Relative hydride, proton, and hydrogen atom transfer abilities of [HM(diphosphine)2]PF6 complexes (M = Pt, Ni)

11432

J. Am. Chem. Soc. 1999, 121, 11432-11447

Relative Hydride, Proton, and Hydrogen Atom Transfer Abilities of

[HM(diphosphine)₂]PF₆Complexes (M ) Pt, Ni)

Douglas E. Berning,^†Bruce C. Noll,^‡and Daniel L. DuBois*^,†

Contribution from the National Renewable Energy Laboratory, 1617 Cole BouleVard,

Golden, Colorado 80401, and Department of Chemistry and Biochemistry, UniVersity of Colorado,

Boulder, Colorado 80309

ReceiVed June 7, 1999

Abstract: A series of [M(diphosphine)₂]X₂, [HM(diphosphine)₂]X, and M(diphosphine)₂complexes have been

prepared for the purpose of determining the relative thermodynamic hydricities of the [HM(diphosphine)₂]X

complexes (M ) Ni, Pt; X ) BF₄, PF₆; diphosphine ) bis(diphenylphosphino)ethane (dppe), bis(diethyl-

phosphino)ethane (depe), bis(dimethylphosphino)ethane (dmpe), bis(dimethylphosphino)propane (dmpp)).

Measurements of the half-wave potentials (E_1/2) for the M(II) and M(0) complexes and pK_ameasurements for

the metal hydride complexes have been used in a thermochemical cycle to obtain quantitative thermodynamic

information on the relative hydride donor abilities of the metal-hydride complexes. The hydride donor strengths

vary by 23 kcal/mol and are influenced by the metal, the ligand substituents, and the size of the chelate bite

of the diphosphine ligand. The best hydride donor of the complexes prepared is [HPt(dmpe)₂](PF₆), a third-

row transition metal with basic substituents and a diphosphine ligand with a small chelate bite. The best hydride

acceptors have the opposite characteristics. X-ray diffraction studies were carried out on eight complexes:

[Ni(dmpe)₂](BF₄)₂, [Ni(depe)₂](BF₄)₂, [Ni(dmpp)₂](BF₄)₂, [Pt(dmpp)₂](PF₆)₂, [Ni(dmpe)₂(CH₃CN)](BF₄)₂,

[Ni(dmpp)₂(CH₃CN)](BF₄)₂, Ni(dmpp)₂, and Pt(dmpp)₂. The cations [Ni(dmpp)₂]²⁺and [Pt(dmpp)₂]²⁺exhibit

significant tetrahedral distortions from a square-planar geometry arising from the larger chelate bite of dmpp

compared to that of dmpe. This tetrahedral distortion produces a decrease in the energy of the lowest unoccupied

molecular orbital of the [M(dmpp)₂]²⁺complexes, stabilizes the +1 oxidation state, and makes the

[HM(dmpp)₂]⁺complexes poorer hydride donors than their dmpe analogues. Another interesting structural

feature is the shortening of the M-P bond upon reduction from M(II) to M(0).

Introduction

in predicting quantitatively which complexes will be protonated

under a given set of conditions. Whether the reactant in a

catalytic cycle is a metal hydride or its deprotonated analogue

can have important chemical consequences.³The free energy

scale resulting from pK_ameasurements can be significantly

extended by measuring the enthalpies of protonation reactions

and making the assumption that the entropy corrections are

constant for all complexes.⁴

Transition-metal hydride complexes are intermediates in a

variety of important stoichiometric and catalytic reactions,¹and

a knowledge of the thermodynamics associated with the

formation and cleavage of the M-H bond is important to

understanding known reactions and designing new ones. The

M-H bond can be cleaved in three different ways, as shown

in reactions 1-3. These three reactions may be regarded as half-

Various approaches for measuring the free energies or

enthalpies of reaction 2 as well as kinetic studies of hydrogen

atom transfer reactions have been reported.^5,6Of particular

L_nMH⁺f L_nM + H⁺

L_nMH⁺f L_nΜ⁺+ Η^•

L_nMH⁺f LnM²⁺+ H^-

(1)

(2)

(3)

(2) (a) Kristja´nsdo´ttir, S. S.; Norton, J. R. In Transition Metal Hy-

drides: Recent AdVances in Theory and Experiment; Dedieu, A., Ed.;

VCH: New York, 1991; pp 309-359. (b) Jordan, R. F.; Norton, J. R. J.

Am. Chem. Soc. 1982, 104, 1255-1263. (c) Moore, E. J.; Sullivan, J. M.:

Norton, J. R. J. Am. Chem. Soc. 1986, 108, 2257-2263. (d) Kristja´nsdo´ttir,

S. S.; Moody, A. E.; Weberg, R. T.; Norton, J. R. Organometallics 1988,

7, 1983-1987. (e) Jessop, P. G.; Morris, R. H. Coord. Chem. ReV. 1992,

121, 155. (f) Bullock, R. M.; Song, J.-S.; Szalda, D. J. Organometallics

1996, 15, 2504-2516. (g) Davies, S. C.; Henderson, R. A.; Hughes, D. L.;

Oglieve, K. E. J. Chem. Soc., Dalton Trans. 1998, 425-431.

(3) (a) Kollar, K.; Sandor, P.; Szalontai, G.; Heil, B. J. Organomet. Chem.

1990, 393, 153. (b) Takahashi, H.; Achiwa, K. Chem. Lett. 1987, 1921-

1922. (c) Chao, T.-H.; Espenson, J. H. J. Am. Chem. Soc. 1978, 100, 129-

133.

(4) (a) Busch, R. C.; Angelici, R. J. Inorg. Chem. 1988, 27, 681-686.

(b) Sowa, J. R., Jr.; Zanotti, V.; Facchin, G.; Angelici, R. J. J. Am. Chem.

Soc. 1991, 113, 9185-9192. (c) Angelici, R. J. Acc. Chem. Res. 1995, 28,

51-60.

(5) (a) Simo˜es, J. A. M.; Beauchamp, J. L. Chem. ReV. 1990, 90, 629-

688. (b) Kiss, G.; Zhang, K.; Mukerjee, S. L.; Hoff, C. D. J. Am. Chem.

Soc. 1990, 112, 5657-5658. (c) Schock, L. E.; Marks, T. J. J. Am. Chem.

Soc. 1988, 119, 7701-7715. (d) Wei, M.; Wayland, B. B. Organometallics

1996, 15, 4681-4683.

reactions analogous to reduction half-reactions in electrochem-

istry. Just as reduction potentials can be used to predict the

thermodynamic driving force for an electron-transfer reaction,

free energy measurements for half-reactions 1-3 can be used

to determine the relative ability of metal hydrides to transfer

H⁺, H^•, or H^-in solution.

Reaction 1 is a simple deprotonation reaction, and measure-

ments of the thermodynamics and kinetic acidity of a number

of transition-metal hydrides have been reported.²The free energy

scale resulting from the thermodynamic measurements is useful

^†National Renewable Energy Laboratory.

^‡University of Colorado.

(1) Transition Metal Hydrides: Recent AdVances in Theory and Experi-

ment; Dedieu, A., Ed.; VCH: New York, 1991.

Published on Web 11/30/1999

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11433

relevance to this work is the measurement of the relative free

energy for the homolytic transfer of a hydrogen atom in solution,

reaction 2, by means of the thermochemical cycle shown in

Scheme 1, reactions 4-7.⁷The free energy change associated

licity, except for some correlation between the nucleophilic

11

character and the position in the Periodic Table.”

To our

knowledge, there has been only one previous report of ther-

modynamic measurements for this reaction for transition-metal

hydrides. Sarker and Bruno have reported free energy measure-

ments for hydride transfer between CpM(CO)₂(L)H complexes

(where M ) Mo, W and L ) CO, PR₃) and triphenylcarbonium

ion derivatives in acetonitrile.¹²The kinetic hydricities of various

metal hydrides have been explored in a few studies. Transition-

metal hydrides have been reacted with substrates such as

alcohols, alkyl halides, acyl halides, aldehydes, ketones, and

the trityl cation or its derivatives.¹³Although the mechanistic

pathways appear to vary depending on the metal hydride and

substrate chosen for study, some general conclusions can be

drawn. For isostructural complexes, those containing second-

and third-row transition metals are better hydride donors than

those containing first-row transition metals, and complexes with

electron-donating ligands are better hydride donors than those

with electron-withdrawing ligands. Also, early-transition-metal

hydrides are generally more hydridic than late-transition-metal

hydrides. However, as pointed out by Darensbourg and co-

workers, using a different substrate as a probe of hydricity can

lead to significantly different ordering of only moderately

structurally diverse series of compounds.^13cThis is not surprising

if the mechanisms for these reactions are dependent on both

the substrate and metal hydride chosen. A thermodynamic scale

of hydricity should not be subject to this problem.

Scheme 1

L_nMH⁺f L_nM + H⁺∆G° ) 1.37(pK_a)

(4)

(5)

(6)

L_nM f L_nM⁺+ e^-

H⁺+ e^-f H^•

∆G° ) 23.06E°(+1/0)

∆G° ) -23.06E°(H⁺/H^•)

L_nMH⁺f L_nM⁺+ H^•∆G° ) 1.37(pK_a) +

23.06E° + 53.6 (7)

with reaction 4 can be expressed in terms of pK_avalues

measured in a given solvent as discussed above. The free energy

associated with reaction 5 is measured by the reversible one-

electron oxidation potential of M(0). Reaction 6 is the reduction

of H⁺to H^•in the solvent of choice. The free energy associated

with reaction 6 in acetonitrile has been estimated using a series

of approximations,^7aand sources of potential errors have been

discussed in the literature.^2a,6c,8More recent work has simply

treated the free energy of this reaction as an adjustable parameter

that can be determined by comparison with the enthalpy of a

standard reaction and making a correction for entropy.^7cThe

result is a free energy scale that should provide a quantitative

method for determining the relative hydrogen atom donor

potential for a large range of compounds in solution. This scale

is also useful in estimating bond dissociation enthalpies, which

has been its major application. A similar thermochemical cycle

has been used by Wang and Angelici to estimate bond

dissociation enthalpies in solution for a large number of

transition-metal hydrides.⁹In their work, enthalpies were

measured for reaction 4 and free energies for reaction 5, and

the assumption was made that the entropy change for reaction

5 is negligible. Although less rigorous, this method allows

compounds to be included for which it is not possible to measure

pK_avalues directly.

The relative free energy for hydride transfer in solution can

be obtained using the thermochemical cycle shown in Scheme

2, reactions 8-11. This cycle is analogous to that used above

Scheme 2

L_nMH⁺f L_nM + H⁺

L_nM f L_nM²⁺+ 2e^-

H⁺+ 2e^-f H^-

∆G° ) 1.37(pK_a)

(8)

(9)

∆G° ) 46.1E°(II/0)

∆G° ) -46.1E°(H⁺/H^-)

(10)

L_nMH⁺f L_nM²⁺+ H^-∆G° ) 1.37(pK_a) +

Reaction 3 is the heterolytic cleavage of the M-H bond to

produce M⁺and H^-. The hydride transfer reaction for transition-

metal hydrides is analogous to hydride transfer reactions of

NADH, which are important in respiration. Hydride transfer

equilibria for NADH analogues have been measured in a number

of different solvents.¹⁰The lack of similar data for transition-

metal hydrides has led Dedieu to comment, “Paradoxically,

much more is known about the acidic behavior of transition

metal hydrides than about their hydridic behavior. In particular,

the factors controlling the hydricity are far from being well-

understood. There is currently no simple and unambiguous

method either for identifying or for measuring the nucleophi-

46.1[E°(II/0)] + 79.6 (11)

(Scheme 1) for obtaining homolytic bond dissociation energies.

The first reaction, (8), is again the measurement of the pK_aof

the hydride. The second reaction, (9), is a two-electron oxidation

of a M(0) complex instead of the one-electron oxidation used

for obtaining homolytic bond dissociation energies. The free

energy associated with this reaction may be obtained by

measuring two reversible one-electron oxidations (the M/M⁺

and M⁺/M²⁺couples) or a single reversible two-electron

oxidation (the M/M²⁺couple). The third reaction in this

(11) Dedieu, A. In Transition Metal Hydrides: Recent AdVances in

Theory and Experiment; Dedieu, A., Ed.; VCH: New York, 1991; pp 381-

387.

(6) (a) Bullock, R. M.; Samsel, E. G. J. Am. Chem. Soc. 1990, 112,

6886-6898. (b) Eisenberg, D. C.; Lawrie, C. J. C.; Moody, A. E.; Norton,

J. R. J. Am. Chem. Soc. 1991, 113, 4888-4895. (c) Eisenberg, D. C.; Norton,

J. R. Isr. J. Chem. 1991, 31, 55-66.

(7) (a) Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1989, 111, 6711-

6717. (b) Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990, 112, 2843. (c)

Parker, V. D.; Handoo, K. L.; Roness, F.; Tilset, M. J. Am. Chem. Soc.

1991, 113, 7493-7498. (d) Wayner, D. D. M.; Parker, V. D. Acc. Chem.

Res. 1993, 26, 287-294.

(8) Bordwell, F. G.; Harrelson, J. A., Jr.; Satish, A. V. J. Org. Chem.

1989, 54, 3101-3105.

(9) Wang, D.; Angelici, R. J. J. Am. Chem. Soc. 1996, 118, 935-942.

(10) (a) Cheng, J.-P.; Lu, Y.; Zhu, X.; Mu, L. J. Org. Chem. 1998, 63,

6108-6114. (b) Anne, A.; Moiroux, J. J. Org. Chem. 1990, 55, 4608-

4614. (c) Klippenstein, J.; Arya, P.; Wayner, D. D. M. J. Org. Chem. 1991,

56, 6, 6736-6737. (d) Kreevoy, M. M.; Ostovic, D.; Lee, I.-S. H.; Binder,

D. A.; King, G. W. J. Am. Chem. Soc, 1988, 110, 524-530.

(12) Sarker, N.; Bruno, J. W. J. Am. Chem. Soc. 1999, 121, 2174-2180.

(13) (a) Labinger, J. A. In Transition Metal Hydrides: Recent AdVances

in Theory and Experiment; Dedieu, A., Ed.; VCH: New York, 1991; pp

361-379. (b) Labinger, J. A.; Komadina, K. H. J. Organomet. Chem. 1978,

155, C25-C28. (c) Kao, S. C.; Spillett, C. T.; Ash, C.; Lusk, R.; Park, Y.

K.; Darensbourg, M. Y. Organometallics 1985, 4, 83-91. (d) Kao, S. C.;

Gaus, P. L.; Youngdahl, K.; Darensbourg, M. Y. Organometallics 1984, 3,

1601-1603. (e) Gaus, P. L.; Kao, S. C.; Youngdahl, K.; Darensbourg, M.

