Proton Relay in Iron Porphyrins for Hydrogen Evolution Reaction

Abstract

The efficiency of the hydrogen evolution reaction (HER) can be facilitated by the presence of proton-transfer groups in the vicinity of the catalyst. A systematic investigation of the nature of the proton-transfer groups present and their interplay with bulk proton sources is warranted. The HERs electrocatalyzed by a series of iron porphyrins that vary in the nature and number of pendant amine groups are investigated using proton sources whose pK_a values vary from ~9 to 15 in acetonitrile. Electrochemical data indicate that a simple iron porphyrin (FeTPP) can catalyze the HER at this Fe^I state where the rate-determining step is the intermolecular protonation of a Fe^III-H⁻ species produced upon protonation of the iron(I) porphyrin and does not need to be reduced to its formal Fe⁰ state. A linear free-energy correlation of the observed rate with pK_a of the acid source used suggests that the rate of the HER becomes almost independent of pK_a of the external acid used in the presence of the protonated distal residues. Protonation to the Fe^III-H⁻ species during the HER changes from intermolecular in FeTPP to intramolecular in FeTPP derivatives with pendant basic groups. However, the inclusion of too many pendant groups leads to a decrease in HER activity because the higher proton binding affinity of these residues slows proton transfer for the HER. These results enrich the existing understanding of how second-sphere proton-transfer residues alter both the kinetics and thermodynamics of transition-metal-catalyzed HER.

Graphical Abstract

INTRODUCTION

Reactions that require both protons and electrons are abundant in nature and are involved in both generation and storage of energy. Of these, reactions like O₂ reduction/generation,^1–6 H₂ generation/oxidation^7–10 and CO_x/SO_x/NO_x reduction^11–18 are of keen contemporary interest because of the ongoing pursuit for clean energy and environment. Efficient catalysts are required for the conversion of electricity from renewable energy sources like solar and wind to chemical energy for storage in the form of chemical fuels. Natural enzymes have evolved to include sophisticated active sites where, apart from the inner coordination sphere of the metal, the noncovalent secondary interactions like hydrophobicity, hydrogen bonding, and electrostatics play a major role in defining the reactivity of these active sites.^19,20 The protein scaffold surrounding the enzyme active sites often controls the bond formation and cleavage by means of electron-transfer, proton-transfer, and atom-transfer reactions, imparting them the desired rate and selectivity.^19,21–23 Emulating these finer attributes of metalloenzyme catalysis in bioinspired small molecules has been a major area of focus in recent records.^24–27 The role of proton relays is being actively investigated in reactions that require multiple protons and electrons, e.g., H⁺ reduction,^28,29 O₂ reduction,^30,31 CO₂ reduction,^32,33 H₂ oxidation,^34–36 and H₂O oxidation.^37–39 These results all stress the importance of the precise positioning of proton-transfer groups in promoting proton transfer, proton-coupled electron transfer, and hydrogen-bond formation, which are involved in the key steps in catalysis.

