Amino acid modified Ni catalyst exhibits reversible H₂ oxidation/production over a broad pH range at elevated temperatures

Significance

Enzymes achieve rapid and reversible H₂ oxidation catalysis by cooperative behavior between the active site and the protein scaffold. To better understand the role of the enzyme scaffold, we have attached amino acids (glycine, arginine, and arginine methyl ester) to an active functional mimic of hydrogenase to give ${\left[\text{Ni}{\left({\text{P}}_2^{\text{Cy}}{\text{N}}_2^{\text{Amino}\hspace{0.17em}\text{acid}}\right)}_2\right]}^{2+}$. The resulting complexes are fully reversible catalysts with the arginine complex exhibiting high activity for both H₂ oxidation/production, functionality achieved by the addition of an outer coordination sphere.

Abstract

Hydrogenases interconvert H₂ and protons at high rates and with high energy efficiencies, providing inspiration for the development of molecular catalysts. Studies designed to determine how the protein scaffold can influence a catalytically active site have led to the synthesis of amino acid derivatives of ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)}_2\right]}^{2+}$ complexes, ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Amino}\hspace{0.5em}\mathrm{acid}}\right)}_2\right]}^{2+}$ (CyAA). It is shown that these CyAA derivatives can catalyze fully reversible H₂ production/oxidation at rates approaching those of hydrogenase enzymes. The reversibility is achieved in acidic aqueous solutions (pH = 0–6), 1 atm 25% H₂/Ar, and elevated temperatures (tested from 298 to 348 K) for the glycine (CyGly), arginine (CyArg), and arginine methyl ester (CyArgOMe) derivatives. As expected for a reversible process, the catalytic activity is dependent upon H₂ and proton concentrations. CyArg is significantly faster in both directions (∼300 s⁻¹ H₂ production and 20 s⁻¹ H₂ oxidation; pH = 1, 348 K, 1 atm 25% H₂/Ar) than the other two derivatives. The slower turnover frequencies for CyArgOMe (35 s⁻¹ production and 7 s⁻¹ oxidation under the same conditions) compared with CyArg suggests an important role for the COOH group during catalysis. That CyArg is faster than CyGly (3 s⁻¹ production and 4 s⁻¹ oxidation) suggests that the additional structural features imparted by the guanidinium groups facilitate fast and reversible H₂ addition/release. These observations demonstrate that outer coordination sphere amino acids work in synergy with the active site and can play an important role for synthetic molecular electrocatalysts, as has been observed for the protein scaffold of redox active enzymes.

Hydrogenases are metalloenzymes that interconvert H₂ and protons (Eq. 1), reactions necessary for biological processes within certain organisms to provide energy by splitting hydrogen, as well as to balance the redox potential in the cell (1–4). Their reversible catalytic behavior is a demonstration of their energy efficiency, indicating that they are operating at the equilibrium potential of the H₂/H⁺ couple. They are also very active under conditions optimized for a particular direction, operating at up to 20,000 s⁻¹ for H₂ production (5) and 10,000 s⁻¹ for H₂ oxidation (6). Under conditions where reversibility is optimized (3, 7–10), net turnover frequencies (TOFs) are typically slower. For hydrogenases attached to electrode surfaces, the TOFs are not always quantitated due to uncertainty in surface coverage (9), but in one example, TOFs for a series of mutants of Desulfovibrio fructosovorans [NiFe]-hydrogenase were reported, ranging from 3 to 500 s⁻¹ for H₂ production and 600 to 1,000 s⁻¹ for H₂ oxidation under one set of conditions (8).

$${\mathrm{H}}_2\mathrm{\rightleftharpoons }2{\mathrm{H}}^{+}+2{\mathrm{e}}^{-}.$$ [1]

The ability of hydrogenases to function reversibly requiring no applied overpotential and with high catalytic TOFs is a direct reflection of their biological optimization, and a hallmark of enzymes. Reversibility requires enzymes that have fast and reversible proton and electron transfer, and that are nearly thermoneutral for H₂ addition/elimination, so that little excess energy is needed for either the production or oxidation of hydrogen.

The ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)}_2\right]}^{2+}$ complexes are hydrogenase mimics that can be tuned to either produce or oxidize H₂. A key feature of the ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)}_2\right]}^{2+}$ complexes is a positioned pendant amine in the second coordination sphere, inspired by [FeFe]-hydrogenase, to assist the metal in forming or heterolytically cleaving H₂ (11, 12). We have shown that tuning ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)}_2\right]}^{2+}$ for H₂ oxidation or production catalysis depends on the free energy of H₂ addition (ΔG°_H2) to these complexes. Complexes with large positive ΔG°_H2 values are generally fast H₂ production catalysts with TOFs of 10⁴ to 10⁵ s⁻¹ at overpotentials (OPs) of 0.5–0.6 V, and those with large negative ΔG°_H2 values are fast H₂ oxidation catalysts with TOFs of ∼50 s⁻¹ at 1.0 atm H₂ and an OP of 0.4 V using triethylamine as a base (13). The large positive or negative values of ΔG°_H2 are associated with moderate to large overpotentials, where overpotential is the extra energy over the equilibrium potential needed to drive the reaction, determined at E_cat/2 (SI Appendix). Using thermoneutral H₂ addition/elimination as a design principle, reversibility was achieved for a complex related to those studied here, ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Ph}}{\mathrm{N}}_2^{\mathrm{CH2CH2OCH3}}\right)}_2\right]}^{2+}$, a complex that has a free energy of hydrogen addition of ΔG_H2 = 0.8 kcal/mol (14). This complex operated with a TOF of <1 s⁻¹ in both directions, the slow TOFs being consistent with what has typically been observed for this family of complexes as the thermodynamic driving force for H₂ addition (ΔG°_H2) approaches 0 kcal/mol. These observations point to features in hydrogenase enzymes that are not typically present in the ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)}_2\right]}^{2+}$ catalysts.

To investigate how the protein scaffold can influence catalysis, we have used the ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)}_2\right]}^{2+}$ complex as a probe molecule upon which to build an enzyme-inspired outer coordination sphere (15, 16). We recently reported the incorporation of the amino acids arginine or glycine into the outer coordination sphere of this class of catalysts to give ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}}\right)}_2\right]}^{8+}$ (CyArg) or ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Glycine}}\right)}_2\right]}^{4+}$ (CyGly), where the amine of the amino acid is a member of the P₂N₂ ring (Fig. 1) (17, 18). Both of these complexes are active, water-soluble H₂ oxidation catalysts with TOFs up to 200 s⁻¹ at 1 atm H₂ and 298 K, operating at low overpotentials (less than 180 mV) (17, 18). Unexpectedly, catalysis was fastest under acidic conditions (from pH = 0 to pH = 1) (17, 18). Both complexes also demonstrated impressive TOFs for H₂ oxidation as a function of pressure, up to 144,000 s⁻¹ at 133 atm H₂, 298 K, and pH = 0.5, albeit at reduced energy efficiency (overpotential = 460 mV) (18). The carboxyl functional group in both complexes is thought to play a key role by coupling proton and electron transfer, facilitated by the ability of the COOH group to deprotonate the pendant amine (17, 18). The CyArg derivative operates at similar overpotentials but with TOFs ∼6 times faster than those of the CyGly catalyst. Thus, both the carboxyl and guanidinium substituents play important roles in modifying catalytic activity and demonstrate the dramatic influence that the amino acid functional groups can have on the active site.

Fig. 1.

Complexes studied for reversible catalysis; ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Glycine}}\right)}_2\right]}^{3+}$ (CyGly), ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}}\right)}_2\right]}^{7+}$ (CyArg), and ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{ArginineOMe}}\right)}_2\right]}^{7+}$ (CyArgOMe).

The low overpotential of the CyGly and CyArg catalysts suggested that they might also function as reversible catalysts for both H₂ production and oxidation. To probe this proposed reversibility in more detail we report in this work the effect of increased temperatures and different hydrogen pressures on the performance of these catalysts. It is shown that both complexes can reversibly interconvert H⁺ and H₂, which requires reversible proton and electron transfers as well as reversible H₂ binding and release. A third complex, CyArgOMe, where ArgOMe is the methyl ester of arginine, has also been synthesized and studied to directly probe the role of the COOH group. Our results demonstrate that amino acids incorporated in the outer coordination sphere can facilitate fast, reversible enzyme-like catalysis for small synthetic molecular electrocatalysts.

Results

H₂ Oxidation at Elevated Temperatures.

The H₂ oxidation activity of CyArg was studied as a function of temperature under 1 atm of H₂ at pH = 1.55. The plateau currents of the waves (i_cat) were scan-rate independent from 293 to 337 K (SI Appendix, Fig. S1) and showed a linear dependence on catalyst concentration (SI Appendix, Fig. S1) as expected for a process that is first order in catalyst. As illustrated in Fig. 2A, the catalytic current increased with increasing temperature. The TOFs and overpotentials were calculated from this electrochemical data (Materials and Methods), with TOFs ranging from 45 s⁻¹ (293 K) to 205 s⁻¹ (337 K). It can be seen from Fig. 2B that the overpotentials, the excess energy beyond the equilibrium potential needed for the reaction, decreased from 95 mV at 293 K to 50 mV at 337 K.^† These TOFs and overpotentials are similar to those observed for hydrogenase enzymes (8).

