Dihydrogen Evolution by Protonation Reactions of Nickel(I)

Article abstract of DOI:10.1021/ic960216v

Nickel-mediated formation of H₂ by protonation of Ni(I) has been established and the kinetics of the process investigated. The diamagnetic complex [Ni^II(psnet)](BF₄)₂ was prepared and reduced to [Ni^I(psnet)](BF₄) with NaBH₄ in THF (psnet = bis(5-(diphenylphosphino)-3-thiapentanyl)amine). Both complexes were structurally characterized by X-ray diffraction. [Ni(psnet)]^I+ was demonstrated to be an authentic Ni(I) complex with a ...(d_z2)^I ground state. Under appropriate conditions, [Ni(psnet)]⁺ reacts with acids in nonaqueous media to give near-quantitative yields of H₂ according to the stoichiometry Ni^I + H⁺ → Ni^II + 1/2H₂. Dihydrogen production was demonstrated to be directly related to Ni(I) oxidation. The reaction system [Ni(psnet)]⁺/HCl/DMF, which gives H₂ yields of ≥90%, was subjected to a kinetics analysis. The overall reaction [Ni(psnet)]⁺ + HCl → [Ni(psnet)Cl]⁺ + 1/2H₂ proceeds by two parallel pathways dependent on chloride concentration. Addition of Bu₄NCl accelerates the reaction, whereas (Bu₄N)(PF₆) decreases the rate. A two-term rate law is presented which includes contributions from both pathways, whose common initial step is protonation of Ni(I). Path A (low chloride concentration) involves the formation and collapse of nickel hydride chloride ion pairs; the rate-determining step is the minimal reaction 2Ni^III-H^- → H₂ + 2Ni^II. Path B (high chloride concentration) includes as the rate-limiting step collapse of a nickel hydride dichloride ion pair followed by the bimolecular reaction of two Ni^III-H^- intermediates or reduction to Ni^II-H^- by Ni^I followed by protonation of the hydride. The relation of these results to the reactions of hydrogenase enzymes is considered.

Full text of DOI:10.1021/ic960216v

4148
Inorg. Chem. 1996, 35, 4148-4161
Dihydrogen Evolution by Protonation Reactions of Nickel(I)
Thomas L. James, Lisheng Cai, Mark C. Muetterties, and R. H. Holm*
Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138
ReceiVed February 28, 1996X
Nickel-mediated formation of H2 by protonation of Ni(I) has been established and the kinetics of the process
investigated. The diamagnetic complex [NiII(psnet)](BF4)2 was prepared and reduced to [NiI(psnet)](BF4) with
NaBH4 in THF (psnet ) bis(5-(diphenylphosphino)-3-thiapentanyl)amine). Both complexes were structurally
characterized by X-ray diffraction. [Ni(psnet)]1+ was demonstrated to be an authentic Ni(I) complex with a
1
+
2
...(dz ) ground state. Under appropriate conditions, [Ni(psnet)] reacts with acids in nonaqueous media to give
near-quantitative yields of H2 according to the stoichiometry NiI + H+ f NiII + /2H2. Dihydrogen production
1
was demonstrated to be directly related to Ni(I) oxidation. The reaction system [Ni(psnet)]+/HCl/DMF, which
gives H2 yields of J90%, was subjected to a kinetics analysis. The overall reaction [Ni(psnet)]+ + HCl f
[Ni(psnet)Cl]+ + /2H2 proceeds by two parallel pathways dependent on chloride concentration. Addition of
1
Bu4NCl accelerates the reaction, whereas (Bu4N)(PF6) decreases the rate. A two-term rate law is presented which
includes contributions from both pathways, whose common initial step is protonation of Ni(I). Path A (low
chloride concentration) involves the formation and collapse of nickel hydride chloride ion pairs; the rate-determining
step is the minimal reaction 2NiIII-H- f H2 + 2NiII. Path B (high chloride concentration) includes as the
rate-limiting step collapse of a nickel hydride dichloride ion pair followed by the bimolecular reaction of two
NiIII-H- intermediates or reduction to NiII-H- by NiI followed by protonation of the hydride. The relation of
these results to the reactions of hydrogenase enzymes is considered.
-
+
H
e
Introduction
NiI + H+ f NiIII-H- 98 NiII-H- 8 NiII + H2 (2a)
2NiIII-H- f 2NiII + H2
Biomimetic research on [NiFe] hydrogenases1,2 is now
extensive3-5 and has concentrated on structure modeling by
synthesis of new complexes, often for comparison with Ni
EXAFS of the enzymes and stabilization and EPR properties
of Ni(III) and Ni(I) states. Our most recent contribution to this
area is the demonstration that the physiological oxidation states
of nickel, Ni(III,II,I), are unlikely to be traversed in any
environment without accompanying alteration(s) in stereochem-
istry accountable in terms of the stereoelectronic preferences
of the metal.6 Functional modeling, on the other hand, is in a
primitive stage of development. Nickel-mediated dihydrogen
evolution, whether or not in the context of hydrogenase, has
been brought about under four approaches: (i) reaction 1,
(2b)
nickel(III) hydrides which then react further to produce Ni(II)
and dihydrogen7b,8,9 by either (a) reduction and protonation
events, the order of which is uncertain, or (b) bimolecular
coupling of two nickel(III) hydrides; (iii) electrocatalysis in
aqueous solution;10-12 (iv) protonation of a binuclear nickel(II)
phosphine-thiolate complex.13 A number of the electrocatalytic
studies were directed toward the reduction of carbon dioxide;11
the potentials employed were far more negative than hydroge-
(8) Lahiri, G. K.; Schussel, L. J.; Stolzenberg, A. M. Inorg. Chem. 1992,
31, 4991.
(9) (a) Vlcˇek, A., Jr.; Vlcˇek, A. A. Inorg. Chim. Acta 1980, 41, 123. (b)
Hsiao, Y.-M.; Chojnacki, S. S.; Hinton, P.; Reibenspies, J. H.;
Darensbourg, M. Y. Organometallics 1993, 12. 870. (Dihydrogen
formation from Ni(I) and acids was stated, but details were not
reported.)
+
H
Ni0 + H+ f NiII-H- 8 NiII + H2
(1)
entailing protonation of Ni(0) complexes to form Ni(II) hydrides
which are then protonated to yield Ni(II) and dihydrogen;7 (ii)
reactions 2, involving protonation of Ni(I) complexes to putative
(10) Efros, L. L.; Thorp, H. H.; Brudvig, G. W.; Crabtree, R. H. Inorg.
Chem. 1992, 31, 1722.
(11) (a) Fisher, B.; Eisenberg, R. J. Am. Chem. Soc. 1980, 102, 7361. (b)
Bradley, M. G.; Tysak, T.; Graves, D. J.; Vlachopoulos, N. A. J. Chem.
Soc., Chem. Commun. 1983, 349. (c) Beley, M.; Collin, J.-P.; Ruppert,
R. P.; Sauvage, J.-P. J. Am. Chem. Soc. 1986, 108, 7461. (d) Grant,
J. L.; Goswani, K.; Spreer, L. O.; Otvos, J. W.; Calvin, M. J. Chem.
Soc., Dalton Trans. 1987, 2105. (e) Collin, J.-P.; Jouaiti, A.; Sauvage,
J.-P. Inorg. Chem. 1988, 27, 1986. (f) Shinoya, M.; Kimura, E.; Iitaka,
Y. J. Am. Chem. Soc. 1990, 112, 9237.
(12) (a) Caˇlus¸aru, A.; Voicu, V. Electroanal. Chem. Interfacial Electrochem.
1973, 43, 257. (b) Caˇlus¸aru, A.; Baˇnicaˇ, F. G. Electroanal. Chem.
Interfacial Electrochem. 1973, 47, 190. (c) Baˇnicaˇ, F. G.; Caˇlus¸aru,
A. J. Electroanal. Chem. 1983, 145, 389. (d) Baˇnicaˇ, F. G. J.
Electroanal. Chem. 1984, 171, 351; Talanta 1985, 32, 1145. (e)
Baˇnicaˇ, F. G.; Diacu, E.; Luca, G. Bull. Soc. Chim. Belg. 1987, 96,
485. (f) Baˇnicaˇ, F. G.; Florea, M.; Moraru, M. J. Electroanal. Chem.
1990, 285, 281. (g) Baˇnicaˇ, F. G.; Diacu, E. Collect. Czech. Chem.
Commun. 1991, 56, 140. (h) Baˇnicaˇ, F. G. Bull. Soc. Chim. Fr. 1991,
128, 697.
X Abstract published in AdVance ACS Abstracts, May 1, 1996.
(1) (a) Cammack, R.; Fernandez, V. M.; Schneider, K. In The Bioinorganic
Chemistry of Nickel; Lancaster, J. R., Jr., Ed.; VCH Publishers, Inc.;
New York, 1988; Chapter 8. (b) Moura, J. J. G.; Teixeira, M.; Moura,
I.; LeGall, J. In The Bioinorgnic Chemistry of Nickel; Lancaster, J.
R., Jr., Ed.; VCH Publishers, Inc.: New York, 1988; Chapter 9. (c)
Cammack, R. AdV. Inorg. Chem. 1988, 32, 297. (d) Cammack, R. In
Bioinorganic Catalysis; Reedijk, J., Ed.; Marcel Dekker, Inc.; New
York, 1993; Chapter 7. (e) Kolodziej, A. F. Prog. Inorg. Chem. 1994,
41, 493.
(2) Albracht, S. P. J. Biochim. Biophys. Acta 1994, 1188, 167.
(3) Baidya, N.; Mascharak, P. K. Trends Inorg. Chem. 1993, 3, 275.
(4) Kovacs, J. A. AdV. Inorg. Biochem. 1994, 9, 173.
(5) Halcrow, M.; Christou, G. Chem. ReV. 1994, 94, 2421.
(6) James, T. L.; Smith, D. M.; Holm, R. H. Inorg. Chem. 1994, 33, 4869.
(7) (a) Tolman, C. A. J. Am. Chem. Soc. 1970, 92, 4217. (b) Sacconi,
L.; Orlandini, A.; Midollini, S. Inorg. Chem. 1974, 13, 2850. (c) Rigo,
P.; Bressan, M.; Basato, M. Inorg. Chem. 1979, 18, 860. (d) Cariati,
F.; Ugo, R.; Bonati, F. Inorg. Chem. 1966, 5, 1128. (Formation of
H2 was implied but not documented.)
(13) Franolic, J. D.; Millar, M. M. Abstracts, 205th National Meeting of
the American Chemical Society, Denver, CO, 1993; INOR 534;
American Chemical Society: Washington, DC, 1993.
S0020-1669(96)00216-9 CCC: $12.00 © 1996 American Chemical Society

