Comprehensive thermodynamic characterization of the metal-hydrogen bond in a series of cobalt-hydride complexes

Article abstract of DOI:10.1021/ja0122804

A detailed structural and thermodynamic study of a series of cobalt-hydride complexes is reported. This includes structural studies of [H₂Co(dppe)₂]⁺, HCo(dppe)₂, [HCo(dppe)₂(CH₃CN)]⁺, and [Co(dppe)₂-(CH₃CN)]²⁺, where dppe = bis(diphenylphosphino)ethane. Equilibrium measurements are reported for one hydride- and two proton-transfer reactions. These measurements and the determinations of various electrochemical potentials were used to determine 11 of 12 possible homolytic and heterolytic solution Co-H bond dissociation free energies of [H₂Co(dppe)₂]⁺ and its monohydride derivatives. These values provide a useful framework for understanding observed and potential reactions of these complexes. These reactions include the disproportionation of [HCo(dppe)₂]⁺ to form [Co(dppe)₂]⁺ and [H₂Co(dppe)₂]⁺, the reaction of [Co(dppe)₂]⁺ with H₂, the protonation and deprotonation reactions of the various hydride species, and the relative ability of the hydride complexes to act as hydride donors.

Full text of DOI:10.1021/ja0122804

Published on Web 03/01/2002
Comprehensive Thermodynamic Characterization of the
Metal-Hydrogen Bond in a Series of Cobalt-Hydride
Complexes
Rebecca Ciancanelli,† Bruce C. Noll,† Daniel L. DuBois,*,‡ and
M. Rakowski DuBois*,†
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Colorado,
Boulder, Colorado 80309, and National Renewable Energy Laboratory, 1617 Cole BouleVard,
Golden, Colorado 80401
Received October 1, 2001. Revised Manuscript Received January 14, 2002
Abstract: A detailed structural and thermodynamic study of a series of cobalt-hydride complexes is reported.
This includes structural studies of [H2Co(dppe)2]+, HCo(dppe)2, [HCo(dppe)2(CH3CN)]+, and [Co(dppe)2-
(CH3CN)]2+, where dppe ) bis(diphenylphosphino)ethane. Equilibrium measurements are reported for one
hydride- and two proton-transfer reactions. These measurements and the determinations of various
electrochemical potentials were used to determine 11 of 12 possible homolytic and heterolytic solution
Co-H bond dissociation free energies of [H2Co(dppe)2]+ and its monohydride derivatives. These values
provide a useful framework for understanding observed and potential reactions of these complexes. These
reactions include the disproportionation of [HCo(dppe)2]+ to form [Co(dppe)2]+ and [H2Co(dppe)2]+, the
reaction of [Co(dppe)2]+ with H2, the protonation and deprotonation reactions of the various hydride species,
and the relative ability of the hydride complexes to act as hydride donors.
Scheme 1
Introduction
Transition metal hydrides are important intermediates in a
wide range of catalytic reactions, and a more complete
understanding of the M-H bond is important to the rational
development of new catalysts and new catalytic processes. The
metal-hydride bond can be cleaved in three different ways -
by loss of a proton, a hydrogen atom, or a hydride anion,
Scheme 1. Any of the reactions or combination of reactions
shown in the scheme may be important for both stoichiometric
and catalytic reactions.
Quantitative information on the energetics of these three
possible reaction pathways can provide insights into the
chemical reactivity and physical properties of these complexes,
and can be quite useful in selecting appropriate conditions to
carry out a desired chemical reaction. For example, the
stereochemistry for the addition of metal hydrides to activated
olefins is dependent on whether the hydride or its conjugate
base is undergoing the reaction.1 Similarly the catalytic hydro-
formylation of alkenes or hydrogenation of ketones involve a
metal-hydride species,2a,b but the conjugate base of a hydride
is the active catalyst in certain C-C bond cleavage reactions.2c
The homolytic bond dissociation energies of metal hydrides and
hydroxides are critical to our understanding of C-H bond
activation reactions by metal radicals and metal-oxo complexes.3
Finally, the hydride donor abilities of metal hydrides have been
used to predict the heterolytic activation of hydrogen4 and the
reduction of carbonyl complexes to formyl complexes.5
Most previous thermodynamic studies of transition metal
hydrides have focused on a single type of bond cleavage
reaction.6-8 However, all three reactions shown in Scheme 1
are related to each other by the electron-transfer reactions
indicated.9 As one traverses Scheme 1 from left to right, the
(3) (a) Wei, M.; Wayland, B. B. Organometallics 1996, 15, 4681-4683 and
references therein. (b) Mayer, J. M. Acc. Chem. Res. 1998, 31, 441-450.
(4) (a) Berning, D. E.; Noll, B. C.; DuBois, D. L. J. Am. Chem. Soc. 1999,
121, 11432-11447. (b) Curtis, C. J.; Miedaner, A.; Ellis, W. W.; DuBois,
D. L. J. Am. Chem. Soc. 2002, 124, ASAP (Feb. 21, 2002).
(5) Ellis, W. W.; Miedaner, A.; Curtis, C. J.; Gibson, D. H.; DuBois, D. L. J.
Am. Chem. Soc. 2002, 124, in press.
(6) Some studies dealing primarily with pKa measurements are: (a) Kristja´n-
sdo´ttir, S. S.; Norton, J. R. In Transition Metal Hydrides: 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. IBID 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. (h) Abdur-Rashid, K.; Fong, T. P.; Greaves, B.;
Gusev, D. G.; Hinman, J. G.; Landau, S. E.; Lough, A. J.; Morris, R. H.
J. Am. Chem. Soc. 2000, 122, 9155-9171.
† University of Colorado.
‡ National Renewable Energy Laboratory.
(1) (a) Pattenden, G. Chem. Soc. ReV. 1988, 17, 361. (b) Schrauzer, G. N.;
Windgassen, R. J. J. Am. Chem. Soc. 1967, 89, 1999.
(2) (a) Kollar, K.; Sandor, P.; Szalontai, G.; Heil, B. J. Organomet. Chem.
1990, 393, 153. (b) Linn, D.; Halpern, J. J. Am. Chem. Soc. 1987, 109,
2969-2974. (c) Hou, Z.; Koizumi, T.; Fujita, A.; Yamazaki, H.; Wakatsuki,
Y. J. Am. Chem. Soc. 2001, 123, 5812-5813.
9
2984 VOL. 124, NO. 12, 2002 J. AM. CHEM. SOC.
10.1021/ja0122804 CCC: $22.00 © 2002 American Chemical Society

Characterization of the Metal−Hydrogen Bond
A R T I C L E S
metal complexes are sequentially reduced by one electron, and
the hydride ion is sequentially oxidized. We have recently
measured the free energies associated with these three bond
cleavage reactions for a series of [HNi(diphosphine)2]+ com-
plexes.3,4 In this paper we demonstrate that thermochemical
cycles similar to the one shown in Scheme 1 can be extended
to dihydride systems to provide a nearly comprehensive
thermodynamic description of these systems as well. By using
[H2Co(dppe)2]+ as an example of a dihydride system, the set
of reactions shown in Scheme 1 can be expanded to include
the effects of oxidation on the different bond cleavage reactions
of monohydride species. In addition, sequential bond cleavage
reactions of dihydrides can be studied. These results provide
the most comprehensive thermodynamic description developed
to date for a transition metal-hydride system.
equivalent phosphorus atoms. The equivalence of the four
phosphorus atoms is also supported by the 31P NMR spectrum
which exhibits a single broad resonance at 72.9 ppm. As
discussed below, the results of an X-ray diffraction study of
this complex indicate a distorted trigonal bipyramidal structure
with the hydride trans to one phosphorus atom and cis to three
others. The apparent equivalence of the phosphorus atoms in
the NMR spectra is attributed to a fluxional process. Fluxionality
is common in five-coordinate d8 complexes.14
The hydride ligand of [HCo(dppe)2(CH3CN)]2+ also exhibits
a broad quintet resonance (-21.1 ppm) in the 1H NMR
spectrum. In this case, the equivalence of the four phosphorus
atoms is a result of the hydride ligand occupying a position
that is cis to the four phosphorus atoms in a nonfluxional
octahedral complex. Again a single broad resonance is observed
in the 31P NMR spectrum consistent with four equivalent nuclei.