Y. J. Am. Chem. Soc. 1985, 107, 2428-2434. (f) Kinney, R. J.; Jones, W.

D.; Bergman, R. G. J. Am. Chem. Soc. 1978, 100, 7902-7915. (g) Martin,

B. D.; Warner, K. E.; Norton, J. R. J. Am. Chem. Soc. 1986, 108, 33-39.

(h) Hembre, R. T.; McQueen, J. S. Angew. Chem., Int. Ed. Engl. 1997, 36,

65-67. (i) Hembre, R. T.; McQueen, S. J. Am. Chem. Soc. 1994, 116,

2141-2142. (j) Cheng, T.-Y.; Brunschwig, B. S.; Bullock, R. M. J. Am.

Chem. Soc. 1998, 120, 13121-13137.

11434 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

thermochemical cycle, (10), is the reduction of a proton in

solution by two electrons to form the hydride ion. The free

energy for reaction 10 in acetonitrile is assigned a value of 79.6

kcal/mol. This value is obtained by adding the free energy of

the reduction of a proton to a hydrogen atom in acetonitrile

(53.6 kcal/mol in reaction 6) to that for the reduction of a

hydrogen atom to H^-in acetonitrile. The reduction potential of

a hydrogen atom in acetonitrile has been assigned a value of

-1.128 V versus the ferrocene/ferrocenium couple,^7d,14which

leads to a ∆G°(H^•/H^-) value of 26.0 kcal/mol. The sources of

error in determining the free energy value of 79.6 kcal/mol for

reaction 10 should be similar to those encountered in determin-

ing the value of 53.6 kcal for the free energy of reaction 6.

Free energies of hydride transfer obtained using the cycle shown

in Scheme 2 should provide a quantitative measure of the

relatiVe hydride donor abilities for a series of compounds in

acetonitrile even if the constant of 79.6 kcal/mol is not

completely accurate.

various parameters influence the ability of [HM(diphosphine)₂]⁺

to function as hydride donors.

Results

Synthesis and Spectral Characterization of Complexes.

Ni(II), Ni(0), Ni-H, Pt(II), Pt(0), and Pt-H complexes were

prepared to probe the relationship between first- and third-row

transition-metal complexes as hydride donors. In some cases,

Pd(II) and Pd(0) complexes were prepared as well, but as

discussed below we were unable to synthesize the corresponding

palladium hydride complexes. Four diphosphine ligandss1,2-

bis(dimethylphosphino)ethane (dmpe), 1,2-bis(diethylphos-

phino)ethane (depe), 1,2-bis(diphenylphosphino)ethane (dppe),

and 1,3-bis(dimethylphosphino)propane (dmpp)swere studied

for both the Ni and Pt complexes. Three different phosphine

substituents, methyl, ethyl, and phenyl, are used to probe the

electron-donating effect of substituents, whereas dmpe and dmpp

are used to probe the effect of chelate bite size. In some

instances, other ligands were used to complement these four.

The reaction of [Ni(CH₃CN)₆](BF₄)₂with the appropriate

diphosphine ligand produces five-coordinate [Ni(diphosphine)₂-

(CH₃CN)](BF₄)₂complexes, as shown in reaction 13. Two

In previous work, we have shown that [HM(L)₂]⁺complexes

(where M ) Ni, Pt and L is a diphosphine ligand) can transfer

a hydride ligand to coordinated CO, as shown in reaction 12.¹⁵

[HM(L)₂]⁺+ [M′L′_n(CO)]⁺f [M(L)₂]²⁺+ [M′L′_n(CHO)]

[Ni(CH₃CN)₆]²⁺+ 2L f [Ni(L)₂(CH₃CN)]²⁺(13)

L ) dppe, dmpe, dmpp, depe

(12)

This is an interesting reaction for several reasons. (1) Hydride

transfer is occurring from a positively charged hydride complex.

In other studies, hydride transfers have been observed only for

anionic or neutral transition-metal hydrides. (2) These hydride

donors involve late transition metals, as opposed to early

transition metals for which more hydridic character is expected.

(3) These metal hydride complexes can be generated by

electrochemical reduction of the corresponding [M(diphos-

phine)₂]²⁺complexes in the presence of a proton source. In this

respect, these [HM(diphosphine)₂]⁺/[M(diphosphine)₂]²⁺couples

closely resemble NADH/NAD⁺couples in biological systems.

(4) A carbonyl ligand is reduced to a formyl ligand, a reaction

that generally requires a main-group-metal hydride as the

reductant,¹⁶although a few transition-metal hydride donors are

known for this reaction.¹⁷(5) The ability of reaction 12 to

proceed as shown is strongly dependent on the nature of both

the hydride donor complexes and the carbonyl complexes. (6)

The [M(diphosphine)₂]²⁺complexes undergo either two one-

electron reductions or a single two-electron reduction. This

suggests that if pK_ameasurements can be made for the

[HM(diphosphine)₂]⁺complexes, a quantitative thermodynamic

scale of relative hydride donor abilities could be made for these

complexes using the thermochemical cycle outlined above. In

this paper, we describe such measurements and examine how

infrared bands are observed for [Ni(dmpe)₂(CH₃CN)](BF₄)₂and

[Ni(dmpp)₂(CH₃CN)](BF₄)₂in the solid state (Nujol mulls),

which are assigned to CN stretching and combination modes¹⁸

(see Table 1 for IR and ³¹P NMR data for these complexes).

The acetonitrile ligand is weakly coordinated, and it can be

removed to produce four-coordinate complexes by applying a

vacuum or by crystallizing the complexes from a noncoordi-

nating solvent. [Pd(diphosphine)₂](BF₄)₂complexes and [Pt-

(diphosphine)₂](PF₆)₂were prepared by similar methods (see

Experimental Section), and spectral data for these complexes

are also listed in Table 1.

The Ni(0) complexes, Ni(diphosphine)₂, shown in Table 1

were prepared by reacting Ni(COD)₂with 2 equiv of the

diphosphine ligand. The Pt(0) and Pd(0) complexes were

prepared by reducing the corresponding Pt(II) or Pd(II) com-

plexes with sodium naphthalenide in tetrahydrofuran (THF).

These complexes are quite air sensitive, but they are thermally

stable, and the complexes containing dmpe and dmpp ligands

can be sublimed.

The most convenient route to the [HNi(diphosphine)₂](PF₆)

complexes is to prepare the Ni(0) complexes in situ as described

above and add NH₄PF₆to protonate the metal. This method

does not work for the more basic platinum analogues, because

protonation of the Pt(0) derivatives with NH₄PF₆results in

formation of [Pt(diphosphine)₂](PF₆)₂and hydrogen evolution.

The platinum hydrides are readily prepared by reduction of

[Pt(diphosphine)₂](PF₆)₂with NaBH₄on alumina. Spectral data

for these complexes are summarized in Table 1. The crystal

structure of [HPt(depe)₂](PF₆) has been reported previously.¹⁵

Metal hydride complexes can also be prepared by reacting

hydrogen gas with 1:1 mixtures of M(0) and M(II) complexes

in benzonitrile. For example, [Ni(depe)₂](BF₄), generated by

comproportionation of Ni(depe)₂and [Ni(depe)₂](BF₄)₂, reacts

cleanly with hydrogen to form [H(Ni(depe)₂](BF₄). Similarly,

addition of hydrogen gas to a 1:1 mixture of Pt(dmpe)₂and

[Pt(dmpe)₂](PF₆)₂results in formation of [HPt(dmpe)₂](PF₆). In

(14) Zhang, X.-M., Bruno, J. W.; Enyinnaya, E. J. Org. Chem. 1998,

63, 4671-4678.

(15) Miedaner, A.; DuBois, D. L.; Curtis, C. J.; Haltiwanger, R. C.

Organometallics 1993, 12, 299-303.

(16) (a) Casey, C. P.; Andrews, M. A.; McAlister, D. R.; Rinz, J. E. J.

Am. Chem. Soc. 1980, 102, 1927. (b) Sweet, J. R.; Graham, W. A. G. J.

Am. Chem. Soc. 1982, 104, 2811. (c) Tam, W.; Lin, G.-Y.; Wong, W.-K.;

Kiel, W. A.; Wong, V. K.; Gladysz, J. A. J. Am. Chem. Soc. 1982, 104,

141.

(17) (a) Wolcanski, P. T.; Bercaw, J. E. Acc. Chem. Res. 1980, 13, 121.

(b) Labinger, J. A.; Wong, K. A.; Sheidt, W. R. J. Am. Chem. Soc. 1978,

100, 3254. (c) Fagan, P.; Moloy, K.; Marks, T. J. J. Am. Chem. Soc. 1981,

103, 6959. (d) Nelson, G. O.; Sumner, C. E. Organometallics 1986, 5, 1983.

(e) Farnos, M. D.; Woods, B. A.; Wayland, B. B. J. Am. Chem. Soc. 1986,

108, 3659. (f) Zhang, X.-X.; Parks, G. F.; Wayland, B. B. J. Am. Chem.

Soc. 1997, 119, 7938-7944. (g) Dombek, B. D. Organometallics 1985, 4,

1707. (h) Darensbourg, M. Y.; Ash, C. E.; Arndt, L. W.; Janzen, C. P.;

Youngdahl, K. A.; Park, Y. K. J. Organomet. Chem. 1990, 383, 191.

(18) Storhoff, B. N.; Lewis, C. H., Jr. Coord. Chem. ReV. 1977, 33, 1.

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11435

Table 1. Selected Spectral Data for M(II), M-H, and M(0) Complexes

³¹P NMR^a(δ)

f

IR (cm^-1

)

[ML₂]X₂

(J_Pt-P, Hz)

[ML₂]

(J_Pt-P, Hz)

[HML₂]X

¹H NMR (δ)

metal

Ni

L

L′

(J_Pt-P, Hz)

M-H (J_Pt-H, J_P-H, Hz)

ν(MH)

ν(CN)

dmpe

depe

dmpp

dppe

dmpe

depe

dmpp

dppe

dmpe

41.8^c

63.3^c

15.0^b

24.6

-14.02 (8.5)

-14.16

1905

1924

1940

2264, 2300

40.5^b

46.2^c

-15.7^c

42.5^d

-13.6^c

-19.9^b

-14.33

2300, 2335

54.7^c

-12.87 (2.4)

1943, 1964

Pd

Pt

38.3^c

-5.3^b

35.0 (2169)

58.5 (2125)

-29.6^c(2079)

-12.6^b(3714)

21.5^b(3614)

-52.1^b(3661)

-7.3 (2184)

-11.55 (726, 30)

-12.12 (650, 29)

2015

2048

2072

2070

22.1 (2256)

-54.4^c(2295)

20.4^c(2363)

depe

32.1^e(3643, 62)

-15.9 (3665, 62)

26.5^e(3469, 50)

-54.6 (3765, 50)

-9.8^e(3504, 55)

-45.4 (3834, 55)

32.2^e(2388, 72)

-15.3 (2122, 72)

38.5^e(2256, 66)

-56.0 (2146, 66)

6.2^e(2551, 68)

-51.3 (2142, 68)

depe

dmpp

dmpe

^aSpectra recorded in CD₃CN unless otherwise noted. ^bIn toluene-d₈. ^cIn CD₃NO₂. ^dIn CD₂Cl₂. ^eIn benzonitrile. ^fIR spectra were recorded as

Nujol mulls. For [HNi(dppe)₂](PF₆), the two IR bands appear to arise from a solid-state effect, because only one band is observed in dichloromethane

solutions.

these reactions, an equilibrium concentration of the M(I)

complexes homolytically cleaves hydrogen. None of the

M(diphosphine)₂or [M(diphosphine)₂]²⁺complexes react with

hydrogen in benzonitrile. In only one case is a reaction between

a M(II) species and hydrogen observed. [Ni(dmpp)₂](BF₄)₂

heterolytically cleaves hydrogen in dimethylformamide to form

[HNi(dmpp)₂](BF₄) and H⁺. This reaction does not occur in

benzonitrile or acetonitrile.¹⁹

Hydride Transfer Reactions. Measuring equilibrium con-

stants for reactions of [HM(diphosphine)₂]⁺complexes with

[M′(diphosphine′)₂]²⁺complexes, as shown in reaction 14 for

[HPt(dmpe)₂]⁺and [Ni(dmpe)₂]²⁺, would be the simplest

method for determining the relative hydride-donor abilities of

[HM(diphosphine)₂]⁺cations. This reaction is simply a competi-

equilibrium constants. However, it is clear that all of the hydrides

discussed in the preceding paragraph can reduce N-benzyl-

nicotinamide hexafluorophosphate, which indicates that these

metal hydrides are better hydride donors than N-benzylnicotin-

amide hexafluorophosphate.

pK_aMeasurements. Norton and co-workers have measured

the pK_avalues of a number of transition-metal hydrides in

acetonitrile. The choice of acetonitrile was made according to

Norton for the following reasons: “(1) CH₃CN is a polar solvent

(ꢀ ) 36) in which almost all hydride transition-metal complexes

are soluble and in which their anions should undergo less ion

pairing than in solvents such as THF; (2) CH₃CN has a low

self-ionization constant (K_{autoprotolysis}) 3 × 10^-29) and is thus

suitable for experiments over a wide range of acid and base

strengths; (3) there exists a fair body of thermodynamic acid/

base data for organic compounds in this solvent.”^2bUnfortu-

nately, the M(diphosphine)₂complexes which result from

deprotonation of the corresponding hydrides are almost totally

insoluble in acetonitrile. To overcome this solubility problem

for the M(0) complexes while maintaining solvent properties

similar to those of acetonitrile, we chose benzonitrile as the

solvent for both pK_aand electrochemical measurements. To

establish that the relatiVe pK_avalues of metal hydrides would

not differ significantly between benzonitrile and acetonitrile,

we measured the equilibrium constants for the reactions of

[HNi(dppe)₂]⁺with pyridine and [HPt(dppe)₂]⁺with triethyl-

amine in both acetonitrile and benzonitrile by ³¹P NMR

spectroscopy. Using a pK_avalue of 12.3 determined for pyridine

in acetonitrile²⁰for both acetonitrile and benzonitrile, we

measured pK_avalues of 14.2 ( 0.3 and 14.7 ( 0.3 for [HNi-

(dppe)₂]⁺in acetonitrile and benzonitrile, respectively. Similarly,

using a pK_avalue of 18.5 determined for triethylamine in

acetonitrile,²⁰pK_avalues of 22.0 ( 0.2 and 22.2 ( 0.1 were

measured for [HPt(dppe)₂]⁺in acetonitrile and benzonitrile,

respectively. Thus, a ∆pK_avalue (pK_a(Pt) - pK_a(Ni)) of 7.8 in

acetonitrile corresponds to a ∆pK_aof 7.5 in benzonitrile. Within

the accuracy of our measurements, these values are the same.

On the basis of these results, we conclude that relatiVe ∆pK_a

values measured in these two solvents are not significantly

different, although the absolute pK_ascales may be different.