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4149
commercial samples and were not purified further. The compounds
2-(diphenylphosphino)ethanethiol18 (psh), and [Ni(NS3tBu)H](BPh4)19
tBu
(NS3 ) tris[(tert-butylthio)ethyl]amine) were prepared by literature
methods. Solvents were dried by standard methods and thoroughly
degassed prior to use.
Bis(5-(diphenylphosphino)-3-thiapentanyl)amine (psnet). A solu-
tion of Na(psh), prepared from 0.55 g (23.9 mmol) of sodium and 5.88
g (23.9 mmol) of psh in 100 mL of ethanol, was added dropwise to a
solution of 2.13 g (11.9 mmol) of bis(2-chloroethyl)amine (generated
from the hydrochloride with NaOEt) in 100 mL of ethanol at 0 °C.
The reaction mixture was allowed to warm to room temperature with
stirring and was stirred for an additional 12 h, during which time a
white precipitate formed. The mixture was filtered, and the filtrate
was concentrated in vacuo. The resulting white paste was dissolved
in toluene to remove residual NaCl, the mixture was filtered, and the
filtrate was taken to dryness in vacuo. The product was obtained as
5.5 g (82%) of a sticky white solid; analytical data indicate an ethanol
Figure 1. Minimal scheme for the nickel-mediated reaction 2H+
+
2e- a H2. B is a neighboring base; if the scheme were to be related
to hydrogenase, the Ni-C designation would best apply to 3 or 4 with
monosolvate. FAB-MS (3-nitrobenzyl alcohol): m/z 562 (M + H+
-
EtOH). 1H NMR (C6D6): δ 2.25-2.28 (m, 4), 2.34 (t, 4), 2.44 (t, 4),
2.52-2.55 (m, 4), 5.02 (s, 1), 7.00-7.05 (m, 12), 7.34-7.38 (m, 8).
Anal. Calcd for C34H43ONP2S2: C, 67.19; H, 7.13; N, 2.30. Found:
C, 67.30; H, 6.85; N, 2.34.
1
S ) /2. A number of related reaction cycles have been proposed.1
nase potentials or the H+/H2 potential at pH 7 (-655 mV vs
SCE), and dihydrogen was formed as a byproduct. These
reductive conditions may subsume approaches i and ii, but in
no case was the reactive nickel species identified.11,12 With
several recent exceptions,10,12h approach iii was not pursued in
relation to hydrogenase.
[Bis(5-(diphenylphosphino)-3-dithiapentanyl)amine]nickel(II) Tet-
rafluoroborate ([Ni(psnet)](BF4)2). A solution of 3.33 g (9.80 mmol)
of [Ni(OH2)6](BF4)2 in 100 mL of ethanol was added dropwise to a
solution of 5.50 g (9.80 mmol) of psnet in 100 mL of ethanol. A deep
purple precipitate formed immediately. The reaction mixture was stirred
for 2 h, and the solid was collected by filtration. This material was
washed thoroughly with ether, toluene, tetrahydrofuran (THF), and again
with ether; it was recrystallized from methanol/ether. The product was
obtained as 4.7 g (61%) of a purple microcrystalline solid. FAB-MS
(3-nitrobenzyl alcohol): m/z 706 (M - BF4-). Visible spectrum
(propylene carbonate): λmax (ꢀM, M-1 cm-1) 526 (1490), 607 (sh, 1160)
nm. 1H NMR (acetone): δ 1.45 (t, 4), 3.05 (t, 4), 3.31 (br, 4), 3.44 (t,
4), 7.46 (m, 8), 7.66 (m, 12). Anal. Calcd for C32H37B2F8NNiP2S2:
C, 48.40; H, 4.70; N, 1.76; Ni, 7.39. Found: C, 48.58; H, 5.06; N,
1.64; Ni, 7.28. The compound was additionally identified by an X-ray
structure determination.
[Bis(5-(diphenylphosphino)-3-thiapentanyl)amine]nickel(I) Tet-
rafluoroborate ([Ni(psnet)](BF4)). To a solid mixture of 1.72 g (2.20
mmol) of [Ni(psnet)](BF4)2 and 0.099 g (2.60 mmol) of NaBH4 was
added 100 mL of THF with rapid stirring. A purple slurry formed
initially; over 1 h the reaction mixture became a deep brown-green
solution, from which a light green powder precipitated over the next 5
h. The precipitate was allowed to settle overnight and was collected
by filtration. To remove excess NaBF4, the green solid was dissolved
in acetonitrile, the mixture was filtered, and the solvent was removed
in vacuo. The crude product was washed extensively with ether and
recrystallized from THF/ether to afford the pure product as 0.92 g (60%)
of a green microcrystalline solid. Absorption spectrum (DMF): λmax
(ꢀM, M-1 cm-1) 259 (13 200), 357 (5500), 949 (107) nm. Anal. Calcd.
for C32H37BF4NNiP2S2: C, 54.35; H, 5.27; N, 1.98; Ni, 8.30. Found:
C, 54.20; H, 5.15; N, 1.88; Ni, 8.19. The compound was additionally
identified by an X-ray structure determination.
Chloro[bis(5-(diphenylphosphino)-3-thiapentanyl)amine]nickel-
(II) Chloride ([Ni(psnet)Cl]Cl). A solution of 0.86 g (2.6 mmol) of
[Ni(OH2)6]Cl2 in 100 mL of ethanol was added dropwise to a solution
of 2.0 g (3.6 mmol) of psnet in 100 mL of ethanol. The resultant deep
purple solution was stirred for 2 h; the solvent was removed in vacuo,
yielding a purple glass which was washed thoroughly with ether. The
solid was dissolved in ca. 75 mL of dimethylformamide (DMF),
forming a pastel green solution. Slow addition of ether followed by
filtration afforded the product as 1.85 g (75%) of light green powder.
FAB-MS (DMF): m/z (M - Cl) 654. Absorption spectrum (DMF):
λmax (ꢀM, M-1 cm-1) 259 (14 600), 630 (21), 968 (sh, 24), 1050 (27)
nm. Anal. Calcd for C32H37Cl2NNiP2S2: C, 55.60; H, 5.39; N, 2.03.
Found: C, 55.50; H, 5.29; N, 2.00.
Our present investigation centers on nickel-mediated dihy-
drogen evolution, for which a working hypothesis is represented
by the scheme in Figure 1. The catalytic cycle contains reaction
2a and thus approach ii-a to dihydrogen evolution. There has
been considerable recent progress in the generation, EPR
properties, and redox chemistry of various Ni(I) com-
plexes;3-5,14-16 however, very few of these complexes have been
isolated in substance16 nor has their reactivity toward protic acids
been tested. Initial complexes in reactions 2 were generated in
situ; in addition, assertions about dihydrogen formation in
reactions 2 are still speculative. We have sought routes to
Ni(I) species reactive toward Bro¨nsted acids but sufficiently
stable to be isolated. The approach taken in this work exploits
the redox versatility of nickel complexes formed with confor-
mationally flexible ligands containing mixed hard-soft donor
sets.6 Here we have utilized fully characterized Ni(II,I)
complexes of the pentadentate ligand bis(5-diphenylphosphino)-
3-thiapentanyl)amine and show that the Ni(I) species is capable
of stoichiometric dihydrogen evolution. Inasmuch as the H2-
evolving system developed here includes neither a Ni-Fe
complex nor thiolate ligation at the nickel site, properties of
the active site of DesulfoVibrio gigas (D. gigas) hydrogenase
established by crystallography,17 it cannot be construed, and is
not intended, as a close functional analogue of hydrogenase. It
is, instead, a first-generation experimental representation of
nickel-mediated dihydrogen evolution whose reactivity and
mechanism indicate certain features obligatory to a functional
analogue.
Experimental Section
Preparation of Compounds. Unless otherwise noted, all operations
were carried out under a pure dinitrogen atmosphere. Reagents were
(14) (a) Baidya, N.; Olmstead, M.; Mascharak, P. K. Inorg. Chem. 1991,
30, 929. (b) Baidya, N.; Olmstead, M.; Mascharak, P. K. J. Am. Chem.
Soc. 1992, 114, 9666. (c) Baidya, N.; Noll, B. C.; Olmstead, M. M.;
Mascharak, P. K. Inorg. Chem. 1992, 31, 2999. (d) Baidya, N.;
Olmstead, M. M.; Whitehead, J. P.; Bagyinka, C.; Maroney, M. J.;
Mascharak, P. K. Inorg. Chem. 1992, 31, 3612.
(15) Farmer, P. J.; Reibenspies, J. H.; Lindahl, P. A.; Darensbourg, M. Y.
J. Am. Chem. Soc. 1993, 115, 4665.
(16) Suh, M. P.; Kim, H. K.; Kim, M. J.; Oh, K. Y. Inorg. Chem. 1992,
31, 3620.
(17) Volbeda, A.; Charon, M.-H.; Piras, C.; Hatchikian, E. C.; Frey, M.;
Fontecilla-Camps, J. C. Nature 1995, 373, 580.
(18) Chatt, J.; Dilworth, J. R.; Schmutz, J. A.; Zubieta, J. A. J. Chem.
Soc., Dalton Trans. 1979, 1595.
(19) Stavropoulos, P.; Muetterties, M. C.; Carrie´, M.; Holm, R. H. J. Am.
Chem. Soc. 1991, 113, 8485.