Fig. 2.

Electrocatalytic data shows enhanced TOFs and lower overpotentials at increasing temperatures. (A) Electrocatalytic behavior of 45 μM CyArg under 1 atm H₂ at temperatures from 293 to 337 K at pH = 1.55, 0.1 M Na₂SO₄. The data collected for the same sample under N₂ at 293 K is shown in light gray. The vertical dotted line denotes the equilibrium potential. The horizontal black arrow indicates the initial scanning direction. Data were collected at a scan rate of 0.1 V/s. (B) The TOF (black squares) and overpotential (blue circles) for H₂ oxidation for CyArg as a function of temperature shows an increase in TOF with a corresponding decrease in overpotential.

Reversible Catalysis: Effect of Decreased H₂ Pressure and Elevated Temperatures on H₂ Oxidation and Production.

Under 1 atm H₂, the catalytic response is dominated by H₂ oxidation (Fig. 2). The low observed overpotentials suggest the complex can operate reversibly, and that catalysis of proton reduction to H₂ should be observed at lower H₂ concentrations. To provide more favorable conditions under which to observe reversible H₂ oxidation/H₂ production, a mixture of 25% H₂/Ar was used. Fig. 3A shows the electrocatalytic behavior at five pH values at 348 K under 1.0 atm of a 25% H₂/Ar mixture. Both H₂ production and H₂ oxidation are observed at all pH values operating at the equilibrium potential of the H₂/H⁺ couple, consistent with a truly reversible system. The fastest TOFs in both directions are observed for pH values under 2. Scan rate independence was confirmed for the catalytic waves in both directions (SI Appendix, Fig. S2). Reversible catalysis was not observed below ∼320 K.

Fig. 3.

Reversible H₂ oxidation/production catalysis is observed for CyArg at various pH conditions under 1 atm 25% H₂/Ar. (A) Cyclic voltammograms for 50 μM CyArg at 348 K in 0.1 M Na₂SO₄: pH 0 (black trace), pH 1.0 (red trace), pH 2.5 (blue trace), pH 5.0 (magenta trace), and pH 6.0 (green trace). The corresponding dotted vertical lines correspond to the equilibrium potential for each pH condition. The horizontal dotted line demarcates the zero current line and the horizontal black arrow indicates the initial scanning direction. Data were collected at a scan rate of 0.1 V/s. (B) TOFs for H₂ production (black) and H₂ oxidation (blue) as a function of pH at 348 K.

The fastest TOFs (largest currents) were observed for CyArg at pH 1 and below: 300 s⁻¹ for H₂ production and 20 s⁻¹ for H₂ oxidation, with little enhancement at lower pH values (Fig. 3B, Table 1). The H₂ oxidation TOFs showed a counterintuitive increase of ∼3 times as the pH decreased from 6.0 to 0.0, similar to previous observations at room temperature (7). For H₂ production, a ∼50-fold increase in TOF is observed as the pH decreases from 6.0 to 1.0. The observed half wave potentials for H₂ oxidation and H₂ production were within 15 mV of each other in all cases, as expected for a fully reversible catalytic process.

Table 1.

Maximum rates of H₂ oxidation and production under reversible conditions (348 K, 1 atm)

			H2 production TOF ± 5% (s−1)
Complex	pH	H2 oxidation TOF ± 5% (s−1) Under 25% H2/Ar	Under 25% H2/Ar	Under N2
CyArg	6.0	6	7	21
	5.0	11	10	32
	2.5	14	15	36
	1.0	20	300	403
	0	20	304	403
CyArg(OMe)	0	7	35	124
CyGly	1.0	4	3	12

To ensure that the observed reactivity was due to the CyArg complex, two control experiments were performed. In one experiment (SI Appendix, Fig. S3A) no complex was used and no catalysis was observed as expected. In a second experiment (SI Appendix, Fig. S3B), [Ni(CH₃CN)₆](BF₄)₂ was tested for catalytic activity at pH = 0 and 348 K. No significant current was observed, indicating that the observed response was due to the CyArg complex.