The dihydride complex [H2Co(dppe)2]+ has been assigned
as the cis isomer on the basis of the NMR spectra. The resonance
for the hydrides appears as two broad and overlapping quartets
at -13.2 ppm in the 1H NMR spectrum, and two broad
resonances at 69.7 and 80.5 ppm are observed in the 31P NMR
spectrum. The 31P NMR resonances for all the cobalt complexes
are quite broad, and this is attributed to the quadrupole of the
Results
Synthesis and Spectral Characterization of Co Complexes.
The series of complexes HCo(dppe)2, [HCo(dppe)2]+, and [HCo-
(dppe)2(CH3CN)]2+ 11,12 (where dppe ) bis(diphenylphosphino)-
ethane) is particularly attractive for obtaining information on
the effects of oxidation on the thermodynamics of the M-H
bond, because they represent the sequential oxidation from Co-
(I) to Co(III). In addition, [Co(dppe)2(CH3CN)]2+ and [H2Co-
(dppe)2]+ are also known12,13 and provide the opportunity for
studying successive proton, hydrogen atom, and hydride-transfer
reactions. The complexes were all synthesized by modifications
of previous synthetic methods as described in the Experimental
Section. A more complete spectral characterization is provided
for the products, since several of the spectroscopic techniques
were not routinely available when these complexes were first
reported.
7
59Co nucleus which has a spin of /2.
The paramagnetic Co(II) hydride [HCo(dppe)2]+ has been
characterized by EPR spectroscopy. The spectrum, recorded in
acetonitrile at room temperature, exhibited a doublet with a g
value of 2.1 and a hyperfine coupling of 80 G. These results
are consistent with the low-spin magnetic moment of 2.2 µB
reported previously for this complex.12 The data are similar to
those observed for previously characterized trigonal bipyramidal
complexes such as [Co(PP3)(CH3CN)]2+ (g ) 2.11, A ) 90
G)15 (where PP3 is tris(diphenylphosphinoethyl)phosphine),
[HCo(PP3)](PF6), 2.01 µB,16 and HCo(CP3)PEt3](BPh4) (where
CP3 is CH3C(CH2PPh2)3, 2.15 µB, and g ) 2.20.17
All of the diamagnetic hydride complexes have been char-
acterized by 1H and 31P NMR spectroscopy. The hydride region
of the 1H NMR spectrum of HCo(dppe)2 exhibits a broad quintet
at -14.9 ppm. The quintet pattern arises from coupling to four
Structural Studies. Single crystals of HCo(dppe)2 and the
cations [Co(dppe)2(CH3CN)]2+, [HCo(dppe)2(CH3CN)]2+, and
[H2Co(dppe)2]+ were characterized by X-ray diffraction, and
perspective drawings of the neutral and cationic cobalt structures
are shown in Figure 1. Tables 1 and 2 contain selected bond
lengths and bond angles, respectively, for these complexes. It
can be seen from Figure 1a that HCo(dppe)2 is best described
as a distorted trigonal bipyramidal complex with one phosphorus
atom and the hydride occupying axial positions. The H1-Co-
P1 bond angle for this complex is 163.2°. The deviation from
180° is caused by the small P1-Co-P2 angle (87.0°) typical
of a chelating dppe ligand. The average of the three angles
between the axial hydrogen atom and the three equatorial phos-
phorus atoms is 82.2°, significantly less than 90°. This bending
of cis ligands toward hydrogen is expected for both steric and
electronic reasons.17 The four Co-P bond distances (2.12-2.16
Å, av ) 2.15 Å) and the Co-H bond distance of1.46 Å are
similar to values reported for other Co(I) hydrides.15,16
(7) Some studies dealing primarily with homolytic M-H bond cleavage are:
(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, 110, 7701-7715. (d) Tilset, M.; Parker, V. D. J. Am. Chem. Soc.
1989, 111, 6711-6717. (e) IBID 1990, 112, 2843. (f) Parker, V. D.;
Handoo, K. L.; Roness, F.; Tilset, M. J. Am. Chem. Soc. 1991, 113, 7493-
7498. (g) Ryan, B. O.; Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990,
112, 2618-2626.
(8) Some studies dealing primarily with the thermodynamics and kinetics of
hydride transfer are: (a) Sarker, N.; Bruno, J. W. J. Am. Chem. Soc. 1999,
121, 2174-2180. (b) Labinger, J. A. In Transition Metal Hydrides: Recent
AdVances in Theory and Experiment; Dedieu, A., Ed.; VCH: New York,
1991; pp 361-379. (c) Labinger, J. A.; Komadina, K. H. J. Organomet.
Chem. 1978, 155, C25-C28. (d) Kao, S. C.; Spillett, C. T.; Ash, C.; Lusk,
R.; Park, Y. K.; Darensbourg, M. Y. Organometallics 1985, 4, 83-91. (e)
Kao, S. C.; Gaus, P. L.; Youngdahl, K.; Darensbourg, M. Y. Organome-
tallics 1984, 3, 1601-1603. (f) Gaus, P. L.; Kao, S. C.; Youngdahl, K.;
Darensbourg, M. Y. J. Am. Chem. Soc. 1985, 107, 2428-2434. (g) Kinney,
R. J.; Jones, W. D.; Bergman, R. G. J. Am. Chem. Soc. 1978, 100, 7902-
7915. (h) Martin, B. D.; Warner, K. E.; Norton, J. R. J. Am. Chem. Soc.
1986, 108, 33-39. (i) Hembre, R. T.; McQueen, J. S. Angew. Chem., Int.
Ed. Engl. 1997, 36, 65-67. (j) Hembre, R. T.; McQueen, S. J. Am. Chem.
Soc. 1994, 116, 2141-2142. (k) Cheng, T.-Y.; Brunschwig, B. S.; Bullock,
R. M. J. Am. Chem. Soc. 1998, 120, 13121-13137.
(9) These relationships are clearly presented in: Wayner, D. D. M.; Parker,
V. D. Acc. Chem. Res. 1993, 26, 287-294.
(14) English, A. D.; Ittel, S. D.; Tolman, C. A.; Meakin, P.; Jesson, J. P. J. Am.
Chem. Soc. 1977, 99, 117-120.
(10) Berning, D. E.; Miedaner, A.; Curtis, C. J.; Noll, B. C.; Rakowski DuBois,
M.; DuBois, D. L. Organometallics 2001, 20, 1832-1839.
(15) (a) Bianchini, C.; Innocenti, P.; Meli, A.; Peruzzini, M.; Zanobini, F.
Organometallics 1990, 9, 2514-2522. (b) Orlandini, A.; Sacconi, L. Cryst.
Struct. Commun. 1975, 4, 157-161.
(11) (a) Sacco, A.; Ugo, R. J. Chem. Soc. 1964, 3274-3278. (b) Holah, D. G.;
Hughes, A. N.; Maciaszek, S.; Magnuson, V. R.; Parker, K. O. Inorg. Chem.
1985, 24, 3956-3962. (c) Bianco, V. D.; Doronzo, S.; Gallo, N. Inorg.
Synth. 1980, 20, 206-208.
(16) Bianchini, C.; Masi, D.; Mealli, C.; Sabat, M. Gazz. Chim. Ital. 1986, 116,
201-206.