[HPt(dmpe)₂]⁺+ [Ni(dmpe)₂]²⁺

f

[Pt(dmpe)₂]²⁺+ [HNi(dmpe)₂]⁺(14)

tion between two [M(diphosphine)₂]²⁺species for the hydride

ligand, and it can be followed by ³¹P NMR spectroscopy.

However, for the [HM(diphosphine)₂]⁺and [M′(diphos-

phine′)₂]²⁺complexes synthesized in this study, the equilibrium

constants are too large; i.e., all of the reactions are quantitative.

The results obtained from these competition reactions produced

a relative ordering of hydride donor abilities for these complexes

of [HPt(dmpe)₂]⁺> [HPt(depe)₂]⁺> [HPt(dmpp)₂]⁺> [HNi-

(dmpe)₂]⁺> [HNi(depe)₂]⁺> [HNi(dmpp)₂]⁺. Although this

information is quite useful, it does not provide a quantitative

measure of the driving force for these reactions.

Reactions of these six hydrides with N-benzylnicotinamide

hexafluorophosphate, an NAD analogue, result in some hydride

transfer in all cases. For [HPt(dmpe)₂]⁺, this reaction goes to

completion in less than 1 h at room temperature. For [HNi-

(depe)₂]⁺and [HNi(dmpp)₂]⁺, the reaction requires weeks for

appreciable hydride transfer to occur. The inability of these

reactions to reach equilibrium in a reasonable amount of time,

coupled with a slow decomposition of the nickel hydrides under

the same conditions in the absence of N-benzylnicotinamide

hexafluorophosphate, prevented a reliable measurement of these

(19) A reviewer has suggested that a cooperative action of M(0) and

M(II) species in the cleavage of hydrogen is also possible. Additional

experiments will be required to distinguish between these two alternatives.

We would note that hydrogen is not activated by a mixture of [Pt(dmpe)₂]²⁺

and Ni(dmpe)₂.

(20) (a) Coetzee, J. F. Prog. Phys. Org. Chem. 1967, 4, 45. (b) Kolthoff,

I. M.; Chantooni, M. K., Jr.; Bhowmik, S. J. Am. Chem. Soc. 1968, 90,

23-28. (c) Coetzee, J. F.; Padmanabhan, G. R. J. Am. Chem. Soc. 1965,

87, 5005.

11436 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

Kristja´ndo´ttir and Norton have proposed that the relatiVe pK_a

values for a series of hydrides should be the same in water,

acetonitrile, and dichloromethane.^2a

Two methods were used to measure the relative acidity of

[HNi(depe)₂]⁺, [HNi(dmpp)₂]⁺, and [HNi(dmpe)₂]⁺. The first

was to measure equilibrium values for reactions of the hydrides

with tetramethylguanidine (TMG), as shown in reaction 15. The

HML₂⁺+ TMG f ML₂+ TMGH⁺

(15)

K_{eq 15}) [ML₂][TMGH⁺]/[HML₂][TMG]

+

M

L

K_{eq 15}

pK_a

Ni

depe

dmpp

dmpe

depe

0.30 ( 0.13

0.20 ( 0.15

23.8 ( 0.2

24.0 ( 0.3

24.4 ( 0.2

27.1 ( 0.3

0.087 ( 0.04

Pt

(1.6 ( 0.8) × 10^-4

Figure 1. ³¹P NMR spectrum of the reaction of [Pt(depe)₂H]PF₆and

Pt(dmpp)₂in benzonitrile. The ¹⁹⁵Pt satellites have not been labeled

for clarity. The labels are as follows: (A) [Pt(depe)₂H]PF₆; (B) Pt-

(depe)₂; (C) [HPt(dmpp)₂]PF₆; (D) Pt(dmpp)₂; (E) [HPt(depe)(dmpp)]-

PF₆; (F) Pt(depe)(dmpp). The spectrum has been enlarged to facilitate

the viewing of the smaller peaks, and the top portion of peak B which

corresponds to Pt(depe)₂has been cropped.

K_{eq 15}values are the result of five separate experiments using

different ratios of reactants. The uncertainty shown is 1 standard

deviation. To convert the equilibrium constants, K_{eq 15}, to pK_a

values, the pK_avalue of 23.3 for tetramethylguanidine²⁰is added

to the pK_eqvalue. The second method used for measuring the

relative acidity of the nickel hydrides was to determine the

equilibrium constants for the competition reactions between

nickel hydride complexes and Ni(0) complexes, as shown in

reaction 16. Again, five separate experiments were carried out

measured for the reaction of [HPt(depe)₂]⁺with a large excess

of tetramethylguanidine, a pK_avalue for this complex of 27.1

could be established (reaction 15, last entry). The equilibrium

constants for the reaction of [HPt(depe)₂]⁺with Pt(dmpe)₂and

Pt(dmpp)₂were then used to calculate the pK_avalues for the

last two complexes. Using this approach, the pK_avalues for

the six hydrides shown in reactions 15 and 16 are all referenced

to TMG.

For the Pt complexes, there is an exchange of the diphosphine

ligands to produce [HPt(L)(L′)]⁺and Pt(L)(L′), which ac-

companies the proton exchange reaction, 16. That is, equilibria

17 and 18 also occur. This results in fairly complex ³¹P NMR

+

HML₂⁺+ ML′₂f ML₂+ HML′₂

(16)

+

K_{eq 16}) [ML₂][HML′₂⁺]/[HML₂][ML′₂]

+

M

L

L′

K_{eq 16}

pK_a(HML′₂

)

Ni

depe

dmpp

depe

dmpp

dmpe

dmpp

dmpe

1.27 ( 0.2

3.40 ( 0.5

2.41 ( 0.2

5.16 ( 0.5

27.6 ( 6

23.9

24.3

Pt

27.8

28.5

HPtL₂⁺+ HPtL′₂⁺f 2HPtL′L⁺

(17)

5.62 ( 0.4

K_{eq 17}) [HPtL′L⁺]²/[HPtL₂][HPtL′₂

]

+

using different stoichiometric ratios of the reactants. To ensure

that equilibrium was being obtained, measurements of K_eqwere

made starting from opposite sides of the equilibrium. For

example, K_{eq 16}was determined both for the reaction of [HNi-

(depe)₂]⁺with Ni(dmpe)₂and for the reverse reaction of

Ni(depe)₂with [HNi(dmpe)₂]⁺. The pK_avalues for [HNi-

(dmpe)₂]⁺and [HNi(dmpp)₂]⁺were then calculated by adding

23.8 (obtained from reaction 15) to log K_{eq 16}for the reactions

of [Ni(dmpp)₂] and [Ni(dmpe)₂] with [HNi(depe)₂]⁺. The pK_a

values for [HNi(dmpp)₂]⁺and [HNi(dmpe)₂]⁺calculated using

this approach are in good agreement with those determined by

direct reaction with TMG (reaction 15). The relative pK_avalues

resulting from competition reactions between two different metal

complexes are more accurate than those resulting from reaction

15, because all of the species in reaction 16 are spectroscopically

observable in the same experiment. The temperature dependence

of equilibrium 16 has been studied by ³¹P NMR for the reactions

of [HNi(dmpp)₂]⁺with Ni(dmpe)₂, of [HNi(dmpp)₂]⁺with

Ni(depe)₂, and of [HNi(depe)₂]⁺with Ni(dmpe)₂. Plots of ∆G°

versus T give ∆S° values of 6 ( 2, 5 ( 2, and -4 ( 2 cal/(K

mol), respectively, for these three reactions. These values

indicate the entropy changes associated with these proton

exchange reactions are small.

L

L′

K_{eq 17}

depe

dmpp

dmpe

27.6 ( 2.7 (30.8 ( 1.9)

16.7 ( 1.1 (17.2 ( 3.6)

39.9 ( 3.5 (38.1 ( 2.3)

PtL₂+ PtL′₂f 2PtL′L

(18)

K_{eq 18}) [PtL′L]²/[PtL₂][PtL′₂]

L

L′

K_{eq 18}

depe

dmpp

dmpe

2.36 ( 0.02 (2.30 ( 0.12)

21.3 ( 1.2 (26.3 ( 2.4)

13.9 ( 0.3 (14.0 ( 0.8)

spectra, as shown in Figure 1 for the reaction of [HPt(depe)₂]⁺

with Pt(dmpp)₂. Because of equilibria 17 and 18, all three

possible hydride species ([HPt(depe)₂]⁺, [HPt(dmpp)₂]⁺, and

[HPt(depe)(dmpp)]⁺) and all three possible Pt(0) species (Pt-

(depe)₂, Pt(dmpp)₂, and Pt(depe)(dmpp)) are present. In Figure

1, the four singlets (A-D) with their associated Pt satellites

are assigned to two Pt(0) and two platinum hydride species

containing two identical diphosphine ligands. The two mixed

species each have A₂X₂spin systems that appear as two triplets

with associated Pt satellites. The resonances with negative

chemical shift values (approximately -55 ppm) in Figure 1 are

The pK_aof tetramethylguanidine is not large enough for it to

deprotonate all of the [HPt(diphosphine)₂]⁺complexes at

reasonable concentrations. When the equilibrium constant was

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11437

assigned to the dmpp ligand in [HPt(depe)(dmpp)]⁺and Pt-

(depe)(dmpp), because the chemical shifts are similar to those

of [HPt(dmpp)₂]⁺and Pt(dmpp)₂(see Table 1). The resonances

with positive chemical shifts (approximately +30 ppm) are

assigned to the coordinated depe ligand of [HPt(depe)(dmpp)]⁺

and Pt(depe)(dmpp) on the basis of the similarity of their

chemical shifts with those of [HPt(depe)₂]⁺and Pt(depe)₂. The

large chemical shift difference between diphosphine ligands in

five- and six-membered chelate rings (approximately 80 ppm)

is well-documented.²¹Two criteria were used to assign the triplet

resonances to either the mixed platinum hydride or the mixed

Pt(0) species. First, for the unmixed complexes, ¹J_Pt-Pis larger

for the Pt(0) complexes (approximately 3600-3700 Hz) than

for the platinum hydride complexes (approximately 2200-2300

Hz), as can be seen from Table 1. This should be true for the

mixed-ligand complexes as well. Second, the chemical shifts

for the unmixed hydride complexes are negative of the chemical

shifts of their corresponding Pt(0) analogues; therefore, it is

expected that the triplet resonances of [HPt(depe)(dmpp)]⁺will

occur at more negative chemical shifts than for the triplet

resonances of [Pt(depe)(dmpp)]. These assignments were con-

firmed by ³¹P NMR spectra recorded on reactions between the

various hydride complexes, reaction 17, which produce only

mixed hydride species. Similarly, reaction 18 produces only

mixed Pt(0) complexes. Equilibrium constants for reactions 17

and 18 were calculated from ³¹P NMR data for reactions 17

and 18, and from reaction 16 (values in parentheses for K_{eq 17}

and K_{eq 18}). Both equilibria 17 and 18 are reached rapidly (in

45 min or less) for the Pt complexes, but significant concentra-

tions of the mixed-ligand species are not observed for the Ni

complexes at room temperature. In addition, no significant

exchange of diphosphine ligands is observed for either [Ni-

(diphosphine)₂]²⁺or [Pt(diphosphine)₂]²⁺complexes on the basis

of ³¹P NMR studies of mixtures of [M(L)₂]²⁺and [M(L′)₂]²⁺

complexes.

Figure 2. Cyclic voltammograms of 1 × 10^-3M solutions of (a, thin

solid line) [Pt(depe)₂][PF₆]₂in acetonitrile; (b, line with open circles)

[Pt(depe)₂][PF₆]₂in benzonitrile; (c, thick solid line) [Ni(depe)₂][BF₄]₂

in benzonitrile. The scan rate was 0.05 V/s, and the working electrode

was a 2 mm diameter glassy-carbon electrode. The potentials are

referenced to the ferrocene/ferrocenium couple.

presumably with the formation of H₂. Although these experi-

ments do not provide the data needed for an accurate pK_avalue,

they suggest a value between 23.9 and 27.8. These experiments

also indicate that Pd hydrides are less stable to protonation than

their nickel and platinum analogues (i.e., the pH range over

which they are stable must be small).

Electrochemical Measurements. Cyclic voltammograms of

[Ni(depe)₂]²⁺in benzonitrile and [Pt(depe)₂]²⁺in benzonitrile

and acetonitrile are shown in Figure 2. It can be seen that

[Ni(depe)₂]²⁺undergoes two reversible one-electron reductions,

which is typical for the nickel complexes.^22aThe half-wave

potentials, E_1/2, of the nickel complexes shown in Table 2 have

an estimated accuracy of (0.02 V. Controlled-potential elec-

trolysis of [Ni(dmpp)₂]²⁺at -0.96 V resulted in the passage of

0.8 faraday of charge/mol of complex, consistent with a one-

electron reduction. Plots of the peak current vs the square root

of the scan rate are linear for both of the reduction waves of

the Ni complexes for scan rates between 0.05 and 4.0 V/s,

indicating diffusion-controlled reductions.

As can be seen from Figure 1, although the spectra are

complex, the reactions interconverting the six species involved

are clean. There are no significant unassigned peaks observed

in the ³¹P NMR spectra for these reactions. Using the integration

ratios obtained from such spectra, it is possible to measure

equilibrium constants for the proton transfer between mixed and

unmixed ligand complexes as shown in reaction 19. These

On the basis of the structural and spectroscopic results, the

Ni(II) complexes can exist as either four- or five-coordinate

species, as shown in equilibrium 20. Recrystallization of the

HPtL₂⁺+ PtL′L f HPtL′L⁺+ PtL₂

(19)

K_{eq 19}) [HPtL′L⁺][PtL₂]/[HPtL₂][PtL′L]

+

[Ni(diphosphine)₂(RCN)]²⁺

f

[Ni(diphosphine)₂]²⁺+ RCN (20)

L

L′

K_{eq 19}

pK_a(HML′L⁺)

[Ni(diphosphine)₂]²⁺+ e^-f [Ni(diphosphine)₂]⁺(21)

depe

dmpp

dmpe

8.31 ( 0.20

3.60 ( 1.65

3.77 ( 0.30

28.1

27.8

28.5

complexes in the absence of added acetonitrile or benzonitrile

results in four-coordinate species, indicating that the equilibrium

lies to the right under these conditions. However, recrystalli-

zation in the presence of benzonitrile or acetonitrile results in

the formation of five-coordinate species. As one might expect,

the net result of equilibrium 20 is to shift the reduction potential

for the Ni(II/I) couple (reaction 21) to more negative potentials

as the concentration of acetonitrile or benzonitrile increases.

This is caused by a decrease in the concentration of the more

easily reduced four-coordinate [Ni(diphosphine)₂]²⁺species. The

E_1/2values for the Ni(II/I) couples shift by -0.04 to -0.11 V

equilibrium constants can then be used to calculate pK_avalues

for the mixed-ligand complexes in the same way as they were

calculated for the unmixed complexes using reaction 16.

Combining these results with the equilibrium studies described

above provides pK_avalues for 11 different [HM(diphosphine)₂]⁺

complexes.