To evaluate the importance of the guanidinium groups in CyArg for catalytic reversibility, the analogous water-soluble H₂ oxidation electrocatalyst, CyGly, was studied, where the arginine in the complex shown in Fig. 1 is replaced with a glycine. Under similar conditions, 348 K, 1 atm 25% H₂/Ar, and pH = 1.0, the optimal pH conditions for the CyGly catalyst, reversible catalytic behavior is also observed but at lower TOFs (Table 1 and SI Appendix, Fig. S4). The TOF for H₂ production was 3 and 4 s⁻¹ for H₂ oxidation, significantly slower in both directions compared with CyArg.

A second control complex, the methyl ester of CyArg, ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{ArgOMe}}\right)}_2\right]}^{2+}$ (CyArgOMe), was prepared to test the importance of the COOH groups. The TOFs for this complex were also significantly reduced compared with CyArg, with TOFs at 348 K, 1 atm 25% H₂/Ar, and pH = 1.0 of 7 s⁻¹ for H₂ oxidation and 35 s⁻¹ for H₂ production (Table 1), but faster than those observed for CyGly.

Reversibility under an N₂ Atmosphere: Studies of H₂ Production in the Absence of H₂.

To enhance H₂ production at high temperature, the electrochemical behavior for CyArg was studied under an N₂ atmosphere from pH = 0.0 to pH = 6.0, and temperatures up to 348 K. Under these conditions the catalyst exhibited predominantly H₂ production (Table 1, Fig. 4, and SI Appendix, Fig. S5). The small contribution from H₂ oxidation (Fig. 4, asterisk) is a result of the H₂ produced in the reverse reaction. This wave is not observed in anodic sweeps originating at −0.05 V, providing further evidence of reversibility. Similar results were observed for CyGly at pH = 1.0 (SI Appendix, Fig. S4).

Fig. 4.

CyArg under N₂ shows reversible catalysis dominated by H₂ production. Cyclic voltammograms for 50 μM CyArg under 1 atm N₂ in aqueous 0.1 M Na₂SO₄ at pH 0.0 and temperatures from 293 to 348 K. The vertical gray dotted line denotes the equilibrium potential for reversible catalysis. The horizontal black arrow indicates the initial scanning direction. Scan rates were 0.1 V/s. The asterisk denotes H₂ oxidation observed as a result of the H₂ produced from the H⁺ reduction.

In the electrochemistry for all three complexes, a second wave is observed negative of the reversible electrocatalytic response (H_rev). This wave (H_irr) is most easily seen for CyArg in Fig. 3 and is also scan-rate independent. Based on the lack of catalytic activity for a glassy carbon electrode following a rinse test, this wave appears to represent a second homogeneous process for catalytic H₂ production (SI Appendix, Fig. S6) (18). At lower pH values, H_rev increases relative to H_irr suggesting either enhanced catalytic TOFs for H_rev vs. H_irr, or that the species resulting in H_irr is transforming into the species responsible for H_rev. The mechanism for H_irr is not yet clear and is the subject of further investigation.

H₂ Oxidation at High Pressure and High Temperature.

To investigate the role of temperature on overpotential, catalysis for CyArg was run at 348 K, at 67 atm H₂. Although high pressure alone results in an increase in overpotential (18), operating at elevated temperatures and pressure (SI Appendix, Fig. S7) results in a reduction in the overpotential, from 350 mV at 298 K to 200 mV at 350 K.

Reversibility of H₂ Addition.

Increasing TOFs of H₂ production as a function of decreased H₂ pressure (Fig. 3) suggests that H₂ addition becomes reversible. To further evaluate the reversibility of the addition of H₂ to CyArg, two additional experiments were performed. In one experiment, hydrogen addition is shown to be reversible by purging the hydrogen addition product $\left({\left[\mathrm{N}{\mathrm{i}}^0{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arg}}\mathrm{H}\right)}_2\right]}^{7+}\right)$ with N₂, resulting in the loss of H₂ based on the recovery of the characteristic red color of the Ni^II species. In a second experiment, ³¹P NMR experiments were performed for CyArg at temperatures from 298 to 348 K at pH values of 1.0 and 4.0 at 1 atm H₂. At both pH values, complete conversion to the hydrogen addition products was confirmed. Upon heating, the Ni(II) species was regenerated. Cycling the temperature between 313 and 348 K resulted in the reversible addition (313 K) and loss (348 K) of H₂ as seen in Fig. 5 and SI Appendix, Figs. S8 and S9.

Fig. 5.

