.2 kcal/mol at the basis II level) was obtained.

Considering the Cp*-dppe steric repulsion (greater in the eq,eq

pseudo-octahedral structure), the computational result provides

an even stronger justification for the adoption of the observed

structure by complex 2⁺.

A comparison of the [CpWH₃(H₂PCH₂CH₂PH₂)]ⁿ⁺(n ) 0,

IV; 1, VI) optimized structures shows that the W-H distances

become shorter upon oxidation (average -0.010 Å), while the

W-P and W-CNT distances become longer (+0.085 and

+0.007 Å, respectively). While the shortening of the W-H

distances is as expected on the basis of a contraction of the W

atomic radius upon oxidation, the lengthening of the W-P and

W-CNT distances indicates correspondingly weaker interac-

tions in the oxidized complex. Variations of this kind have been

noted before and attributed to a weakening of the M-P π and

M-Cp δ back-bonding components, respectively.^97-100

The W-H bond strengthening upon oxidation is further

confirmed by the calculation of the bond dissociation energies

according to the processes of eqs 3 and 4. For the neutral system,

each of the three inequivalent hydride ligands H1, H2, and H3

was removed as a neutral H atom from structure IV to give a

different 17-electron CpWH₂(H₂PCH₂CH₂PH₂). The three re-

sulting fragments were calculated at their fixed geometries,

giving rise to energy differences (also called bond energy

terms)¹⁰¹of 77.3, 72.7, and 79.6 kcal/mol (77.5, 72.7, and 79.5

kcal/mol at the basis II level). Subsequently, the lowest energy

geometry was allowed to relax, giving the fully optimized

doublet CpWH₂(H₂PCH₂CH₂PH₂) geometry shown (VII) in

Figure 7, which corresponds to the geometry experimentally

observed for complex Cp*MoCl₂(dppe).¹⁰²The cis isomer for

Figure 7. B3LYP/LANL2DZ-optimized geometries and relative

energies for CpWH₃(H₂PCH₂CH₂PH₂) and CpWH₂(H₂PCH₂CH₂PH₂)

+ H.

the neutral and the cationic complexes. Complex 1, in fact, was

shown to adopt a pseudo-trigonal prismatic geometry which is

unprecedented for compounds with the CpMX₃L₂stoichiome-

try.^81,93For this purpose, the H₂PCH₂CH₂PH₂model rather than

the more typical (PH₃)₂system was chosen to replace the dppe

ligand. In addition, a comparison between the energy changes

for eqs 4 and 5 could provide information on the M-H bond

strength dependence on the metal oxidation state.

CpWH₃(H₂PCH₂CH₂PH₂) f

CpWH₂(H₂PCH₂CH₂PH₂) + H (4)

[CpWH₃(H₂PCH₂CH₂PH₂)]⁺f

[CpWH₂(H₂PCH₂CH₂PH₂)]⁺+ H (5)

All geometry optimizations were carried out with a standard

LANL2DZ basis set (basis I) for all atoms. Single-point

calculations were then carried out at these optimized geometries

by using an augmented basis set (basis II) with triple-ú functions

for the hydridic hydrogen atoms. The effect of introducing a

third contraction for the description of the hydride atomic orbital

is a lowering of the total energy by a few kilocalories per mole

for each system. However, the relative energies at the basis II

level do not differ from those at the basis I level by more than

0.3 kcal/mol.

The results of the geometry optimizations are summarized

in Figures 7 and 8. The global minimum for the neutral

trihydride complex corresponds to a pseudo-octahedral structure

(structure III in Figure 7), while the next most favored structure

(IV) corresponds to the experimentally observed (for 1) pseudo-

trigonal prismatic structure. The discrepancy between the

theoretical prediction and the experimental result is probably

due to the steric repulsion between the Cp* and the dppe ligands.

The calculated bond distances and angles for the W model

system IV are in good agreement with those observed for the

Mo system 1 (the calculated geometrical parameters of systems

III-IX have been deposited as Supporting Information). The

calculated hydrogen positions are remarkably close to those

determined by the single-crystal X-ray analysis, although the

(94) Sta¨rker, K.; Curtis, M. D. Inorg. Chem. 1985, 24, 3006-3010.

(95) Owens, B. E.; Poli, R. Inorg. Chim. Acta 1991, 179, 229-237.

(96) Abugideiri, F.; Fettinger, J. C.; Pleune, B.; Poli, R.; Bayse, C. A.;

Hall, M. B. Organometallics 1997, 16, 1179-1185.

(97) Orpen, A. G.; Connelly, N. G. Organometallics 1990, 9, 1206-

1210.

(98) Orpen, A. G.; Connelly, N. G. J. Chem. Soc., Chem. Commun. 1985,

1310-1311.

(99) Krueger, S. T.; Poli, R.; Rheingold, A. L.; Staley, D. L. Inorg. Chem.

1989, 28, 4599-4607.

(100) Keogh, D. W.; Poli, R. J. Chem. Soc., Dalton Trans. 1997, 3325-

3333.

(93) Fettinger, J. C.; Pleune, B.; Poli, R. J. Am. Chem. Soc. 1996, 118,

4906-4907.

(101) Martinho Simo˜es, J. A.; Beauchamp, J. L. Chem. ReV. 1990, 90,

629-688.

2218 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

Figure 8. B3LYP/LANL2DZ-optimized geometries and energies for [CpWH₃(H₂PCH₂CH₂PH₂)]⁺and [CpWH₂(H₂PCH₂CH₂PH₂)]⁺+ H.

the dihydride system was optimized to a local minimum 2.5

kcal/mol higher (3.0 kcal/mol at the basis II level) than the trans

isomer.

theses) are 1828 (1), 1862 (0.79), and 1951 (0.30) cm^-1for n

) 0 and 1885 (0.75), 1920 (1), and 1960 (0.33) cm^-1for n )

1. The frequencies, relative intensities, and normal modes are

also shown in Figure 2b, in comparison with the experimental

IR spectra. The correspondence between calculated and experi-

mental spectra is quite acceptable. This agreement confirms the

strengthening of the W-H bonds upon oxidation, which is

independently indicated by the bond shortening and BDE

increase.