To determine if pK_adata could be obtained for a Pd-H

complex, a solution of Pd(dmpp)₂in benzonitrile was mixed

with a solution of [HPt(dmpp)₂]⁺, but no reaction was observed.

A similar experiment with [HNi(dmpp)₂]⁺resulted in the

formation of [Pd(dmpp)₂]²⁺and Ni(dmpp)₂in a 1:2 ratio,

(22) (a) Miedaner, A.; Haltiwanger, R. C.; DuBois, D. L. Inorg. Chem.

1991, 30, 417-427. (b) Alyea, C. E.; Meek, D. W. Inorg. Chem. 1972, 11,

1029.

(21) Garrou, P. E. Chem. ReV. 1981, 81, 229.

11438 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

Table 2. Electrochemical Data for M(diphosphine)₂Complexes

∆S°_rc(I/0)

∆S°_rc(II/I)

∆S°_rc(II/0)

(cal/(K mol))

b

E

_1/2(II/I)^aE_1/2(I/0)^aE_1/2(II/0)^a

∆G°₂₂

∆S°₂₂

∆H°₂₂

compd

(V)

(kcal/mol) (cal/(K mol)) (kcal/mol) exptl^ccalcd^dexptl^ccalcd^dexptl^ccalcd^d

25.4

Ni(dppp)₂

(-0.19)

[-0.12]

-0.70

(-0.70)

[-0.66]

-0.86

(-0.89)

[-0.76)

-1.13

(-1.16)

[-1.04]

(-0.91)

[-0.95]

-0.90

(-0.88)

[-0.95]

-1.32

(-1.33)

[-1.35]

-1.30

Ni(dppe)₂

Ni(dmpp)₂

Ni(depe)₂

36.9

50.2

56.0

11

40.1

58.2

63.0

6.7

16

30.5

22

37.2

38

27.3

23.7

15.8

(23)

18

(16)

37.7

(40)

28

(26)

53.5

(63)

46

(42)

12.2

(17)

18

(16)

37.7

(36)

28

(26)

49.9

(53)

46

(42)

(-1.29)

[-1.34]

Pt(dppe)_2]

Ni(dmpe)₂

-1.24

-1.35

57.2

62.3

20.5

29.5

64.6

71.0

46.7

55.7

(62)

38

46

(42)

(-1.39)

[-1.32]

-1.51

(-1.53)

-1.63

(-1.65)

-1.73

Pt(dmpp)₂

Pt(depe)₂

Pt(dmpe)₂

69.6

75.2

79.8

14.7

22.9

41

73.9

81.9

91.8

40.9

(40)

49.1

(59)

67

46

(42)

46

(42)

46

^aHalf-wave potentials vs FeCp₂for the II/I, I/0, and II/0 couples in benzonitrile; potentials in acetonitrile are given in parentheses, and those in

dichloromethane or dichloroethane are given in brackets. ^bFree energy for reaction 22 in benzonitrile calculated from ∆G° ) -nFE° assuming E°

) ¹/₂[E_1/2(II/I) + E_1/2(I/0)] and n ) 2. ^cEntropy changes associated with II/I, I/0, and II/0 redox couples in benzonitrile; entropy changes in acetonitrile

are given in parentheses. ^dCalculated values were obtained by using eq 24 with AN values of 15.5 for benzonitrile and 19.3 for acetonitrile and

r values of 4.0 and 7.0 for complexes with methyl and phenyl substituents, respectively.

upon changing from noncoordinating solvents, such as di-

chloroethane or dichloromethane (shown by the values in

brackets in Table 2), to pure benzonitrile or acetonitrile (values

in parentheses in Table 2). These shifts correspond to free energy

changes of 0.9-2.5 kcal/mol or to equilibrium constants for

reaction 20 between 0.2 and 5 at room temperature.²³Titration

of a dichloroethane solution of [Ni(dmpp)₂](BF₄)₂with aceto-

nitrile was followed by UV-visible spectroscopy to determine

an equilibrium constant of 0.8 for reaction 20.²⁴These values

are consistent with spectroscopic measurements, which indicate

that the acetonitrile ligands of the five-coordinate [Ni-

(diphosphine)₂(CH₃CN)]²⁺species are fully dissociated in

noncoordinating solvents.

[Pt(depe)₂]²⁺undergoes a quasi-reversible (∆E_p) 68 mV at

50 mV/s) two-electron reduction in benzonitrile and a nearly

reversible two-electron reduction in acetonitrile (∆E_p) 46 mV

at 50 mV/s). A comparison of the slopes of the chronoampero-

metric plots for potential steps across the reduction wave of

[Pt(depe)₂]²⁺and the two reduction waves of [Ni(dmpp)₂]²⁺in

benzonitrile gives a ratio of 1.1:1.0. Because the slopes are

proportional to the number of electrons involved in the charge

transfer, these results are consistent with a two-electron reduction

of [Pt(depe)₂]²⁺. Pt(dmpp)₂and Pt(dmpe)₂exhibit quasi-

reversible oxidations in benzonitrile with peak-to-peak separa-

tions of 106 and 210 mV, respectively, at a scan rate of 10

mV/s. The failure to observe completely reversible couples for

Pt(dmpp)₂and Pt(dmpe)₂results in larger possible errors for

the E_1/2values shown in Table 2 for these two complexes

(estimated to be ( 0.05 V). The two-electron oxidations

observed for the Pt(diphosphine)₂complexes give E_1/2values

that are average values for the II/I and I/0 couples. No

information can be obtained for the E_1/2values of the individual

II/I and I/0 couples.

The temperature dependence of E_1/2values of redox couples

can provide information on the entropy changes associated with

the oxidation or reduction of a species in solution. Weaver and

co-workers used nonisothermal cells to obtain ∆S° values for a

number of redox couples in a wide variety of solvents.²⁵Koval

and co-workers later proposed using ferrocene as a redox

standard for variable-temperature measurements of the redox

potentials. The measured entropy changes can then be corrected

for the known entropy change of the ferrocene couple.²⁶

Equation 23 gives the relationship between the entropy change

for reaction 22, E° (approximated by E_1/2), and temperature.

+

nFeCp₂+ [ML₂]²⁺f [ML₂]^(2-n)++ nFeCp₂

d(∆G°)/dT ) -nFd(E°)/dT ) ∆S°₂₂

(22)

(23)

The ∆S°₂₂values shown in Table 2 were determined by

measuring the E_1/2values as a function of temperature by

differential-pulse and cyclic voltammetry. ∆G°₂₂, ∆S°₂₂, and

∆H°₂₂values for the reduction of [M(diphosphine)₂]²⁺com-

plexes by ferrocene at 294 K are listed in Table 2.

To obtain the experimental values for the entropy changes

associated with individual redox couples (∆S°_rc) of the [M(diphos-

phine)₂]²⁺complexes, the entropy of the ferrocene redox couple

(11.5 cal/(K mol) in acetonitrile, 13.1 cal/(K mol) in benzo-

nitrile)²⁵is multiplied by n (the number of electrons transferred

(23) (a) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Funda-

mentals and Applications; Wiley: New York, 1980; p 35. (b) Brown, E.

R.; Large, R. F. In Physical Methods of Chemistry Part II: Electrochemical

Methods; Weissberger, A., Rossiter, B. W., Eds.; Techniques of Chemistry

I; Wiley: New York, 1971; p 483.

(25) (a) Yee, E. L.; Cave, R. J.; Guyer, K. L.; Tyma, P. D.; Weaver, M.

J. J. Am. Chem. Soc. 1979, 101, 1131-1137. (b) Sahami, S.; Weaver, M.

J. J. Electroanal. Chem. Interfacial Electrochem. 1981, 122, 155-170. (c)

Hupp, J. T.; Weaver, M. J. Inorg. Chem. 1984, 23, 3639-3644. (d) The

value of 13.1 cal/(K mol) for the entropy of the ferrocene redox couple in

benzonitrile was extrapolated from Figure 6 of ref 25c using an acceptor

number of 15.5.

(24) Qualitative data for titration of [Ni(depe)₂]²⁺in dichloromethane

with acetonitrile are consistent with an equilibrium constant for this complex

of approximately 1: Solar, J. M.; Ozkan, M. A.; Isci, H.; Mason, R. Inorg.

Chem. 1984, 23, 758.

(26) Koval, C. A.; Gustafson, R. M.; Reidsema, C. M. Inorg. Chem.

1987, 26, 950-952.

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11439

Table 3. Thermodynamics of Proton (pK_a), Hydrogen Atom

in reaction 22) and added to ∆S°₂₂. Experimental values of ∆S°_rc

for the II/I, I/0, and II/0 couples for [M(diphosphine)₂]²⁺

complexes are also listed in Table 2. Hupp and Weaver have

proposed the semiempirical relationship shown in eq 24 for

calculating the entropy of a one-electron reduction.^25cThe

•

-

(∆H°_H), and Hydride (∆G°_H) Transfer Reactions for

[HM(diphosphine)₂]⁺Complexes

a

b

-

•

M

L

pK_a

E_1/2(II/0)

∆H°_H

∆G°_H

Ni

dppe

14.7^c

(14.2)^c

23.9

-0.80

(-0.79)

-1.09

58.9

(58.8)

62.1

(61.6)

62.1

(62.4)

62.8

(62.7)

62.1

(61.2)

56.0

(56.0)

52.8

(52.5)

50.7

(48.9)

48.1

(47.2)

41.6

(40.8)

38.8

(38.8)

dmpp

depe

∆S°_rc) 21.9 - 0.58(AN) + 20.7(Z_ox²- Z_red²)/r (24)

(-1.11)

-1.22

23.8

(-1.22)

-1.24

acceptor number (AN) is an intrinsic property of the solvent

that reflects both the electrophilicity and polarity of the solvent,²⁷

Z_oxand Z_redare the charges on the complexes in their oxidized

and reduced forms, respectively, and r is the radius of the

molecule in angstroms as defined by Sutin et al.²⁸∆S°_rcvalues

calculated using this equation are listed in Table 2 following

the experimental values. The values used for r are 4.0 Å for

the dmpe, dmpp, and depe complexes of nickel and platinum

and 7.0 Å for dppe complexes. These values are based on

crystallographic data and/or molecular mechanics calculations

and represent an average value expected for the II and 0

oxidation states. From eq 24 and the calculated values of ∆S°_rc

in Table 2, it can be seen that ∆S°_rcfor the I/0 couple is expected

to be smaller than that of the II/I couple because of the larger

charges involved in the latter couple. This is true where the

values of ∆S°_rccould be obtained for both couples. The

experimental ∆S°_rcvalues tend to be larger than the calculated

∆S°_rcvalues for the II/I couple. As a result, the experimental

∆S°_rcvalues are generally larger than the calculated values for

the reduction from M(II) to M(0) (Table 2). Equation 24 was

developed for complexes in which the ligands prevented direct

interactions between the metal and solvent and for which

solvent-solvent interactions were more important than solvent-

metal interactions.^25cBecause the solvent can approach and even

coordinate to the metal center in the case of the Ni complexes,

the agreement between the experimental and calculated values

is surprising.

Pt

Ni

Pt

dppe

dmpe

dmpp

depe

22.2

(22.0)

24.3

(-1.24)

-1.35

61.7

(60.8)

(-1.39)

-1.51

27.8

27.1

28.5

(-1.53)

-1.63

(-1.65)

-1.73

dmpe

^aCalculated using the expression ∆H°_H) 1.37pK_a+ 23.06E°(1+/

0) + 59.5.^7c^bCalculated using eq 11. A value of 26.0 kcal/mol was

used for ∆G° for reaction 10 in benzonitrile (the same value as for

acetonitrile). Although this absolute value may not be correct for

•

-

benzonitrile, the relative values calculated for ∆G°_Hshould be accurate.

^cValues for benzonitrile solutions; values for acetonitrile solutions are

given in parentheses.

tetrahedral distortion, which is slight for these compounds. The

bite angles, P-Ni-P, of the bidentate ligands are approximately

86°, which is typical for Ni(II) complexes containing chelating

phosphines with a two-carbon backbone.²²All of the Ni-P bond

lengths in these two cations are close to the average value of

2.21 Å, which is in the normal range for square-planar Ni(II)

compounds.²²

In contrast to [Ni(dmpe)₂]²⁺and [Ni(depe)₂]²⁺, the [Ni-

(dmpp)₂]²⁺cation exhibits a significant tetrahedral distortion.

The trans P-Ni-P angles are approximately 149°, and the

dihedral angle, â, between the planes defined by the two dmpp

ligands and Ni is 43.7°. [Pt(dmpp)₂]²⁺shows a smaller

tetrahedral distortion, with trans P-Pt-P angles of approxi-

mately 171° and a dihedral angle â of 11.5°. The average bite

angle of the dmpp ligand is 91.5° for [Ni(dmpp)₂]²⁺and 88.3°

for [Pt(dmpp)₂]²⁺. The tetrahedral distortions appear to have

Thermodynamics of Hydride Transfer. The relative free

energies of hydride transfer for [HM(diphosphine)₂]⁺cations

can be calculated using the thermochemical cycle shown in

Scheme 2. The pK_avalues for the [HM(diphosphine)₂]⁺cations

and the E_1/2values for the II/0 couple (or the average of the

II/I and I/0 couples for those complexes having two one-electron

transfers) were measured for the [M(diphosphine)₂]²⁺and

M(diphosphine)₂complexes as discussed above. ∆G°_Hvalues

calculated according to eq 11 are listed in Table 3, along with

E_1/2values for the II/0 couples or the average of the II/I and I/0

couples. The pK_avalues and E_1/2values for the I/0 couples can

also be used to determine the relative enthalpies for hydrogen

little effect on the M-P bond distances. For [Ni(dmpp)₂]²⁺

,

the average Ni-P bond distance is 2.21 Å, which is the same

as that observed for the two Ni(II) complexes described above.

For [Pt(dmpp)₂]²⁺, the average Pt-P bond distance is 2.33 Å,

in the normal range expected for Pt(II) complexes.²⁹

The drawings shown in Figure 4 illustrate that the two five-

coordinate Ni(II) complexes [Ni(dmpe)₂(CH₃CN)]²⁺and

[Ni(dmpp)₂(CH₃CN)]²⁺have significantly different structures.

[Ni(dmpe)₂(CH₃CN)]²⁺is best described as a square-pyramidal

complex with the acetonitrile ligand coordinated in the axial

position, whereas [Ni(dmpp)₂(CH₃CN)]²⁺has a trigonal-

bipyramidal geometry with the acetonitrile coordinated in the

equatorial position. The four N-Ni-P bond angles for

[Ni(dmpe)₂(CH₃CN)]²⁺are all within 1° of the average value

of 98.4°. In contrast, the average N-Ni-P bond angle for the

two axial phosphine ligands of [Ni(dmpp)₂(CH₃CN)]²⁺is 85.1°,

and the average N-Ni-P bond angle for the two equatorial

phosphorus atoms is 119.1°. The average of the Ni-P bond

distances for both complexes is 2.23 Å, but the two Ni-N bond

distances are quite differents2.28 Å for [Ni(dmpe)₂(CH₃CN)]²⁺

•

atom transfer (∆G°_H, Table 3) using Scheme 1.