H₂ binding to CyArg is reversible above 313 K. ³¹P{¹H} NMR spectra for CyArg at pH = 4.0 in water. (A) The starting Ni(II) (under N₂) looks identical from 298 to 348 K (shown); (B) exposed to 1 atm 25% H₂ at 298 K, conversion to the H₂ addition products, [Ni(0)(P^Cy₂N^Arg₂H)₂], is observed, including the e/e (15.5 ppm; 60%) and e/x (e/x: 14 ppm and e/x: −9 ppm; 40%) isomers. These isomers are the only observed species up to 313 K (shown) and the e/e and e/x peaks merge at 323 K (SI Appendix, Fig. S9). (C) Heating to 348 K, Ni(II) was observed (11 ppm; 10%), demonstrating H₂ release. (D and E) Cycling the temperature of the solution between 313 and 348 K demonstrates reversible H₂ addition from the H₂ addition products. Integrals are shown.

Stability of CyArg at High Temperature and Low pH.

The stability of CyArg was tested by cyclic voltammetry and NMR spectroscopy. Sequential cyclic voltammograms showed no degradation in catalytic peak current over 20 min at 348 K (SI Appendix, Fig. S10). The ³¹P{¹H} NMR spectra at elevated temperature also showed minimal degradation (∼5%) after 90 min under similar solution conditions (SI Appendix, Figs. S8 and S9).

Protonation State of the Complex.

The extent of protonation of CyGly was examined in MeCN by ³¹P{¹H} NMR. Under N₂, no pendant amines were protonated and no hydrides were observed, as expected (Fig. 6, bottom spectrum). Introducing 1 atm H₂ generated the hydrogen addition product (15.5 ppm) in the ¹H NMR spectrum (Fig. 6, middle spectrum).^‡ Addition of acid results in a shift in the hydrogen addition product to 13.8 ppm, and a hydride species at −11.4 ppm (Fig. 6, top spectrum). The integrations are consistent with a species with three protons, instead of only two as expected if the only source was the addition of H₂.

Fig. 6.

The ¹H NMR spectroscopy indicates protonation of the complex CyGly in MeCN. (Bottom) Under N₂ at neutral pH, no protonated amines or hydrides are observed in the ¹H NMR spectrum. (Middle) Adding H₂ at neutral pH results in the hydrogen addition product, x/x in MeCN (15.6 ppm). (Top) The addition of three equivalents of HTFSI results in an additional resonance at −11.4 ppm, consistent with a hydride, along with an observed upfield shift in the x/x resonance to 13.8 ppm. The integrations are consistent with a species with three protons.

Discussion

The TOFs and lack of overpotential of the CyArg complex presented here are comparable to those of the [NiFe]-hydrogenase enzymes. In the following discussion we will examine some of the features responsible for this enzymelike performance, including the specific roles of H₂ pressure, highly acidic solutions, ligand structure, and temperature.

The Effect of H₂ Pressure.

The catalytic TOFs for H₂ oxidation for CyArg depend linearly on the H₂ pressure (18), and H₂ pressures of 1.0 atm or greater inhibit H₂ production (Fig. 2). For lower partial pressures of H₂ (0.25 or 0.0 atm, Figs. 3 and 4, respectively) reversible catalysis is observed by cyclic voltammetry experiments with large currents for both H₂ oxidation and production. These results indicate the reversibility of H₂ binding, and also imply that the TOFs for these complexes are determined by the rate of H₂ addition, and by extension of microscopic reversibility, hydrogen elimination.

The Role of Aqueous, Acidic Solutions.

As can be seen from Fig. 3, the catalytic TOFs increase dramatically below pH 2. As demonstrated by NMR experiments, the catalyst is more highly protonated under acidic conditions (Fig. 6), likely resulting in a different catalytic mechanism. A proposed mechanism is shown in Fig. 7 (clockwise for H₂ oxidation, counterclockwise for H₂ production), with the standard mechanism for the nonprotonated species shown in SI Appendix, Fig. S11. This mechanism is initiated by the protonation of the doubly charged Ni(II) complex, 1, to form the triply charged Ni(II) complex, 2. Protonation produces a positive shift of the Ni(II/I) couple (Fig. 3), indicating a lower-energy lowest unoccupied molecular orbital (LUMO). The lower-energy LUMO of 2 will likely also lower the associated barriers for H₂ addition/elimination (19), contributing to faster and more reversible H₂ addition for complex 2 compared with complex 1. We hypothesize that this proton is positioned endo, based on the observation from ³¹P{¹H} NMR that the catalyst preferentially positions protons endo to the Ni in aqueous solution (Fig. 5), the necessary location for H₂ release.