An analogous procedure was repeated for the cationic

complex (see Figure 8), the 16-electron dihydrido complex

[CpWH₂(H₂PCH₂CH₂PH₂)]⁺being calculated in both singlet

(VIII) and triplet (IX) spin states, the latter yielding a more

stable structure. It is interesting to note that the optimized triplet

geometry is a four-legged piano stool, qualitatively similar to

those experimentally observed by X-ray crystallography for the

spin triplet Cp*MoCl₃L (L ) PMe₃, PMePh₂) and [CpMoCl₂-

(PMe₃)₂]⁺complexes.^99,103The optimized singlet geometry, on

the other hand, is better viewed as a pseudo-octahedral structure

with a missing equatorial position. In this structure, the chelating

H₂PCH₂CH₂PH₂ligand occupies an equatorial and the axial

position, as experimentally observed for CpMoCl₃(dmpe).⁹⁵

Other geometric arrangements gave higher energy local minima

for both singlet and triplet configurations. Similarly to the

trihydride system mentioned above, the dihydride system also

shows a shortening of the W-H distances and a lengthening

of the W-P and W-CNT distances upon oxidation.

As shown by the calculations, the BDE for the W-H bonds

(with respect to the relaxed dihydride geometry) increases from

66.1 kcal/mol for the neutral trihydride system to 69.1 kcal/

mol for the cationic complex (relative to the triplet ground state).

This increase is paralleled, as expected, by an increase of overlap

population for the three W-H bonds: the values increase from

0.232 to 0.258 for W-H1, from 0.227 to 0.233 for W-H2,

and from 0.241 to 0.245 for W-H3, respectively.

Discussion

(a) M-H Bond Strength. The W-H bond strengthening

upon oxidation of 1 and 2 is clearly indicated by three

independent experimental and/or computational results: (i) the

shortening of all W-H bonds calculated for the [CpWH₃(H₂-

PCH₂CH₂PH₂)]ⁿ⁺system; (ii) the increase of the calculated BDE

for the same model system; and (iii) the blue shift of the Mo-

H(D) and W-H stretching vibrations upon oxidation of 1 (1-

d₃) and 2, the latter being reproduced by the calculations on

the model systems. From simple theory, the vibrational fre-

quency correlates directly with the bond energy and inversely

with the bond length.¹⁰⁴Although exceptions to this rule have

been presented for bonds that contain multiple components (e.g.,

ligand-metal σ bonding + π back-bonding), whereupon

stronger bonds are occasionally found not to be the shortest,¹⁰⁵

a similarly unusual behavior is not expected for a bond as simple

as a terminal M-H bond.

This M-H bond strengthening upon oxidation is a novel

finding in transition metal hydride chemistry. Unfortunately,

most 17-electron hydride complexes are too unstable for IR

studies. IR studies of [Cp*FeH(dppe)]⁺, [WH₂Cl₂(PMe₃)₄]⁺, and

their respective neutral precursors indicated either no change

or a slight red shift in the spectra upon oxidation.^8,9The use of

thermodynamic cycles involving pK_aand electrochemical

A normal-mode analysis was carried out on the [CpWH₃(H₂-

PCH₂CH₂PH₂)]ⁿ⁺optimized geometries that correspond to the

experimentally observed structures for 1 and 2⁺. The calculated

normal modes of vibration for the [WH₃]ⁿ⁺moiety reproduce

the experimentally observed increase of frequency upon oxida-

tion. The calculated frequencies (relative intensities in paren-

(102) Fettinger, J. C.; Keogh, D. W.; Pleune, B.; Poli, R. Inorg. Chim.

Acta 1997, 261, 1-5.

(103) Abugideiri, F.; Gordon, J. C.; Poli, R.; Owens-Waltermire, B. E.;

Rheingold, A. L. Organometallics 1993, 12, 1575-1582.

(104) Infrared and Raman Spectroscopy: Methods and Applications;

Schrader, B., Ed.; VCH: Weinheim, 1995; p 693.

(105) Ernst, R. D.; Freeman, J. W.; Stahl, L.; Wilson, D. R.; Arif, A.

M.; Nuber, B.; Ziegler, M. L. J. Am. Chem. Soc. 1995, 117, 5075-5081.

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2219

Scheme 1

Scheme 2

of polarity. For a M-H^δ+system, oxidation should definitely

lead to a weakening of the ionic component. Notice, however,

that the sharp acidity increase following oxidation of hydride

complexes¹⁰⁸is not, per se, an indication of M-H^δ+bond

polarity. Experimental and theoretical studies of charge distribu-

tion in low-valent, carbonyl-based transition metal hydrides

always indicate a M-H^δ-bond polarization, even for highly

acidic complexes.^109-111For instance, XPS studies lead to the

estimation of a hydride formal charge of -0.3, -0.8, and -0.75

for H₂Fe(CO)₄, HMn(CO)₅, and HCo(CO)₄, respectively.¹⁰⁹

The net effect of oxidation on the M-H bond strength may

thus be either a strengthening or a weakening, although

strengthening is expected for M-H^δ-systems with very

electron-rich fragments (high-energy M frontier orbital) and

when the bonding energetics is dominated by the covalent

component. Compounds 1 and 2 contain ligands that have low

electronegativities, are strong electron donors, and do not have

strong π-accepting capabilities. Consequently, the Cp*M(dppe)

(M ) Mo, W) fragments have a low “group electronegativity”.

One final consideration may be made on the basis of the

thermodynamic cycle (Scheme 2) that was introduced by Tilset

et al.^72,106,107to estimate bond strength variations on the basis

of electrochemical data. According to this cycle, a M-H bond

weakening is associated with systems for which the oxidation

of M^•occurs at a less positive potential (easier oxidation) relative

to that for the oxidation of MH. We want to offer a word of

caution about the use of this cycle. The 16-electron [M]⁺species

(and to a lesser and variable extent also the 17-electron M^•and

[MH]^•+species) is highly likely to establish strong interactions

with the polar solvent or with the supporting electrolyte used

for the electrochemical measurements, skewing the calculation

toward the result of a MH bond weakening following oxidation.

In other words, the calculated difference F[E_ox(MH) - E_ox(M^•)]

should also reflect variations in the extent of metal-medium

interactions. For our system, we cannot experimentally carry

out this study because of the unavailability of Cp*MH₂(dppe)

(M ) Mo, W). The DFT calculations, however, yield a ∆E of

-3.0 kcal/mol for reaction 6. Assuming negligible ∆(PV) and

measurements led to the deduction of a M-H bond weakening

by 8-12 kcal/mol upon one-electron oxidation for compounds

CpCrH(CO)₂L (L ) P(OMe₃)₃, PPh₃, PEt₃), Cp*CrH(CO)₃,

TpMH(CO)₃, Tp*MH(CO)₃(M ) Cr, Mo, W), and Cp*FeH-

(dppe).^72,106,107

To visualize the expected effect of oxidation on the M-H

bond strength, let us qualitatively analyze the M-H bond as

having a covalent and an ionic component, according to Pauling.