Structural Studies. Crystallographic data for four four-

coordinate M(II) complexes ([Ni(dmpe)₂](BF₄)₂, [Ni(depe)₂]-

(BF₄)₂, [Ni(dmpp)₂](BF₄)₂, and [Pt(dmpp)₂](PF₆)₂), two five-

coordinate Ni(II) complexes ([Ni(dmpe)₂(CH₃CN)](PF₆)₂and

[Ni(dmpp)₂(CH₃CN)](BF₄)₂), and two M(0) complexes (Ni-

(dmpp)₂and Pt(dmpp)₂) are given in Table 4. Selected bond

distances and bond angles are given in Tables 5 and 6,

respectively. Drawings of the four four-coordinate M(II) cations

are shown in Figure 3. The [Ni(dmpe)₂]²⁺and [Ni(depe)₂]²⁺

cations are nearly planar, as measured by trans P-Ni-P angles

of 178 and 180°, respectively. The dihedral angle, â, between

the two planes formed by the Ni atom and the two phosphorus

atoms of each bidentate ligand is 3.4° for [Ni(dmpe)₂]²⁺and

0.0° for [Ni(depe)₂]²⁺. This angle is a useful measure of the

(29) (a) Gieren, V. A.; Bru¨ggeller, P.; Hofer, K.; Hu¨bner, T.; Ruiz-Pe´rez,

C. Acta Crystallogr. 1989, C45, 196-198. (b) Bru¨ggeller, P. Inorg. Chem.

1990, 29, 1742-1750. (c) Tulip, T. H.; Yamagata, T.; Yoshida, T.; Wilson,

R. D.; Ibers, J. A.; Otsuka, S. Inorg. Chem. 1979, 18, 2239-2250.

(27) Gutmann, V. Electrochim. Acta 1976, 21, 661-670.

(28) Brown, G. M.; Sutin, N. J. Am. Chem. Soc. 1979, 101, 883-891.

11440 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

Table 4. Crystallographic Data for [Ni(dmpe)₂](PF₆)₂, [Ni(depe)₂](BF₄)₂, [Ni(dmpp)₂](BF₄)₂, [Pt(dmpp)₂](PF₆)₂, [Ni(dmpe)₂(CH₃CN)](PF₆)₂,

[Ni(dmpp)₂(CH₃CN)](BF₄)₂, Ni(dmpp)₂, and Pt(dmpp)₂

[Ni(dmpe)₂](PF₆)₂

[Ni(depe)₂](BF₄)₂

[Ni(dmpp)₂](BF₄)₂

[Pt(dmpp)₂](PF₆)₂

empirical formula

formula mass, amu

cryst syst

space group

a (Å)

b (Å)

c (Å)

C₁₃H₃₅F₁₂NNiO₂P₆

709.95

monoclinic

P2₁/c

17.7374(2)

9.064

17.9013(2)

90

96.1870(10)

90

2861.32(5)

4

C₂₀H₅₀B₂F₈NiOP₄

662.81

orthorhombic

Pccn

17.7301(2)

10.91290(10)

15.77990(10)

90

C₁₄H₃₆B₂F₈NiP₄

560.64

monoclinic

P2₁/n

9.8371(13)

16.764(3)

14.8801(18)

90

101.136(7)

90

C_15.61H_39.73F₁₂N_0.27O_0.53P₆Pt

848.70

monoclinic

P2₁/c

17.5851(17)

9.3130(10)

18.412(3)

90

94.269(6)

90

R (deg)

â (deg)

90

γ (deg)

V (Å³)

3053.20(5)

4

2407.7(6)

4

3007.0(6)

4

Z

T (K)

150(2)

162(2)

160(2)

141(2)

R index^a(I > 2σ(I))

R index^a(all data)

R1 ) 0.0315

R1 ) 0.0434,

wR2 ) 0.0805

a ) 0.0373, b ) 1.0395

1.047

R1 ) 0.0316

R1 ) 0.0438,

wR2 ) 0.0731

a ) 0.0398, b ) 1.4490

0.914

R1 ) 0.0286

R1 ) 0.0354,

wR2 ) 0.0716

a ) 0.0308, b ) 1.7342

1.034

R1 ) 0.0332

R1 ) 0.0412,

wR2 ) 0.0805

a ) 0.0351, b ) 5.0718

1.047

weighting coeff^b

goodness of fit^con F²

[Ni(dmpe)₂(CH₃CN)](PF₆)₂

[Ni(dmpp)₂(CH₃CN)](BF₄)₂

Ni(dmpp)₂

Pt(dmpp)₂

empirical formula

formula mass, amu

cryst syst

space group

a (Å)

b (Å)

c (Å)

C₁₆H₃₈F₁₂N₂NiP₆

731.01

monoclinic

P2₁/c

19.9406(4)

9.0706(2)

16.7066(3)

90

97.49

90

2996.03(10)

4

C₁₆H₃₉B₂F₈NNiP₄

601.69

triclinic

C₁₄H₃₆NiP₄

387.02

monoclinic

C2/c

17.5777(17)

9.5357(11)

14.7122(16)

90

123.458(5)

90

C₁₄H₃₆P₄Pt

523.40

monoclinic

C2/c

17.481(6)

9.892(3)

14.524(5)

90

123.879(7)

90

P1h

10.7503(12)

14.2268(19)

17.8531(15)

93.236(6)

92.155(6)

92.390(5)

2721.6(5)

4

R (deg)

â (deg)

γ (deg)

V (Å³)

2057.4(4)

4

2085.3(13)

4

Z

T (K)

170(2)

169(2)

152(2)

160(2)

R index^a(I > 2σ(I))

R index^a(all data)

R1 ) 0.0560

R1 ) 0.1062,

wR2 ) 0.1349

a ) 0.0313, b ) 4.6821

1.144

R1 ) 0.0528

R1 ) 0.0899,

wR2 ) 0.1336

a ) 0.0482, b ) 4.9849

1.021

R1 ) 0.0442

R1 ) 0.0638,

wR2 ) 0.0874

a ) 0.0227, b ) 1.4966

1.463

R1 ) 0.0218

R1 ) 0.0223,

wR2 ) 0.0510

a ) 0.0308, b ) +¹/₂

1.113

weighting coeff^b

goodness of fit^con F²

^aR1 ) ∑||F_o| - |F_c||/∑|F_o|; wR2 ) [∑w(F_o- F_c²)²/∑w(F_o)²]^1/2

.

^bw^-1) [σ²(F_o) + (aP)²+ bP], where P ) (F_o+ 2F_c²)/3. ^cGOF ) S )

2

[∑w(F_o- F_c²)²/(M - N)]^1/2, where M is the number of reflections and N is the number of parameters refined.

Table 5. Selected Bond Distances (Å)

2

compd

M-P(1)

M-P(2)

M-P(3)

M-P(4)

M-N

[Ni(dmpe)₂](PF₆)₂

[Ni(depe)₂](BF₄)₂

[Ni(dmpp)₂](BF₄)₂

[Pt(dmpp)₂](PF₆)₂

[Ni(dmpe)₂

(CH₃CN)](PF₆)₂

[Ni(dmpp)₂

2.2133(4)

2.2358(3)

2.1952(4)

2.3042(10)

2.224(2)

2.2127(4)

2.2290(4)

2.2277(12)

2.3080(9)

2.229(2)

2.2197(4)

2.2134(4)

2.2210(4)

2.3243(12)

2.236(2)

2.2013(4)

2.3178(11)

2.225(2)

2.278(19)

1.992(3)

2.2120(11)

2.2131(11)

2.2206(12)

2.2617(11)

(CH₃CN)](BF₄)₂

[Ni(dmpp)₂]

[Pt(dmpp)₂]

2.1479(6)

2.2721(8)

2.1443(6)

2.2709(8)

and 1.99 Å for [Ni(dmpp)₂(CH₃CN)]²⁺. In the former a dis-

order problem involving the acetonitrile ligand probably con-

tributes to the long bond length. In [Ni(dppm)₂(CH₃CN)]²⁺the

Ni-N bond distance is 2.10 Å.^22aThis would indicate that the

Ni-N bond length is significantly shorter for an equatorially

bound acetonitrile as compared to an axially bound aceto-

nitrile. Both of the acetonitrile ligands are slightly bent at the

N atom by 10-15°, presumably due to steric interactions for

[Ni(dmpp)₂(CH₃CN)]²⁺and packing forces for [Ni(dmpe)₂-

each diphosphine ligand and the metal are 87.4 and 89.7° for

Ni(dmpp)₂and Pt(dmpp)₂, respectively. For the Ni complex,

the average Ni-P bond distance is 2.15 Å. For the Pt complex

the average Pt-P bond distance is 2.27 Å. Although reduction

generally results in a lengthening of metal-ligand bonds, these

values are 0.06 Å shorter than the corresponding values of their

Ni(II) and Pt(II) analogues. However, a similar bond shortening

occurs during the reduction sequence [Ir(dppf)₂]⁺/Ir(dppf)₂/

[Ir(dppf)₂]^-.³⁰These data suggest that bond shortening will

generally occur upon reduction of square-planar d⁸complexes

to form tetrahedral d¹⁰complexes.

(CH₃CN)]²⁺

.

Drawings of the two M(0) complexes Ni(dmpp)₂and Pt-

(dmpp)₂are shown in Figure 5. These two complexes are both

slightly distorted tetrahedral complexes. The dihedral angle

between the two planes defined by the phosphorus atoms of

(30) Longato, B.; Riello, L.; Bandoli, G.; Pilloni, G. Inorg. Chem. 1999,

37, 2818.

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11441

Table 6. Selected Bond Angles (deg)

Discussion

[Ni(dmpe)₂](PF₆)₂

As pointed out by Labinger, the transfer of a hydride ligand

will almost always produce a coordinatively unsaturated species.

This led him to conclude that measurement of hydricity for

transition-metal hydrides would need to rely on more “phe-

nomenological tests.”^13aHowever, there are stable 16-electron

species such as the square-planar d⁸transition-metal complexes.

It was the stability of [M(diphosphine)₂]²⁺complexes of Ni and

Pt that suggested to us that [HM(diphosphine)₂]⁺complexes

might function as useful hydride donors for the reduction of

metal carbonyl complexes.¹⁵The objective of this work is to

demonstrate a quantitative measure of the relative hydricity for

[HM(diphosphine)₂]⁺complexes and to delineate some of the

features that determine their hydridic character.

P(2)-Ni-P(1)

P(1)-Ni-P(4)

P(1)-Ni-P(3)

85.933(16) P(2)-Ni-P(4)

93.826(16) P(2)-Ni-P(3)

177.701(18) P(4)-Ni-P(3)

177.613(18)

93.998(16)

86.338(16)

[Ni(depe)₂](BF₄)₂

P(2)-Ni-P(2A) 180.0

P(2)-Ni-P(1A)

95.222(14)

P(2)-Ni-P(1)

84.778(14)

[Ni(dmpp)₂](BF₄)₂

99.926(18) P(1)-Ni-P(3)

89.18(3)

P(1)-Ni-P(4)

P(1)-Ni-P(2)

P(4)-Ni-P(3)

149.834(17)

93.34(3)

147.74(4)

P(3)-Ni-P(2)

93.851(19) P(4)-Ni-P(2)

[Pt(dmpp)₂](PF₆)₂

P(1)-Pt-P(2)

P(1)-Pt-P(4)

P(2)-Pt-P(4)

88.06(4)

93.62(5)

172.25(5)

P(1)-Pt-P(3)

P(2)-Pt-P(3)

P(4)-Pt-P(3)

171.62(5)

91.29(4)

88.14(5)

Structural Studies. Diffraction studies of several of these

complexes were performed to probe structure-activity relation-

ships. The Ni(II) complexes can be isolated as either four- or

five-coordinate species. The gross geometry of the five-

coordinate complexes depends on the bite size of the diphos-

phine ligand. For [Ni(dmpe)₂(CH₃CN)]²⁺a square-pyramidal

geometry is observed with an apical acetonitrile, whereas

[Ni(dmpp)₂(CH₃CN)]²⁺is a trigonal bipyramid in which the

acetonitrile ligand occupies an equatorial position. The small

chelate bites observed for [Ni(dmpe)₂]²⁺and [Ni(depe)₂]²⁺

produce nearly planar complexes. The larger chelate bites

observed for [Ni(dmpp)₂]²⁺and [Pt(dmpp)₂]²⁺result in signifi-

cant tetrahedral distortions. The LUMO of these complexes is

shown for planar (1) and tetrahedrally distorted geometries (2).

[Ni(dmpe)₂(CH₃CN)](PF₆)₂

P(1)-Ni-P(4)

P(4)-Ni-P(2)

P(4)-Ni-P(3)

P(1)-Ni-N

91.69(10)

163.9(5)

85.08(10)

99.1(3)

P(1)-Ni-P(2)

P(1)-Ni-P(3)

P(2)-Ni-P(3)

P(4)-Ni-N

85.21(8)

162.5(5)

93.14(10)

97.4(5)

P(2)-Ni-N

98.7(3)

P(3)-Ni-N

98.4(4)

[Ni(dmpp)₂(CH₃CN)](BF₄)₂

N-Ni-P(1)

P(1)-Ni-P(2)

P1-Ni-P(3)

N-Ni-P(4)

P(2)-Ni-P41

82.55(10)

88.76(4)

169.91(4)

110.35(10)

121.81(4)

N-Ni-P(2)

127.74(10)

87.70(10)

95.19(4)

96.20(4)

89.61(4)

N-Ni-P(3)

P(2)-Ni-P(3)

P(1)-Ni-P(4)

P(3)-Ni-P(4)

[Ni(dmpp)₂]

P(2)-Ni-P(2A) 113.41(4)

P(2)-Ni-P(1A) 115.41(2)

P(1)-Ni-P(1A) 110.72(4)

P(2)-Ni-P(1)

101.24(2)

[Pt(dmpp)₂]

P(2)-Pt-P(2A)

P(2)-Pt-P(1)

115.85(4)

99.75(3)

P(2)-Pt-P(1A)

P(1)-Pt-P(1A)

114.73(3)

112.82(4)

reduction potentials of the [M(diphosphine)₂]²⁺complexes. It

can be seen from Table 2 that the influence of the substituents

on the redox potentials follows the expected order of electron-

donating ability, Me > Et > Ph. Changing from a methyl to a

phenyl substituent produces a 0.49 V difference between the

reduction potential of [Pt(dmpe)₂]²⁺and [Pt(dppe)₂]²⁺and a

0.55 V shift for the average reduction potentials of the analogous

Ni complexes. The reduction potentials for the Pt complexes

are approximately 0.45 V more negative than the average of

the Ni(II/I) and Ni(I/0) couples for the same ligand. The effect

of the chelate bite size is somewhat different than changing the

metal or the ligand substituents, because it has a significant

effect on the potentials of the II/I couples but not on the po-

tentials of the I/0 couples. The Ni(II/I) couple for [Ni(dmpp)₂]²⁺

The antibonding overlap between the metal d orbital and the

ligand orbitals is decreased in 2 compared to 1, and LUMO 2

is stabilized with respect to LUMO 1, as described in detail in

a previous publication.^22aThis simple molecular orbital picture

provides a useful approach to understanding the thermodynamic

properties of the [M(diphosphine)₂]²⁺complexes and their

hydride analogues.