Fig. 7.

Altered mechanism due to protonation. The first step in this mechanism is the protonation of the Ni(II). The carboxylic acids, as well as the R′ groups could contribute in several hydrogen bonding and protonation/deprotonations, one of which is shown. The additional amino acids are not shown for simplicity. Note that for simplicity, the charges shown do not account for charges on the amino acid side chains, hence the overall charge will depend on the nature of the amino acid.

H₂ addition generates a hydride (Figs. 6 and 7) and protonates one pendant amine, likely on the ligand that was not initially protonated. This ligand will have more basic pendant amines and will thus be the favored site for heterolytic H₂ cleavage (16, 20) as shown in Fig. 7. The overall mechanism is supported by the hydride intermediate observed using ³¹P{¹H} NMR spectroscopy (Fig. 6) that is consistent with complex 4, an intermediate not observed for nearly all other ${\left[\mathrm{Ni}\left({\mathrm{P}}_2^{\mathrm{R}}{\mathrm{N}}_2^{\mathrm{R}\prime }\right)\right]}^{2+}$ complexes (16). Of interest is that once complex 4 is formed, there are two equivalent deprotonation sites (i.e., the molecule is now symmetric, assuming fast mobility of the hydride, which has been observed before for the nonprotonated complex) (21). This effectively provides two sites for deprotonation, and may contribute to the high efficiency for this reversible process. Multiple proton channels have been proposed for both [FeFe]- and [NiFe]-hydrogenases (22–26), and it may be that the enzymes use multiple channels to achieve maximum efficiency. Structural studies and studies evaluating the movement of protons are ongoing to provide additional details of the mechanism for the protonated complex.

The Effect of the Carboxyl Group.

The carboxyl group of the amino acid plays an important role in the catalytic mechanism. As discussed previously (17, 18), the introduction of a carboxylic acid in the outer coordination sphere results in rapid proton transfer, a critical component in eliminating the overpotentials required for reversibility. The endo positioned carboxyl group may also play a role in stabilizing the dihydrogen intermediate, 3, of Fig. 7. The observation that the CyArg catalyst is 3 times faster than CyArgOMe for H₂ oxidation and 8 times faster for H₂ production, coupled with the observation that H₂ addition is rate limiting, indicates that the OH group of the carboxyl substituent facilitates H–H bond cleavage as shown in structure 3. This interaction is less likely for CyArgOMe, because the larger Me group of the ester will interact more strongly with the R substituent on P, thereby reducing the interaction between the carboxyl group and the dihydrogen ligand. This possibility is currently under investigation using computational methods.

The Effect of the Amino Acid Side Chain.

A comparison of the TOFs of CyGly with CyArg for H₂ oxidation and production at pH 1 (Table 1) indicates that the Arg substituent enhances H₂ oxidation by a factor of five and H₂ production by a factor of 100. This demonstrates the essential influence of the amino acid side chain on the activity, and because H₂ addition is rate limiting, on the rate of H₂ addition and elimination in particular. This has been attributed previously to the CyArg complex exhibiting Arg–Arg interactions that alter the Ni…N distance to facilitate H₂ addition (18). Another possible contribution is a fine tuning of the positioning of the carboxyl group due to interactions between the amino acid side chains, optimizing its role. These postulates will be aided by ongoing structural studies.

The Effect of Temperature.

As shown in Figs. 2 and 4, increasing the temperature results in faster TOFs for H₂ oxidation combined with lower overpotentials. In contrast, previous studies have shown that when H₂ pressure alone was used to increase the TOFs, i.e., in the absence of elevated temperature, the overpotential also increased, from 180 to 460 mV (17, 18). The higher overpotential under high pressure conditions in the absence of elevated temperature is consistent with slow electron and/or proton transfer, because a more positive potential is required for the electron transfer rate to match the rate of addition of H₂. Therefore, the observed decrease in overpotential as a function of temperature suggests that temperature is facilitating the electron and/or proton transfers shown in Fig. 7. At sufficiently high temperatures, a Nerstian shape is observed for the reversible electrocatalytic waves and the potential shifts linearly as a function of pH (60 mV per pH unit), all of which are consistent with rapid, reversible electron and proton transfers under these conditions. For efficient electrocatalysis (i.e., operating at low or no overpotential), electron-transfer and proton-transfer rates must be fast and reversible, and at least as fast as the H₂ addition and elimination steps. CyArg achieves this with a combination of factors both internal to and external to the complex.