(12) Pilloni, G.; Schiavon, G.; Zotti, G.; Zecchin, S. J. Organomet. Chem. 1977,
134, 305-318.
(17) (a) Elian, M.; Hoffmann, R. Inorg. Chem. 1975, 14, 1058-1076. (b)
Wander, S. A.; Miedaner, A.; Noll, B. C.; Barkley, R. M.; DuBois, D. L.
Organometallics 1996, 15, 3360-3373.
(13) DuBois, D. L.; Miedaner, A. Inorg. Chem. 1986, 25, 4642-4650.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 2985

A R T I C L E S
Ciancanelli et al.
Figure 1. Ortep plots of (a) HCo(dppe)2; (b) [HCo(dppe)2(CH3CN)]2+; (c) [H2(Co(dppe)2]+; (d) [Co(dppe)2(CH3CN)]2+. Thermal ellipsoids are drawn at
50% probability.
Table 1. Selected Bond Distances (Å)
compd
M−P(1)
M−P(2)
M−P(3)
M−P(4)
M−N
M−H
HCo(dppe)2
[HCo(dppe)2CH3CN][PF6]2
[H2Co(dppe)2][BF4]
2.1519(5)
2.278(2)
2.2138(10)
2.1242(5)
2.270(2)
2.1793(10)
2.1623(5)
2.295(2)
2.2158(10)
2.1523(5)
2.272(2)
2.1687(10)
1.46(3)
1.34(6)
1.42(3)
1.47(4)
1.929(8)
[Co(dppe)2CH3CN][BF4]2
2.2843(14)
2.2988(15)
2.3376(15)
2.3037(14)
2.017(4)
The [HCo(dppe)2(CH3CN)]2+ cation (Figure 1b) is an octa-
hedron with trans hydride and acetonitrile ligands. The H1-
Co-N1 bond angle of 173° is nearly linear, and the cis
phosphorus ligands again bend away from the nitrogen atom
of acetonitrile (average N-Co-P ) 93.4°) toward the hydride
ligand. The average Co-P bond distance (2.28 Å) is signifi-
cantly longer than that observed for HCo(dppe)2 discussed
above, and the Co-N distance of 1.93 Å is less than that
observed for [Co(dppe)2(CH3CN)]2+ (2.02 Å) discussed below.
The cationic dihydride complex, [H2Co(dppe)2]+ (Figure 1c),
has a distorted cis octahedral structure. The P-Co-P angles
that are constrained by the chelate rings have angles near 90°
(89.2° average), but the four remaining angles deviate signifi-
cantly from the 180° and 90° expected for an ideal octahedral
complex. This distortion is again a result of the movement of
the phosphorus atoms toward the hydride ligands. The two
Co-P bond distances for the phosphorus atoms that are trans
to the hydride ligands are slightly longer than the Co-P
distances for the two trans phosphorus atoms. The average
Co-P bond distance for this complex (2.20 Å) is intermediate
between those of HCo(dppe)2 and [HCo(dppe)2(CH3CN)]2+
.
The [Co(dppe)2(CH3CN)]2+ cation (Figure 1d) is best de-
scribed as a square pyramid in which acetonitrile occupies the
apex. Three of the four N-Co-P bond angles are very close
9
2986 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Characterization of the Metal−Hydrogen Bond
Table 2. Selected Bond Angles (deg)
A R T I C L E S
HCo(dppe)2
P(2)-Co-P(1)
P(1)-Co-P(4)
P(1)-Co-P(3)
86.98(2)
107.72(2)
105.67(2)
P(2)-Co-P(4)
P(2)-Co-P(3)
P(4)-Co-P(3)
137.95(2)
P(2)-Co-H(1)
78.1(11)
79.2(11)
P(1)-Co-H(1)
P(3)-Co-H(1)
163.2(11)
89.2(10)
123.41(2)
91.06(2)
P(4)-Co-H(1)
[HCo(dppe)2CH3CN]2+
N(1)-Co-P(2)
P(2)-Co-P(4)
P(2)-Co-P(1)
N(1)-Co-P(3)
88.9(2)
173.76(10)
83.68(8)
93.5(2)
N(1)-Co-P(4)
N(1)-Co-P(1)
P(4)-Co-P(1)
P(2)-Co-P(3)
97.3(2)
93.8(2)
95.57(9)
97.42(8)
P(4)-Co-P(3)
82.54(9)
173(3)
85(3)
P(1)-Co-P(3)
P(2)-Co-H(1)
P(1)-Co-H(1)
172.61(10)
89(3)
93(3)
N(1)-Co-H(1)
P(4)-Co-H(1)
P(3)-Co-H(1)
80(3)
[H2Co(dppe)2]+
P(4)-Co-P(2)
P(2)-Co-P(1)
P(2)-Co-P(3)
P(4)-Co-H(1)
152.89(4)
89.80(2)
109.57(4)
80.9(11)
P(4)-Co-P(1)
P(4)-Co-P(3)
P(1)-Co-P(3)
P(2)-Co-H(1)
107.06(4)
88.72(4)
99.90(4)
79.2(11)
P(1)-Co-H(1)
P(4)-Co-H(2)
P(1)-Co-H(2)
H(1)-Co-H(2)
86.2(11)
75.2(14)
170.8(14)
85.3(18)
P(3)-Co-H(1)
P(2)-Co-H(2)
P(3)-Co-H(2)
169.1(11)
85.0(14)
89.0(14)
[Co(dppe)2CH3CN]2+
N(1)-Co-P(1)
P(1)-Co-P(2)
P(1)-Co-P(4)
89.89(12)
82.70(5)
96.32(5)
N(1)-Co-P(2)
N(1)-Co-P(4)
P(2)-Co-P(4)
91.33(12)
88.34(12)
178.97(6)
N(1)-Co-P(3)
P(2)-Co-P(3)
105.52(12)
97.18(5)
P(1)-Co-P(3)
P(4)-Co-P(3)
164.58(6)
83.85(5)
to 90° with the fourth angle being significantly larger at 105.5°.
The Co-N bond length of 2.02 Å is 0.09 Å longer than that
for [HCo(dppe)2(CH3CN)]2+. This is attributed to the unpaired
electron occupying an orbital with significant Co-N antibonding
character for [Co(dppe)2(CH3CN)]2+. The average Co-P bond
length (2.30 Å) is the longest of the complexes studied and is
consistent with an increase in the Co-P bond length as the
charge on the complex increases, that is, HCo(dppe)2 (2.15 Å)
< [H2Co(dppe)2]+ (2.20 Å) < [HCo(dppe)2(CH3CN)]2+ (2.28
Å) < [Co(dppe)2(CH3CN)]2+ (2.30 Å). This trend has been
observed for a number of metal-phosphine complexes.3,10,18 The
trend would be consistent with either increased π back-bonding
or increased M-L bond covalency as the positive charge
decreases.