Dissociation of the acetonitrile ligands from the five-

coordinate [Ni(diphosphine)₂(CH₃CN)]²⁺complexes to produce

four-coordinate complexes is facile, and the energy differences

between the four- and five-coordinate complexes are small (1-3

kcal/mol). As a result of the small energy differences between

the four- and five-coordinate species, the energetics of these

five-coordinate complexes can be understood in terms of the

corresponding four-coordinate complexes.

Both Ni(dmpp)₂and Pt(dmpp)₂have dihedral angles between

the two planes defined by the diphosphine ligand and the metal

of nearly 90°. This indicates that during reduction from

[M(diphosphine)₂]²⁺to M(diphosphine)₂these metal complexes

undergo a distortion from square-planar structures to tetrahedral

structures. This twisting motion is larger for the Ni and Pt

complexes with diphosphine ligands with small bite sizes

because the M(II) complexes are more planar. Reduction is also

accompanied by a contraction of the M-P bond lengths of

approximately 0.06 Å.

is 0.5 V positive of the Ni(II/I) couple for [Ni(dmpe)₂]²⁺

.

Similarly the redox potential for the II/I couple of [Ni(dppp)₂]²⁺

(where dppp is 1,3-bis(diphenylphosphino)propane) is 0.54 V

positive of that of [Ni(dppe)₂]²⁺in dichloromethane.^22aHow-

ever, the E_1/2values for the I/0 couples of [Ni(dppe]₂]²⁺and

[Ni(dppp)₂]²⁺are nearly the same, and this is also true for

[Ni(dmpe)₂]²⁺and [Ni(dmpp)₂]²⁺. Comparison of the redox

potentials of Pt(dmpe)₂and Pt(dmpp)₂shows that changing from

a diphosphine ligand with a two-carbon backbone to one with

a three-carbon backbone produces a change in the II/0 couple

of approximately +0.25 V. This value is what would be

expected if the Pt(II/I) couples shifted to more positive potentials

by approximately 0.5 V and the potentials of the I/0 couples

remained the same. As discussed above, the tetrahedral distortion

observed for [M(diphosphine)₂]²⁺complexes containing diphos-

phine ligands with large chelate bites leads to a lower energy

of the LUMO by decreasing antibonding interactions. This

results in more positive potentials for the II/I couples of

Electrochemical Studies. The substituents on the diphosphine

ligands, the nature of the metal, and the size of the chelate bite

of the diphosphine ligand all have comparable effects on the

11442 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

Figure 3. Drawings of [Ni(dmpe)₂]²⁺, [Ni(depe)₂]²⁺, [Ni(dmpp)₂]²⁺, and [Pt(dmpp)₂]²⁺cations showing the atom-numbering schemes.

Figure 4. Drawings of [Ni(dmpe)₂(CH₃CN)]²⁺and [Ni(dmpp)₂(CH₃CN)]²⁺cations showing the atom-numbering schemes.

Figure 5. Drawings of Ni(dmpp)₂and Pt(dmpp)₂showing the atom-numbering schemes.

complexes containing diphosphine ligands with larger bite sizes.

The potential of the I/0 couple does not depend on the chelate

bite size, because the Ni(I) and Pt(I) complexes are expected

to have a nearly tetrahedral geometry.^30,31

The observation that chelate bite size has a significant effect

on the II/I couple but not on the I/0 couple has some interesting

consequences. Large chelate bites can be used to stabilize the

M(I) oxidation state. For example, [Pd(dppx)₂]⁺(where dppx

is R,R′-bis(diphenylphosphino)-o-xylene) can be observed as a

stable species, as discussed in a previous publication.^22aThe

stabilization of Pd(I) and Pt(I) species would suggest that radical

type chemistry could play important roles for these complexes.

(31) Zotti, G.; Zecchini, S.; Pilloni, G. J. Organomet. Chem. 1983, 246,

61.

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11443

A possible example of this radical chemistry is the reaction of

[Pt(dmpe)₂]⁺(produced by comproportionation of Pt(II) and

Pt(0) complexes) with H₂to form [HPt(dmpe)₂]⁺in benzonitrile,

which was described above. This reaction is more rapid for the

dmpp analogue, as would be expected from a higher concentra-

tion of the M(I) species. Radical chemistry of [M(diphos-

phine)₂]⁺complexes should be favored by large chelate bites,

large substituents, and high temperatures.

mean that the five-coordinate Pt-H and four-coordinate Pt(0)

complexes undergo Pt-P bond cleavage reactions to form four-

coordinate hydrides and three-coordinate Pt(0) intermediates.

These unsaturated species can react with dangling phosphine

ligands from a second intermediate to produce ligand exchange.

Formation of analogous intermediates from the four-coordinate

M(II) complexes is more difficult, and hence, scrambling is not

observed in this case.

Because the reductions of the [M(diphosphine)₂]²⁺complexes

produce large structural distortions, the role that entropy plays

in determining the E_1/2values could be significantly larger than

that expected for complexes in which the structural changes

are not so dramatic. However, the experimental ∆S°_rc(II/0)

values shown in Table 2 are in reasonable agreement with the

values calculated using the method of Hupp and Weaver,^25c

which are also shown in Table 2. The higher experimental values

for ∆S°_rc(II/I) compared to the calculated values for the nickel

complexes may be partly attributed to the loss of a nitrile ligand

upon reduction, which would be expected to increase the entropy

in addition to that associated with the change in charge. These

results indicate that the structural distortions are not leading to

abnormally large entropy contributions. Consequently, the order

of the free energy changes and the enthalpy changes for these

redox couples are the same, as shown in Table 2.

pK_aMeasurements. In the pK_ameasurements discussed

above, all of the nickel and platinum complexes containing

diphosphine ligands with methyl or ethyl substituents can be

referenced back to a common organic base, tetramethylguani-

dine. Therefore, the relative free energies of proton transfer

between the M(0) and M-H complexes for these complexes

are unambiguously determined for these complexes in benzo-

nitrile. Because the acidities of [HNi(dppe)₂]⁺and [HPt(dppe)₂]⁺

are significantly different from those of the other complexes

studied, pyridine and triethylamine, respectively, were used as

bases. The factors which determine the pK_avalues of the

[HM(diphosphine)₂]⁺complexes are, in order of decreasing

importance: the substituents of the diphosphine ligands, the

metal, and the chelate bite size of the diphosphine ligand. The

pK_avalues decrease by 5-10 pK_aunits if the methyl substituents

are replaced with phenyl substituents, by 3-8 units if Pt is

replaced with Ni, and by 0.3-0.7 unit upon changing from a

five- to a six-membered chelate ring.

The proton transfer between Pt and Ni complexes containing

the same diphosphine ligands and between all of the Ni

complexes likely proceeds via M-H-M′ bridges.³²In the case

of two different Pt complexes, the role of hydride exchange

between two Pt metals is less clear. Reactions between

[HPt(L)₂]⁺and Pt(L′)₂complexes (where L and L′ are two

different diphosphine ligands) resulted in mixed [HPt(L)(L′)]⁺

and Pt(L)(L′) complexes. The exchange of diphosphine ligands

permits the determination of pK_avalues for platinum hydride

complexes containing two different diphosphine ligands, but it

is not clear whether the apparent hydride transfer involves the

transfer of a hydride ligand, the transfer of diphosphine ligands,

or both. Studies of the reactions of [HPt(L)₂]⁺with [HPt(L′)₂]⁺

and of Pt(L)₂with Pt(L′)₂indicate that rapid diphosphine ligand

scrambling reactions occur at room temperature for both Pt-H

and Pt(0) species. Similar reactions are not observed for the

analogous Ni(0) or Ni-H complexes at room temperature, and

no diphosphine ligand scrambling is observed for either Ni(II)

or Pt(II) complexes. We have interpreted these observations to

The temperature dependence of the equilibrium constant as

measured by ³¹P NMR spectroscopy for the proton-transfer

reaction between [HNi(dmpp)₂]⁺and Ni(dmpe)₂was studied

to determine if there is a significant entropy contribution to this

reaction. The measured value of ∆S° for this reaction is 6 ( 2

cal/(K mol). This result indicates that there are no large entropy

contributions to the proton-transfer reactions resulting from

differences in ring size of the chelating diphosphine ligands.

Similar studies of the reactions between [HNi(depe)₂]⁺and

Ni(dmpe)₂and between [HNi(dmpp)₂]⁺and Ni(depe)₂gave ∆S°

values of -4 ( 2 and 5 ( 2 cal/(K mol), respectively. These

data suggest that the ∆S° values for the proton-transfer reactions

in our study are small.

Tilset and Parker have assumed that S° values of transition-

metal hydrides and their corresponding radicals are the same,

i.e., S°(MH⁺) ) S°(M⁺).^7cSubstitution of S°(M⁺) for S°(MH⁺)

for the entropy change associated with the proton exchange

reaction (16) leads to the conclusion that ∆S°_{eq 16}should equal

the differences in the ∆S°_rc(I/0) values observed in Table 2 for

ML₂and ML′₂(∆S°_eq) ∆∆S°_rc(I/0)). For the reaction of

16

[Ni(dmpp)₂H]⁺with Ni(depe)₂, the difference between the

∆S°_rc(I/0) values is 5 ( 4 cal/(K mol) (see Table 2) compared

to a ∆S° value of 5 ( 2 cal/(K mol) for the proton exchange

reaction between [HNi(dmpp)₂]⁺and Ni(depe)₂. This agreement

suggests that S° values of M⁺and MH⁺complexes are the same

in our study. Hopefully, variable-temperature measurements of

proton exchange reactions and electron exchange reactions will

be made on other systems so that the validity of the assumption

that S° is the same for a transition-metal hydride (MH) and its

radical (M^•) may be tested more rigorously. It is important to

establish that S° for MH and M^•are the same, because the

accuracy of the M-H bond dissociation energies derived from

the thermochemical cycle shown in Scheme 1 depends on the

validity of this assumption. However, if the M-H bond is

strongly polarized to produce a dipole which collapses upon

H^•removal, it is not clear that the contribution to ∆S° resulting

from the difference between MH and M^•should be zero.

Hydride Transfer Potentials. The reaction of [HM(diphos-

phine)₂]⁺complexes with [M(diphosphine)₂]²⁺complexes re-

sulted in a qualitative ordering of the hydride complexes in terms

of their ability to donate hydride ligands. However, to obtain

quantitative information, it was necessary to use the thermo-

chemical cycle shown in Scheme 2. It can be seen from Table

-

3 that the order of ∆G°_Hvalues, or hydride transfer potentials,

parallels the E_1/2values for the II/0 couples, but the order does

not follow the sequence of pK_avalues. More electron-donating

substituents increase the hydride donor potential with Me > Et

> Ph. For example, [HNi(dmpe)₂]⁺and [HPt(dmpe)₂]⁺are

approximately 14 kcal/mol better hydride donors than [HNi-

(dppe)₂]⁺and [HPt(dppe)₂]⁺, respectively. For the same diphos-

phine ligand, platinum is a better hydride donor than Ni by 10-

15 kcal/mol. A smaller chelate bite also appears to favor hydride

transfer, because the hydride donor abilities of [HNi(dmpe)₂]⁺

and [HPt(dmpe)₂]⁺are approximately 10 kcal/mol greater than

those of [HNi(dmpp)₂]⁺and [HPt(dmpp)₂]⁺, respectively. This

is a direct result of the influence of the chelate bite size on the

(32) Numerous compounds containing M-H-M bridges are known: (a)

Venanzi, L. M. Coord. Chem. ReV. 1982, 43, 251-274. (b) Knobler, C.

B.; Kaesz, H. D.; Minghetti, G.; Bandini, A. L.; Banditelli, G.; Bonati, F.

Inorg. Chem. 1983, 22, 2324-2331.

11444 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

potential of the II/I redox couple. Of the complexes studied,

the best hydride donor is [HPt(dmpe)₂]⁺, a third-row-metal

complex containing a diphosphine ligand with a small chelate

bite and electron-donating methyl groups.

reactions proceed too far to completion to allow direct measure-

ment of the equilibrium constants. An alternative approach,

illustrated for the hydride exchange between [HNi(dmpp)₂]⁺and

[Ni(dmpe)₂]²⁺in benzonitrile (reaction 27), is to use the entropy

changes associated with reactions 25 and 26 to calculate the

entropy change for reaction 27. ∆S°_{eq 25}is the entropy associated

Although a quantitative relationship could not be developed

because of the slowness of the hydride transfer reaction, the

observation that [HNi(depe)₂]⁺and [HNi(dmpp)₂]⁺can transfer

a hydride ligand to N-benzylnicotinamide hexafluorophosphate

indicates that the [HM(diphosphine)₂]⁺complexes studied are

more powerful hydride donors than most NADH analogues.

The [HM(diphosphine)₂]⁺complexes in Table 3 are much

more powerful hydride donors than the CpM(CO)₂(L)H com-

plexes (M ) Mo, W) studied by Sarker and Bruno.¹²[HNi-

(dppe)₂]⁺, the poorest hydride donor in our study, is a stronger

hydride donor by approximately 16 kcal/mol than CpMo(CO)₂-

(PMe₃)H, the best hydride donor in the CpM(CO)₂(L)H series;

[HPt(dmpe)₂]⁺is a stronger hydride donor by 40 kcal/mol. These

data emphasize the important role of the stability of the

unsaturated complexes that result from hydride transfer. For the

CpM(CO)₂(L)H complexes, hydride transfer is accompanied by

coordination of the solvent, and the hydride transfer ability of

these complexes should be strongly solvent dependent. For the

[HM(diphosphine)₂]⁺complexes, stable square-planar com-

plexes are formed. The stability of the square-planar complexes

is more important than the charge on the hydride complexes or

the position of the transition metal in the Periodic Table.

