Report

A Synthetic Nickel Electrocatalyst with a Turnover Frequency Above 100,000 s−1 for H2 Production

See allHide authors and affiliations

Science  12 Aug 2011:
Vol. 333, Issue 6044, pp. 863-866
DOI: 10.1126/science.1205864

Abstract

Reduction of acids to molecular hydrogen as a means of storing energy is catalyzed by platinum, but its low abundance and high cost are problematic. Precisely controlled delivery of protons is critical in hydrogenase enzymes in nature that catalyze hydrogen (H2) production using earth-abundant metals (iron and nickel). Here, we report that a synthetic nickel complex, [Ni(PPh2NPh)2](BF4)2, (PPh2NPh = 1,3,6-triphenyl-1-aza-3,6-diphosphacycloheptane), catalyzes the production of H2 using protonated dimethylformamide as the proton source, with turnover frequencies of 33,000 per second (s−1) in dry acetonitrile and 106,000 s−1 in the presence of 1.2 M of water, at a potential of –1.13 volt (versus the ferrocenium/ferrocene couple). The mechanistic implications of these remarkably fast catalysts point to a key role of pendant amines that function as proton relays.

Electrocatalysts that efficiently convert the energy from electricity into chemical bonds in fuels (such as hydrogen), or the reverse, converting chemical energy to electrical energy, will play a vital role in future energy storage and energy delivery systems. Hydrogenase enzymes (1, 2) efficiently catalyze both the production and the oxidation of hydrogen using earth-abundant metals (nickel and iron). Detailed information about catalytic reactions of enzymes has been obtained from protein film voltammetry (3), but enzymes are difficult to obtain in sufficient amounts to adapt for commercial applications, and their stability is often limited outside of their native environment (4). Platinum is an excellent catalyst for hydrogen oxidation and production, but the scarcity and high cost of precious metals pose serious limitations to widespread use. These considerations have led to efforts to design molecular catalysts that employ earth-abundant metals. Synthetic complexes of nickel (58), cobalt (912), iron (1315), or molybdenum (16, 17) have been developed recently as electrocatalysts for the production of hydrogen.

In nature, [FeFe] hydrogenase enzymes catalyze the formation of H2 from water, with reported rate constants as high as 9000 s−1 at 30°C (18). Crystallographic and spectroscopic studies have led to the proposal of the structure of the active site of the [FeFe] hydrogenase enzyme shown in structure 1 of Fig. 1 (1). The amine base positioned near the iron center has been proposed to function as a proton relay that facilitates the formation or cleavage of the H-H bond. Substantial progress on the preparation of structural mimics of the active site (13, 14, 19) provides insight into mechanistic features of the catalytic reaction. For example, Rauchfuss and co-workers have recently demonstrated that a pendant amine in a structural model closely related to 1 plays a key role in the production of H2 (20). Efforts in our laboratory have focused on developing mononuclear complexes of Fe, Co, and Ni that contain an amine base in the second coordination sphere, adjacent to a vacant coordination site or a hydride ligand on the metal center (2123). Some of these complexes are very effective electrocatalysts for H2 formation or H2 oxidation, and they provide functional models for the proton relay functions of the hydrogenase enzymes. For example, a series of [Ni(PPh2NC6H4X2)2]2+ catalysts with structure 2 have been studied. Under optimized conditions, [Ni(PPh2NC6H4Br2)2]2+, with bromophenyl substituents on the amine ligand, catalyzed the formation of H2 with turnover frequencies as high as 1040 s−1 and an overpotential of ~290 mV (7).

Fig. 1

Proposed structure of the [FeFe] hydrogenase active site (1) based on crystallographic data (1); synthetic Ni complex (2) that catalyzes H2 formation, with pendant amines that function as proton relays (7); proposed transition state (24) (3) for production of hydrogen catalyzed by 2.

Mechanistic (5, 7) and theoretical (24) studies of the [Ni(PR2NR′2)2]2+ catalysts suggest that the transition state for H2 production has structure 3 shown in Fig. 1, in which the formation of the H-H bond occurs in a heterolytic manner, with the pendant amine acting as a proton donor and the nickel functioning as a hydride donor. The formation of transition state 3 involves the reduction of the starting complex 2 by two electrons and the addition of two protons. Previous nuclear magnetic resonance spectroscopic studies on reduction and protonation of closely related analogs of 2 have shown that intermediates formed before transition state 3 are isomeric Ni(0) complexes in which an amine in each ligand is protonated. The double protonation reaction of Ni0(PPh2NBn2)2 (where Bn is benzyl) forms structure 4 as the kinetic product (Fig. 2) (25). The kinetic product isomerizes to form an equilibrium mixture of structures 4, 5, and 6 by an intermolecular process that involves deprotonation by a base in solution and reprotonation at an endo position. This process is assisted by water. Only isomer 6 can readily convert to transition state 3 by a simple intramolecular proton transfer from nitrogen to nickel, and therefore only isomer 6 is active in the catalytic cycle. These studies led us to propose that only a fraction of the doubly protonated Ni(0) intermediates formed during hydrogen production are in the catalytically active form, isomer 6, and that significantly higher catalytic rates might be achieved if the stabilized exo-protonated structures could be avoided. We also proposed that the observed enhancement of turnover frequencies by water is a result of its role in converting the exo-protonated isomers 4 and 5 to the endo-endo isomer 6 (7).