Electrochemical Studies. Electrochemical studies of [Co-
(dppe)2(CH3CN)]2+ and HCo(dppe)2 have been described previ-
ously.12 However, these studies were performed in different
solvents or solvent mixtures with different reference and
working electrodes. To obtain accurate thermodynamic data, it
is important to obtain all measurements under the same
conditions. Therefore, we have repeated the cyclic voltammo-
grams using acetonitrile and benzonitrile as solvents, glassy
carbon as the working electrode, and the ferrocene/ferrocenium
couple as an internal reference.19
A cyclic voltammogram of [Co(dppe)2(CH3CN)]2+ in aceto-
nitrile is shown in Figure 2, trace a. It consists of three
reversible, diffusion-controlled, one-electron reductions at -0.70,
-1.56, and -2.03 V, assigned to the Co(II/I), Co(I/0), and Co-
(0/-I) couples as reported previously. The separation between
the cathodic and anodic peaks of each wave is near the 60 mV
value expected for a reversible one-electron process. Because
of the reversible nature of these one-electron processes, the
thermodynamic data based on the values for these couples
should be very reliable. As discussed above, the [Co(dppe)2-
(CH3CN)]2+ cation is five-coordinate with a weakly bound
acetonitrile ligand. This ligand is lost during the sequential
reduction to the Co(-I) species. The ligand loss most likely
occurs during the reduction from Co(II) to Co(I) as the latter
Figure 2. (a) Cyclic voltammogram of a 1 × 10-3 M solution of [Co-
(dppe)2(CH3CN)][BF4]2 in 0.3 M Bu4NBF4/acetonitrile. (b) Cyclic volta-
mmogram of a 1 × 10-3 M solution of HCo(dppe)2 in 0.3 M Bu4NBF4/
benzonitrile. (c) Cyclic voltammogram of a 1 × 10-3 M solution of
[HCo(dppe)2(CH3CN)][PF6]2 in 0.3 M Bu4NBF4/acetonitrile. In each case,
the scan rate was 0.05 V/s, and the potentials are shown vs the ferrocene/
ferrocenium couple.
d8 species is expected to form a square-planar complex. The
geometry of the square-planar Co(I) complex is expected to
change to a distorted tetrahedron upon further reduction on the
basis of the structures of other four-coordinate d8, d9, and d10
complexes.3,10,18 These rearrangements are rapid on the CV time
scale at scan rates of 50-1000 mV/sec.
The cyclic voltammograms of the cobalt(I)-, cobalt(II)-, and
cobalt(III)-hydride derivatives have also been obtained. Studies
of HCo(dppe)2 were performed in benzonitrile because it is
insoluble in acetonitrile. Studies of [HCo(dppe)2(CH3CN)]2+ and
[HCo(dppe)2]+ have been carried out in both benzonitrile and
acetonitrile to confirm that the electrochemical responses are
(18) Longato, B.; Riello, L.; Bandoli, G.; Pilloni, G. Inorg. Chem. 1999, 37,
2818.
(19) (a) Gagne´, R. R.; Koval, C. A.; Lisensky, G. C. Inorg. Chem. 1980, 19,
2854. (b) Gritzner, G.; Kuta, J. Pure Appl. Chem. 1984, 56, 462. (c) Hupp,
J. T. Inorg. Chem. 1990, 29, 5010.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 2987

A R T I C L E S
Ciancanelli et al.
very similar in both solvents. Cyclic voltammograms of HCo-
(dppe)2 and [HCo(dppe)2(CH3CN)]2+ in benzonitrile and ac-
etonitrile are shown by traces b and c of Figure 2. It can be
seen that the HCo(I)/HCo(II) couple involves a reversible
electron transfer near -1.2 V in both solvents. A second
irreversible oxidation wave is observed at -0.53 V (50 mV/S
scan rate in acetonitrile) that is assigned to the oxidation of
[HCo(dppe)2]+ to form [HCo(dppe)2(CH3CN)]2+. An irrevers-
ible cathodic peak at -1.02 V corresponding to the HCo(III)/
HCo(II) reduction is associated with this wave.
and unprotonated anisidine. Integrations reached constant values
over a period of 6 days. To confirm that the system is indeed
at equilibrium, the reaction was also followed in the reverse
direction, that is, starting from [Co(dppe)2]+. The results of these
experiments gave the same value for the equilibrium constant
within experimental error (Keq 2 ) 1.0 ( 0.2). The pKa value
of [HCo(dppe)2(CH3CN)]2+ in acetonitrile (11.3) can be deter-
mined by subtracting the log of the equilibrium constant for
reaction 2 from the pKa value of the anisidinium ion in the same
solvent (11.3).20
Because the [HCo(dppe)2(CH3CN)]2+/[HCo(dppe)2]+ couple
is irreversible on the time scale of a cyclic voltammmetry
experiment, its potential was determined from potentiometric
measurements on acetonitrile solutions containing 50:50 mix-
tures of the isolated Co(III) and Co(II) hydride derivatives. If
these measurements are made within 30 min after mixing fresh
solutions of [HCo(dppe)2]+ and [HCo(dppe)2(CH3CN)]2+, a
stable and reproducible potential of -0.83 V is obtained.
However, measurements taken over longer time periods of 1-2
h were irreproducible. Additional studies revealed that solutions
of [HCo(dppe)2]+ slowly disproportionate upon standing in
solution to form [Co(dppe)2]+ and [H2Co(dppe)2]+ (reaction 1).
The formation of [Co(dppe)2]+ can be followed by the slow
Similarly, [H2Co(dppe)2]+ reacts with tetramethylguanidine
in acetonitrile-d3 and benzonitrile solutions to form HCo(dppe)2.
Again, an excess of a mixture of tetramethylguanidine and its
protonated form was used to establish a fixed ratio of these
two species. In this case, the reaction was monitored by 1H NMR
spectroscopy using the hydride resonance of the two cobalt
complexes. This reaction can also be studied by 31P NMR
spectroscopy. The reverse reaction was studied in benzonitrile
because of the low solubility of HCo(dppe)2 in acetonitrile. The
equilibrium constant measured in acetonitrile for the deproto-
nation reaction (2.7 ( 0.3) is within experimental error of the
value measured in benzonitrile for the reverse reaction (2.5).
The pKa value calculated for [H2Co(dppe)2]+ in acetonitrile is
22.8 based on a pKa value of 23.3 for protonated tetrameth-
ylguanidine.21
2[HCo(dppe)2]+ f [H2Co(dppe)2]+ + [Co(dppe)2]+ (1)
To determine the hydride donor ability of HCo(dppe)2,
attempts were made to observe equilibrium hydride-transfer
reactions between HCo(dppe)2 and [Ni(diphosphine)2]2+ com-
plexes of known hydride acceptor ability. However, these studies
were complicated by exchange of the diphosphine ligands and
decomposition of some of the complexes during the course of
the reaction. A different method was therefore used to determine
the hydride donor ability of [H2Co(dppe)2]+. Hydride donor
values can also be obtained from equilibria involving the
heterolytic cleavage of hydrogen in the presence of a base of
known strength.4 [HCo(dppe)2(CH3CN)]2+ was found to react
reversibly with hydrogen and 2,4-dichloroaniline to form [H2-
Co(dppe)2]+ and 2,4-dichloroanalinium (pKa ) 8.0 in acetoni-
trile)20 (reaction 3), and the equilibrium constant for reaction 3
(Keq 3 ) 1.1 ( 0.2) was determined by NMR spectroscopy. The
activity of hydrogen at 1.0 atm was taken as 1.0 in reaction 3
as this is the reference state for hydrogen for the normal
hydrogen electrode.
appearance of the three cyclic voltammetric waves associated
with this compound. Approximately 5% of the expected
stoichiometric concentration of [Co(dppe)2]+ was observed in
1 h, and crystals of [H2Co(dppe)2](PF6)2 were isolated from
acetone solutions of [HCo(dppe)2](PF6) over a period of several
days. Identical disproportionation reactions have been reported
to be fast for [HRh(dppe)2]+ and [HIr(dppe)2]+.12
No reversible or quasi-reversible waves are observed for
cyclic voltammograms of [H2Co(dppe)2]2+ in acetonitrile solu-
tions. As a result, it is not possible to obtain information on the
effects of the oxidation state of Co on the Co-H bond cleavage
energies for this dihydride complex.
Equilibrium Studies of Proton and Hydride Transfers.