Good hydride acceptors have the opposite characteristics of

good hydride donors. From Table 3, it appears that [Ni(dppe)₂]²⁺

and [Ni(dmpp)₂]²⁺should be the best hydride acceptors based

[HNi(dmpp)₂]⁺+ Ni(dmpe)₂f

[HNi(dmpe)₂]⁺+ Ni(dmpp)₂(25)

[Ni(dmpe)₂]²⁺+ Ni(dmpp)₂f

[Ni(dmpp)₂]²⁺+ Ni(dmpe)₂(26)

[HNi(dmpp)₂]⁺+ [Ni(dmpe)₂]²⁺

f

[HNi(dmpe)₂]⁺+ [Ni(dmpp)₂]²⁺(27)

with proton exchange between [HNi(dmpp)₂]⁺and Ni(dmpe)₂,

for which a value of 6 ( 2 cal/(K mol) was measured. ∆S°_{eq 26}

is the difference of ∆S°_rc(II/0) values for the reaction of

Ni(dmpp)₂with [Ni(dmpe)₂]²⁺(2 ( 2 cal/K mol). The sum of

∆S° contributions from reactions 25 and 26 gives a value of 8

( 4 cal/(K mol) for ∆S°_{eq 27}. This corresponds to a T∆S

contribution of 2.4 kcal/mol to the free energy of reaction 27

at room temperature. From Table 3 the free energy of reaction

27 is 11.6 kcal/mol, which gives an enthalpy contribution to

reaction 27 of 14 kcal/mol. Therefore, [HNi(dmpe)₂]⁺is a better

hydride donor than [HNi(dmpp)₂]⁺because of a larger enthalpic

driving force. The entropic contribution at room temperature is

small and opposite to the enthalpic contribution.

-

on ∆G°_H. To evaluate the ability of these complexes to act as

hydride acceptors, we studied the reaction of [Ni(dppe)₂]²⁺and

[Ni(dmpp)₂]²⁺with hydrogen in dimethylformamide. It was

observed that [Ni(dmpp)₂]²⁺heterolytically cleaves hydrogen

to form [HNi(dmpp)₂]⁺. In this case, the metal complex is a

sufficiently good hydride acceptor that it can cleave the H-H

bond, which illustrates the usefulness of the free energy data

shown in Table 3 for suggesting feasible reactions. However,

[Ni(dppe)₂]²⁺does not react with hydrogen under the same

conditions, which implies there is a kinetic constraint for this

complex.

Using the same approach, ∆S° calculated for the hydride

exchange between [HNi(depe)₂]⁺and [Ni(dmpp)₂]²⁺is 0 ( 4

cal/(K mol). Assuming S° for MH⁺is equal to S° for M⁺and

the same is true for M′H⁺and M′⁺, the entropy change

associated with the hydride transfer between complexes M′H⁺

and M²⁺should be the difference in the ∆S°_rc(II/I) values

associated with complexes M and M′. For the hydride exchange

between [HNi(depe)₂]⁺and [Ni(dmpp)₂]²⁺, this leads to a

predicted ∆S° value of 0 compared to the experimental 0 ( 4

cal/(K mol). In conclusion, the entropy changes associated with

the hydride transfer reactions in this study are small; conse-

quently, the free energies for these reactions are determined

primarily by enthalpy contributions.

In previous studies, [Pd(ttpE)(DMF)]²⁺(where ttpE is bis-

(3-(diethylphosphino)propyl)phenylphosphine and DMF is di-

methylformamide) heterolytically cleaved hydrogen to form a

hydride, but [Pd(etpE)(DMF)]²⁺(where etpE is bis(2-(dieth-

ylphosphino)ethyl)phenylphosphine) did not.³³The only dif-

ference in these complexes is that ttpE has a three-carbon chain

backbone while etpE has a two-carbon chain backbone. This

result also indicates the chelate bite size is important for

heterolytic activation of hydrogen. We believe that this will

prove to be a rather general phenomenon.

From the perspective of predicting the relative hydride donor

or acceptor abilities of complexes in solution, the relative free

energy of the hydride transfer reaction is the most important

thermodynamic parameter. However, because the structural

changes associated with hydride transfer are large (e.g., [HPt-

(depe)₂]⁺is a distorted tetrahedron¹⁵while [Pt(depe)₂]²⁺is

expected to be square planar on the basis of the structure of

[Ni(depe)₂]²⁺), it is of interest to consider the entropy changes

associated with the hydride transfer reactions. Ideally, the

temperature dependence of the equilibrium constant for the

transfer of the hydride ligand between M′H⁺and M²⁺would

provide this information. However, as discussed above, these

A useful approach to understanding the features responsible

for promoting the hydride transfer reaction is to view this

reaction in a stepwise manner, in which the hydride ligand is

first transferred without structural change followed by relaxation

of the tetrahedral M(II) complex to a square-planar complex.

This approach is frequently used for delineating the factors that

promote hydrogen atom transfer.^5aIf the hydride transfer step

occurs without distortion, it will be favored by more basic

substituents on the diphosphine ligands. Following hydride

transfer, the tetrahedral [M(diphosphine)₂]²⁺complexes will

relax to square-planar structures. Because a tetrahedral Pt(II)

complex is less stable than a tetrahedral Ni(II) complex, the

driving force to form a square-planar platinum complex will

be greater than for a nickel complex, and platinum complexes

will be better hydride donors than nickel complexes. Similarly,

complexes with large chelate bites will not be able to obtain a

completely square-planar geometry. As a consequence, the

driving force for hydride transfer will be less for complexes

with large chelate bites.

(33) Wander, S. A.; Miedaner, A.; Noll, B. C.; Barkley, R. M.; DuBois,

D. L. Organometallics 1996, 15, 3360-3373.

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11445

[Ni(dmpe)₂](PF₆)₂‚CH₃NO₂. X-ray-quality crystals of [Ni(dmpe)₂]-

(PF₆)₂were grown from nitromethane/methanol. Data collection was

at 150 K using 30 s scans. Structure solution and refinement proceeded

normally. In addition to the Ni salt, the asymmetric unit contains one

molecule of CH₃NO₂disordered at two sites. One PF₆anion exhibits

rotational disorder.

[Ni(depe)₂](BF₄)₂‚H₂O. Data collection was at 162 K using 30 s

scans. Data beyond 0.75 Å resolution were discarded in the final

refinement, due to weak scattering at high angle. Nickel is positioned

on a crystallographic inversion center. The asymmetric unit is 0.5 Ni

cation, one BF₄anion, and 0.5 water. No disorder is present.

[Ni(dmpp)₂](BF₄)₂. Data collection was at 160 K using 10 s scans.

The asymmetric unit consists of the Ni cation and two BF₄anions.

One dmpp ligand is disordered, showing inversion about the central

carbon atom of the propane backbone.

[Pt(dmpp)₂](PF₆)₂‚0.53CH₃CN‚0.53CH₃OH. Crystals of [Pt(dmpp)₂]-

(PF₆)₂were grown from acetonitrile/methanol at -20 °C. Data

collection was at 141 K using 30 s scans. The asymmetric unit is the

Pt cation, two PF₆anions, and 0.53 molecule each of methanol and

acetonitrile. Both dmpp ligands are disordered, manifested as inversion

about the central carbon of the propane backbone. One PF₆displays

rotational disorder.

[Ni(dmpe)₂(CH₃CN)](PF₆)₂‚CH₃CN. Crystals of [Ni(dmpe)₂(CH₃-

CN)](PF₆)₂were grown from acetonitrile/methanol at -20 °C. Data

collection was at 170 K using 30 s scans. The asymmetric unit is

comprised of the Ni cation, two PF₆anions, and an additional molecule

of acetonitrile. The Ni is out of the plane formed by the four P atoms.

Ni is disordered across this plane. The two Ni sites are separated by

0.58 Å. A molecule of acetonitrile is present on either side of this plane,

and each is within bonding distance (2.28 Å) when Ni is on the same

side, and just out of distance (3.37 Å) when Ni is on the opposite side.

Site occupancy is 0.59 for the major site and 0.41 for the minor site.

[Ni(dmpp)₂(CH₃CN)](BF₄)₂. Crystals of [Ni(dmpp)₂(CH₃CN)](BF₄)₂

were grown from acetonitrile/methanol at -20 °C. Data collection was

at 169 K using 30 s scans over a full sphere of reciprocal space. The

asymmetric unit contains two crystallographically independent sets of

cations and anions. Disorder is present at the central carbon of a single

propane backbone. No additional molecules are present. The system

was examined for higher space group symmetry. No additional

symmetry was found.

Ni(dmpp)₂. Crystals of Ni(dmpp)₂were grown from toluene at -85

°C. Data collection was at 152 K using 90 s scans. The asymmetric

unit is 0.5 molecule, with Ni on a 2-fold rotation axis. There is no

disorder, and there are no additional molecules.

Pt(dmpp)₂. Crystals of Pt(dmpp)₂were grown from toluene at -85

°C. Data collection was at 160 K using 30 s scans. The asymmetric

unit is 0.5 molecule, with Pt on a 2-fold rotation axis. There is no

disorder, and there are no additional molecules.

Electrochemical Studies. All electrochemical experiments were

carried out under an atmosphere of N₂in 0.3 M Bu₄NBF₄in benzonitrile

or 0.3 M Et₄NBF₄in acetonitrile. Cyclic voltammetry experiments were

carried out on a Cypress Systems computer-aided electrolysis system.

The working electrode was a glassy-carbon disk (2 mm diameter), and

the counter electrode was a glassy-carbon rod. A platinum wire

immersed in a permethylferrocene/permethylferrocenium solution was

used as a pseudoreference electrode to fix the potential. Ferrocene was

used as an internal standard, and all potentials are referenced to the

ferrocene/ferrocenium couple. Cell temperatures were maintained with

a NESLAB RTE-110 refigerated bath/circulator containing a 50/50

mixture of ethylene glycol and water.

Syntheses. 1,2-Bis(dimethylphosphino)ethane (dmpe), 1,2-bis(di-

ethylphosphino)ethane (depe), bis(1,5-cyclooctadiene)nickel(0), and

dichloro(1,5-cyclooctadiene)platinum(II) were purchased from Strem

Chemical Co. and used without further purification. 1,3-Bis(dimethyl-

phosphino)propane (dmpp) was purchased from Organometallics, Inc.,

and used without further purification. Benzonitrile, acetonitrile, am-

monium hexafluorophosphate, 1,1,3,3-tetramethylguanidine, triethyl-

amine, pyridine, tetrabutylammonium tetrafluoroborate, and tetraethyl-

ammonium tetrafluoroborate were purchased from Aldrich Chemical

Co. and used as received. Tetrahydrofuran was purchased from Aldrich

Chemical Co. and distilled over Na/benzophenone prior to use. [Ni(CH₃-

Summary and Conclusions

A thermochemical cycle has been used to measure the relative

hydride donor abilities of a series of [HM(diphosphine)₂]⁺

complexes. On the basis of hydride transfer experiments, these

complexes are better hydride donors than CpM(CO)₂(L)H

complexes (where M ) Mo, W and L ) CO, PR₃) and most

NADH analogues. The hydride donor abilities of the [HM-

(diphosphine)₂]⁺complexes (M ) Ni, Pt) in this study and the

CpM′(CO)₂(L)H compounds (M′ ) Mo, W) studied by Sarker

and Bruno span a range of 50 kcal/mol. For the [HM-

(diphosphine)₂]⁺cations, the stability of the square-planar

complexes formed on hydride transfer contributes more to the

hydride donor ability than their position in the Periodic Table.

Structural differences between [HM(diphosphine)₂]⁺, M(diphos-

phine)₂, and [M(diphosphine)₂]²⁺complexes are important for

understanding differences in hydride donor ability. In this study,

we have shown that electrochemical reduction of [M(diphos-

phine)₂]²⁺to M(diphosphine)₂results in a twisting of the

diphosphine ligands about the metal to produce a tetrahedral

structure with shorter M-P bonds. Protonation of the M(diphos-

phine)₂complexes leads to [HM(diphosphine)₂]⁺cations which

have distorted tetrahedral M-P₄structures and M-P bond

distances comparable to [M(diphosphine)₂]²⁺complexes (i.e.,

the bonds lengthen on protonation). Hydride transfer is ac-

companied by a tetrahedral to square-planar structural change

without appreciable change in the M-P bond distance. Increas-

ing the bite size of the diphosphine ligand of the [M(diphos-

phine)₂]²⁺complexes produces a tetrahedral distortion of the

square-planar complexes. This distortion makes these complexes

easier to reduce, causes them to be better hydride acceptors,

and stabilizes the +1 oxidation state. For [HM(diphosphine)₂]⁺

hydrides, the features favoring hydride donation are a third-

row transition metal containing a diphosphine ligand with basic

substituents and a small chelate bite.

Experimental Section

Physical Measurements and General Procedures. ¹H and ³¹P NMR

spectra were recorded on a Varian Unity 300 MHz spectrometer at

299.95 and 121.42 MHz, respectively. ¹H chemical shifts are reported

relative to tetramethylsilane using residual solvent protons as a

secondary reference. ³¹P chemical shifts are reported relative to external

phosphoric acid. For the ³¹P spectra of the platinum complexes, the

spectra appear as a singlet flanked by satellites due to coupling to the

¹⁹⁵Pt (33.8% abundance) isotope. The chemical shift of the singlet is

1

reported followed by J_Pt-Pobtained from the satellites. In the

experiments where accurate, relative integrations of the ³¹P signals were

needed, the repetition rate of the spectrometer was set to 11.96 s.

Infrared spectra were recorded on a Nicolet 510P spectrometer as Nujol

mulls. Absorption experiments were carried out on a Carey 5E UV-

vis spectrometer. Elemental analyses were performed by Schwarzkopf

Laboratories, Woodside, NY, or Galbraith Laboratories, Inc., Knoxville,

TN. All syntheses were carried out using Schlenk and drybox

techniques.

X-ray Diffraction Studies. Single-crystal X-ray data for eight

complexes were collected at low temperature, ca. 150 K, on a Siemens

SMART CCD diffractometer using Mo KR radiation (λ ) 0.710 73

Å). An arbitrary hemisphere of data to 0.68 Å resolution was collected

for each sample. A full sphere was collected for [Ni(dmpp)₂(CH₃CN)]-

(BF₄)₂, which crystallized in triclinic space group P1h. Each data frame

was 0.3° in ω; frames were a correlated scan comprised of two

exposures at half the total scan time. All data were corrected for Lorentz

and polarization effects, as well as for absorption. Structure solution

was by direct methods except for Ni(dmpp)₂, which was solved using

the Patterson function. Anisotropic thermal parameters were refined

for all non-hydrogen atoms. Details for specific compounds are given

in the following paragraphs.

11446 J. Am. Chem. Soc., Vol. 121, No. 49, 1999

Berning et al.

CN)₆](BF₄)₂was prepared according to literature methods.³⁴[Ni(dmpe)₂]-

(BF₄)₂and [Ni(depe)₂](BF₄)₂were prepared by modified literature

procedures,³⁵which are similar to those of [Ni(dmpp)₂](BF₄)₂discussed

below. [Ni(dmpe)₂H](PF₆), [Pt(dmpe)₂](PF₆)₂,³⁵[Pt(depe)₂](PF₆)₂,³⁶

[Pt(dmpe)₂H](PF₆), and [Pt(depe)₂H](PF₆) were prepared according to

the procedures in ref 15. Ni(dmpe)₂,³⁵Ni(depe)₂,³⁷and Pt(dmpe)₂³⁸were

prepared by modified literature procedures, and representative syntheses

are described below. N-Benzylnicotinamide hexafluorophosphate was

also prepared by a modified literature procedure.³⁹

(m, 24 H, P(CH₂CH₃)₂), 1.31 (m, 8 H, PCH₂CH₂P), 1.44 (m, 16 H,

P(CH₂CH₃)₂). ³¹P NMR (toluene-d₈): δ 40.5 (s).