4150 Inorganic Chemistry, Vol. 35, No. 14, 1996
James et al.
Bis[1,2-bis(diphenylphosphino)ethane]hydridonickel(II) Tetra-
fluoroborate ([Ni(diphos)2H](BF4). The following procedure is
somewhat simpler and affords a higher yield than that originally
reported.20 To a stirred solution of 0.95 g (1.1 mmol) of [Ni(diphos)2]21
in 40 mL of THF was added dropwise 220 µL (1.2 mmol) of HBF4‚
Et2O (85%, Aldrich). The resulting deep orange solution was stirred
for 30 min, during which a microcrystalline orange solid separated from
the solution. The solid was isolated by filtration, washed with ether,
and recrystallized from acetone/ether. The product was obtained as
0.75 g (72%) of an orange crystalline solid. 1H NMR (CD2Cl2): δ
-13.03 (quintet, 1), 2.38 (s, 8), 7.14 (br, 16), 7.19 (t, 16), 7.37 (t, 8).
Reactions of [Ni(psnet)](BF4) with Acids. All reactions were
carried out in 20 mL sealed vials using butyl rubber caps. Solutions
containing 15-20 mg (ca. 5 mM) of [Ni(psnet)](BF4) in 5 mL of
propylene carbonate (PC) or DMF were transferred to the vials and
sealed under dinitrogen. Acids were added using a gastight syringe;
reaction mixtures were rapidly stirred. Dihydrogen production was
monitored with a Varian Model 3700 gas chromatograph equipped with
a thermal conductivity detector, strip chart recorder, and integrator.
Resolution of headspace gases was accomplished with a 2-m column
of 0.6-cm-i.d. aluminum tubing packed with 60/80 mesh 5-Å molecular
sieves, operating at ambient temperature. Argon was used as the carrier
gas. Samples of headspace gas (200 µL) were removed from the sealed
vials with a Pressure-Lok syringe and injected directly into the gas
chromatograph. Dihydrogen yields were calculated by comparison of
integrated peak areas with standards prepared under identical conditions.
In experiments that determined the time course of dihydrogen evolution,
aliquots were removed from the reaction solutions at 30-45 s for the
first ca. 4 min of the reaction and at 2-9 min intervals thereafter. The
decrease in headspace pressure caused by the removal of multiple
aliquots was considered in the calculation of dihydrogen yields.
Kinetics Measurements. The kinetics of dihydrogen formation in
the reaction of [Ni(psnet)]+ with HCl in DMF were investigated. All
reactions were performed under strictly anaerobic conditions. Solvent
DMF (Aldrich, 99+%, 1.1 mM water as determined by Karl Fischer
titration) was stored over 3-4-Å molecular sieves in an inert atmosphere
box prior to use. The acid source was 1.0 M HCl in diethyl ether
(Aldrich, anhydrous) added as aliquots to DMF solutions; the titer of
the ether solution was monitored periodically. Before use, (Bu4N)-
(PF6) (Aldrich, 98%) was recrystallized from absolute ethanol and dried
in vacuo for 48 h, and Bu4NCl (Fluka, 99%) was dried in vacuo for 24
h.
Reactions were monitored with a Varian Cary 3 spectrophotometer
equipped with a cell compartment thermostated at 25.0 ( 0.5 °C.
Dihydrogen evolution was followed by monitoring the decay of [Ni-
(psnet)]+ absorbance at 380 nm, where overlap with product absor-
bances is minimized. The tight isosbestic point at 272 nm indicates
no measurable accumulation of intermediates. Initial concentrations
in the reaction systems were [[Ni(psnet)]+] ) 0.420-2.70 mM and
[HCl] ) 5.01-48.9 mM, depending on the reaction rate. In reactions
designed to determine the effects of Bu4NCl and (Bu4N)(PF6) on
reaction rate, aliquots of stock solutions of these compounds in DMF
were added to a solution of [Ni(psnet)]+ prior to the addition of HCl.
Initial concentrations were [Bu4NCl] ) 0.00120-0.600 M and [(Bu4N)-
(PF6)] ) 0.0125-0.508 M. Computations in data analysis were
performed with the commercial programs Kaleidagraph and Igor.
Collection and Reduction of X-ray Data. Diffraction-quality
crystals of [Ni(psnet)](BF4)2 (purple prisms) and [Ni(psnet)](BF4) (green
elongated plates) were obtained by diffusion of ether into solutions of
the compounds in methanol and THF, respectively. Crystals were
mounted in Paratone-N oil and attached to glass fibers. Diffraction
data were collected using graphite-monochromatized Mo KR radiation
on a Siemens R3m/v four-circle automated diffractometer equipped
with a Siemens LT-2 cryostat operating at 233 K. Unit cell parameters
were obtained by least-squares refinement of machine-centered reflec-
tions. Intensities of three standard reflections monitored every 97
reflections indicated no significant decay during the data collections.
Crystallographic data are summarized in Table 1. Corrections for
Lorentz and polarization effects were applied using the program XDISK
Table 1. Summary of Crystallographic Dataa
[Ni(psnet)](BF4)2
[Ni(psnet)](BF4)
formula
fw
cryst syst
space group
Z
C32H37B2F8N1Ni1P2S2
794.0
monoclinic
P21/c
C32H37B1F4N1Ni1P2S2
707.2
triclinic
P1h
2
4
a, Å
b, Å
c, Å
10.887 (3)
14.331 (3)
23.057 (5)
10.701 (3)
12.978 (3)
13.138 (3)
109.97 (2)
96.53 (2)
102.77 (2)
1636.3 (8)
1.44
R, deg
â, deg
γ, deg
V, Å3
103.43 (2)
3499.0 (14)
1.51
0.83
F
calc, g/cm3
µ, mm-1
0.87
4.38, 4.26
R,b Rw,c %
6.12, 6.48
a Obtained at T ) 233 K with graphite-monochromatized Mo KR
radiation (λ ) 0.710 69 Å). b R ) ∑|Fo| - |Fc|/∑|Fo|. c Rw ) {∑[w(|Fo|
2
2
- |Fc| )]/∑[w|Fo| ]}1/2
.
of the SHELXTL PLUS program package; empirical absorption
corrections were made with XEMP. Assignments of space groups from
statistics and systematic absences were confirmed by successful
solutions and refinements of the structures.
Solution and Refinement of Structures. Initial structure solutions
were obtained either by direct methods or Patterson maps. All non-
hydrogen atoms not located in these solutions were found by successive
Fourier or difference Fourier maps with intervening cycles of least-
squares refinement. All non-hydrogen atoms were described aniso-
tropically in both structures. In the final stages of the refinements,
hydrogen atoms were placed 0.96 Å from bonded carbon atoms and
0.90 Å from bonded nitrogen atoms and were assigned a temperature
factor of 0.08 Å2. Final difference Fourier maps generally exhibited
residual electron density only near anions, indicating possible disorder
that was not refined further. Final R-values are set out in Table 1.22
Other Physical Measurements. Absorption spectra were recorded
on Varian 2390 and Perkin-Elmer Lambda 4C spectrophotometers.
NMR spectra were obtained on a Bruker AM-400 spectrometer. EPR
spectra were determined at 120 K on a Bruker ESP 300-E spectrometer
operating at X-band frequencies. Spin concentrations were determined
from a calibration curve derived from the integrated spectra of standard
solutions of CuSO4 in glycerol:water(1:1 (v/v)). Integrated Ni(I) spectra
were corrected for transition probability differences23 between sample
and standard. Electrochemical measurements were performed under a
dinitrogen atmosphere with use of an EG&G Model 263 potentiostat.
For cyclic voltammetry experiments, 1-2 mM solutions containing 0.1
M Bu4NPF6 supporting electrolyte and a Pt disc working electrode were
used. Potentials were determined vs a SCE reference electrode.
Conductance measurements were performed with a Fisher Scientific
digital conductivity meter equipped with a Pt probe.
Results and Discussion
Nickel(II,I) Complexes. (a) Synthesis and Properties. The
objective of this investigation required a Ni(I) complex suf-
ficiently stable to permit full structural and electronic charac-
terization and reactive toward protic sources. The procedure
employed utilizes pentadentate ligand psnet and is outlined in
Figure 2. Upon recrystallization from methanol/ether, [Ni-
(psnet)](BF4)2 was obtained as deep purple prisms that are
soluble in acetonitrile, acetone, propylene carbonate, and DMF.
The compound is air-stable for days in solution and indefinitely
stable in the solid state. The complex [Ni(psnet)]2+ is diamag-
netic in acetone (1H NMR) and exhibits an absorption spectrum
in PC, shown in Figure 3, that resembles that of low-spin
Ni(II) species in distorted trigonal bipyramidal environments.24
(22) See paragraph at the end of this article concerning Supporting
Information.
(23) Aasa, R.; Va¨nnga˚rd, T. J. Magn. Reson. 1975, 19, 308.
(20) Schunn, R. A. Inorg. Chem. 1970, 9, 394.
(21) Chatt, J.; Hart, F. A.; Watson, H. R. J. Chem. Soc. 1962, 2537.

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4151
Figure 2. Preparation of Ni(II,I) complexes utilized in this work. The structures of [Ni(psnet)]2+,1+ have been established by X-ray determinations;
the structure of [Ni(psnet)Cl]1+ is proposed.
is not reliable. Structural and other electronic properties of [Ni-
(psnet)]+ are described below.
The complex [Ni(psnet)Cl]+ was required to identify the
nickel-containing product of the protonation reaction of [Ni-
(psnet)]+ with HCl. It was obtained as the light green chloride
salt which in ethanol solution exists as purple low-spin [Ni-
(psnet)]2+ according to the absorption spectrum. However, a
DMF solution has the light green color of the solid, and chloride
binding was demonstrated by FAB-MS and the absorption
spectrum (Figure 3). The latter is typical of octahedral Ni(II),24
3
3
exhibiting a split band of parentage A2g f T2g at 968 and
3
3
1050 nm and the A2g f T1g transition at 630 nm. The third
octahedral feature at higher energies was obscured by charge-
transfer absorption. The solid state magnetic moment of 3.12
µB (room temperature) is also consistent with an octahedral
structure.
(b) Structures. The crystallographically determined struc-
tures of five-coordinate [Ni(psnet)]2+,1+ are presented in Figure
4; selected metric parameters are set out in Table 2. These are
the first structures of nickel complexes with NP2S2 coordination
spheres. Both approach idealized C2 microsymmetry, with the
pseudosymmetry axis coincident with the Ni-N bond axis. The
structure of [Ni(psnet)]2+ is distorted from either the square
pyramidal (SP) or trigonal bipyramidal (TBP) limit. The bond
angles P(1)-Ni-N(1) ) 103.4(3)° and P(2)-Ni-N(1) )
112.3(3)° are between the limits for the two structures, as is
P(1)-Ni-P(2) ) 144.3(1)°. The angle S(1)-Ni-S(2) )
176.1(1)° is consistent with either geometry; in the TBP limit
the sulfur atoms are in axial positions. The Ni-N and mean
Ni-P and Ni-S bond distances are comparable to those in other
low-spin Ni(II) complexes with related donor sets.26
Upon reduction of [Ni(psnet)]2+ to [Ni(psnet)]+, the P(1)-
Ni-N and P(2)-Ni-N angles expand to 115.6(1) and
119.8(1)°, respectively, while the P(1)-Ni-P(2) angle contracts
to 124.6(1)°. Although the S(1)-Ni-S(2) angle decreases by
ca. 10° (to 166.5(1)°), the overall angular changes are in the
direction of TBP coordination, in which sulfur atoms are axially
positioned. Because of the small number of structurally
Figure 3. Absorption spectra of Ni(II,I) complexes: (‚‚‚) [Ni(psnet)]-
(BF4)2 in propylene carbonate; (- - -) [Ni(psnet)](BF4) in DMF; (s)
[Ni(psnet)Cl]Cl in DMF. Band maxima are indicated.
The spectrum in PC solution reveals a characteristically intense
1
1
1
1A1′ f E′ transition at 526 nm; the low-intensity A1′ f E′′
feature expected at higher energies is obscured by charge-
transfer transitions. A shoulder near 600 nm presumably reflects
the lowering of symmetry from idealized D3h coordination.
Reduction of [Ni(psnet)]2+ with NaBH4 in THF affords green
-
[Ni(psnet)]+, obtained in 60% yield as the BF4 salt. The
compound is soluble in acetonitrile, acetone, PC, THF, and
DMF; in solution it is oxidized within minutes to [Ni(psnet)]2+
but is stable to air for hours in the solid state. The solid state
magnetic moment of 2.10µB (room temperature) indicates an S
1
) /2 ground state. A single broad band at 949 nm in the
absorption spectrum in DMF solution (Figure 3) is reminiscent
of the unresolved d-d spectra of certain trigonal bipyramidal
Cu(II) complexes.24a,25 Because d-d spectral/structure correla-
tions of Ni(I) complexes are incompletely developed, an
assignment of stereochemistry based on the absorption spectrum
(26) (a) Haugen, L. P.; Eisenberg, R. Inorg. Chem. 1969, 8, 1072. (b)
Dapporto, P.; Sacconi, L. J. Chem. Soc. A 1970, 1804. (c) Bianchi,
A.; Dapporto, P.; Fallani, G.; Ghilardi, C. A.; Sacconi, L. J. Chem.
Soc., Dalton Trans. 1973, 641. (d) DiVaira, M.; Sacconi, L. J. Chem.
Soc., Dalton Trans. 1975, 493. (e) Orpen, A. G.; Brammer, F. H.;
Allen, O. K.; Watson, R. T. J. Chem. Soc., Dalton Trans. 1989, Suppl.
S1.
(24) (a) Lever, A. B. P. Inorganic Electronic Spectroscopy; Elsevier: New
York, 1984, Chapter 6. (b) Sacconi, L.; Mani, F.; Bencini, A. In
ComprehensiVe Coordination Chemistry; Wilkinson, G., Ed.; Perga-
mon Press: Oxford, U.K., 1987; Vol. 5, pp 1-347.
(25) Tyagi, S.; Hathaway, B.; Kremer, S.; Stratemeier, H.; Reinen, D. J.
Chem. Soc., Dalton Trans. 1984, 2087.