Reactions of the Co(III) hydride complexes with appropriate
bases were studied to determine acid strengths of these
complexes. For example, [HCo(dppe)2(CH3CN)]2+ is deproto-
nated by anisidine in acetonitrile to form an equilibrium mixture
as shown in reaction 2
[HCo(dppe)2(CH3CN)]2+ + B /
[Co(dppe)2]+ + BH+ + CH3CN (2)
[HCo(dppe)2(CH3CN)]2+ + B + H2 /
[H2Co(dppe)2]+ + BH+ + CH3CN (3)
BH+ f B + H+
H+ + H- f H2
(4)
(5)
where
B ) CH3OC6H4NH2
This reaction can be conveniently followed by 1H NMR
spectroscopy. Integration of the resonances of the methylene
protons of the dppe ligand in each complex provides a
convenient method for determining the ratios of [HCo(dppe)2-
(CH3CN)]2+ and [Co(dppe)2]+. The ratios of protonated anisi-
dine and anisidine were determined by using a large excess
(approximately 10-fold) of each reagent in a predetermined ratio.
Under these conditions, the deprotonation of [HCo(dppe)2(CH3-
CN)]2+ makes a negligible contribution to the ratio of protonated
[HCo(dppe)2(CH3CN)]2+ + H- f
[H2Co(dppe)2]+ + CH3CN (6)
∆G°H- ) -1.37pKeq 3 - 1.37pKa(4) + 76.0 kcal/mol (7)
(20) Edidin, R. T.; Sullivan, J. M.; Norton, J. R. J. Am. Chem. Soc. 1987, 109,
3945-3953.
(21) (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.
9
2988 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Characterization of the Metal−Hydrogen Bond
A R T I C L E S
Table 3. Thermodynamic Data (∆G° Values in kcal/mol) for
details). These values are ultimately determined by the open
circuit potential of -0.83 V for the HCo(III/II) couple and the
free energies associated with the deprotonation of [HCo(dppe)2-
(CH3CN)]2+ and the hydride donor ability of [H2Co(dppe)2]+
which were determined from equilibria 2 and 3, respectively.
These equilibria were chosen because they are uncomplicated
by the insolubility of HCo(dppe)2 in acetonitrile, and both Co-
(III) species are relatively stable and easily purified. The reaction
of [H2Co(dppe)2]+ with tetramethylguanidine in benzonitrile
provides a cross check on the free energy values for the hydride,
hydrogen atom, and proton-transfer reactions of Table 3.
Cobalt-Hydride Derivatives
+
•
−
complex
pK
a
∆G°H
∆G°H
∆G°H
[H2Co(dppe)2]+
HCo(dppe)2
22.8
38.1
23.6
11.3
31.3
52.3
32.4
15.5
58.0
59.1
49.9
52.9
65.1
49.1
59.7
[HCo(dppe)2]+
[HCo(dppe)2(CH3CN)]+2
Reaction 3 can be combined with reactions 4 and 5 to give
information on the heterolytic cleavage of the Co-H bond of
[H2Co(dppe)2]+ in acetonitrile (the reverse of reaction 6). The
free energies associated with reactions 3-5 can be used in eq
Discussion
-
7 to calculate that the hydride donor ability, ∆G°H , of [H2Co-
(dppe)2]+ is 65.1 kcal/mol.
Factors Influencing the pKa Values of the Co-H Com-
plexes. The free energy values for the successive deprotonation
of [H2Co(dppe)2]+ to HCo(dppe)2 and [Co(dppe)2]- are 31.3
kcal/mol (pKa ) 22.8) and 52.3 kcal/mol (pKa ) 38.1),
respectively. This corresponds to a change of 15.3 pKa units
for these two sequential deprotonation reactions. This ∆pKa is
much larger than the value of 8.7 for the sequential deproto-
nation of the isoelectronic compound H2Fe(CO)4 in water.22
Similarly, H3Re3(CO)12 has been deprotonated sequentially to
[Re3(CO)12]3- with ∆pKa1 ) 7 and ∆pKa2 ) 15.23 The carbonyl
ligands of H2Fe(CO)4 are good pi-acceptors, and as a conse-
quence sequential deprotonation for this complex does not
produce a large negative charge localized on the Fe atom. The
diphosphine ligand, dppe, is not able to delocalize the negative
charge as well as carbonyl ligands, and a larger negative charge
is produced on the cobalt atom. As a result, the pKa values as
well as the ∆pKa for [H2Co(dppe)2]+ are significantly larger
than that for H2Fe(CO)4.
Free Energy Relationships. The reduction potentials and
equilibrium constants discussed above for the cobalt-dppe
derivatives have been used to determine free energy values of
the three different Co-H bond cleavage reactions for the four
cobalt species as summarized in Table 3. Equations 8-11 show
-
a thermodynamic cycle that can be used to calculate ∆G°H for
HCo(dppe)2. The scheme makes use of the reduction potentials
for the HCo(III/II) and HCo(II/I) couples (reactions 8 and 9),
the pKa (11.3) of [HCo(dppe)2(CH3CN)]2+ in acetonitrile
(reaction 10), and the ∆G° for the reduction of a proton to the
hydride ion (reaction 11).9 The sum of these steps, reaction 12,
is hydride donation by HCo(dppe)2, and the numerical value of
-
∆G°H for this reaction can be determined by adding the ∆G°
values for steps 8 through 11, as shown in eq 13. The value of
-
∆G°H for HCo(dppe)2 determined by this method is 49.1 kcal/
mol.
HCo(dppe)2 / [HCo(dppe)2]+ + e-
[HCo(dppe)2]+ + CH3CN /
(8)
From the data shown in Table 3, the effect of sequential
oxidation on the Co-H cleavage reactions of HCo(dppe)2 can
also be assessed. The pKa value of HCo(dppe)2 decreases from
38.1 to 23.6 for the monocation and to 11.3 for the dication.
The ∆pKa values of ∼15 and 12 units, respectively, are
somewhat smaller than those reported for CpM(CO)2(L)H
hydrides (where M ) Cr, Mo, and W, and L ) CO or PR3)
and their oxidized partners, which are in the range of 20-25
pKa units.24,25 The pKa values are useful in assessing the potential
stability of an oxidized hydride to deprotonation. The fact that
[HCo(dppe)2]+ is relatively stable to deprotonation probably
contributes to the reversibility of the HCo(dppe)2/[HCo(dppe)2]+
couple.
[HCo(dppe)2(CH3CN)]2+ + e- (9)
[HCo(dppe)2(CH3CN)]2+
/
[Co(dppe)2]+ + CH3CN + H+ (10)
H+ + 2e- f H-
(11)
(12)
HCo(dppe)2 f [Co(dppe)2]+ + H-
∆G°H (12) ) 23.06E°HCo(I/II) + 23.06E°HCo(II/III)
+
-
•
1.37pKa(10) + 79.6 (13)
Factors Influencing the ∆G°H Values of the Co-H
Complexes. The successive homolytic bond cleavage reactions
of [H2Co(dppe)2]+ require 58.0 and 49.9 kcal/mol, respectively.
These values indicate that the observed disproportionation of
[HCo(dppe)2]+, reaction 1, is favored by approximately 8 kcal/
This result can be checked by comparison to the value
determined by an alternate thermodynamic cycle. A ∆G°H of
-
49.3 kcal/mol is calculated for HCo(dppe)2 from the sequential
protonation of HCo(dppe)2 to form [H2Co(dppe)2]+ (-31.3 kcal/
mol), removal of H- from [H2Co(dppe)2]+ to form [HCo(dppe)2-
(CH3CN)]2+ (65.1 kcal/mol), and deprotonation of [HCo(dppe)2-
(CH3CN)]2+ to form [Co(dppe)2]+ (15.5 kcal/mol). All of these
values were determined independently as discussed above, and
this approach uses no electrochemical measurements. The two
•
mol. In addition, the sum of ∆G°H for the two homolytic
cleavage reactions is 107.9 kcal/mol as compared to a free
energy of 103.6 kcal/mol for the homolytic cleavage of H2 in
acetonitrile.9 This is consistent with the observed addition of
H2 to [Co(dppe)2]+ to form [H2Co(dppe)2]+. This reaction can
be reversed by purging solutions of [H2Co(dppe)2]+ with
nitrogen.