Ni(dmpp)₂. The complex was synthesized using a procedure similar

to that for Ni(dmpe)₂in 80.3% yield. The product can be sublimed

1

under high vacuum at 110 °C. H NMR (toluene-d₈): δ 1.10 (m, 24

H, P(CH₃)₂), 1.30 (m, 8 H, PCH₂CH₂CH₂P), 1.73 (m, 4 H, PCH₂CH₂-

CH₂P). ³¹P NMR (toluene-d₈): δ -19.9 (s).

[Pt(dmpp)₂](PF₆)₂. Pt(COD)Cl₂(0.80 g, 2.14 mmol) was added as

a solid to dmpp (0.71 g, 4.32 mmol) in acetonitrile (75 mL), and the

resulting mixture was stirred at room temperature for 16 h. The solvent

was removed in vacuo to give a white solid. The solid was dissolved

in water (25 mL), and a solution of NH₄PF₆(1.00 g, 6.13 mmol) in

water (20 mL) was added, which resulted in the formation of a white

precipitate. The precipitate was collected by filtration, washed with

water (3 × 10 mL), and dried under vacuum to give the product in

91.2% yield. Anal. Calcd for C₁₄H₃₆F₁₂P₆Pt: C, 20.67; H, 4.46; P, 22.85.

[Ni(dmpp)₂](BF₄)₂. [Ni(CH₃CN)₆][BF₄]₂(0.61 g, 1.22 mmol) was

added as a solid to a solution of dmpp (0.41 g, 2.49 mmol) in degassed

acetonitrile (100 mL) at room temperature. The solution was stirred at

room temperature for 6 h, and the solvent was removed in vacuo to

give a dark red solid. The solid was washed with methanol (3 × 5

mL) and dried overnight under vacuum to give the desired product in

88.0% yield. Anal. Calcd for C₁₄H₃₆B₂F₈NiP₄: C, 29.99; H, 6.47; P,

1

Found: C, 20.68; H, 5.02; P, 22.47. H NMR (nitromethane-d₃): δ

22.10. Found: C, 29.60; H, 6.71; P, 22.13. H NMR (nitromethane-

1.90 (quintet, 24 H, ²J_P-H) 7.2 Hz, ³J_Pt-H) 27.3 Hz, P(CH₃)₂), 2.30

d₃): δ 1.76 (quintet, 24 H, ²J_P-H) 7.8 Hz, splitting 1.9 Hz, P(CH₃)₂),

2.05 (m, 8 H, PCH₂CH₂CH₂P), 2.23 (m, 4 H, PCH₂CH₂CH₂P).³¹P NMR

(nitromethane-d₃): δ -13.6 (s).

(m, 8 H, PCH₂CH₂CH₂P), 2.40 (m, 4 H, PCH₂CH₂CH₂P). ³¹P NMR

1

(nitromethane-d₃): δ -29.6 (s, J_Pt-P) 2079 Hz).

[Pt(dmpp)₂H](PF₆). Sodium borohydride on basic alumina (1.16 g

of material with 10% NaBH₄content, 3.06 mmol) was added as a solid

to a solution of [Pt(dmpp)₂](PF₆)₂(1.00 g, 1.23 mmol) in acetonitrile

(50 mL). The resulting mixture was stirred at room temperature

overnight. The alumina was removed by filtration, and the solvent was

removed from the filtrate by applying a vacuum. The solid that formed

was washed with ethanol (3 × 5 mL) to remove excess sodium

borohydride and dried overnight under vacuum to give the product in

72.1% yield. Anal. Calcd for C₁₄H₃₇F₆P₅Pt: C, 25.11; H, 5.57. Found:

C, 25.25; H, 5.56. IR (Nujol): ν_Pt-H2072 cm^-1. ¹H NMR (acetonitrile-

d₃): δ -12.55 (quintet, 1 H, ¹J_Pt-H) 642 Hz, ²J_P-H) 29 Hz, Pt-H).

[Ni(depe)₂H](PF₆)₂. A solution of depe (1.05 g, 5.09 mmol) in THF

(125 mL) was cooled to -78 °C, and Ni(COD)₂(0.70 g, 2.54 mmol,

COD ) 1,5-cyclooctadiene) was added as a solid. The solution was

warmed to room temperature and stirred for an additional 16 h to give

a clear, yellow solution of Ni(depe)₂. Ammonium hexafluorophosphate

(0.85 g, 5.21 mmol) in THF (50 mL) was added, which resulted in the

formation of a precipitate. The volume was reduced to approximately

10 mL in vacuo, and the suspension was filtered. The resulting solid

was washed with water (3 × 10 mL) followed by diethyl ether (3 ×

10 mL) and dried overnight in vacuo to give the product in 93.5%

yield. Anal. Calcd for C₂₀H₄₉F₆P₅Ni: C, 38.95; H, 8.01. Found: C,

39.16; H, 7.89. IR (Nujol): ν_Ni-H1924 cm^-1. ¹H NMR (nitromethane-

d₃): δ -14.2 (s, 1 H, Ni-H), 1.12 (m, 24 H, P(CH₂CH₃)₂), 1.71-1.91

(m, 24 H, PCH₂CH₂P, P(CH₂CH₃)₂). ³¹P NMR (nitromethane-d₃): δ

46.2 (s).

1

³¹P NMR (acetonitrile-d₃): δ -54.4 (s, J_Pt-P) 2295 Hz).

Pt(dmpe)₂. Sodium naphthalenide (0.1 M) in THF was added

dropwise to a suspension of [Pt(dmpe)₂][PF₆]₂(0.96 g, 1.22 mmol) in

THF (100 mL) until the green color of the sodium naphthalenide

persisted. The solution was stirred at room temperature for 2 h, and

the solvent was removed by applying a vacuum. The resulting solid

was washed with acetonitrile (3 × 10 mL) and dried overnight under

a vacuum to give the product in 87.0% yield. The product can be

sublimed under high vacuum at 110 °C. Anal. Calcd for C₁₂H₃₂Pt: C,

[Ni(dmpp)₂H](PF₆). The complex was synthesized using a procedure

similar to that for [Ni(depe)₂H](PF₆) in 94.1% yield. Anal. Calcd for

C₁₄H₃₇F₆P₅Ni: C, 31.55; H, 7.01. Found: C, 31.56; H, 7.00. IR

(Nujol): ν_Ni-H1940 cm^-1. ¹H NMR (nitromethane-d₃): δ -14.3 (s, 1

H, Ni-H), 1.43 (m, 24 H, P(CH₃)₂), 1.72 (m, 8 H, PCH₂CH₂CH₂P),

2.00 (m, 4 H, PCH₂CH₂CH₂P). ³¹P NMR (nitromethane-d₃): δ -15.7

(s).

1

29.10; H, 6.51. Found: C, 29.03; H, 6.77. H NMR (toluene-d₈): δ

3

1.32 (m, 8 H, PCH₂CH₂P), 1.42 (m, 24 H, J_Pt-H) 22 Hz, P(CH₃)₂).

1

³¹P NMR (toluene-d₈): δ -12.6 (s, J_Pt-P) 3714 Hz).

Ni(dmpe)₂. A solution of dmpe (1.10 g, 7.33 mmol) in THF (50

mL) was cooled to -78 °C, and Ni(COD)₂(1.01 g, 3.67 mmol) was

added as a solid. The solution was warmed to room temperature and

stirred for an additional 16 h to give a clear, yellow solution of

Ni(dmpe)₂. The solvent was removed in vacuo, and the product was

washed with acetonitrile (3 × 10 mL). The solid was then dried under

vacuum overnight to give the product in 72.9% yield. The product can

be sublimed under high vacuum at 110 °C. ¹H NMR (toluene-d₈): 1.24

(s, 24 H, P(CH₃)₂), 1.38 (m, 8 H, PCH₂CH₂P). ³¹P NMR (toluene-d₈):

15.0 (s).

Pt(depe)₂. The complex was synthesized using a procedure similar

to that for Pt(dmpe)₂in 59.4% yield. H NMR (toluene-d₈): δ 0.99

1

(m, 24 H, P(CH₂CH₃)₂), 1.24 (m 8 H, PCH₂CH₂P), 1.46 (m, 16 H,

1

P(CH₂CH₃)₂). ³¹P NMR (toluene-d₈): δ 21.5 (s, J_Pt-P) 3614 Hz).

Pt(dmpp)₂. The complex was synthesized using a procedure similar

to that for Pt(dmpe)₂in 58.2% yield. The product can be sublimed

under high vacuum at 110 °C. Anal. Calcd for C₁₄H₃₆Pt: C, 32.13; H,

6.93. Found: C, 32.16; H, 7.08. ¹H NMR (toluene-d₈): δ 1.47 (m, 24

H, ³J_Pt-H) 21 Hz, P(CH₃)₂), 1.53 (m, 8 H, PCH₂CH₂CH₂P), 1.87 (m,

1

4 H, PCH₂CH₂CH₂P). ³¹P NMR (toluene-d₈): δ -52.1 (s, J_Pt-P

)

Ni(depe)₂. The complex was synthesized using a procedure similar

3661 Hz).

1

to that for Ni(dmpe)₂in 78.7% yield. H NMR (toluene-d₈): δ 1.02

Relative Hydride Transfer Potentials (Reaction 14). These

reactions were performed by adding 20-30 mg of [M(L)₂]²⁺(M )

Ni, Pt; L ) dmpe, depe, dmpp) and 1 equiv of [M(L)₂H]⁺(M ) Ni,

Pt; L ) dmpe, depe, dmpp) to NMR tubes in a drybox followed by

0.7 mL of CD₃CN. The solutions were allowed 24 h to reach

equilibrium. The relative concentration of each species present was

calculated by integration of the ³¹P NMR signals.

Relative Proton-Transfer Potentials (Reaction 16). These experi-

ments were performed by preparing 1.4 × 10^-2M solutions of the

metal hydride and M(0) complexes in benzonitrile in a drybox. Various

aliquots were then added to NMR tubes to give a constant volume of

0.7 mL. The mixtures were allowed 24 h to reach equilibrium, and the

relative concentration of each species present was calculated by

integration of the ³¹P NMR signals.

(34) Hathaway, B. J.; Holah, D. G.; Underhill, A. E. J. Chem. Soc. 1962,

2444.

(35) (a) Hope, E. G.; Levason, W.; Powell, N. A. Inorg. Chim. Acta

1986, 115, 187. (b) von Kozelka, J.; Ludwig, W. HelV. Chim. Acta 1983,

66, 902. (c) Gulliver, D. J.; Levason, W.; Smith, K. G. J. Chem. Soc., Dalton

Trans. 1981, 2153.

(36) (a) Ittel, S. D. Inorg. Synth. 1990, 28, 98. (b) Ittel, S. D. Inorg.

Synth. 1977, 17, 117. (c) Tolman, C. A.; Seidel, W. C.; Gosser, L. W. J.

Am. Chem. Soc. 1974, 96, 53.

(37) (a) Mastrorilli, P.; Moro, G.; Nobile, C. F.; Latronico, M. Inorg.

Chim. Acta 1992, 192, 183. (b) Gavero, G.; Frigo, A.; Turco, A. Gazz.

Chim. Ital. 1974, 104, 869.

(38) (a) Chaloner, P. A.; Broadwood-Strong, G. T. L. J. Chem. Soc.,

Dalton Trans. 1996, 1039. (b) Nuzzo, R. G.; McCarthy, T. J.; Whitesides,

G. M. Inorg. Chem. 1981, 20, 1312.

(39) (a) Ostovic, D.; Han Lee, I.-S.; Roberts, R. M. G.; Kreevoy, M. M.

J. Org. Chem. 1985, 50, 4206. (b) Karrer, P.; Stare, F. J. HelV. Chim. Acta

1937, 20, 418.

Equilibrium Constants for Diphosphine Ligand Transfer (Reac-

tions 17 and 18). These experiments were performed by preparing 1.4

× 10^-2M solutions of the platinum hydride and platinum(0) complexes

[HM(diphosphine)₂]PF₆Complexes

J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11447

in benzonitrile in a drybox. Various aliquots of different hydride

complexes (reaction 17) or different Pt(0) complexes (reaction 18) were

then added to NMR tubes to give a constant volume of 0.7 mL. The

mixtures were allowed 24 h to reach equilibrium, and the concentration

of each species present was calculated by integration of the ³¹P NMR

signals.

of [M(diphosphine)₂]X₂and M(diphosphine)₂complexes in benzonitrile

and bubbling hydrogen through the resulting solutions.

Acknowledgment. This work was supported by the U.S.

Department of Energy, Office of Science, Chemical Sciences

Division.

pK_aMeasurements. These experiments were performed by preparing

1.4 × 10^-2M samples of the metal hydride complexes in benzonitrile

in a drybox. Various aliquots were then added to NMR tubes, followed

by the appropriate base (tetramethylguanidine, pyridine, or triethyl-

amine), to give a constant volume of 0.7 mL. The mixtures were

equilibrated for 20 min, and the relative concentrations of the species

present were calculated by integration of the ³¹P NMR signals.

Hydrogen Activation. These reactions were performed by dissolving

20-30 mg of [M(L)₂]²⁺(M ) Ni, Pt; L ) dmpe, depe, dmpp) in

dimethylformamide-d₇(0.7 mL) and bubbling hydrogen through the

solution. The reaction was monitored by ³¹P and ¹H NMR spectroscopy.

Similar experiments were performed by dissolving equimolar amounts

Supporting Information Available: Tables of crystal data,

data collection parameters, structure solution and refinement,

atomic coordinates and equivalent isotropic displacement pa-

rameters, bond lengths, bond angles, anisotropic thermal

parameters, and hydrogen coordinates and isotropic displacement

parameters for [Ni(dmpe)₂](PF₆)₂, [Ni(depe)₂](BF₄)₂, [Ni(dmpp)₂]-

(BF₄)₂, [Pt(dmpp)₂](PF₆)₂, [Ni(dmpe)₂(CH₃CN)](PF₆)₂, [Ni-

(dmpp)₂(CH₃CN)](BF₄)₂, Ni(dmpp)₂, and Pt(dmpp)₂. This ma-

terialisavailablefreeofchargeviatheInternetathttp://pubs.acs.org.

JA991888Y

Article Doi

Relative hydride, proton, and hydrogen atom transfer abilities of [HM(diphosphine)2]PF6 complexes (M = Pt, Ni)

DOI: 10.1021/ja991888y

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja991888y

Full text of DOI:10.1021/ja991888y

Products guided by the article

R&D Labs maybe for 491838-79-0

Relevant to this article

Hot Product