4152 Inorganic Chemistry, Vol. 35, No. 14, 1996
James et al.
Figure 5. Cyclic voltammogram (50 mV/s) of [Ni(psnet)]2+ in
propylene carbonate solution. Peak potentials are indicated (E1/2
-0.21 V vs SCE).
)
TBP Ni(I) complex with thioether ligation.19 However, the most
important structural result is the lengthening of Ni-S bond
lengths by 0.24-0.25 Å, while the Ni-N and Ni-P distances
are nearly constant, their changes being j0.03 Å. This requires
that the added electron occupy an antibonding MO concentrated
along the S-Ni-S axial portion of the coordination unit (vide
infra).
(c) Electron Transfer. The cyclic voltammogram of [Ni-
(psnet)]2+ in PC, shown in Figure 5, reveals a one-electron
reduction at E1/2 ) -0.21 V that is electrochemically quasi-
reversible and chemically reversible. The peak current function
ip/ν1/2 is independent of scan rate, and ipa/ipc ≈ 1 at scan rates
of 20-200 mV/s. Similar behavior was observed in acetonitrile
and DMF solutions using both glassy carbon and Pt disc
electrodes. Coulometric reduction at a Pt gauze electrode at
-0.37 V (E1/2 ) -0.17 V) induces the passage of 0.96 F, while
subsequent oxidation at 0.03 V results in the transfer of 0.95 F,
indicating essentially no decomposition over the 1-h time scale
of the experiment. Indeed, the electrolytic cycle may be
repeated four times in sequence without appreciable deterioration
of the electroactive species. The cyclic voltammogram of
[Ni(psnet)]+ is identical to that in Figure 5.
Figure 4. Structures of [Ni(psnet)]2+ (top) and [Ni(psnet)]+ (bottom)
showing 40% and 50% probability ellipsoids, respectively, the atom
labeling scheme, and selected bond distances which illustrate the
lengthening of the Ni-S bonds upon reduction.
Table 2. Selected Interatomic Distances (Å) and Angles (deg) in
[Ni(psnet)](BF4)2 (I) and [Ni(psnet)](BF4) (II)
A second quasi-reversible one-electron reduction (not shown)
of [Ni(psnet)]2+, attributed to a Ni(I)/Ni(0) couple, is observed
at E1/2 ) -1.36 V in PC solution. A notable feature of this
couple is the large peak-to-peak separation of 340 mV at the
moderate scan rate of 50 mV/s, a property consistent with a
rate-limiting structural change29 upon reduction to Ni(0). Given
the tendency of Ni(0) to form tetrahedral complexes with soft
donor sets, the most reasonable structural change associated with
this couple is decoordination of the secondary amine, affording
a four-coordinate complex with P2S2 ligation. We have recently
demonstrated the decoordination of two pyridyl ligands from a
six-coordinate Ni(II) complex upon reduction to Ni(I), leaving
a P2S2 coordination sphere.6 Oxidation of [Ni(psnet)]2+ in PC
and acetonitrile is irreversible, with Epa J 1.0 V.
I
II
Ni(1)-S(1)
Ni(1)-S(2)
Ni(1)-P(1)
Ni(1)-P(2)
Ni(1)-N(1)
2.183(3)
2.189(3)
2.230(3)
2.218(3)
2.144(9)
Ni(1)-S(1)
Ni(1)-S(2)
Ni(1)-P(1)
Ni(1)-P(2)
Ni(1)-N(1)
2.429(1)
2.426(1)
2.245(2)
2.242(2)
2.128(4)
S(1)-Ni(1)-S(2)
S(1)-Ni(1)-P(1)
S(2)-Ni(1)-P(1)
S(1)-Ni(1)-P(2)
S(2)-Ni(1)-P(2)
P(1)-Ni(1)-P(2)
S(1)-Ni(1)-N(1)
S(2)-Ni(1)-N(1)
P(1)-Ni(1)-N(1)
P(2)-Ni(1)-N(1)
176.1(1)
89.3(1)
92.1(1)
92.3(1)
88.8(1)
144.3(1)
88.3(3)
87.8(3)
103.4(3)
112.3(3)
S(1)-Ni(1)-S(2)
S(1)-Ni(1)-P(1)
S(2)-Ni(1)-P(1)
S(1)-Ni(1)-P(2)
S(2)-Ni(1)-P(2)
P(1)-Ni(1)-P(2)
S(1)-Ni(1)-N(1)
S(2)-Ni(1)-N(1)
P(1)-Ni(1)-N(1)
P(2)-Ni(1)-N(1)
166.5(1)
87.2(1)
100.3(1)
97.6(1)
87.4(1)
124.6(1)
83.6(1)
83.1(1)
115.6(1)
119.8(1)
(d) Ground State of [Ni(psnet)]1+. The EPR spectrum of
[Ni(psnet)]1+ in frozen 2-MeTHF/acetonitrile solution, presented
in Figure 6, is rhombic with g1 ) 2.21, g2 ) 2.13, and g3 )
2.01. Each line is split into a triplet by coupling with two
equivalent 31P nuclei (a1 ) 45 G, a2 ) 40 G, a3 ) 44 G); no
14N hyperfine splitting was observed. Spin quantitation gave
0.97 spin/Ni. The spectrum of [Ni(psnet)]+ in fluid THF
solution at ambient temperature is a triplet with gav ) 2.12 and
aav ) 42 G.
characterized TBP Ni(I) complexes with similar ligands,
dimensional comparisons are limited. We note that the mean
Ni-P distance in [Ni(psnet)]+ (2.244 Å) is close to the mean
Ni-P equatorial distance in [Ni{P(CH2CH2PPh2)3}]+ (2.219
Å)27 and [Ni{P(CH2CH2PPh2)3}I] (2.268 Å).28 Also, the mean
Ni-S axial distance of 2.428 Å is 0.07 Å longer than the mean
Ni-S equatorial distance in the only other structurally defined
The EPR spectra demonstrate metal-centered reduction, and
the observation that g1, g2 > g3 ≈ 2 is consonant with Ni(I)
(27) Cecconi, F.; Midollini, S.; Orlandini, A. J. Chem. Soc., Dalton Trans.
1983, 2263.
(28) Dapporto, P.; Sacconi, L. Inorg. Chim. Acta 1980, 39, 61.
(29) Geiger, W. E. Prog. Inorg. Chem. 1985, 33, 275.

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4153
Table 3. Yields of Dihydrogen Evolution Reactions
%
equiv
yield
H2
b
compound
acid
H+ solvent timea
[Ni(psnet)](BF4)
HBF4‚Et2Oc
8
1
8
1
8
8
8
8
8
8
8
8
8
8
DMF
PC
PC
DMF
DMF
PC
DMF
PC
DMF
PC
DMF
PC
DMF
PC
d
<5
91
55
5 min
d
120 min 88
45 min
45 min
24 h
12 h
d
12 h
d
HCl/Et2Oe
90
92
86
94
61
95
57
94
51
97
aq HClf
aq HBrf
aq HIf
6 h
d
24 h
f
aq H2SO4
Figure 6. EPR spectrum of [Ni(psnet)]+ in 7/3 2-MeTHF:MeCN (v/
v) at 120 K; g-values and a(31P) values are indicated.
standards
NaHB(OMe)3
[Ni(diphos)2H](BF4) HBF4‚Et2O
aq HCl
8
1
8
1
8
THF
CH2Cl2 2 h
CH2Cl2 30 min
THF
THF
30 min
97
87
94
85
90
with the odd electron in the dz orbital.30 This in turn is
2
[Ni(NS3tBu)H]BPh4 HCl/Et2O
1 h
15 min
consistent with the ca. 0.24-Å expansion of the Ni-S bonds
upon reduction to [Ni(psnet)]+. This observation has several
precedents. For example, the axial Ni-N and Ni-I bonds in
[Ni{N(CH2CH2PPh2)3}I]28 are longer by 0.13 and 0.30 Å,
respectively, compared to the corresponding bond lengths in
[Ni{N(CH2CH2PPh2)3}I]+.26b A similar change in axial Ni-N
bond lengths was found in [Ni{N(CH2CH2AsPh2)3}I],28 as
compared with the Ni(II) compound [Ni{N(CH2CH2AsPh2)3}-
Ph](BPh4).31 Lastly, extended Hu¨ckel calculations32 based on
the crystallographic structure of [Ni(psnet)]+ substantiate the
a Approximate times at which maximum yields were obtained. For
times e1 h, gas samples were removed at 5, 15, 30, 45, 60, and 120
min; for longer reactions, gas samples were removed every 6 h. b Yield
basis: 1H2/2Ni for [Ni(psnet)]+, 1 H2/1 BH or 1 Ni for standards. c 85%,
used as the neat liquid. d Several days for low-yield reactions. e 1.0 M
HCl in ether. f 1.0 M aqueous acid.
[Ni(diphos)2H]+ + HBF4 f H2 + [Ni(diphos)2](BF4)2 (3)
[Ni(NS3tBu)H]+ + HCl f H2 + [Ni(NS3tBu)Cl]+ (4)
1
2
ground state configuration ...(dz ) . The HOMO is a σ* orbital
2.32 eV above the last filled d-type MO. Its metal component
2
is nearly entirely 3dz (33%; 2% 4py) with significant contribu-
tions from phosphorus (10%) and sulfur (39%). The nitrogen
component is 4%; all other contributions are <1%.
NiII-H- + H+ f NiII + H2
(5)
Dihydrogen Evolution Reactions. With [Ni(psnet)]+ es-
tablished as an authentic Ni(I) complex of known structure,
ground state, and redox capacity, its reactions with Bronsted
acids were next investigated. Previously, stoichiometric dihy-
drogen evolution from Ni(II) hydrides7 and boron hydrides33
in the presence of acids had been documented. Three mono-
hydride sources, NaBH(OMe)3, [Ni(diphos)H](BF4),20 and [Ni-
(NS3tBu)H](BPh4)19 were chosen as primary standards for the
purposes of verifying the efficacy of the H2 quantitation
procedure and probing the proton-reducing ability of well-
characterized Ni(II) hydrides. With reference to Table 3,
reaction of all three compounds with excess strong acid produces
g90% yields of dihydrogen. For the two Ni(II) complexes,
dihydrogen is evolved in reactions 3 and 4, minimally repre-
sented as reaction 5. When one equivalent of HBF4‚Et2O was
used, [Ni(diphos)2](BF4)2 was isolated in quantitative yield from
a large-scale reaction. When reaction 4 was conducted with 1
equiv of ethereal HCl, the known product chloride complex19
was formed in an in situ yield of 84% determined spectropho-
tometrically. With this demonstration of high-yield stoichio-
metric protonations of nickel(II) hydrides, attention was directed
to the mediation of dihydrogen formation by an initial Ni(I)
reactant.
Under appropriate conditions determined empirically, [Ni-
(psnet)]+ reacts with ethereal HCl, HBF4·Et2O, and aqueous HCl,
HBr, HI, and H2SO4 to yield dihydrogen (Table 3). The rate
of H2 production is highly dependent on the acid used, with
nonaqueous acids reacting much faster than aqueous acids in
DMF and PC solutions. For example, 1 equiv of HBF4‚Et2O
reacts almost instantaneously with [Ni(psnet)]+ in PC, causing
a color change from green to purple and affording a 91% yield
of H2 based on the overall stoichiometry of reaction 6.
NiI + H+ f NiII + 1/2H2
(6)
Similarly, ethereal HCl gave yields of 88-92% in 45 min or
less. In contrast, maximum H2 production with aqueous acids
(used in excess) is reached after some hours. Under constant
conditions, the rate of H2 evolution with hydrohalic acids stands
in the order of aqueous acid strength.
(30) Maki, A. H.; Edelstein, N.; Davison, A.; Holm, R. H. J. Am. Chem.
Soc. 1964, 86, 4580.
(31) Sacconi, L.; Dapporto, P.; Stoppioni, P. Inorg. Chem. 1976, 15, 325.
(32) Extended Hu¨ckel calculations employed the program CACAO (PC
version 4.0): Mealli, C.; Proserpio, D. M. J. Chem. Educ. 1990, 67,
399. In these calculations, the z-axis is in the Ni-S direction and the
x-axis in the Ni-N direction. Parameters: (a) Summerville, R. H.;
Hoffmann, R. J. Am. Chem. Soc. 1976, 98, 7240. (b) Lauher, J. W.;
Elian, M.; Summerville, R. H.; Hoffmann, R. J. Am. Chem. Soc. 1976,
98, 3219. (c) Chen, M. M. L.; Hoffmann, R. J. Am. Chem. Soc. 1976,
98, 1647.
(33) (a) Brown, H. C.; Schlesinger, H. I.; Sheft, I.; Ritter, D. M. J. Am.
Chem. Soc. 1953, 75, 192. (b) Cotton, F. A.; Wilkinson, G. AdVanced
Inorganic Chemistry, 5th ed.; Wiley-Interscience: New York, 1988,
Chapter 6.
Interestingly, certain other strong acids, such as p-toluene-
sulfonic acid in DMF and aqueous HNO3 and HBF4 in DMF,
produce no H2 after incubation with [Ni(psnet)]+ for over 72
h. As shown spectrophotometrically, p-toluenesulfonic acid
quantitatively oxidizes [Ni(psnet)]+ to [Ni(psnet)]2+ within
minutes. Nitric acid causes a slow color change from green to
brown, but we could not identify the reaction product. Aqueous
HBF4 shows no reaction with [Ni(psnet)]+ in DMF.
The nickel-containing products of all protonation reactions
tested are EPR-silent, identifying them as Ni(II) species. As