Let us first focus on the covalent component. A M-H^δ-bond

polarity corresponds to a situation where the M frontier orbital

is higher in energy than the H frontier orbital (1s) (see Scheme

1, left). An increase of formal positive charge on the metal center

(such as, for instance, following an oxidation process) has two

main effects: (i) it lowers the energy of the metal-based frontier

orbital used for the M-H interaction and (ii) it contracts the

metal orbitals. The first effect has a positive consequence on

the M-H bond strength if the frontier orbital ends up closer in

energy to the H 1s orbital. An excessive energy lowering, on

the other hand, would weaken the covalent interaction. The

second effect has two consequences: it lowers the overlap

between the frontier orbitals of M and H, weakening the

interaction, but it also reduces the metal radius, allowing the H

atom to approach the metal at a closer distance. All in all, the

orbital contraction may provide a positive contribution to bond

strengthening. The DFT calculations on the [CpWH₃(H₂PCH₂-

CH₂PH₂)]ⁿ⁺systems show an increase of M-H overlap pop-

ulation upon oxidation (see Results). For a system having a

M-H^δ+bond polarity (Scheme 1, right), the increase of formal

charge on the metal may have again a positive effect on the

orbital overlap as a consequence of the radius contraction, but

the energy separation between the frontier orbitals becomes

larger, decreasing the strength of the interaction.

CpWH₃(H₂PCH₂CH₂PH₂) +

Concerning the ionic component of the M-H bond, for a

metal-based HOMO (a typical situation for dⁿconfigurations

with n > 0), oxidation results in an increased effective positive

charge (or a decreased negative charge) for the metal, as well

as an increased “group electronegativity” for the M fragment

(viz., the metal with all its ligands except the hydride). The

latter effect leads to a greater draw of electron density from the

H ligand, reducing its effective negative charge or increasing

its effective positive charge. The two effects combine to yield

any possible result (strengthening or weakening of the ionic

interaction) for a M-H^δ-bond system, including an inversion

[CpWH₂(H₂PCH₂CH₂PH₂)]⁺f

[CpWH₃(H₂PCH₂CH₂PH₂)]⁺+ CpWH₂(H₂PCH₂CH₂PH₂)

(6)

T∆S factors, this corresponds to the free energy change ∆G,

allowing us to calculate a less positive potential (by 0.13 V)

(108) Ryan, O. B.; Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990,

112, 2618-2626.

(109) Chen, H.-W.; Jolly, W. L.; Kopf, J.; Lee, T. H. J. Am. Chem. Soc.

1979, 101, 2607-2610.

(106) Skagestad, V.; Tilset, M. J. Am. Chem. Soc. 1993, 115, 5077-

5083.

(107) Tilset, M.; Hamon, J.-R.; Hamon, P. J. Chem. Soc., Chem.

Commun. 1998, 765-766.

(110) Sweany, R. L. In Transition Metal Hydrides; Dedieu, A., Ed.;

VCH: 1992; pp 65-101.

(111) Bauschlicher, C. W., Jr.; Langhoff, S. R. In Transition Metal

Hydrides; Dedieu, A., Ed.; VCH: 1992; pp 103-126.

2220 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

Scheme 3

for the [CpWH₃(H₂PCH₂CH₂PH₂)]^0/+couple relative to that for

the [CpWH₂(H₂PCH₂CH₂PH₂)]^0/+couple, in accord with the

bond strengthening picture. This calculated potential shift refers,

of course, to gas-phase, isolated species and is not contaminated

by interactions with an electrochemical medium. It is to be

observed that eq 6 is nothing other than the difference between

eqs 4 and 5.

(b) Hydride Scrambling Process. As shown in the Results

section, the equivalence of the P and H hyperfine couplings in

the EPR spectrum of 1⁺clearly indicates that fluxional processes

are occurring, averaging both the H and P ligands in the system.

Due to the broadness of the EPR signal, a similar fluxional

process is not unequivocally identified for complexes 2⁺and

2⁺-d₃(see Figure 1). As shown by the X-ray investigation and

backed up by the DFT geometry optimization, the structure of

2⁺is pseudo-octahedral, with a significant twist toward a

pseudo-trigonal prismatic geometry. The structure of 1⁺can be

reasonably assumed to be identical to that of 2⁺. Thus, it is

very likely that the fluxional process involves the fast inter-

conversion of pseudo-octahedral and pseudo-trigonal prismatic

structures in solution, involving a pseudo-Bailar twist. The same

scrambling mechanism was previously proposed for the precur-

sor complexes 1 and 2.^81,96

(c) Stability of [Cp*MH₃(dppe)]⁺(M ) Mo, W). Stable

and unambiguous examples of paramagnetic transition metal

polyhydride complexes are exceedingly rare. Structurally char-

acterized examples are TaCl₂H₂L₄⁷and [WCl₂H₂(PMe₃)₄]⁺.⁸A

factor involved in the stability of these and other paramagnetic

polyhydride species may be the presence of π-donor ligands

such as Cl which are capable of establishing interactions with

the half-filled metal-based HOMO. The stabilizing effect of

halide for hydride substitution has been pointed out by Kochi

et al. in a study involving the lifetimes of various Cp₂WX₂

radical cations (X ) H, Cl).⁵⁸Complexes 1⁺and 2⁺, on the

other hand, do not have π-donor-stabilizing ligands; thus, other

reasons must be found for their relative stability.

Since one of the principal decomposition pathways for 17-

electron monohydride complexes is deprotonation by virtue of

the increased acidity relative to that of the 18-electron precur-

sors,¹⁰⁸an increased stability may be expected through the use

of donor ligands via a buffering effect of this acidity. At the

same time, donor ligands render the 18-electron precursor more

easily oxidizable, as experimentally verified for several systems,

such as the CpRe(PAr₃)₂H₂(Ar ) p-C₆H₄X; X ) H, Me, F,

OMe)⁶³and Cp*Ru(L₂)H (L₂) dppm, dppp, (PPh₃)₂, (PMe-

Ph₂)₂, (PMe₂Ph)₂, (PMe₃)₂)⁴³series. A qualitative correlation

between oxidation potential and stability is, indeed, typically

observed. The substitution of a more electron-rich metal for one

less so will also result in a similar effect on the oxidation

potential of the complex and a dampening of the hydride

acidity.^106,108However, the 18-electron hydride precursor is a

possible base for the deprotonation pathway of decomposition

of its own oxidation product, and its basicity is raised by the

same factors that lower the acidity of the radical cation. Thus,

the overall proton-transfer thermodynamics should remain

approximately unaffected. The stabilization of 17-electron

hydride complexes by donor ligands must be more a kinetic

than a thermodynamic phenomenon. A lowering of the kinetic

acidity as a result of steric protection has also been invoked to

explain the stability of complexes [Cp*FeH(dppe)]⁺and [Cp-

MoH(PMe₃)₃]⁺.^9,112

When proton transfer is kinetically slow, the starting 18-

electron hydride can be completely consumed by the oxidant,

and the decomposition mechanism by proton transfer can be

eliminated if sufficiently strong external bases are absent. Under

these conditions, the radical complex can still decompose by a

disproportionation mechanism.^57,75,112Steric protection, however,

should also slow a decomposition process by disproportionation,

since this is usually initiated by the coordination of a donor

molecule to the 17-electron complex. These simple arguments

rationalize well the observed stability of complexes 1⁺and 2⁺.