The temperature also affects H₂ addition and elimination. As shown in Fig. 5, and SI Appendix, Figs. S8 and S9, H₂ addition is reversible above 313 K. This is consistent with the temperature at which reversible catalysis is initially observed (323 K), consistent with the need for reversible H₂ binding and release before reversible activity can be observed.

Summary and Conclusions

Achieving reversibility for a catalytic cycle necessitates achieving reversibility for all of the steps that occur during catalysis. These steps can be grouped into three categories: (i) e⁻ transfer steps, (ii) H⁺ transfer steps, and (iii) H₂ addition and elimination. For the complexes described above, fully reversible catalysis of H₂ oxidation and production has been achieved by an interrelated combination of solution conditions (pH, H₂ pressure, and temperature) and careful catalyst design.

CyArg displayed reversible catalytic H₂ oxidation and production behavior at a broad pH range (0–6) at elevated temperatures (>323 K) and subatmospheric H₂ pressure. The complex showed the maximal activity for both directions at low pH conditions (pH < 1), and H₂ production was inhibited under higher pressures of H₂. Catalyst design is also critical to the achieved catalytic functionality and our studies demonstrate that reversible H₂ addition and release from the catalyst at pressures close to 1.0 atm H₂ begins with a suitable choice of the pendant amine (a component of the second coordination sphere) and the electronic properties of the Ni center (determined by the first coordination sphere). In addition, it is also clear that the carboxylates and side chains introduced into the outer coordination sphere by incorporation of amino acids exert an essential influence on the catalytic properties of the active center in much the same way that the protein matrix is thought to influence the catalytic activity of enzymes. In this sense, the amino acid substituted ${\left[\mathrm{Ni}\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Amino}\hspace{0.5em}\mathrm{acid}}\right)\right]}^{2+}$ complexes described here, that are beginning to rival the TOFs and overpotentials of enzymes, represent a remarkable opportunity to probe in detail how amino acids and the protein matrix can influence the activity of catalytically active sites in synthetic electrocatalysts.

Although the mechanistic details controlling fast, reversible catalysis in hydrogenases are likely different, the Léger group demonstrated that catalytic bias in [NiFe]-hydrogenase is controlled not by the redox properties of the active site, but by residues remote from the active site that control either H₂ release (H₂ production) or electron transfer (H₂ oxidation) (8). In this model system, we demonstrate that we are controlling H₂ addition and proton transfer with groups remote from the active site, providing a conceptual parallel between the role of the outer coordination sphere in the molecular model and the role of the protein scaffold in the enzymatic system.

It is notable that the absence of overpotentials and the high catalytic TOFs observed for both H₂ production and oxidation at 348 K for CyArg are comparable to those of the [NiFe]-hydrogenase enzymes, but offer significant advantages in stability under a broader set of reaction conditions. CyArg is not inhibited by CO (18), is stable for extended periods at high acid concentrations (18), and as shown in this work, is reasonably stable under high acid at elevated temperatures. These conditions are optimal for functioning in fuel cells, and were achieved by using an enzyme-inspired design, including first, second, and outer coordination spheres.

Materials and Methods

General Procedures.

All samples were prepared under an N₂ atmosphere either using a standard single manifold Schlenk line or glove box. Ultra-pure water, 18.2 MΩ • cm, was obtained from a Millipore unit.

Preparation of complexes.

The ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Glycine}}\right)}_2\right]}^{2+}$ (CyGly) (17) and ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}}\right)}_2\right]}^{2+}$ (CyArg) (18) were prepared as previously reported. ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}\hspace{0.5em}\mathrm{methyl}\hspace{0.5em}\mathrm{ester}}\right)}_2\right]}^{2+}$ (CyArgOMe) is a new complex and was prepared by the same synthetic procedure, using arginine methyl ester HCl as the amine precursor.

$\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}\hspace{0.5em}\mathrm{methyl}\hspace{0.5em}\mathrm{ester}}\right)$: ¹H NMR (CD₃OD): δ 1.15–1.97 (H-Cy, Arg-H^δ, and Arg-H^ε, 30H, m); 3.17–3.41 ppm (-PCH₂N, Arg-H^α, and Arg-H^γ, 14H, m), 3.72 (Arg-COOCH₃, 3H, s). ³¹P{¹H} NMR (CD₃OD): δ −41.6 ppm (broad), -45.2 ppm (sharp). ESI MS (+ve mode) (in methanol): m/z ${\left[\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}\hspace{0.5em}\mathrm{methyl}\hspace{0.5em}\mathrm{ester}}\right)+\mathrm{H}\right]}^{+}$: 657.42 (calcd: 656.78).