Fig. 2

Doubly protonated Ni(0) intermediates (isomers 4, 5, and 6) involved in the production of H2 catalyzed by 2. For clarity, organic substituents on the N and P are not shown.

Here, we report the synthesis of [Ni(PPh2NPh)2](BF4)2, 7, which was isolated as orange crystals in 75% yield from the reaction shown in Scheme 1 (26). The new complex contains two seven-membered cyclic diphosphine ligands, PPh2NPh (where PPh2NPh = 1,3,6-triphenyl-1-aza-3,6-diphosphacycloheptane). Because this ligand has just one pendant amine, it precludes the possibility of forming nonproductive pinched isomers analogous to 4 and 5 upon reduction and protonation of 7. Although the reduced form of 7 can be protonated either endo or exo with respect to Ni, stabilization by a second N•••H interaction in an exo-exo form (e.g., as in 4) is not possible. The molecular structure of 7 determined by x-ray diffraction (Fig. 3) is a distorted square planar complex with all four Ni-P bond distances nearly equal at 2.21 to 2.22 Å. Complexation of both phosphorus atoms in each PPh2NPh ligand in 7 forms one five-membered chelate ring and a six-membered ring that incorporates the pendant amine. In contrast, both of the eight-membered cyclic ligands in 2 form two six-membered chelate rings upon binding to the metal. As a result, the P-Ni-P bond angle (79.83°) for each of the two diphosphine ligands in 7 is smaller than the 82° to 84° typically observed for the P-Ni-P bonds in derivatives of 2 (5, 7). This smaller bite angle leads to a decrease in the steric interactions between the phenyl substituents on adjacent phosphorus atoms of the two ligands in 7. As a result, 7 is much more planar than the corresponding derivatives of 2. For [Ni(PPh2NC6H4Me2)2]2+, the dihedral angle between the two planes defined by the Ni atom and the two phosphorus atoms of each diphosphine ligand is 24.16° (7), whereas for 7, it is 0.0°. Thus, the replacement of two eight-membered cyclic ligands with two seven-membered cyclic ligands in these nickel complexes has significant structural consequences.

Fig. 3

Solid-state molecular structure of [Ni(PPh2NPh)2](BF4)2∙2CH3CN, 7. The BF4 counterions, CH3CN solvent molecules, and H atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level.

The cyclic voltammogram of an acetonitrile solution of [Ni(PPh2NPh)2]2+, 7, shows one reversible reduction wave assigned to overlapping Ni(II/I)/Ni(I/0) couples at –1.13 V versus the ferrocenium/ferrocene couple (simulated as two one-electron waves at –1.09 and –1.16 V). (See Supporting Online Material for a cyclic voltammogram of 7, fig. S3, and a detailed analysis.) Figure 4 shows successive cyclic voltammograms of 7 recorded in acetonitrile with increasing concentrations of protonated dimethylformamide {[(DMF)H]OTf, pKa = 6.1 in acetonitrile} (27). A catalytic wave for the reduction of protons is observed with a half-wave potential of –1.13 V in the presence of 0.20 M acid, corresponding to an overpotential of ~625 mV determined using the method of Evans (28). The catalytic production of H2 was confirmed by a gas chromatographic analysis of the gas produced during a controlled potential electrolysis experiment (–1.4 V, current efficiency 99 ± 5% for H2 production, 11 turnovers with no observable decomposition). At high acid concentrations relative to the catalyst, Eq. 1 can be used to calculate pseudo first-order rate constants, kobs, for H2 evolution. icatip=20.446RTkobsFυ(1)In this equation, icat is the catalytic current, ip is the peak current measured in the absence of acid, 2 is the number of electrons involved in the catalytic reaction, kobs is the observed first-order rate constant, R is the universal gas constant, T is the temperature in Kelvin, F is Faraday’s constant, υ is the scan rate, and 0.446 is a constant determined by numerical solution of the diffusion equations (29). The value of kobs increases linearly with acid concentration (fig. S9), indicating a first-order dependence of the catalytic rate on acid concentration. At the highest acid concentration studied (0.43 M), a value of icat/ip = 38 was measured, corresponding to a turnover frequency of 33,000 s−1 at 22°C.