The reduction of protons to produce molecular hydrogen by means of storing energy is catalyzed most efficiently by the natural enzyme hydrogenase.^34,40 To attain efficient HER and hydrogen oxidation reaction (HOR), several groups have attempted to incorporate proton shuttles in mononuclear complexes using first-row transition metals (e.g., iron, nickel, and cobalt).^41–45 Such efforts are inspired by the bridging azadithiolate ligand present in the binuclear active site of [Fe–Fe]hydrogenase.^34,46,47 Few research groups have investigated the kinetic and thermodynamic advantage of utilizing outer-coordination-sphere proton relays in artificial complexes during catalysis.^42,48,49 Conceivably, a proton-transfer residue can shuttle protons, lowering the barrier, and/or act as a local source of proton, lowering the entropic cost involved in protonating the catalytic metal center. Thus, these phenomena entail interplay between the pK_a values of the metal center, of these second-sphere residues, and of the bulk acid source (pH in water). DuBois and coworkers had demonstrated that the incorporation of a nitrogenous base in a diphosphine ligand backbone, forming the complex [Ni(PNP)₂]²⁺, readily binds hydrogen and shows faster electrochemical hydrogen oxidation at lower overpotential than the analogous [Ni(depp)₂]²⁺ complex with no tethered base.⁴³ Later, these ligands were modified by increasing the chain length and number of bases in [Ni(P^R₂N^R’₂)₂]²⁺ to allow better access of the nitrogenous proton shuttle to the nickel center, resulting in very high turnover rates for electrocatalytic HER in the nickel catalyst [Ni(P^Ph₂N^Ph)₂](BF₄)₂ (P^Ph₂N^Ph = 1,3,6-triphenyl-1-aza-3,6-diphosphacycloheptane).^50–52 Shaw and coworkers established that the arginine amino acid in a water-soluble [Ni(P^Cy₂N^Arg₂)₂]⁸⁺ complex enhances the electrocatalytic activity when these residues were protonated.³⁶ In the above case, the guanidium groups on the outer sphere of the molecule were proposed to shuttle protons via the Grotthuss mechanism.³⁶ Graham and Nocera introduced a hanging carboxylic acid/pendant base in the second coordination sphere of iron porphyrins, hangman porphyrins, to channel protons to the active site to exhibit enhanced catalytic HER activity compared to nonhangman analogues.⁵³ The thermodynamic advantage of reducing the overpotential for HER was also evidenced by having a hanging carboxylic acid in a cobalt hangman porphyrin.⁴⁸ Similarly, while iron porphyrins were reported to catalyze HER in their formal Fe⁰ state,⁸ the presence of four hydrogen-bond donor triazole moieties (pK_a in acetonitrile is ~8)⁵⁴ along with four ferrocene groups in a mononuclear iron porphyrin FeFc₄ allowed the protonation of iron(I) porphyrins, reducing the overpotential of HER by 50% in the presence of strong acids in both aqueous and organic solvents.⁵⁵ Very recently, Artero and co-workers reported HER catalyzed by a [Co(bapbpy)Cl]⁺ complex containing a redox-active bipyridine ligand with associated pendant proton relays.⁵⁶ The electrocatalytic behavior in response to systematic variation in pK_a of the acid source suggests that the Co^I state becomes activated for HER depending upon pK_a of the acid sources used. Conversely, systematic variations in pK_a and the number of the pendant bases along with variation of pK_a of the acid source are necessary to complement these investigations and advance the current understanding of the roles played by these proton shuttles.

In this paper, electrocatalytic HER catalyzed by a series of mononuclear iron meso-tetraphenylporphyrin (FeTPP) derivatives (Figure 1) using acid sources over a range of pK_a = 8.64–14.98 in acetonitrile is investigated. It has been suggested computationally that intramolecular proton transfer results in a reduced energy barrier.⁵⁷ Using unsubstituted FeTPP as a control, electrocatalysis by the iron porphyrins with pendant bases is investigated. The iron porphyrins are appended with arms bearing basic residues, which vary in both their pK_a values and numbers. The results obtained suggest that pendant basic residues, which get protonated in the presence of bulk acid sources, can act as local proton sources and lower the barrier of protonation of the Fe^I state of the catalyst. The effect of variation in the pK_a value and number of pendant bases is discussed in detail using a series of iron porphyrins, where both pK_a value of the pendant base and number of bases are systematically varied.

Figure 1.

Representative molecular structures of the iron porphyrins [FeTPP (A), FeL3 (B), FeL2 (C), and ⁶LFe (D)] studied here.

RESULTS AND DISCUSSION

Electrocatalytic Hydrogen Evolution by FeTPP.