4154 Inorganic Chemistry, Vol. 35, No. 14, 1996
James et al.
Figure 8. Time course of the consumption of [Ni(psnet)]+ ) Ni(I)
and production of H2 in a reaction system in DMF initially 2.67 mM
in Ni(I) and 30 mM in HCl: 0, micromoles of H2 determined by GC;
O, [Ni(I)] calculated from H2 yields under the stoichiometry of reaction
6. The solid line is [Ni(I)] determined spectrophotometrically.
Figure 7. Spectrophotometric display of the time course of a reaction
system in DMF initially 1.05 mM in [Ni(psnet)]+ and 25 mM in HCl.
The spectra were recorded over 1 h; band maxima and an isosbestic
point are indicated.
357 nm diminishes, while the band at 259 nm due to [Ni(psnet)-
Cl]+ increases in intensity as the reaction proceeds. A clean
isosbestic point is observed at 272 nm, indicating negligible
accumulation of intermediates. The final UV-visible spectrum
is identical with that of authentic [Ni(psnet)Cl]+ (Figure 3).
Presented in Figure 8 are the time courses for H2 production
(determined by GC) and [Ni(psnet)]+ consumption (monitored
spectrophotometrically) in a typical reaction. Both sets of points
are derived from the reactions performed with the same DMF
stock solution of [Ni(psnet)]+. Also shown is a plot of
[Ni(psnet)]+ concentration as calculated from H2 yields assum-
ing the stoichiometry of reaction 6. The near-coincidence of
calculated and observed Ni(I) concentrations indicates that H2
evolution tracks very closely with Ni(I) consumption, from
which we conclude that H2 production is directly coupled to
Ni(I) oxidation. For both the observed and calculated values
of [NiI], plots of 1/[NiI] vs time are linear to three half-lives,
showing that both data sets support kinetics second order in
Ni(I) (vide infra).
(d) HCl Equilibria in DMF. Reported values for the pKa
of HCl in DMF occur in the range 1.4-3.6.34 The water content
of the solvent used in the determination of these constants varied
from 5 mM (pKa ) 3.2-3.6) to 100 mM (pKa ) 1.4). Because
pKa values in nonaqueous media may vary significantly ac-
cording to solvent preparation and water content,34 we have
redetermined the dissociation constant of HCl in the DMF
solvent lot used for kinetics measurements. In addition to the
intrinsic acidity of reaction 7, Kolthoff et al.35 have shown that
determined spectrophotometrically, reaction of [Ni(psnet)]+ with
1 equiv of HBF4·Et2O in PC produces [Ni(psnet)]2+ in 90%
yield. Further, when 1-20 equiv of ethereal HCl are used in
DMF solution, [Ni(psnet)Cl]+ is formed in 89-94% yields. The
reaction products were additionally identified by FAB-MS. The
solution structures of Ni(II) products formed in other reaction
systems (Table 3) were not determined.
Kinetics of Dihydrogen Evolution. We have noted above
several minimal pathways that have been proposed for the
reduction of protons by low-valent nickel complexes. However,
no detailed mechanistic investigation of nickel-based proton
reduction with fully characterized reactants and products has
been reported. Given the finding that several protic sources
react with [Ni(psnet)]+ to produce H2 according to the stoichi-
ometry of reaction 6, these H2 evolution systems based on
protonation of [Ni(psnet)]+ present an excellent opportunity for
such an investigation.
(a) Selection of Reaction System. The system [Ni(psnet)]+/
ethereal HCl in DMF has been selected for kinetics analysis.
Its reaction is essentially quantitative (Table 3) and proceeds at
a rate convenient for spectrophotometric monitoring. On the
basis of a much faster reaction rate and lower H2 yield with
excess acid, the system with HBF4‚Et2O as the protic source in
PC is considerably less attractive. Systems containing aqueous
acids in DMF or PC solvent are not optimal, given the sensitivity
of acid dissociation constants in nonaqueous media to small
fluctuations in water content;34 these constants may enter into
a kinetics analysis (vide infra). Finally, DMF was chosen over
PC as the solvent because quantitative chloride binding in
[Ni(psnet)Cl]+ was not observed in the latter. Solutions of
[Ni(psnet)Cl]Cl in PC invariably contain mixtures of low-spin
[Ni(psnet)]2+ and high-spin [Ni(psnet)Cl]+, thereby complicating
the absorption spectra. All observations hereafter refer to the
reaction system [Ni(psnet)]+/1.0 M HCl in Et2O/DMF.
(b) Intermediates. Several reactions were performed by
incubating [Ni(psnet)]+ (ca. 1 mM) with excess acid in an EPR
tube at -10 to -40 °C for 5 s to 5 min followed by immersion
in liquid nitrogen. In all cases, the EPR spectrum of unreacted
[Ni(psnet)]+ exhibited diminished intensity compared with the
initial solution, and no other signals were observed. While not
detecting an odd-spin intermediate, the procedure is slow and
the possibility of a short-lived NiIII-H- intermediate remains
open.
HCl h H+ + Cl-
Ka
(7)
(8)
(9)
-
2HCl h H2Cl2 h H+ + HCl2
-
HCl + Cl- h HCl2
Kf
the acidity of HCl in DMF solutions is influenced by the
homoconjugation equilibrium 8. They estimated the formation
-
constant of HCl2 in reaction 9 to fall in the range 100-200
M-1. Equilibrium constants Ka and Kf are implicated in the
kinetics analysis that follows.
The equilibrium constants for dissociation and homoconju-
gation of HCl in DMF were determined conductimetrically in
(34) Izutsu, K. Acid-Base Dissociation Constants in Dipolar Aprotic
SolVents; IUPAC Chemical Data Series No. 35; Blackwell Scientific:
Oxford, U.K., 1990.
(35) Kolthoff, I. M.; Chantooni, M. K., Jr.; Smagowski, H. Anal. Chem.
1970, 42, 1622.
(c) Time Course of H2 Evolution. Spectral changes at 250-
450 nm accompanying the reaction of [Ni(psnet)]+ with ethereal
HCl are depicted in Figure 7. The feature of [Ni(psnet)]+ at

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4155
Figure 10. Dependence of reaction rate on the concentration of HCl
at [NiI] ) 0.55-1.00 mM, [HCl]0 ) 7.5-29.5 mM, and 25 °C. For
the plot with points b, [Cl-]0 ) 0.122 M; the solid line is a fit of the
data to the equation kobs ) 0.0015 + 99.9[HCl]0 (R ) 0.999). For the
plot with points 9, [Cl-]0 ) 0.090 M; the solid line is a fit of the data
to the equation kobs ) 0.0018 + 91.3[HCl]0 (R ) 0.998).
Figure 9. Dependence of reaction rate on the concentration of HCl at
[NiI] ) 1.00 mM, [HCl]0 ) 12-49 mM, and 25 °C. The solid line is
a fit of the data to the equation ln kobs ) 12.7 + 2.18 ln[HCl]0 (R )
0.996).
solutions that contained <5% (v/v) of ether.36 A plot of molar
conductance Λm vs [HCl]1/2 exhibited marked departure from
-
linearity when [HCl] > 0.01 M, consistent with HCl2 forma-
tion. From data in the regime [HCl] < 1.6 × 10-3 M, where
homoconjugation is minimal, the dissociation constant was
determined to be Ka ) 1.3 × 10-5 M (pKa ) 4.9). This marked
increase in the pKa value compared with the cited literature
values is attributed to the lower water content in the solvent lot
used here (e1.1 mM) and the presence of 0-5% (v/v) of ether.
-
The formation constant for HCl2 was estimated by plotting
Λm([HCl]([HCl] + 1/Kf))1/2 vs [HCl] for various values of Kf.35
Linear plots were obtained for the values Kf ) 20-250 M-1
,
with the best fits obtained for Kf ) 40-80 M-1
.
(e) Reaction Order and Anion Effect. The kinetics of
reaction 6 (NiI ) [Ni(psnet)]+) were investigated spectropho-
tometrically by following the diminution of the [Ni(psnet)]+
absorbance at 380 nm. Protonation reactions performed with
>8 equiv of HCl exhibited clean second order kinetics with
respect to Ni(I), as indicated by linear plots of {(A0 - A∞)/
(At - A∞)[NiI]0} vs time (not shown). Reaction rates at various
HCl concentrations were determined from the slopes of these
plots. Within a concentration range that gave reasonable initial
absorbances of [Ni(psnet)]+ (>0.42 mM), the systems displayed
pseudo-second-order kinetics for [HCl]0/[NiI] ) 8-100. At the
ratios [HCl]0/[NiI] . 100, the reaction rates prohibited observa-
tion of the initial phase of the reaction (ca. 25% completion)
by the method employed. The dependence of observed rates
vs [HCl]0 under the conditions [HCl]0/[NiI] ≈ 10-50 is shown
in the logarithmic plot of Figure 9. On the basis of the
relationship ln kobs ) ln k + a ln[HCl]0, the data were
satisfactorily fitted with a ) 2.18, corresponding to a reaction
apparently mixed order in HCl.
The addition of chloride as Bu4NCl was found to accelerate
reaction 6 dramatically. Further, the apparent reaction order
in Ni(I) was found to be dependent on the added chloride
concentration. Three regimes were identified: clean first-order
behavior at high [Cl-]0; mixed first-order-second-order at
intermediate values of [Cl-]0; clean second-order behavior at
low [Cl-]0. More specifically, at [NiI] ) 0.42 mM and [Cl-]0
> 30 mM, the reaction becomes first order in Ni(I) as
demonstrated by linear plots of ln{(At - A∞)/(A0 - A∞)} vs
time (not shown), while, at [Cl-]0 < 1.6 mM, the reaction
Figure 11. Dependence of reaction rate on the concentration of [Bu4-
NCl] at [NiI] ) 0.42-0.55 mM, [Bu4NCl]0 > 30 mM, and 25 °C. For
the plot with points b, [HCl]0 ) 25 mM; the solid line is a fit of the
data to the equation kobs ) 91.8[Cl-]/{(1 + 0.26[Cl-] + 12.1[Cl-]2)(1
+ 21[Cl-])}. For the plot with points 9, [HCl]0 ) 15 mM; the solid
line is a fit of the data to the equation kobs ) 56.3[Cl-]/{(1 + 0.28[Cl-]
+ 12.1[Cl-]2)(1 + 22[Cl-])}.
exhibits a second-order dependence on Ni(I). In the regime
[Cl-]0 > 30 mM, the reaction becomes first order in HCl, as
demonstrated by the linear plots in Figure 10. When [HCl]0 is
held constant, the rate exhibits the complex bell-shaped
dependence on [Cl-] shown in Figure 11. When [HCl]0 was
varied at constant [Bu4NPF6] ) 0.125 M, the data were
satisfactorily fitted to the equation given above for Figure 9
with a ) 2.09 (not shown), consistent with either a second-
order or mixed-order reaction.
To examine the anion effect further, reactions were performed
at various concentrations of (Bu4N)(PF6) in the presence and
absence of Bu4NCl. In the latter case, second-order kinetics in
Ni(I) were observed. In the presence of both salts, a mixed-
order dependence in Ni(I) was found under conditions where
first-order dependence applies in the absence of (Bu4N)(PF6).
A plot of the observed reaction rate vs [Bu4NPF6] in the absence
(36) Method of data analysis: French, C. M.; Roe, I. G. Trans. Faraday
Soc. 1953, 49, 314.