In fact, complex 1⁺does not decompose by disproportionation,

but rather by H₂reductive elimination (except in acetonitrile).

A discussion of the decomposition mechanism and a reason for

the faster decomposition of the Mo complex are presented in

the following sections.

(d) Decomposition Mechanisms for Paramagnetic Poly-

hydride Species: An Introduction. Mechanistic studies of the

decomposition of paramagnetic polyhydride systems are less

common than those of the corresponding monohydride systems.

Based on the available literature, the current knowledge may

be summarized as shown in Scheme 3. Deprotonation (by action

of either the residual unoxidized starting material or another

external base, paths a) leads to collapse to dimeric products

(112) Fettinger, J. C.; Kraatz, H.-B.; Poli, R.; Quadrelli, E. A.; Torralba,

R. C. Organometallics 1998, 17, 5767-5775.

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2221

Scheme 4

(path a.1) or to further oxidization (path a.2). It seems that the

former possibility is favored only for relatively unhindered

systems that lead to the formation of strong metal-metal

interactions, e.g., Cp₂WH₂,⁵⁸(C₅H₄SiMe₃)₂NbH₃,⁶⁸and [ReH₆-

(PMePh₂)₂]^-.⁶¹The alternative product of further oxidation (path

a.2) is stabilized by solvent coordination. In addition, under the

influence of the reduced electron density following oxidation,

one or more pairs of hydride ligands may collapse to dihydrogen

ligands and be replaced by additional solvent molecules,

producing a class of products of general formula [MH_n+1-2k(S)_k]⁺.

This path of decomposition has been proven or inferred (from

supporting protonation studies) for a few polyhydride com-

plexes, e.g., IrH₃(PMe₂Ph)₃,⁷¹Cp*RuH₃(PPh₃),⁷⁶and OsH₆-

(PPrⁱ₃)₂.¹¹³

this occurs only upon intervention of an electron-transfer-

catalyzed pathway upon further oxidation.⁶³There have been a

number of other reports dealing with the decomposition of

oxidation products of polyhydride species, but no detailed

mechanistic studies have been carried out, nor have specific

mechanistic proposals been advanced.^59,60,64Generally, these

lead only to the monocationic series of products. An exception

is complex MoH₄(PMe₂Ph)₄, whose reaction with AgBF₄in

MeCN leads to [MoH₂(PMe₂Ph)₄(MeCN)₂]²⁺as the only

product, although it has been suggested that Ag⁺may act as a

hydride abstracting agent rather than an oxidant in this reaction.⁶⁰

Finally, the possibility that oxidation induces a direct reduc-

tive elimination of H₂(path c) should also be considered. A

classical/nonclassical equilibrium for a 17-electron polyhydride

complex has been suggested for [Cp*RuH₃(PPh₃)]⁺and [OsH₆-

(PPrⁱ₃)₂]⁺,^76,113but a deprotonation (and not a H₂loss) has been

shown or suggested as the initial step for their decomposition.

On the other hand, oxidation of dialkyl complexes is proven to

favor alkyl-alkyl reductive elimination.^114-116

The disproportionation mechanism (paths b) is much less

common and has been unambiguously shown, to the best of

our knowledge, only for monohydride species.^57,75,112On the

basis of the scheme established for the monohydride species,

the disproportionation products MH_nand [MH_n(S)]²⁺should

+

(e) Decomposition Mechanisms of [Cp*MoH₃(dppe)]⁺in

THF and CH₂Cl₂. The decomposition of complex 1⁺following

the oxidation of 1 in THF or CH₂Cl₂(eq 3) represents the first

unambiguous demonstration of an oxidatively induced H₂

elimination from a polyhydride complex. The reverse process,

namely the oxidative addition of H₂to a stable paramagnetic

complex, had been established for the 15-electron complexes

MCl₂L₄(M ) Nb, Ta; L ) PMe₃or L₂) dmpe).⁷It should

also be mentioned that an H₂elimination was proposed to occur

from Cp₂NbH₂(generated in situ by H atom abstraction from

Cp₂NbH₃), to generate a poorly characterized 15-electron Cp₂-

Nb transient, which was observed only as a fleeting intermediate

by EPR spectroscopy.¹¹⁷

rearrange by proton transfer to afford a mixture of [MH_n+1

]

and [MH_n-1(S)]⁺. Donor solvents may again induce loss of H₂

as indicated in Scheme 3 (path b.1). An additional possibility,

however, is reductive elimination of H₂at the level of the

dicationic intermediate [MH_n(S)]²⁺(path b.2). This would lead

to an entirely new series of products that are characterized by

a double positive charge and by the same parity for the number

of hydride ligands as the starting material, i.e., complexes having

the general formula [MH_n-2k(S)_k+1]

²⁺. This latter pathway is

not available for the disproportionation of monohydride com-

plexes and thus represents an important chemical probe for the

decomposition mechanism. Care, however, should be excercised

because the [MH_n-2k(S)_k+1]

²⁺complexes may also be obtained

by path a.2 if the product [MH_n+1-2k(S)_k]⁺(rather than the

starting material or another external base) serves as proton

acceptor (B). Notably, the previously reported acidolysis of 1

in MeCN affords a mixture of [Cp*MoH₂(dppe)MeCN)]⁺and

[Cp*MoH(dppe)(MeCN)₂]²⁺, even when using less than 1 equiv

of acid.⁸¹Analogously, the acidolysis of MoH₄(PMe₂Ph)₄and

OsH₆(PPrⁱ₃)₂in MeCN with excess HBF₄affords [MoH₂(PMe₂-

Ph)₃(MeCN)₃]²⁺and [OsH₄(PPrⁱ₃)₂(MeCN)₂]²⁺, respectively.^60,113

In turn, the absence of the dicationic products cannot be taken

as evidence ruling out a disproportionation mechanism, because

path b.2 may be too slow relative to path b.1 to significantly

contribute to the generation of products.