-
calculated ∆G°H values for HCo(dppe)2 (49.1 and 49.3 kcal/
mol) are identical within our total estimated uncertainties for
these measurements of (2.5 kcal/mol.
(22) Gamembeck, F.; Krumholz, P. J. Am. Chem. Soc. 1971, 93, 1909.
(23) Kaez, H. D. Chem. Br. 1973, 9, 344.
By using thermodynamic cycles similar to those described
above, it is possible to calculate all of the thermodynamic
parameters shown in Table 3 (see Supporting Information for
(24) Tilset, M. J. Am. Chem. Soc. 1992, 114, 2740-2741.
(25) (a) Ryan, O. B.; Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990, 112,
2618-2626. (b) Skagestad, V.; Tilset, M. J. Am. Chem. Soc. 1993, 115,
5077-5083.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 2989

A R T I C L E S
Ciancanelli et al.
Values for the homolytic H-Co bond cleavage reactions have
also been determined for each of the three oxidation states of
[HCo(dppe)2]n+. Unlike the pKa trends, the ∆G°H values do
•
not show a regular correlation with oxidation state, and the
•
decrease in ∆G°H with increasing charge actually appears to
be relatively minor. For example, oxidation of HCo(dppe)2 to
[HCo(dppe)2]+ results in a 9-10 kcal decrease in the free
energy, but further oxidation to form [HCo(dppe)2(CH3CN)]2+
Figure 3. Thermodynamic relationships between [Co(dppe)2]n+ and [Hx-
Co(dppe)2]m+ where n ) -1, 0, 1, 2, x ) 1, 2, and m ) 0, 1, 2. Numbers
in bold are reduction potentials in volts vs Fc. Numbers near arrows are
∆G° values for the indicated proton, hydrogen atom, or hydride-transfer
reaction in kcal/mol.
•
results in an increase in ∆G°H of approximately 3 kcal/mol.
Three of the four homolytic bond cleavage reactions in Table
3 result in the formation of a metal centered radical as well as
a hydrogen atom. The free energies associated with all of these
reactions are larger than that observed for the homolytic cleavage
of the bond for [HCo(dppe)2]+ which results in the formation
of the low spin 16-electron complex, [Co(dppe)2]+. The forma-
tion of a metal-based radical appears to influence the free energy
associated with the homolytic cleavage of the Co-H bond more
than the oxidation state of the metal or the charge on the metal-
hydride complex.
transition metal-dihydride systems in a manner similar to those
developed previously for monohydride systems. Our results for
the [H2Co(dppe)2]+ system selected for this study are sum-
marized in Figure 3. The figure illustrates the relationships
between the various thermodynamic properties of the hydrides
(and non-hydrides) and helps shift the perspective from a single
•
-
pKa, ∆G°H , or ∆G°H measurement for a single compound to
one of an integrated system of compounds that are interrelated
by the various simple reactions shown. The relationships of
Figure 3 are very useful in developing a quantitative under-
standing of how the cobalt complexes will respond to changes
in the acidity or basicity of the environment, the introduction
of a hydrogen atom or hydride acceptor, or a change in the
electrochemical potential of the solution. We believe these
responses are best understood when viewing the complexes as
part of a total thermodynamic system. The approach developed
here suggests that it should be possible to obtain detailed
thermodynamic descriptions of the metal-hydrogen bond in a
variety of metal-hydride complexes with nonlabile ligands,
including polyhydride and polynuclear systems.
The data in Table 3 indicate that the homolytic cleavage of
molecular hydrogen is thermodynamically favored by all three
odd electron cobalt species, and the driving force increases in
the order [Co(dppe)2(CH3CN)]2+ < [HCo(dppe)2]+ < Co(dppe)2
with the latter two being nearly equal. The free energy values
shown in Table 3 can be converted to enthalpies by correcting
for the entropy contribution for the hydrogen atom in acetonitrile
(approximately 4.6-4.8 kcal/mol).7d,9 This correction is based
on the assumptions that S° for the hydride and that for its
conjugate radical are equal and that the entropy of solvation
for the hydrogen atom and that for the hydrogen molecule are
also equal. The enthalpy values of 64 kcal/mol for HCo(dppe)2
and 63 kcal/mol for [H2Co(dppe)2]+ can be compared to the
values of 65 and 67 kcal/mol reported previously for HCo-
(CO)3PPh3 and HCo(CO)4, respectively.7d,e
Experimental Section
-
1
Materials and Instrumentation. H NMR and 31P NMR spectra
Factors Influencing the ∆G°H Values of the Co-H
were recorded on a Varian Unity 500 MHz spectrometer. 1H chemical
shifts are reported relative to tetramethylsilane using residual solvent
protons as a secondary reference. A capillary filled with phosphoric
acid was used as an external reference for 31P NMR spectra. All 31P
NMR spectra were proton decoupled. Solvent suppression studies were
recorded on a Bruker AM-400 MHz spectrometer using a presaturation
technique. In experiments where accurate integration of the hydride
was necessary, spectra were collected using a repetition rate of 0.100
s with a tip angle of 43° to allow for the relaxation of the hydride
which has a T1 ) 0.135 s. Infrared spectra of Nujol mulls were recorded
on a Nicolet 510P spectrometer. Electron spin resonance spectra were
recorded on a Varian E109 X-band spectrometer. Air-sensitive solutions
of 2 mM concentration in acetonitrile were transferred to an EPR tube
under a N2 atmosphere in a glovebox.
All electrochemical experiments were carried out under an atmo-
sphere of N2 in 0.3 M Bu4NBF4 solutions of benzonitrile or acetonitrile.
Cyclic voltammetry studies were carried out on a Cypress Systems
computer aided electrolysis system. The working electrode was a glassy
carbon disk (2 mm diameter). The counter electrode was a glassy carbon
rod. An Ag/AgCl reference electrode was used to fix the potential.
Ferrocene was used as an internal standard, and all potentials are
referenced to the ferrocene/ferrocenium couple. Chronopotentiometric
studies were carried out using the same electrodes and instrumentation.
1,2-Bis(diphenylphosphinoethane) (dppe), cobalt powder, ferroce-
nium hexafluorophosphate, ferrocenium tetrafluoroborate, sodium boro-
hydride, tetrafluoroboric acid, and tetrabutylammonium tetrafluoroborate
were purchased from Aldrich Chemical Co. Acetonitrile and methylene
chloride were dried by distillation from calcium hydride under nitrogen.
Complexes. The hydride donor ability of the two cationic
species, [H2Co(dppe)2]+ and [HCo(dppe)2]+, are 65.1 and 59.7
kcal/mol, respectively. These values lie within the range (67-
51 kcal/mol) previously observed for [HNi(diphosphine)2]+
complexes.3,4,10 The neutral hydride, HCo(dppe)2, is a much
more powerful hydride donor than either of these by 10-15
kcal/mol, and its hydride donor value of 49.1 kcal/mol is
comparable to the value observed for [HPt(dppe)2]+ (52.5 kcal/
mol).3 It appears that changing from Ni to Co has approximately
the same effect on hydride donor potentials as changing from
Ni to Pt for isoelectronic and isostructural species. In previous
work on [HNi(diphosphine)2]+ complexes, we have shown that
a linear free energy relationship exists between the hydride donor
8
9
-
ability of a complex (∆G°H ) and the potential of the (d /d )
couple of the conjugate hydride acceptor, [Ni(diphosphine)2]2+ 10
.
-
The calculated ∆G°H value of 49.1 kcal/mol for HCo(dppe)2
agrees well with the value of 47.7 kcal/mol calculated using
the equation for the best-fit line for the Ni complexes. This result
suggests that it may be possible to extend the relationship
developed previously for [HNi(diphosphine)2]+ complexes to
provide a useful estimate of the hydricity of other five-coordinate
first row metal-hydride complexes.