4156 Inorganic Chemistry, Vol. 35, No. 14, 1996
James et al.
Scheme 1
the species in Scheme 1 do not carry oxidation states (e.g., Ni-
H2+ ) NiIII-H-).38
Figure 12. Dependence of reaction rate on the concentration of (Bu4N)-
(PF6) at [NiI] ) 0.55 mM, [HCl]0 ) 25 mM, [Bu4NPF6] ) 0-0.5 M,
The most reasonable initial step is protonation of the Ni(I)
center, as in eq 10. The product dication reacts with 1 or 2
equiv of chloride in eq 11 to form the corresponding ion pairs.
The concentrations of these species (and their reaction products)
in different [Cl-] regimes dictate the chloride-dependent effects
on rate and reaction order. Following their formation, two
reaction pathways A and B promote H2 production, with path
A prevailing at low chloride concentrations and path B at high
chloride concentrations.
and 25 °C. The solid line is a fit of the data to the equation kobs
)
(10.6 + 130.7[Bu4NPF6])2/(1 + 19.9[Bu4NPF6] + 38.2[Bu4NPF6]2)2.
of added chloride (second order in Ni(I)) is presented in Figure
12. The PF6- anion slows the reaction markedly compared to
systems in which [Ni(psnet)](BF4) and HCl are the only sources
of anions.
The influence of anion concentration on the rate of reaction
6 cannot be adequately explained by a general effect of ionic
strength on the nature of the reaction medium. Such a model
Path A. Both the mono- and dichloride ion pairs collapse
to bound nickel(III) chloride-hydride species (reactions A1 and
A2). (Given the metal oxidation state, one or both phosphine
groups are likely to decoordinate to maintain a coordination
number e 6.) Subsequent rate-determining reaction of two such
species in the three possible combinations (reactions A3-A5)
affords H2, NiII-Cl-, and free chloride in two cases. These
steps correspond to minimal reaction 2b.
Path B. At high chloride concentration, the dichloride ion
pair undergoes rate-determining collapse with partial ligand
decoordination to a bound nickel(III) hydride-dichloride (reac-
tion B1). Thereafter, H2 formation can occur bimolecularly
(reaction B2, an elaboration of reaction 2b) or by an electron-
transfer sequence (reactions B3-B5); these routes cannot be
distinguished past the rate-determining step. Derivations of the
rate laws in the sections which follow are outlined in the
Appendix.
The two-term rate law corresponding to the reactions in
Scheme 1 is given by eq 12; Kn ) kn/k-n (n ) 1-5). Several
simplifying assumptions have been made in its derivation.
Following protonation reaction 10, the ion pair equilibria of eq
11 are established immediately, such that the Ni-H2+ concen-
tration before reaction can be expressed as a sum of the
concentrations of the three hydride species. The ion-pair status
of such species does not importantly influence the rate of their
bimolecular collapse to form H2 and Ni-Cl+ (k6 ) k7 ) k8).
The first and second terms describe the reaction order at low
would predict a monotonic increase or decrease in reaction rate
- 37
as a function of anion concentration for both Cl- and PF6
.
The quadratic dependence of [Cl-] contradicts this prediction.
Also, the change from first-order kinetics in Ni(I) at high [Cl-]/
[PF6-] to second-order behavior at low concentration ratios
suggests that a specific nickel-anion interaction occurs in the
rate-determining step. Given the propensity of charged species
to form ion pairs in nonaqueous media, the data strongly imply
that specific nickel-chloride ion pair(s) are responsible for the
observed chloride effect. The rate-retarding effect of (Bu4N)-
(PF6) at high [PF6-]/[Cl-] may be attributed to the displacement
-
of Cl- with PF6 in the ion pair.
(f) Rate Law and Proposed Mechanism. The overall
reaction 6 can be variously formulated in terms of coupled
elementary reactions, such as the minimal steps of eq 2a,b. These
may include metal protonation, electron transfer, hydride
protonation, and bimolecular collapse of two NiIII-H- species.
In the systems at hand, other elementary steps including ligand
dissociation, ligand protonation, chloride binding, and ion
pairing were considered. Various combinations of elementary
steps have been formulated, and the rate law for each has been
derived under the rapid-equilibrium or steady-state approxima-
tion. The number of possible reaction sets is too large for a
full description. We begin with that shown in Scheme 1, which
fits satisfactorily all experimental data, and subsequently show
that two other seemingly viable schemes do not. For simplicity,
(38) In the notation of Schemes 1-4, a vertical line signifies an ion pair,
with the species to the right of the line uncoordinated (e.g., Ni-
H2+|Cl- ) nickel(III)-hydride chloride 1:1 ion pair). The absence
of a vertical line implies that all species are coordinated (e.g., Ni-
(H)Cl2 ) nickel(III)-hydride/bis chloride complex (requiring some
decoordination of the psnet ligand)).
(37) Moore, J. W.; Pearson, R. G. Kinetics and Mechanism; Wiley: New
York, 1981; Chapter 7.

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4157
d[NiI]
dt
2
-
) k6Ka2K1 (K2K4 +
Ka1/2K2K3K5[HCl]01/2)2[NiI]2[HCl]0
+
2
k9KaK1K2K3[Cl-][NiI][HCl]0
(1 + K2[Cl-] + K2K3[Cl-]2)(1 + Kf[Cl-])
(12)
and high chloride concentrations, respectively. The first term
does not contain the chloride dependence present in the more
general expression (eq 9A, Appendix) because its contribution
at low [Cl-] is negligible.
At intermediate chloride concentrations, a combination of the
two terms accounts for the observed rate profile (vide infra).
The first term predicts second-order behavior in Ni(I) with HCl
as the only chloride source. It further anticipates the apparent
noninteger order in HCl described in Figure 9; expansion of
the first term affords an order between 2 and 3 for this reactant.
Shown in Figure 13 is a fit of the observed rate constants at
varying [HCl]0 to the general expression 13, where R and â are
products of constant [NiI] and rate and equilibrium constants
of elementary reactions appearing in the first term of eq 12.
2
1/2 2
Figure 13. Dependence of reaction rate on the concentration of HCl
at [NiI] ) 1.00 mM, [HCl]0 ) 12-49 mM, and 25 °C. The solid line
is a fit of the data to the equation kobs ) [HCl]02(290.4 + 817.4[HCl]01/2)2.
kobs ) [HCl]0 (R + â[HCl]0
)
(13)
The second term predicts first-order kinetics in both Ni(I)
and HCl in the high-[Cl-] region, consistent with the observed
reaction orders of these species. In addition, it accounts for
the bell-shaped dependence of the observed rate on chloride
concentration. Illustrated in Figure 11 are two fits of observed
rate constants at constant [HCl]0 and varying [Cl-] to the general
eq 14, in which γ, δ, and ꢀ are constant terms.39 Highly
satisfactory fits were obtained from eqs 13 and 14, consistent
with the operation of Paths A and B.
predominates in the high chloride region. Equation 3A (Ap-
pendix) was used to calculate values of [Cl-] in experiments in
which HCl was the only source of chloride. The relative values
of the first and second terms are strongly influenced by the value
of Kf for homoconjugation reaction 9 used in the calculation,
for this value modulates the chloride concentration in the
numerator of the second term.
The rate behavior in the intermediate range of chloride
concentration (ca. 1.6-30 mM), in which a mixed-order
dependence on [NiI] is observed, is most readily evaluated using
the integrated form 15 of rate law 12. For this purpose, the
kobs ) γ[Cl-]/(1 + δ[Cl-] + ꢀ[Cl-]2)(1 + Kf[Cl-]) (14)
The values K2 ) 0.26-0.28 M-1 (δ) and K3 ) 43-46 M-1
derived from fits of the data to eq 14 indicate that formation of
the neutral dichloride ion pair is decidedly favored over its
monochloride counterpart (Scheme 1). Inasmuch as the dichlo-
ride is the entry point to H2 formation via path B, its preferential
formation may explain why this reaction pathway begins to
dominate at chloride concentrations as low as ca. 30 mM.
The first order dependence on HCl displayed in Figure 10
requires comment. Although the added chloride concentration
is constant in a given set of experiments (0.090 or 0.122 M),
homoconjugation equilibrium 9 causes the actual chloride
concentration to vary (ca. 20%) over the range of [HCl] used.
However, the term [Cl-]/(1 + K2[Cl-] + K2K3[Cl-]2/(1 +
Kf[Cl-]) remains constant to within 2% over the range of [HCl]
used at both concentrations of added chloride, leading to the
observed linear dependence of rate on [HCl]0.
ln([NiI]/(ki[NiI] + kii)) ) -kiit + ln([NiI]0/(ki[NiI]0 + kii))
(15)
latter may be written as -d[NiI]/dt ) ki[NiI]2 + kii[NiI], in which
the terms ki and kii refer to the rate constants for the first and
second terms of the rate law, respectively. Data from this
concentration range were fitted to the equation t ) -1/(kii ln-
([NiI]/(ki[NiI] + kii)) + C′, using ki, kii, and C′ as parameters.
Values of ki and kii obtained from the fits were compared with
observed or calculated values of these constants. For data in
the low end of the range ([Cl-]0 ) 1.6-3.0 mM), Kf[Cl-] is
negligible and ki is approximated by the kobs value for the
corresponding experiment in which no chloride was added. From
eq 12 we find kobs ) 112 M-1 min-1, vs ki ) 104-114 M-1
min-1 from the fits. Within this low range, values of kii were
calculated from eq 12. At higher [Cl-]0 values within the range,
ki is determined by the general rate law (eq 9A, Appendix) and
could not be meaningfully calculated owing to insufficient
knowledge of certain constants in the equation. From the form
of this equation, ki is expected to decrease slightly (ca. 20%) at
higher [Cl-]0. Although the entire intermediate region could
not be tested, the ki and kii values obtained from the fits matched
the expected values to within 20%. This result offers additional
support for the dual path A/B mechanism of Scheme 1.
Calculation of initial rates described by eq 12 at varying
chloride concentrations confirmed that the first term controls
the rate in the low chloride regime, while the second term
(39) The value of Kf (eq 9) used in the fits requires comment. The [Cl-]
values in the second term of eq 12 correspond to concentrations after
homoconjugation is considered. They were calculated using eqs 1A
and 2A of the Appendix. An iterative approach in which [Cl-] was
calculated from a given value of Kf with γ, δ, and ꢀ treated as fitting
parameters failed to produce a self-consistent value of Kf. Thereafter,
eq 14 was fit for several fixed values of Kf with γ, δ, and ꢀ as variables.
Reasonable fits were obtained for Kf ) 5-25 M-1; these values
overlap those obtained by conductivity measurements in the range 20-
25 M-1. The equations plotted in Figure 11 represent the best fits to
the data within this latter range of Kf values.
(g) Retardation of H2 Formation. Set out in Scheme 2 is
the proposed mechanism by which PF6-, taken as a noncoor-
dinating anion in DMF solution, retards the rate of H2 production
in reaction 6. After the Ni(I) protonation reaction 10, all