The cyclic voltammetry (CV) of FeTPP is well documented in the literature^58–60 and exhibits three responses, corresponding to the Fe^III/II, Fe^II/I, and Fe^I/0 redox couples at −0.68, −1.5, and −2.07 V (Figure S1), respectively (with respect to Fc⁺/Fc⁰), with peak-to-peak separations (ΔE_p) of 61, 67, and 71 mV, respectively, at a 100 mV/s scan rate (ν) in acetonitrile. In the presence of acids with a wide range of pK_a values [e.g., tosic acid (TsOH; pK_a = 8.64), p-bromoanilinium tosylate (BA-H⁺; pK_a = 9.43), diisopropylanilinium tosylate (DIPA-H⁺; pK_a = 10.62), N,N-diethylanilinium tosylate (DEA-H⁺; pK_a = 11.42), and collidinium tosylate (Col-H⁺; pK_a = 14.98)] into the acetonitrile solution of 1 mM FeTPP, an electrocatalytic HER is observed (the tosylate salts of the conjugate acids are generated in situ by adding TsOH to the corresponding amine).^61–63 The Fe^III/II potential is not affected in the presence of these bases because they are all noncoordinating due to the presence of bulky groups around the potentially coordinating nitrogen atom (Figure S1). This is in contrast to recent investigations using iron porphyrins, where coordinating axial ligands were demonstrated to substantially affect the kinetics of the HER.⁶⁴

The data suggest that the electrochemical response along with the HER current is dependent on the pK_a value of the external acid source used. For the addition of a strong acid like TsOH (pK_a = 8.64) in 1 mM FeTPP [0.1 M tetrabutylammonium perchlorate (TBAP) in acetonitrile], the Fe^II/I redox wave becomes irreversible and HER is observed with a peak potential of −1.50 V. This HER current corresponds to the process where a Fe^I species is protonated to form Fe^III-H⁻, which is then reduced to Fe^II-H⁻ and becomes further protonated to evolve H₂ similar to previously reported steps in related systems.^48,65–67 There are irreversible processes between the Fe^III/II and Fe^II/I redox couples (Figure 2A), which are due to protonation of the meso-carbon in the porphyrin ligand (also known as phlorin) by a strong acid.^8,68,69 However, closer inspection of the data reveal that the HER current overlays the reduction of Fe^II to Fe^I at −1.50 V, which indicates H₂ formation via protonation of the Fe^III-H⁻ species.

Figure 2.

CV diagrams of 1 mM FeTPP in an acetonitrile solution with a glassy carbon working electrode using 0.1 M TBAP as the electrolyte for the HER using four different acid sources: (A) TsOH; (B) BA-H⁺; (C) DIPA-H⁺; (D) DEA-H⁺; (E) Col-H⁺. The scan rate is 100 mV/s. A platinum wire is used as a counter electrode. All potentials are reported with respect to the Fc⁺/Fc⁰ couple using ferrocene as an internal standard.

When the acid source is changed from TsOH to weaker acids like BA-H⁺ (pK_a = 9.43), DIPA-H⁺ (pK_a = 10.62), DEA-H⁺ (pK_a = 11.42), and Col-H⁺ (pK_a = 14.98), the catalytic process originating from the Fe^I oxidation state at −1.30 V is retained, but the irreversible electrochemical response, due to protonation of the porphyrin ring, is not observed. Furthermore, the catalytic current arising from protonation of the Fe^I species at the Fe^II/I potential gets saturated; i.e., the HER current becomes independent of the substrate [BH]⁺ concentration. For example, with BA-H⁺ and DIPA-H⁺, the catalytic HER current from the Fe^I state becomes saturated at 5 equiv of acid. The further addition of acid into the same solution results in another redox process at more negative potential corresponding to HER by the protonation of Fe^II-H⁻ species (Figure 2B,C) produced by the reduction of its Fe^III-H⁻ precursor. When the acid source is changed to DEA-H⁺ (pK_a = 11.42), which has higher pK_a value than DIPA-H⁺, the catalytic activity of the Fe^I state is substantially reduced, as indicated by the diminished catalytic current at the Fe^II/I potential, and the catalytic current corresponding to HER obtained by the Fe^II-H⁻ species dominates (Figure 2D).