Proposals for the intimate mechanism of this transformation

can be extrapolated from the well-studied diamagnetic coun-

terparts as shown in Scheme 4.¹¹⁸The EPR signal decay is faster

for 1⁺-d₃relative to that for 1⁺, yielding an inVerse isotope

effect of 0.50(3). This strongly suggests a ground-state classical

structure for complex 1⁺, namely [Cp*Mo(H)₃(dppe)]⁺. A

similar inverse isotope effect (0.4 ( 0.2) has been reported for

the H₂elimination from the classical dihydride complex [Ir-

(H)₂(PPh₃)₂(nbd)]⁺(nbd ) norbornadiene).¹¹⁹The elimination

of H₂from a dihydrogen complex, i.e., of type [Cp*MoH(H₂)-

(114) Pedersen, A.; Tilset, M. Organometallics 1993, 12, 56-64.

(115) Pedersen, A.; Tilset, M. Organometallics 1994, 13, 4887-4894.

(116) Fooladi, E.; Tilset, M. Inorg. Chem. 1997, 36, 6021-6027.

(117) Elson, I. H.; Kochi, J. K. J. Am. Chem. Soc. 1975, 97, 1262-

1264.

The decomposition mechanism of [CpReH₂(PAr₃)₂]⁺(Ar )

p-C₆H₄X; X ) H, Me, F, OMe) has been studied in detail, but

(113) Smith, K.-T.; Tilset, M.; Kuhlman, R.; Caulton, K. G. J. Am. Chem.

Soc. 1995, 117, 9473-9480.

(118) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789-805

and references therein.

2222 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

(dppe)]⁺(XI), should lead to a normal (i.e., g1) isotope

effect.¹²⁰A classical structure is shown by the X-ray analysis

for the W analogue 2⁺and is backed up by the DFT calculations

on the [CpWH₃(H₂PCH₂CH₂PH₂)]⁺model system. As exten-

sively documented for diamagnetic polyhydride systems, reduc-

tive elimination of H₂from a 5d metal should be thermody-

namically disfavored relative to that from a 4d metal.¹²¹It can

thus be imagined that X or XI is more easily accessible for Mo

than for W, rationalizing the much greater stability of complex

2⁺relative to that of 1⁺toward decomposition.

than a transition state (e.g., X). This pattern is well established

from the previous studies on paramagnetic monohydride com-

plexes.^57,75,112The deprotonation mechanism is not a viable one

if the oxidation is carried out stoichiometrically and if complex

1⁺is stable in MeCN on the time scale of the oxidation process,

leaving no residual 1 to deprotonate 1⁺. Furthermore, the

stronger donor MeCN will make the adduct XIII a weaker acid

relative to the same molecule having S ) THF or CH₂Cl₂. Thus,

a hypothetical deprotonation of XIII should be slower for S )

MeCN. Since the stability of XIII when S ) THF or CH₂Cl₂is

sufficiently long to allow the complete oxidation of 1 and the

decomposition of 1⁺by H₂elimination, it is extremely unlikely

that 1⁺tranfers a proton to 1 before the latter is completely

oxidized in MeCN.

The chemical nature of the products obtained from the rapid

and stoichiometric oxidation of 1 in MeCN, notably the

stereochemistry of the monohydride product, gives important

information. The deprotonation mechanism would require that

[Cp*MoH(dppe)(MeCN)₂]²⁺is obtained as a secondary product

by protonation of [Cp*MoH₂(dppe)(MeCN)]⁺(Scheme 3). The

protonation of [Cp*MoH₂(dppe)(MeCN)]⁺by HBF₄‚OEt₂,

however, was shown⁸¹to yield a kinetic mixture of two different

isomers of the dication, the thermodynamically less stable isomer

being favored over the more stable one by a 3:1 margin. A

rationalization of this result⁸¹was based on a competition of

the two inequivalent hydrides in a pseudo-trigonal prismatic

structure for the proton, as indicated in XIV.

Path a involves a direct, one-step solvent-induced reductive

elimination. This mechanism would be consistent with both the

inverse isotope effect and the solvent dependence. However, it

is difficult to rationalize, on the basis of this mechanism, the

mechanistic diversion in MeCN (vide infra). The other pathways

involve the intervention of the nonclassical intermediate XI in

a first step, which could either be rate-determining or a rapid

preequilibrium followed by a rate-determining replacement of

the dihydrogen ligand by a solvent molecule. In the second case,

the observed isotope effect on the decomposition rate would

correspond to the combination of an equilibrium isotope effect

for the first step and a kinetic isotope effect for the rate-

determining exchange step. Inverse isotope effects were similarly

interpreted for the reductive elimination of RH (via a R-H σ

complex) from Cp*Ir(PMe₃)(C₆H₁₁)(H) and from Cp₂W(CH₃)-

(H).^122,123The rate of decomposition should be solvent inde-

pendent either if the first step is rate-determining or if the

substitution of the H₂ligand by the solvent is dissociative (via

intermediate XII, path b in Scheme 4), whereas a solvent

dependence is expected for a rate-determining associative

exchange, either of the interchange type (I_a) or via a well-defined

19-electron solvent adduct XIII (path c). The faster decomposi-

tion process in CH₂Cl₂relative to that in THF (see Results)

suggests a rate-determining associative exchange (path c). On

the basis of the better donor capability of THF relative to that

of CH₂Cl₂, one would expect the opposite order of relative

reactivity (THF > CH₂Cl₂). However, since the CH₂Cl₂solvate

of 3 has a 19-electron configuration (see Results), we assume

that the acceleration effect for the decomposition in CH₂Cl₂is

due to the intervention of the second Cl atom in an interchange

associative mechanism, facilitating the expulsion of the H₂

ligand.