Summary and Conclusions
The major objective of this research was to demonstrate that
comprehensive thermodynamic schemes can be developed for
9
2990 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Characterization of the Metal−Hydrogen Bond
A R T I C L E S
Tetrahydrofuran, toluene, and diethyl ether were dried by distillation
from sodium benzophenone ketyl under nitrogen. Absolute ethanol and
reagent grade benzonitrile were used as received. All solvents were
purged with nitrogen for 15 min before use. Acetonitrile-d3, benzene-
d6, and acetone-d6 were freeze-pump-thawed, dried over molecular
sieves, and stored under inert atmosphere in a glovebox. All reactions
were performed using standard Schlenk techniques under nitrogen. [Co-
(CH3CN)6][BF4]2 and [Co(dppe)2(CH3CN)][BF4]2 were prepared
according to literature methods. Synthetic methods and elemental
analyses have been reported for the synthesis of HCo(dppe)2,11a [HCo-
(dppe)2][BPh4],12 [HCo(dppe)2(CH3CN)][BPh4]2,12 [Co(dppe)2][BPh4],27
and [H2Co(dppe)2][ClO4].11a The following synthetic methods have been
slightly modified from the reported procedures. The spectroscopic,
electrochemical, and structural studies indicate that they are the same
complexes as those identified earlier.
acetonitrile (100 mL). The solution was stirred for 36 h. The resulting
dark brown solution was filtered to remove the excess zinc. The solvent
was removed in vacuo. The solid was washed with water (2 × 20 mL)
followed by a diethyl ether wash (2 × 30 mL). The resulting brown
solid is unstable; attempts at recrystallization were unsuccessful. The
yield of the crude product was 71.3% (0.308 g). 1H NMR (acetonitrile-
d3): δ 7.08-7.30 (m, PPh2), 2.32 (s, 8H, PCH2CH2). No 31P NMR
signal was detected. E° (II/I) ) -0.607 V, E° (I/0) ) -1.62 V, E°
(0/-I) ) -2.07 V.
26
13
pKa Measurement for [HCo(dppe)2(CH3CN)](BF4)2. Aliquots (0.5
mL each) of a 1.28 × 10-2 M solution of [HCo(dppe)2(CH3CN)](BF4)2
(0.051 g) in acetonitrile-d3 were syringed into each of five NMR tubes.
This was followed by aliquots of different volumes of a 0.286 M
solution of anisidine in acetonitrile-d3 and a 0.286 M solution of
anisidinium tetrafluoroborate in acetonitrile-d3 to produce a constant
1
volume of 1.0 mL. The solutions were monitored by H NMR for 6
HCo(dppe)2. A slurry of [Co(dppe)2(CH3CN)](BF4)2 (4.92 g, 4.6
mmol) in ethanol (200 mL) was heated to 65 °C. A solution of sodium
borohydride (0.522 g, 13.8 mmol) in 95% aqueous ethanol was added
slowly over an hour, which resulted in the formation of a red precipitate.
The slurry was stirred overnight. The precipitate was collected by
filtration and washed with ethanol (2 × 30 mL) to remove excess
sodium borohydride. The crude product was recrystallized from a
mixture of tetrahydrofuran and ethanol. Yield: 58% (2.29 g). 1H NMR
(benzene-d6): δ 6.90-7.40 (m, 40H, PPh2), 2.15 (s, 8H, PCH2CH2),
-14.9 (quintet, 1H, Co-H). 31P NMR (benzene-d6): δ 72.9 (s). IR
(Nujol mull): νCo-H ) 1882 cm-1. E° (II/I) ) -1.16 V.
[HCo(dppe)2(CH3CN)](PF6)2. A solution of ferrocenium hexafluo-
rophosphate (0.537 g, 1.64 mmol) in acetonitrile (100 mL) was added
to a slurry of HCo(dppe)2 (0.700 g, 0.82 mmol) in acetonitrile (150
mL). The resulting light brown solution was stirred for 3 days. The
solvent was removed in vacuo to give an orange-brown solid. The solid
was washed with toluene (4 × 25 mL) to remove ferrocene. The crude
product was recrystallized from a mixture of methylene chloride and
ethanol to give a yellow solid. Yield: 74% (0.72 g). This procedure
was repeated with ferrocenium tetrafluoroborate to produce the tet-
rafluoroborate derivative. 1H NMR (acetonitrile-d3): δ 6.67-7.59 (m,
5H, PPh2), 2.39 (s, 4H, PCH2CH2), 3.00 (s, 4H, PCH2CH2), -21.1
(quintet, 1H, Co-H). 31P NMR (acetonitrile-d3): δ 66.6 (s). IR (Nujol
mull): νCo-H ) 1977 cm-1, νC-N ) 2282 cm-1. E° (II/I) ) -1.16 V.
days until the ratio of [HCo(dppe)2(CH3CN)](BF4)2 to [Co(dppe)2](BF4)
remained constant. The ratio of the concentrations of the two metal
complexes was calculated by integration of the 1H NMR signals of the
methylene groups of the dppe ligands. The reverse reaction was studied
by mixing a 2.5 × 10-2 M solution of [Co(dppe)2](BF4) in acetonitrile-
d3 (0.5 mL) and 0.250 M solutions of anisidine (0.25 mL) and
anisidinium tetrafluoroborate (0.25 mL) in acetonitrile-d3. The solution
was monitored by 1H NMR for 6 days, until the ratio of products
remained constant.
pKa Measurement for [H2Co(dppe)2](BF4). Aliquots (0.5 mL each)
of a 1.48 × 10-2 M solution of [H2Co(dppe)2](BF4) in benzonitrile
were syringed into each of four NMR tubes followed by known amounts
of tetramethylguanidine (TMG) in benzonitrile to produce a total volume
of 0.65 mL. CD3CN (0.07 mL) was added to each NMR tube to provide
a lock signal. The solutions were monitored by solvent suppression 1H
NMR for 4 days until the ratio of products remained constant as
determined by integration of the hydride 1H NMR resonances. To study
this reaction in the reverse direction, aliquots (0.4 mL) of a 2.2 × 10-3
M solution of HCo(dppe)2 in benzonitrile were syringed into three NMR
tubes, and aliquots of different known volumes of 0.252 M TMG and
0.252 M [TMGH](BF4) solutions in benzonitrile were added to give a
constant volume of 0.65 mL. The solutions were monitored by 1H NMR
for 4 days until the ratio of the products remained constant.
Equilibrium Studies of the [H2Co(dppe)2](BF4)/[HCo(dppe)2-
(CH3CN)](BF4)2 System with Hydrogen and Base. Aliquots (0.4 mL)
of a 6.4 × 10-2 M solution of [H2Co(dppe)2](BF4) in acetonitrile-d3
were syringed into each of four NMR tubes, and various aliquots of
1.27 M solutions of 2,4-dichloroaniline and 2,4-dichloroanilinium
tetrafluoroborate in acetonitrile-d3 were added to produce a constant
volume of 0.8 mL. The solutions were purged with hydrogen (and each
day afterward) and monitored by 1H NMR for 8 days until the ratio of
[H2Co(dppe)2](BF4) to [HCo(dppe)2CD3CN][BF4]2 remained constant
as determined by integration of the hydride 1H NMR signals. The
reverse reaction was studied by adding aliquots (0.4 mL) of a 1.28 ×
10-2 M solution of [HCo(dppe)2CD3CN][BF4]2 in acetonitrile-d3 to five
NMR tubes. Various aliquots of 0.256 M solutions of 2,4-dichloroa-
niline and 2,4-dichloroanilinium tetrafluoroborate (0.0830 g) in aceto-
nitrile-d3 were then added to give a constant volume of 0.8 mL. All of
the NMR tubes were purged with hydrogen, and the solutions were
monitored by 1H NMR for 8 days, until the ratio of the products
remained constant.