4158 Inorganic Chemistry, Vol. 35, No. 14, 1996
James et al.
Scheme 2
state for hard anionic ligands,40 chloride binding may stabilize
the Ni-H2+ intermediate as Ni-HCl ) NiIII-H-Cl- or Ni-
HCl2 ) NiIII-H-Cl-2. We have already noted that the psnet
ligand alone does not stabilize Ni(III) sufficiently to afford a
reversible [Ni(psnet)]2+,3+ couple. Decoordination of phosphine
groups would effect a coordination number of 6 or less in
chloride-bound species and would also lessen steric congestion
and thus facilitate bimolecular reactions of the Ni-HCl+
intermediate in steps A3-A5 of path A. Collapse of the
dichloride ion pair to Ni-HCl2 is expected to engender the rate
enhancement observed in the presence of exogenous chloride.
Binding of two chlorides to the nickel(III)-hydride necessitates
the displacement of both phosphine groups to maintain a
reasonable coordination number. The energy required to
decoordinate the two groups may explain why collapse of the
ion pair in reaction B1 could be rate-limiting in path B.
Bimolecular reaction B2 could be enhanced as a result of
decreased steric hindrance at the Ni(III) center. However,
opening up of the coordination sphere could also enhance the
rate of reaction B3 by chloride-bridged inner-sphere electron
transfer, resulting in H2 formation by steps B3 and B4. As noted
above, we cannot resolve path B after the rate-determining step.
Related Steps in H2 Formation in Other Systems. The
protonation of low-valent metals to form metal hydrides,
followed in some cases by H2 evolution upon reaction with
additional protons, is precedented for a number of transition
metals. The situation has been most recently summarized by
Koelle.41 Of particular relevance to the system at hand are the
reactivities of certain cobalt-hydride complexes. Thus [HCo-
(CN)5]3- evolves H2 in bimolecular reaction 21,42 analogous to
-
possible Ni-Cl--PF6 ion pairs are formed (eqs 11 and 16).
Collapse of the Ni-Cl- ion pairs affords three Ni-HCl+
intermediates (eqs 17), which subsequently react in six possible
combinations to form H2, Ni-Cl+, PF6-, and free chloride (eqs
18). These steps are equivalent to minimal reaction 2b. Rate
eq 19 is derived from Scheme 2, in which all possible
-d[NiI]
) k6Ka2K1 [NiI]2[HCl]0
×
2
2
2[HCo(CN)5]3- f H2 + 2[Co(CN)5]3-
(21)
dt
-
2
K2K4 + K2K3K5[Cl-] + K10K12K13[PF6 ]
(19)
[HCo(dmgH)2(B)] + H3O+ f H2 + [Co(dmgH)2(B)(OH2)]
-
-
- 2
(
)
1 + (K10 + K10K12[Cl ])[PF6 ] + K10K11[PF6 ]
(22)
intermediate species are included because none can be excluded
on the basis of experiment. Development of this equation
involves assumptions similar to those for eq 12. The Ni-H2+
concentrations can be written as a sum of all Ni-PF6--Cl-
ion pairs. The rate of Ni-H2+ species collapse to form H2 and
Ni-Cl+ is independent of ion-pair status (k6 ) k7 ) k8 ) k14
) k15 ) k16). Certain terms in [Cl-] present in the general
expression of the rate law (eq 12A, Appendix), negligible at
the low chloride concentrations used to discern the effect of
[Bu4NPF6] on rate, are omitted.
[CoI(dmgBF2)2]- + H+ f [HCoIII(dmgBF2)2] (23a)
2[HCoIII(dmgBF2)2] f H2 + [CoII(dmgBF2)2] (23b)
[HCoIII(dmgBF2)2] + H+ f H2 + [CoIII(dmgBF2)2]+ (23c)
steps A3-A5 and B2 (Scheme 1). Decomposition of a
hydridocobaloxime complex was found to proceed in part by
the heterolytic pathway 22 (B ) base).43 Further, cobalt-
catalyzed evolution of H2 has been obtained by a reaction
sequence very probably including the fast reactions 23.44 The
relationship of these reactions to those in Scheme 1 and minimal
reactions 2 is evident. In reaction 23c, hydride protonation
precedes electron transfer, a possible sequence noted for reaction
2a and the reverse of reactions B3 and B4. All of these reactions
were carried out in aqueous solution, where ion pairing is an
improbable influence on the kinetics.
The rate law predicts second-order kinetics in Ni(I) and HCl,
consistent with the reaction orders described above. Shown in
Figure 12 is a fit of the observed rate constants to the general
eq 20. Both of the coupled reactions of Scheme 2 and the
kobs ) (ú + η[Bu4NPF6])2/(1 + θ[Bu4NPF6] +
κ[Bu4NPF6]2)2 (20)
Other Reaction Pathways. Schemes 3 and 4 illustrate two
simpler alternative routes to H2 formation with a common initial
step of Ni(I) protonation, reaction 10. In Scheme 3, H2 is
derived rate law show how PF6- slows the rate of H2 formation.
This anion competes with chloride to form nonproductive ion
pairs that cannot collapse to active Ni-HCl2+ intermediates
(Scheme 1). In the development of the kinetics scheme, the
(40) (a) Nag, K.; Chakravorty, A. Chem. ReV. 1980, 33, 87. (b) Lappin,
A. G.; McAuley, A. AdV. Inorg. Chem. 1988, 32, 241.
(41) Koelle, U. New J. Chem. 1992, 16, 157.
(42) (a) Halpern, J.; Pribanic´, M. Inorg. Chem. 1970, 9, 2616. (b) Halpern,
J. Inorg. Chim. Acta 1983, 77, L105. (c) Mooiman, M. B.; Pratt, J.
M. J. Mol. Catal. 1984, 27, 367.
(43) Chao, T.-H.; Espenson, J. H. J. Am. Chem. Soc. 1978, 100, 129. dmgH
) dimethylglyoximate(1-).
(44) Connolly, P.; Espenson, J. H. Inorg. Chem. 1986, 25, 2684. dmgBF2
) difluoroboryl dimethylglyoximate(1-).
-
observation of rate retardation by PF6 and change in Ni(I)
reaction order were significant for these effects suggested the
intervention of ion pairing in the mechanism of H2 evolution.
We interpret our data on anion effects to implicate the
formation and collapse of mono- and dichloride ion pairs as
prerequisite to H2 formation in the [Ni(psnet)]+/HCl/DMF
system. Given the obvious preference of the Ni(III) oxidation