$[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}\hspace{0.5em}\mathrm{methyl}\hspace{0.5em}\mathrm{ester}}\right)}_2{\left(\mathrm{B}{\mathrm{F}}_4\right)}_2$ : ¹HNMR (D₂O): δ 1.18–1.99 (H-Cy, Arg-H^δ, and Arg-H^ε, 30H, m); 3.15–3.32 ppm (-PCH₂N, Arg-H^α, and Arg-H^γ, 14H, m), 3.75 (Arg-COOCH₃, 3H, s). ³¹P{¹H} NMR (D₂O): 15.6 ppm, broad. ESI MS (+ve mode) (in methanol): m/z ${\left[\mathrm{Ni}{\left({\mathrm{P}}_2^{\mathrm{Cy}}{\mathrm{N}}_2^{\mathrm{Arginine}\hspace{0.5em}\mathrm{methyl}\hspace{0.5em}\mathrm{ester}}\right)}_2{\left(\mathrm{B}{\mathrm{F}}_4\right)}_2-\mathrm{B}{\mathrm{F}}_4+\mathrm{C}{\mathrm{H}}_3\mathrm{OH}\right]}^{+}$: 1489.78 (calcd: 1489.13).

Electrochemistry.

Cyclic voltammetry was performed with solutions of 0.1 M Na₂SO₄ electrolyte in water using a glassy-carbon electrode (1 mm diameter), polished with 0.25 μm MetaDi diamond polishing paste (Buehler). The buffer solutions with pH values of 5.0 and 6.0 were prepared using 0.1 M 2-morpholinoethanesulfonic acid (Mes) buffer followed by adjusting the pH using dilute aqueous NaOH and dilute acetic acid solutions. Solutions with pH 0.0 and 1.0 were prepared by adding appropriate amounts of 1.16 M HClO₄, and pH 1.55 were prepared by adding bis(trifluoromethane)sulfonamide acid (HTFSI). All solution pHs were constant throughout the experiment as monitored by pH meter. Cyclic voltammetry experiments were performed from 298 to 348 K on a CH Instruments 600D electrochemical analyzer using a standard three-electrode configuration. A glassy carbon rod was used as the counterelectrode and an AgCl-coated Ag wire (in saturated KCl) separated from the analyte solution by a Vycor frit was used as the reference electrode. All couples were referenced to the equilibrium potential observed at each pH condition along with the internal standard hydroxymethyl ferrocenium/ferrocene couple at 380.0 mV vs. SHE (see SI Appendix for determination of referencing using the open circuit potential method, SI Appendix, Fig. S12).

Turnover frequencies were determined with Eq. 2, as previously reported (17, 18):

$$TOF=\frac{1}{D}{\left(\frac{i_{cat}}{nFA\left[cat\right]}\right)}^2.$$ [2]

In Eq. 2, i_cat is the catalytic current (measured from the zero current to the plateau of the wave for either H₂ oxidation or H₂ production); n is the number of electrons involved in catalysis (two in this case); F is Faraday’s constant; A is the surface area of the electrode (17); [cat] is the catalyst concentration; and D is the diffusion coefficient for the catalyst. The diffusion coefficients were determined previously (17, 18) and extrapolated to higher temperatures as described in the SI Appendix. The overpotentials were calculated by taking the difference between the E_cat/2 (the potential corresponding to 1/2 the catalytic current, i_cat/2) and the equilibrium potential (SI Appendix, Fig. S7), measured using the open circuit potential method (SI Appendix) (27).

NMR Studies.

H₂ addition/elimination.

The 50 μM solution of CyArg complex was prepared in pH 4.0 solution and purged with 25% H₂ in Ar for 15 min; a ³¹P{¹H} NMR spectrum of the resulting sample was recorded at 313 K. Then the sample was heated and cooled between 348 and 313 K 2 times, recording ³¹P{¹H} NMR spectra at each temperature. In a separate experiment, a ³¹P{¹H} NMR spectrum was recorded for 50-μM CyArg under an N2 atmosphere and similar conditions (pH 4.0 solution at 348 K).

Protonation state of CyGly.

The ¹H NMR spectrum of 1.0 mM solution of CyGly* (Gly*: ¹³C/¹⁵N labeled glycine) complex in CD₃CN was recorded under N₂. The sample was purged with H₂ for 15 min and the changes in ¹H NMR spectra were monitored. Another spectrum was measured following the addition of three equivalents of HTFSI to the resulting solution under H₂ atmosphere. All of the measurements were performed at room temperature (298 K).