Fig. 4

Successive cyclic voltammograms of 1.0 mM 7 in CH3CN (0.10 M [Bu4N][PF6]) at increasing concentrations of [(DMF)H]OTf, followed by addition of small aliquots of H2O. Conditions: 1 mm glassy-carbon working electrode; 22°C; scan rate 10 V/s.

At 0.43 M [(DMF)H]OTf, cyclic voltammetry and ultraviolet-visible absorption spectra indicated that less than 5% decomposition of the catalyst occurred over the 0.5 hours required for the catalytic studies. However, higher acid concentrations could not be used to determine reliable turnover frequencies because of catalyst decomposition under more strongly acidic conditions. The catalyst is stable in the presence of H2O, and the addition of H2O (1.2 M) to the mixture results in a further enhancement of the catalytic current beyond the maximum icat achieved with acid alone, resulting in a value of icat/ip = 74, corresponding to a turnover frequency of 106,000 s−1. Simulations of the voltammograms obtained under catalytic conditions are in agreement with the experimentally measured icat/ip = 74 (fig. S10 and table S3). The additional rate enhancement in the presence of water is attributed to the ability of water to enhance the rate of conversion of exo to endo isomers.

The turnover frequencies measured for 7 substantially exceed the highest turnover frequency reported for the formation of H2 catalyzed by the [FeFe] hydrogenase enzyme, 9000 s−1 at 30°C (18). This comparison, however, must be tempered by the recognition that the overpotential of the natural hydrogenase is thought to be less than 100 mV, significantly lower than the overpotential for 7. The exceptionally high turnover frequency observed for 7 is about two orders of magnitude larger than that observed for 2 under comparable conditions (590 s−1 with an overpotential of 300 mV, increasing to 720 s−1 when water is added) (7). Most important, our results clearly indicate that properly designed pendant amine bases in the second coordination sphere that function as proton relays are capable of promoting extremely fast H-H bond formation and proton transfer from solution to the metal center.

The observed first-order dependence on acid is consistent with a one- or two-electron reduction of 7, followed by rate-determining protonation of the reduced complex. In contrast, 2 exhibits catalytic rates that are second-order in acid at concentrations below ~0.1 M and independent of acid above ~0.2 M concentrations (7). Although the overall catalytic mechanisms for 2 and 7 are likely quite similar, changing the rate-determining step has significant effects on the observed catalytic rates. For derivatives of 2, the overall catalytic rate is predominantly controlled by two factors, the formation of endo-endo isomers such as 6 and H2 elimination via transition state 3. For 7, the endo-protonation of a reduced intermediate is the rate-determining event. This implies that the rate of H2 elimination via transition state 3 must be faster for 7 than for 2. This is likely the result of a more hydridic Ni-H bond for 7, which has been shown to correlate with more negative Ni(II/I) reduction potentials in these [Ni(diphosphine)2]2+ complexes (30). The more negative Ni(II/I) potential for 7 (–1.13 V) compared with that of 2 (–0.84 V) is expected on the basis of their structural differences (greater planarity of 7), and this modification contributes to the significantly larger overpotential for H2 production observed for 7.

In conclusion, the observation of an extremely high turnover frequency for 7 clearly demonstrates that its positioned proton relays support high rates for all of the individual catalytic steps, that is, inter- and intramolecular proton transfer, electron transfer (or proton-coupled electron transfer steps), and heterolytic formation of H2. These results highlight the substantial promise that designed molecular catalysts hold for the electrocatalytic production of hydrogen. The high sensitivity of catalytic rates to the incorporation and arrangement of pendant amines in the ligand suggests a potentially broader role that proton relays may play in other important multi-proton, multi-electron catalytic reactions, such as the reduction of oxygen and oxidation of water.

Supporting Online Material

www.sciencemag.org/cgi/content/full/333/6044/863/DC1

Materials and Methods

SOM Text

Figs. S1 to S11

Tables S1 to S4

References (3133)

  • * Sabbatical visitor at Pacific Northwest National Laboratory, 2010–2011.

References and Notes

  1. Materials and methods and spectroscopic data are available as supporting material on Science Online.
  2. Acknowledgments: This research was supported as part of the Center for Molecular Electrocatalysis, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences. Pacific Northwest National Laboratory is operated by Battelle for the U.S. Department of Energy. Structural parameters for 4 are available free of charge from the Cambridge Crystallographic Data Centre under CCDC 828010. We thank J. A. S. Roberts for help with the simulations of the cyclic voltammograms.
View Abstract

Stay Connected to Science

Navigate This Article