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4159
Scheme 3
7. The mechanism of reaction 26 is interpreted in terms of
two sets of reactions (Scheme 1); the initial metal-centered
protonation step NiI + H+ f NiIII-H- is common to both. In
the low [Cl-] regime, path A involves formation and collapse
of ion pairs, with the rate-determining step being the minimal
bimolecular reaction 2NiIII-H- f H2 + 2NiII (eq 2b). In the
high [Cl-] regime, path B implicates the rate-determining
collapse of a dichloride ion pair to NiIII-H-Cl-2 with attendant
partial decoordination of the psnet ligand (most likely the
diphenylphosphino groups). Because these paths fit all kinetics
Scheme 4
-
data satisfactorily, including the rate-decelerating effect of PF6
(described by eq 19), and no other schemes satisfactory in this
respect could be devised, paths A and B are taken as the most
probable descriptions of overall reaction 26.
Hydrogenase in Context. This investigation of Ni-medi-
ated H2 formation has been motivated by the unknown reaction
pathway of [NiFe] hydrogenases. While there is substantial
spectroscopic information pointing toward an intimate involve-
ment of nickel in catalysis,1,2,45-50 the role of iron in influencing
the electronic and reactivity properties of the enzymes is
presently unknown. As emphasized at the outset, the system
studied here cannot be a close model of the enzyme active site.
The pentadentate molecule psnet is not wholly a simulator of
ligand binding in a protein, the complexes [Ni(psnet)]2+,1+ lack
a bridged Ni-Fe center recently discovered in the crystal
structure of D. gigas hydrogenase,17 and the existence of anion
effects in enzymatic proton reduction is problematic. Given
the complexity of Ni-mediated H2 formation, at least in terms
of alternative pathways, we are less concerned with fidelity to
native coordination at this stage of the problem than with
demonstration of the protic reaction chemistry of a fully
characterized Ni(I) molecule. This investigation has shown that
an authentic mononuclear Ni(I) complex of known structure and
a mildly negative redox potential can stoichiometrically evolve
H2 from protic sources. The reaction paths considered most
probable on the basis of kinetics analysis involve steps of protic
oxidative addition to Ni(I) to generate NiIII-H- and, in one
version of path B, electron transfer to Ni(III) followed by
protonation of bound hydride to afford H2. The last of these
steps has been independently demonstrated (eqs 3 and 4). These
events are all found in Figure 1 depicting a minimal scheme
for enzymic H+/H2 interconversion. Note that hydrogenases
catalyze the isotope exchange reaction D2O + H2 a HDO +
HD, indicating that H2 is cleaved heterolytically and emphasiz-
ing the probability of a hydride intermediate, which has been
widely proposed.1
produced by the coupling of two nickel(III) hydrides, as in
minimal reaction 2b, without the formation of nickel chloride
ion pairs or nickel hydride-chloride intermediates. In Scheme
4, the reaction proceeds by electron transfer between the Ni(I)
reactant and the nickel(III) hydride intermediate, followed by
protonation of the nickel(II) hydride, again without formation
of ion pairs. This scheme is composed of the minimal reactions
2a. The derived rate laws for these schemes are eqs 24 and 25.
Having related forms, neither can account for all observed
reaction orders or the rate-reducing effect of (Bu4N)(PF6).
Similar limitations were encountered for all other reaction sets
surveyed except those in Schemes 1 and 2.
Summary. The following are the principal results and
conclusions of this investigation.
1. The diamagnetic complex [Ni(psnet)]2+ is readily prepared
and can be reduced electrochemically or chemically to [Ni-
-
(psnet)]+; both complexes were isolated as BF4 salts and
display distorted trigonal bipyramidal structures.
2. Structural distortions relative to [Ni(psnet)]2+ and EPR
properties demonstrate that [Ni(psnet)]+ is an authentic Ni(I)
1
2
complex with a ...(dz ) ground state.
3. In the presence of one or more equivalents of strong acid,
the well-characterized Ni(II) hydrides [Ni(diphos)2H]+ and [Ni-
(NS3tBu)H]+ give essentially quantitative yields of H2 based on
the stoichiometry NiII-H- + H+ f NiII + H2 (eq 5), thereby
demonstrating protonation of nickel-bound hydride.
4. Under appropriate conditions, [Ni(psnet)]+ reacts with a
series of strong acids in aprotic media to produce H2, in some
cases in >90% yield. Dihydrogen production was demonstrated
to be directly coupled to Ni(I) oxidation under the stoichiometry
The formation of H2 by the coupling of two nickel(III)
hydrides (path A) and the effect of anions on reaction rates are
presumably artifacts of the synthetic system. In synthetic
systems, it is often difficult to separate metal centers sufficiently
to avoid nonphysiological bimolecular reactions while maintain-
ing other design features essential to reactivity, such as open
coordination sites for substrate binding and stereochemical
1
NiI + H+ f NiII + /2H2 (eq 6).
5. From these results in (4), the system [Ni(psnet)]+/1.0 M
HCl in Et2O/DMF, which affords J90% H2 yields, was selected
for a study of reaction kinetics and mechanism.
(45) Scott, R. A.; Wallin, S. A.; Czechowski, M.; DerVartanian, D. V.;
LeGall, J.; Peck, H. D., Jr.; Moura, I. J. Am. Chem. Soc. 1984, 106,
6864.
(46) Fan, C.; Teixeira, M.; Moura, J.; Moura, I.; Huynh, B.-H.; LeGall, J.;
Peck, H. D., Jr.; Hoffman, B. M. J. Am. Chem. Soc. 1991, 113, 20.
(47) (a) Barondeau, D. P.; Roberts, L. M.; Lindahl, P. A. J. Am. Chem.
Soc. 1994, 116, 3442. (b) Roberts, L. M.; Lindahl, P. A. Biochemistry
1994, 33, 14339.
(48) Roberts, L. M.; Lindahl, P. A. J. Am. Chem. Soc. 1995, 117, 2565.
(49) Guigliarelli, D.; More, C.; Fournel, A.; Asso, M.; Hatchikian, E. C.;
Williams, R.; Cammack, R.; Bertrand, P. Biochemistry 1995, 34, 4781.
(50) (a) van der Zwaan, J. W.; Albracht, S. P. J.; Fontijn, R. D.; Slater, E.
C. FEBS Lett. 1985, 179, 271. (b) Whitehead, J. P.; Gurbiel, R. J.;
Bagyinka, C.; Hoffman, B. M.; Maroney, M. J. J. Am. Chem. Soc.
1993, 115, 5629.
6. The system in (5) evolves H2 by two parallel pathways
determined by the chloride concentration and described by a
two-term rate law (eq 12); H2 formation is stimulated by added
chloride (as Bu4NCl) and retarded by added (Bu4N)(PF6) in the
overall reaction 26. At low chloride concentration, the rate
exhibits a second-order dependence on [NiI] and a non-integer
dependence on [HCl]0 between second and third order. At high
chloride concentration the rate exhibits the first-order behavior
[NiI][HCl]0 with a quadratic dependence on [Cl-].
[NiI(psnet)]+ + HCl f [Ni(psnet)Cl]+ + 1/2H2 (26)

4160 Inorganic Chemistry, Vol. 35, No. 14, 1996
James et al.
flexibility for the stabilization of various metal oxidation states.51
In the hydrogenase context, future ligand design must address
these issues as well as the inclusion of a binuclear metal site,
and stable aqueous systems must be sought where anion effects
may be absent and H/D isotope exchange can be examined. As
it stands, however, the [Ni(psnet)]+ system serves as the most
thoroughly scrutinized system for Ni-mediated formation of H2.
Whether nickel alone, or in necessary concert with iron, is the
site of hydrogen activation is unknown and will be difficult to
determine.
-d[NiI]
dt
)
(K2K4 + Ka1/2K2K3K5[HCl]1/2)2[NiI]2[HCl]2
k6Ka2K1
+
2
(1 + K2[Cl-] + K2K3[Cl-]2)2
k9KaK1K2K3[Cl-][NiI][HCl]
(1 + K2[Cl-] + K2K3[Cl-]2)
(6A)
eq 7A and eq 9 are used to express [HCl] in terms of the initial
HCl concentration, [HCl]0, in eq 8A. Substitution for [HCl]
Acknowledgment. This research was supported by NSF
Grants CHE 92-08387 and 94-23830. We thank Professor W.
H. Armstrong for use of a magnetic susceptibility balance, and
Drs. J. D. Franolic, J. R. Long, M. J. Scott, and A. Tyler for
experimental assistance.
-
[HCl]0 ) [HCl] + [HCl2 ] ) [HCl](1 + Kf[Cl-]) (7A)
[HCl] ) [HCl]0/(1 + Kf[Cl-])
(8A)
Appendix
into eq 6A, using eq 8A, yields eq 9A. Equation 9A represents
Determination of [Cl-]. Equations to determine the actual
chloride concentration after homoconjugation equilibrium 9 is
considered are presented below. Equations 1A and 2A apply
to experiments in which HCl and Bu4NCl are present, while eq
3A applies to experiments in which HCl is the only chloride
source.
-d[NiI]
2
) k6Ka2K1 {(K2K4 + Ka1/2K2K3K5[[HCl]0/(1 +
dt
Kf[Cl-])]1/2)2[NiI]2([HCl]0/(1 + Kf[Cl-]))2/(1 + K2[Cl-] +
k9KaK1K2K3[Cl-][NiI][HCl]0
K2K3[Cl-]2)2} +
When [Cl-]0 > [HCl]0:
(1 + K2[Cl-] + K2K3[Cl-]2)(1 + Kf[Cl-])
[Cl-] ) ([Cl-]0 - [HCl]0) + {[(Kf[Cl-]0 - Kf[HCl]0 +
(9A)
1)2 + 4Kf[HCl]0]1/2 - (Kf[Cl-]0 - Kf[HCl]0 + 1)/2Kf}
the general form of the rate law. At the low chloride
concentrations that prevail when HCl is the sole chloride source,
Kf[Cl-] , 1 and K2[Cl-] + K2K3[Cl-] , 1 (from the data fitted
in Figure 11). Therefore, eq 9A simplifies to eq 12.
(1A)
When [HCl]0 > [Cl-]0:
Kinetics Analysis of Scheme 2. Equation 10A employs the
simplifying assumption k6 ) k7 ) k8 ) k14 ) k15 ) k16.
[Cl-] ) {[(Kf[HCl]0 - Kf[Cl-]0 + 1)2 + 4Kf[Cl-]0]1/2
-
(Kf[HCl]0 - Kf[Cl-]0 + 1)/2Kf} (2A)
-d[NiI]/dt ) k6([Ni-HCl+] + [Ni-HCl+|Cl-] +
When [Cl-]0 ) 0:
[Cl-] )
-
[Ni-HCl+|PF6 ])2 (10A)
1/2
Ka[HCl]0
Substitution for the ion-pairing equilibria in eqs 11 and 15 in
(3A)
[
]
terms of [Ni-H2+] yields eq 11A. Substitution for [Ni-H2+
]
1 + Kf[HCl]0
-d[NiI]/dt ) k6[Ni-H2+]2[Cl-]2((K2K4 + K2K3K5[Cl-] +
Kinetics Analysis of Scheme 1. Equation 4A employs the
simplifying assumption k6 ) k7 ) k8. Substitution for the ion-
-
K10K12K13[PF6 ])/(1 + K2[Cl-] + K2K3[Cl-]2 +
-
- 2
-
K10[PF6 ] + K10K11[PF6 ] + K10K12[PF6 ][Cl-]))2 (11A)
-d[NiI]/dt ) k6([Ni-HCl+] + [Ni-HCl|Cl-])2 +
k9[Ni-H2+|2Cl-] (4A)
in terms of [NiI] and [HCl]0 using eqs 7, 10, and 8A yields eq
12A. Using the assumptions outlined in the last step of the
pairing equilibria in eq 11 in terms of [Ni-H2+] yields eq 5A.
2
2
-d[NiI]/dt ) [k6Ka2K1 [NiI]2[HCl]0 /(1 +
(K2K4 + K2K3K5[Cl-])2[Ni-H2+]2[Cl-]2
(1 + K2[Cl-] + K2K3[Cl-]2)2
k9K2K3[Cl-]2[Ni-H2+]
-d[NiI]
dt
-
Kf[Cl-])2]((K2K4 + K2K3K5[Cl-] + K10K12K13[PF6 ])/(1 +
) k6
+
-
- 2
K2[Cl-] + K2K3[Cl-]2 + K10[PF6 ] + K10K11[PF6 ] +
-
K10K12[PF6 ][Cl-]))2 (12A)
(5A)
(1 + K2[Cl-] + K2K3[Cl-]2)
derivation for Scheme 1, one sees that eq 12A simplifies to eq
19.
Kinetics Analysis of Scheme 3.
Substitution for [Ni-H2+] in terms of [HCl] and [NiI] using
eqs 7 and 10, respectively, affords eq 6A. The conservation
-d[NiI]/dt ) k17[Ni-H2+]2
(13A)
(51) Note that undesired bimolecular reactions, such as those which form
FeIIIOFeIII and OdMoVOMoVdO bridges instead of the desired
FeO2 and MoIVO/MoVIO2
units, respectively, have been over-
52a
52bc
Substitution for [Ni-H2+] in terms of [NiI] and [H+] using eq
10 yields eq 14A. The conservation eq 15A and eqs 7-9 lead
to eq 16A. Substitution for [H+] in eq 14A leads to the final
form of the rate law described in eq 24.
come by the use of sterically encumbered ligands.
(52) (a) Momenteau, M.; Reed, C. A. Chem. ReV. 1994, 94, 659. (b) Holm,
R. H. Coord. Chem. ReV. 1990, 100, 183. (c) Schultz, B. E.; Gheller,
S. F.; Muetterties, M. C.; Scott, M. J.; Holm, R. H. J. Am. Chem.
Soc. 1993, 115, 2714.

H2 Evolution by Ni(I) Protonation Reactions
Inorganic Chemistry, Vol. 35, No. 14, 1996 4161
Substitution for [H+] in eq 18A using eq 16A leads to the final
form of the rate law presented in eq 25.
-d[NiI]/dt ) k17K1[NiI]2[H+]2
(14A)
(15A)
(16A)
-
[H+] ) [Cl-] + [HCl2 ]
-d[NiI]/dt ) k18K1[NiI]2[H+]
(18A)
2 1/2
[H+] ) (Ka[HCl]0 + KaKf[HCl]0 )
Kinetics Analysis of Scheme 4.
-d[NiI]/dt ) k18[NiI][Ni-H2+]
Supporting Information Available: Tables of crystallographic data
for the compounds in Table 1, including crystal data and intensity
collection parameters, and of positional and thermal parameters and
interatomic distances and angles (11 pages). Ordering information is
given on any current masthead page.
(17A)
Substitution for [Ni-H2+] in eq 17A using eq 10 yields eq 18A.
IC960216V