(f) Decomposition Mechanisms of [Cp*MoH₃(dppe)]⁺in

MeCN. The different nature of the final products obtained by

oxidation of 1 in MeCN (eq 2), together with the observation

that 3 is relatively stable in MeCN, unambiguously shows that

the decomposition of 1⁺cannot take place by reductive

elimination of H₂in MeCN as it does in THF and CH₂Cl₂. We

should, therefore, consider the other two more usual pathways

of proton transfer and disproportionation (paths a and b in

Scheme 3). The most reasonable hypothesis for this mechanistic

diversion is a greater susceptibility of the solvent adduct XIII

to be oxidized by unsolvated 1⁺, which results from the greater

donating power of the MeCN ligand. Therefore, it is more likely

that XIII is an intermediate along the reaction coordinate rather

The observed stereochemistry could, in principle, be recon-

ciled with the proton-transfer mechanism by invoking a 100%

selectivity (discriminating ability) of the proton donor 1⁺toward

one of the two inequivalent hydride positions, while the stronger

acid HBF₄‚OEt₂would not discriminate between the two

inequivalent positions. It seems, however, more reasonable to

assume that the decomposition follows the disproportionation

pathway as proposed in Scheme 5. The formation of a single

isomer of the monohydride dicationic product may then be easily

rationalized by assuming that intermediate XV is stereochemi-

cally fluxional (dihydrogen/hydride complexes usually are),

leading to a single most stable stereoisomer, and that the

replacement of H₂with a molecule of MeCN occurs stereose-

lectively. In case intermediate XV is stereochemically rigid, the

observed result may still be rationalized if intermediate XIII is

fluxional (17-electron complexes usually are), leading to a single

isomer at this level.

A difference between the relative energies of the classical

and nonclassical forms for the oxidized trihydride complexes

1⁺and 2⁺rationalizes well the different stability observed for

the cations in MeCN. We propose that MeCN coordination can

only take place for the nonclassical tautomer, which is only

thermally accessible for the Mo system. Coordination to the

classical tautomer to afford intermediate XVI is electronically

(119) Howarth, O. W.; McAteer, C. H.; Moore, P.; Morris, G. E. J. Chem.

Soc., Dalton Trans. 1984, 1171-1180.

(120) Hauger, B. E.; Gusev, D.; Caulton, K. G. J. Am. Chem. Soc. 1994,

116, 208-214 and references therein.

(121) Crabtree, R. H. In ComprehensiVe Coordination Chemistry;

Wilkinson, G., Gillard, R. D., McCleverty, J. A., Eds.; Pergamon Press:

Oxford, 1987; Vol. II, pp 689-714.

(122) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am. Chem.

Soc. 1986, 108, 1537-1550.

(123) Bullock, R. M.; Headford, C. E. L.; Hennessey, K. M.; Kegley, S.

E.; Norton, J. R. J. Am. Chem. Soc. 1989, 111, 3897-3908.

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2223

Scheme 5

to a linear shift of the peak potential (E_p) relative to log(V) and

log(c) in the kinetic zone (low scan rates), according to eq 7 (1

electron is involved in the electrochemical oxidation step).

k

_dispc

E_p) E° + 0.071 - 0.019 log

(CV) (7)

V

The peak potential, E_p, varies with a slope of 19 mV per

log(V) unit and -19 mV per log(c) unit.¹²⁵On the other hand,

a mechanism involving a rate-determining proton transfer to

the starting material (Scheme 6b) leads to a variation of E_p

described by eq 8 (T ) 20 °C).¹²⁶

k

_deprotc

E_p) E° + 0.058 - 0.029 log

(CV) (8)

V

Shifts of E_pagainst log(V) and log(c) now have slopes of 29

and -29 mV, respectively, per log unit. Finally, a hypothetical

H₂reductive elimination mechanism (Scheme 6c), which is

already established to take place in THF and CH₂Cl₂but not in

MeCN, yields a straight line for E_pvs log(V) with a slope of

+29 mV per log unit, but the peak potential is independent of

the concentration.¹²⁷An alternative disproportionation mecha-

nism which involves a rate-determining solvent coordination,

followed by a rapid electron-transfer step (DISP1), would behave

as a simple EC process (Scheme 6c).

Scheme 6

The results are qualitatively the same for the TLCV technique,

except that the slopes are now (29 mV for the DISP2

mechanism (eq 9),¹²⁸(58 mV for the proton-transfer mecha-

nism (eq 10)^129,130and 58 mV (vs log(V)) and 0 (vs log(c)) for

the EC mechanism.^129,131

k

_dispc

E_p) E° + 0.055 - 0.029 log

(TLCV) (9)

(TLCV) (10)

V

k

_deprotc

E_p) E° + 0.075 - 0.058 log

V

The H₂reductive elimination pathway can immediately be

excluded for the decomposition in MeCN, since a concentration

dependence is observed both in CV and in TLCV. Let us

examine the experimental variation of the peak potentials vs

log(V). In the kinetic region, the slope of the straight line is

closer to 19 mV per log(V) unit in the CV (see Table 3),

consistent with a DISP2 mechanism (eq 7). This is confirmed

by the concentration dependence at constant scan rate (see

Results). The TLCV data, however, yield a slope close to 58

mV in the E_pavs log(V) plot (see Table 3), which is consistent

with the proton-transfer mechanism rather than with the DISP2

mechanism.¹³²

This apparently contrasting behavior can be rationalized by

invoking the occurrence of both the disproportionation (DISP2)

and the proton-transfer mechanisms in MeCN. Let us compare

the results in CV with those in TLCV in the case of a DISP2

mechanism. The intercept, log(V_i), of the E_pa) f(log(V)) curve

allowed but may experience a greater barrier because of a greater

steric congestion. If coordination to the classical form were

allowed, there would be no apparent reason for the much greater

stability of 2⁺relative to that of 1⁺in MeCN because Mo and

W have very similar atomic radii.

The proposition that disproportionation is the favored mech-

anism for the decomposition of 1⁺in MeCN is confirmed by

the CV and TLCV studies (see Results). Because of the fast

decomposition rate in MeCN, classical kinetic studies could not

be carried out by EPR spectroscopy. On the other hand, the

occurrence of the chemical decomposition process has a

noticeable effect on the electron-transfer parameters, as measur-

able by a variety of voltammetric techniques, as already shown

for other systems.¹²⁴

(126) Andrieux, C. P.; Nadjo, L.; Save´ant, J. M. J. Electroanal. Chem.

1970, 26, 147-186.

For a CV experiment, according to the established theory, a

mechanism involving a solvent-assisted disproportionation

(127) Nicholson, R. S.; Shain, J. Anal. Chem. 1964, 36, 706.

(128) Laviron, E. J. Electroanal. Chem. 1978, 87, 31-37.

(129) Laviron, E. J. Electroanal. Chem. 1972, 39, 1-23.

(130) Laviron, E. J. Electroanal. Chem. 1972, 34, 463-473.

(131) Laviron, E. J. Electroanal. Chem. 1972, 35, 333-342.