[H2Co(dppe)2](BF4). Zinc powder (0.580 g, 8.87 mmol) was added
to a solution of [Co(dppe)2(CH3CN)](BF4)2 (0.950 g, 0.880 mmol) in
acetonitrile (100 mL). The solution was stirred for 36 h. Hydrogen gas
was bubbled through the solution for 15 min. The resulting orange
solution was filtered to remove the excess zinc. Diethyl ether which
had been purged with hydrogen was added to the solution to precipitate
a yellow solid that was collected by filtration and was washed with
water (2 × 20 mL) followed by H2 saturated diethyl ether (2 × 30
1
mL). Yield: 34% (0.260 g). H NMR (acetonitrile-d3): δ 6.17-7.93
(m, 40H, C6H5), 3.29-3.36 (d, br, 4H, PCH2CH2), 2.26 (s, br, 4H,
PCH2CH2), δ -13.2 (m, 2H, CoH). 31P NMR (acetonitrile-d3): 69.7
(s, br), 80.5 (s, br).
[HCo(dppe)2](BF4). A solution of HCo(dppe)2 (0.290 g, 0.339
mmol) in tetrahydrofuran (80 mL) was added to a solution of [HCo-
(dppe)2CH3CN](PF6)2 (0.395 g, 0.333 mmol) in acetonitrile (30 mL).
The resulting brown solution was stirred for 18 h and then filtered.
The solvent was removed in vacuo. The golden brown solid was
recrystallized from a mixture of methylene choride and ethanol.
Yield: 34%. IR (Nujol mull): νCo-H ) 1874 cm-1. E° (II/I) ) -1.16
V. EPR (CH3CN, 25 °C): g ) 2.1, d, A ) 80 G. (It has not been
established whether the observed doublet results from coupling to the
hydride or to a phosphorus atom.)
X-ray Diffraction Studies. Single-crystal X-ray data for three
complexes were collected at low temperature, ca. 150 K, on a Siemens
SMART CCD diffractometer using Mo KR radiation (l ) 0.71073 Å).
An arbitrary hemisphere of data to 0.68 Å resolution was collected for
each sample. 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. Anisotropic
thermal parameters were refined for all non-hydrogen atoms. All
[Co(dppe)2](BF4). Zinc powder (0.287 g, 4.40 mmol) was added to
a solution of [Co(dppe)2(CH3CN)](BF4)2 (0.470 g, 0.440 mmol) in
(26) Hathaway et al. J. Chem. Soc. 1962, 2444-2448.
(27) Nobile, C.; Rossi, M.; Sacco, A. Inorg. Chim. Acta 1971, 698-700.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 2991

A R T I C L E S
Ciancanelli et al.
Table 4. Crystallographic Data for HCo(dppe)2, [HCo(dppe)2CH3CN](PF6)2, [H2Co(dppe)2](BF4), and [Co(dppe)2CH3CN](BF4)2
[HCo(dppe)2CH CN](PF )
[H Co(dppe)2](BF )
[Co(dppe)2CH CN](BF )
3 4 2
3
6 2
2
4
HCo(dppe)2
C52H49CoP4
856.72
monoclinic
P21/n
9.8888(4)
20.7973(10)
20.9906(10)
90
92.2570(10)
90
‚(CH Cl2)
‚(CH ) CO
‚(CH ) CO
2
3 2
3 2
empirical formula
formula mass, amu
cryst syst
space group
a (Å)
C55H54Cl2CoF12NP6
1272.64
monoclinic
P21/n
19.4102(15)
16.9353(13)
20.121(2)
90
115.407(2)
90
C55H56BCoF4OP4
1002.62
monoclinic
P21/n
13.779(2)
19.320(2)
18.595(2)
90
102.225(15)
90
C57H57B2CoF8NOP4
1128.47
orthorhombic
P212121
11.7261(5)
20.6176(9)
22.2925(10)
90
b (Å)
c (Å)
R (deg)
â (deg)
90
90
γ (deg)
V (Å3)
4313.6(3)
4
5974.3(8)
4
4837.9(11)
4
5389.5(4)
4
Z
T (K)
143(2)
141(2)
143(2)
293(2)
R indexa (I > 2σ(I))
R1 ) 0.0463,
wR2 ) 0.1144
R1 ) 0.0715
wR2 ) 0.1248
a ) 0.0670, b ) 1.1366
1.068
R1 ) 0.0947,
wR2 ) 0.2447
R1 ) 0.1486
wR2 ) 0.2728
a ) 0.1597, b ) 0
1.044
R1 ) 0.0618,
wR2 ) 0.1321
R1 ) 0.1305
wR2 ) 0.1591
a ) 0.0709, b ) 0
0.938
R1 ) 0.0759,
wR2 ) 0.1099
R1 ) 0.2299
wR2 ) 0.1497
a ) 0.0418, b ) 0
0.927
R indexa (all data)
weighting coeffb
GOFc on F 2
2
2
2
2
2
2
2
a R1 ) ∑||Fo| - |Fc||/∑|Fo|; wR2 ) [∑w(Fo - Fc )/∑w(Fo )2]1/2
.
b W -1 ) [σ2(Fo ) + (ap)2 + bp], where P ) (Fo + 2Fc )/3. c GOF ) S ) [∑w(Fo
- Fc )/(M - N)]1/2, where M is the number of reflections, and N is the number of parameters refined.
2
hydrides were located by a difference Fourier map but were freely
refined in subsequent cycles of least-squares refinement. Selected details
of the structure solutions and refinements are given in Table 4 , and
comments for specific compounds are given in the following paragraphs.
HCo(dppe)2. Crystals were grown from a saturated solution of
benzene-d6. This complex crystallized in the centrosymmetric space
group P21/n. Data were collected at 145 K using 30 s scans. There
was no disorder.
complex crystallized in the centrosymmetric space group P21/n. A
molecule of acetone was present in the crystal lattice. There was no
disorder.
[Co(dppe)2CH3CN][BF4]2 (4). Crystals were grown from acetone-
d6. Data were collected at 145 K using 30 s scans. A molecule of
acetone was present in the crystal lattice. There was a small amount of
disorder in the BF4 molecule.
[HCo(dppe)2(CH3CN)](PF6)2. Crystals were grown from methlyene
chloride and ethanol. The complex crystallized in the centrosymmetric
space group P21/n. Data were collected at 145 K using 30 s scans.
Two molecules of solvent are present in the crystal lattice. One molecule
of methylene chloride was disordered over two sites. The volume
occupied by a second highly disordered methylene chloride molecule
was analyzed using SQUEEZE.28 This program found electron density
equivalent to 244 electrons which is close to six molecules of methylene
chloride. This electron density was removed from the file of observed
f 's prior to subsequent cycles of least squares refinement.
Acknowledgment. This research was supported by the
National Science Foundation. D. L. DuBois would also like to
acknowledge the support of the United States Department of
Energy, Office of Basic Energy Sciences, Division of Chemical
Sciences.
Supporting Information Available: Complete details of the
X-ray diffraction studies on HCo(dppe)2, [HCo(dppe)2(CH3CN)]-
(PF6)2, [H2Co(dppe)2](BF4), and [Co(dppe)2(CH3CN)](BF4)2,
and details of the calculations of ∆G° values (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
[H2Co(dppe)2](BF4]). Crystals were grown from a saturated solution
of acetone-d6. Data were collected at 145 K using 30 s scans. This
(28) Spek, A. L. SQUEEZE is a routine included in PLATON, A Multipurpose
Crystallographic Tool; Utrecht University: Utrecht, The Netherlands, 2001.
JA0122804
9
2992 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002