(132) We have taken into account only the points obtained for V < 10^-2

V s^-1for the determination of this slope. Above this value, the current has

a diffusion component (the variation of i_pvs log(V) has a slope smaller

than 1).

(Scheme 6a), whereby solvent coordination is rapid and the

125

disproportionation step is rate determining (DISP2),

leads

(124) Hammerich, O.; Parker, V. D. In Organic Electrochemistry; Lund,

H., Baizer, M. M., Eds.; Marcel Dekker: New York, 1991; pp 121-206

and references therein.

(125) Mastragostino, M.; Nadjo, L.; Save´ant, J. M. Electrochim. Acta

1968, 13, 721-749.

2224 J. Am. Chem. Soc., Vol. 121, No. 10, 1999

Pleune et al.

with the E ) E° axis leads to the chemical rate constant. When

E_pa) E°, eqs 7 and 9 lead to the following:

In CV

log(k_dispc) ) 3.55 + log(V_CV

)

(11)

(12)

i

In TLCV log(k_dispc) ) 1.90 + log(V_TLCV

)

i

The difference between eqs 11 and 12 leads to the relationship

(DISP2 mechanism): log(V_TLCV) ) 1.65 + log(V_CV) (13)

i

Equation 13 shows that, for a pure DISP2 mechanism, the

intercept in TLCV is shifted by 1.65 log(V) units with respect

to that in CV. Analogous considerations for the pure depro-

tonation mechanism (eqs 14 and 15) show that the intercept in

TLCV is shifted by 0.71 log(V) units with respect to that in CV

(eq 16).

In CV

log(k_deprotc) ) 2.00 + log(V_CV

)

(14)

(15)

i

In TLCV log(k_deprotc) ) 1.29 + log(V_TLCV

)

i

(proton-transfer mechanism):

log(V_TLCV) ) 0.71 + log(V_CVi) (16)

i

When the rate constants of the DISP2 reaction and the proton

transfer are not too different, it is therefore possible to have a

limitation by one mechanism in CV and by the other in TLCV.

Experimentally, the DISP2 mechanism is rate-limiting in CV

(slope of 19 mV) and the proton transfer is rate-limiting in

TLCV (slope of 58 mV) (see Figure 3).

Figure 9. Plot of log(kx) vs log(x) (b, x ) V in V s^-1; O, x ) c in mol

L^-1) for (a) CV data and (b) TLCV data (see text).

knowledge, is that of complex [CpMoH(PMe₃)₃]⁺, for which

k

_disp/k_deprot) 0.17(2) in THF.¹¹²

Experiments carried out at different concentrations give the

intercepts indicated in Table 3. At each concentration, a value

for log(k_disp) may be calculated from the CV data and eq 11.

Correspondingly, a value for log(k_deprot) may be calculated for

the TLCV data and eq 15. These values are also shown in Table

3. The results of the E vs log(V) and E vs log(c) experiments

allow us to obtain average values of k_dispand k_deprotaccording

to the following procedure. The representation of log(kV) as a

function of log(V) must give a straight line of slope 1 and

intercept equal to log(k). In the same fashion, the representation

of log(kc) as a function of log(c) must give a straight line of

slope 1 and intercept equal to log(k). We have therefore plotted

the log(kV) and log(kc) data together against log(x) (x ) V or c,

respectively) (see Figure 9). The linear fit of these data with an

imposed slope of 1 gave log(k) as the intercept. The CV data

(Figure 9a) yield a value of log(k_disp) ) 3.60(1), or k_disp) 3.98-

(9) × 10³s^-1M^-1, whereas the TLCV data (Figure 9b) yield

log(k_deprot) ) 2.44(4), or k_deprot) 2.8(2) × 10²s^-1M^-1. The

simulated voltammograms calculated with the DIGISIM 2.1

software,⁷⁸for a mechanism where both DISP2 reaction (k_disp

) 4000) and proton transfer (k_deprot) 280) are involved, lead

to curves that fit well the experimental data, as shown in Figure

3 for one particular concentration.

The value of k_disp, according to Scheme 5, would be equal to

(k₁k₂/k_-1)[S] and is therefore dependent on the nature of the

solvent as previously discussed.⁷⁵The ratio of the two second-

order rate constants is k_disp/k_deprot) 14(1). This ratio indicates

that, when the concentrations of 1 and 1⁺(or rather, the sum of

1⁺and 1(MeCN)⁺in equilibrium) are equivalent, complex 1⁺

is ca. 14 times more likely to remove an electron from

1(MeCN)⁺than to donate a proton to 1. It is rare that both

disproportionation and proton-transfer mechanisms may be

observed at the same time for a 17-electron hydride complex.

The only other previously reported case, to the best of our

The slow addition of a chemical oxidant (ferrocenium) to a

MeCN solution of 1 supposedly simulates the conditions of the

thin-layer electrochemical experiment. Thus, a greater proportion

of 1⁺should decompose by the proton-transfer pathway, yielding

more dihydride product. Indeed, this is experimentally observed

(see Results). The monohydride product [Cp*MoH(dppe)-

(MeCN)₂]²⁺is still obtained as a single isomer. This observation

rules out a proton transfer to the dihydride product, [Cp*MoH₂-

(dppe)(MeCN)]⁺(vide supra). It is logical to expect, however,

that the protons released by 1⁺are captured by 1, which is more

basic and more abundant, to yield [Cp*MoH₂(dppe)(MeCN)]⁺,

rather than by the latter to yield the monohydride complex. Thus,

the monohydride complex derives only from the disproportion-

ation pathway, whereas the dihydride product is at least partially

produced by the deprotonation pathway under conditions of slow

addition of the ferrocenium oxidant.

(g) Structure and Stability of Compound 3. The solution

spectroscopic, conductivity, and electrochemical studies defini-

tively prove that a solvent molecule is coordinated to 3 in

-

solution, giving the ionic formulation [Cp*MoH(dppe)(S)]⁺PF₆.

Thus, the complex belongs to the growing class of four-legged

piano stool derivatives of Mo(III).⁸⁹However, the solvent

molecule must be very loosely coordinated because crystalliza-

tion, even from MeCN, affords the solvent-free compound 3,

presumably containing a coordinated PF₆^-. The EPR spectro-

scopic properties of the CH₂Cl₂adduct at low temperature

(conditions in which the chemical exchange with free solvent

is slow on the EPR time scale) show that the CH₂Cl₂ligand

adopts a chelating coordination mode, resulting in a formal 19-

electron count for the complex. The facile coordination of donor

solvents to 17-electron complexes in an equilibrium process with

the 19-electron adduct has been frequently invoked to rationalize

photochemical, electrochemical, and mechanistic observa-

Half-Sandwich Mo(V) and W(V) Polyhydride Complexes

J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2225

tions^133,134(including those discussed in this contribution, see

section f above). The nature of the CH₂Cl₂adduct of 3 provides

a clear, direct demonstration of the accessibility of 19-electron

complexes.

been measured by a combination of CV and TLCV studies. For

the first time, a paramagnetic (poly)hydride system has been

shown to decompose by one or more of three different

mechanisms depending on conditions. Each of the three

decomposition pathways has been quantitatively assessed.

A final remark concerns the unexpected stability of 3 in all

solvents, especially MeCN. As discussed in section d, cationic

paramagnetic monohydride complexes are usually unstable,

decomposing preferentially by either proton-tranfer or dispro-

portionation mechanisms to afford more stable diamagnetic

products. Since 3 is generated under conditions in which a base

is not present, its decomposition by proton transfer cannot take

place, but the disproportionation route is still viable. As we have

argued in section f, even the disproportionation mechanism can

be blocked when the size of the coordination sphere does not

permit coordination of a solvent molecule and production of a

more easily reduceable 19-electron adduct (rationalizing the

stability of 2⁺). Complex [Cp*MoH(dppe)(MeCN)]⁺, on the

other hand, should be no more encumbered than complex

[Cp*MoH(H₂)(dppe)]⁺, which disproportionates rapidly in

MeCN. In addition, the likelihood of further MeCN coordination

to [Cp*MoH(dppe)(MeCN)]⁺is strongly suggested by the 19-

electron configuration of [Cp*MoH(dppe)(η²-Cl₂CH₂)]⁺. De-

composition of 3 in MeCN does eventually take place to afford

[Cp*MoH(dppe)(MeCN)₂]²⁺(the thermodynamically more stable

isomer), [Cp*MoH₂(dppe)(MeCN)]⁺, and other yet uncharac-

terized products. This decomposition process will be the subject

of further studies.

The large difference in stability between the W (stable) and

Mo (rapid decomposition) complexes in MeCN suggests that a

solvent molecule easily coordinates to the Mo center in a

thermally accessible, relatively open nonclassical structure of

1⁺but does not coordinate to the W center in the more crowded

classical structure of 2⁺.

Complex 2⁺adopts a novel and unusual geometry, which

can be described as intermediate between octahedral and trigonal

prismatic (when considering the Cp* as occupying a single

coordination site) and is reproduced by DFT calculations on

the [CpWH₃(PH₂CH₂CH₂PH₂)]⁺model system. This result

represents circumstantial evidence in favor of an exchange

mechanism which involves the interconversion of these two

limiting structures, as also previously proposed for the neutral

analogue.^81,93,96

While a M-H bond weakening upon oxidation has previously

been reported for other systems,^72,106,107oxidation results in the

strengthening of the metal-hydrogen bonds for both compounds

1 and 2 (and for the trideuteride analogues). This is experi-

mentally shown by IR studies and is backed up by density

functional calculations of bond lengths, BDEs, and M-H

stretching frequencies. On the basis of first principles, we argue

that compounds having a M-H^δ-bond polarity should always

lead to M-H bond strengthening upon oxidation when the metal

is electron-rich and the bond strength is dominated by the

covalent component.

Conclusions

The studies reported in this contribution have raised a number

of interesting issues.

The product of H₂elimination from 1⁺PF₆^-, compound 3,

has a remarkable stability for a relatively unhindered paramag-

netic hydride complex. This compound adds a solvent molecule

when dissolved in MeCN, THF, or CH₂Cl₂. The latter is

unambiguously shown by EPR spectroscopy to adopt a chelating

coordination mode, yielding a 19-electron cation, [Cp*MoH-

(dppe)(η²-Cl₂CH₂)]⁺. All these adducts, however, readily lose

the solvent molecule in the solid state, giving back 3.

The sterically encumbering Cp* and dppe ligands in com-

plexes 1⁺and 2⁺reduce the kinetic acidity of the hydride ligands

and disfavor the addition of donor solvents, stabilizing the

complexes relative to the deprotonation and disproportionation

pathways of decomposition, respectively. However, polyhydride

complexes may decompose by yet another decomposition

pathway, namely reductive elimination of H₂. The greater

electron richness of W relative to that of Mo disfavors this

decomposition for complex 2⁺, which is therefore stable and

can be isolated. Its X-ray structure shows that it is a classical

trihydride complex. Conversely, complex 1⁺decomposes readily

by what constitutes the first unambiguous example of oxidatively

induced H₂reductive elimination from a polyhydride complex.

The inVerse isotope effect for the H₂elimination from 1⁺

suggests that this complex, like 2⁺, adopts a classical structure.

The solvent dependence on the rate of decomposition suggests

a rate-determining associative mechanism for the replacement

of H₂by a solvent molecule.

Acknowledgment. We are grateful to the DOE (Grant

DEFG059ER14230) and to the Re´gion Bourgogne for support

of this work. D.M. thanks the Re´gion Bourgogne for a

postdoctoral fellowship. We thank Mrs. Raveau-Fouquet for

technical assistance.

Supporting Information Available: For the X-ray structures

of compounds [Cp*WH₃(dppe)][PF₆] and [Cp*WH₄(dppe)]-

[PF₆], tables of crystal data and refinement parameters, fractional

atomic coordinates, bond distances and angles, anisotropic

displacement parameters, and H-atom coordinates, and for the

DFT calculations, tables of optimized geometric parameters for

systems III-IX (PDF). Additional crystallographic data are

available in CIF format. This material is available free of charge

via the Internet at http://pubs.acs.org.

In MeCN, the decomposition of complex 1⁺proceeds faster

by a combination of disproportionation and (when the neutral

precursor 1 is available) deprotonation. The relative rate of

disproportionation and deprotonation [k_disp/k_deprot) 14(1)] has

(133) Stiegman, A. E.; Tyler, D. R. Comments Inorg. Chem. 1986, 5,

215-245 and references therein.

(134) Tyler, D. R. Acc. Chem. Res. 1991, 24, 325-331 and references

therein.

JA9834881

Article Doi

Stable paramagnetic half-sandwich Mo(V) and W(V) polyhydride complexes. Structural, spectroscopic, electrochemical, theoretical, and decomposition mechanism studies of [CP*MH3(dppe)]+ (M = Mo, W)

DOI: 10.1021/ja9834881

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja9834881

Full text of DOI:10.1021/ja9834881

Products guided by the article

R&D Labs maybe for 221341-72-6

Relevant to this article

Hot Product