Using the Col-H⁺ as the acid source (pK_a = 14.98), the CV diagram of the Fe^II/I redox process remains unaltered and the HER proceeds mostly through the Fe^II-H⁻ species at −1.82 V under the same condition. CV at very slow scan rates shows a loss of reversibility of the Fe^II/I CV response (Figure 4) and minimal current enhancement. Instead, the HER proceeds via Fe^II-H⁻, implying protonation of the Fe^I state but not the resulting Fe^III-H⁻ species. Thus, the electrochemical data presented here clearly demonstrate that there are two possible routes to the formation of H₂ catalyzed by simple iron porphyrins like FeTPP. For acid sources having pK_a ≤ 11.42 (BA-H⁺, DIPA-H⁺, DEA-H⁺, and TsOH), the Fe^I state is protonated to form a Fe^III-H⁻ species, which then accepts another proton to generate H₂ without further reduction. As was already established, another possible route to HER after saturation of the catalytic current or in the absence of catalysis from the Fe^II/I redox state is reduction of the Fe^III-H⁻ species, generating a Fe^II-H⁻ species, which is further protonated to eliminate H₂. The latter process was investigated in detail earlier^{8,64,66,67,70} and is not the focus of this investigation, which is directed to the HER from the Fe^I state.

Figure 4.

Scan rate dependence plot of 1 mM FeTPP in an acetonitrile solution containing 0.1 M TBAP as the electrolyte with a glassy carbon working electrode and platinum wire as a counter electrode. All potentials were reported versus Fc⁺/Fc using ferrocene as an internal standard using 5 equiv of collidinium (Col-H⁺) as the acid source.

The changes in the electrochemical response in FeTPP with a change of pK_a of the acid source reflect the pK_a values of different protonation events in FeTPP. Considering the discussion further and as depicted in the mechanistic scheme of Figure 3, there are four different protonation events under HER conditions that manifest themselves through distinct electrochemical responses. The CV responses allow an estimation of the approximate pK_a values for these processes. The processes and their effect on the CV response are as follows (note that pK_a of the metal center refers to pK_a of the species before protonation or pK_a of the acid source required to protonate because protonation of the same are irreversible):

1. Protonation of the porphyrin ring (pK_a1) results in irreversible processes between the formal Fe^III/II and Fe^II/I redox couples.

2. Protonation of Fe^I (pK_a2) to form Fe^III-H⁻ results in an irreversible nature of the Fe^II/I redox couple.

3. Protonation of Fe^III-H⁻ (pK_a3) results in a HER current at the Fe^II/I redox potential (−1.48 V).

4. Protonation of the Fe^II-H⁻ species (pK_a4) results in a HER current at a more cathodic potential, that beyond the Fe^II/I redox process (−1.86 V).

Figure 3.

Left: CV overlay of 1 mM FeTPP in an acetonitrile solution with a glassy carbon working electrode using 0.1 M TBAP as the electrolyte at a scan rate of 100 mV/s for HER using four different acid sources: (A) TsOH (pK_a = 8.7; associated with protonation steps 1–3); (B) BA-H⁺ (pK_a = 9.43; associated with protonation steps 2–4); (C) DIPA-H⁺ (pK_a = 10.62; associated with protonation steps 2–4; (D) DEA-H⁺ (pK_a = 11.42; associated with protonation steps 2–4); (E) Col-H⁺ (pK_a = 14.98; associated with protonation steps 2 and 4). Platinum wire was used as a counter electrode, and all potentials were reported versus Fc⁺/Fc using ferrocene as an internal standard. Right: Proposed mechanism of HER for FeTPP with probable pathways in the presence of variable acid sources with different pK_a values.

Protonation of the porphyrin ring can occur in the presence of a strong acid like TsOH (pK_a = 8.64) and does not occur in BA-H⁺ (pK_a = 9.43), indicating that pK_a of the porphyrin ring lies below 9.43 (pK_a1) and above 8.64. Protonation of the Fe^II-H⁻ species (pK_a4), resulting in the catalytic HER current, can be achieved even with Col-H⁺, specifying that the Fe^III-H⁻ species is formed and reduced to generate Fe^II-H⁻ species; the pK_a value of the Fe^II-H⁻ species (>14.98) is greater than those of all of the acid sources investigated here. The Fe^II/I redox process is visible in the presence of Col-H⁺ and becomes irreversible only at very slow scan rates (Figure 4), suggesting that the protonation of Fe^I is slow (does not occur at scan rates higher than 100 mV/s) and pK_a2 is greater than 14.98. Despite protonation of Fe^I at slow scan rates, HER is not observed at the Fe^II/I potential via the protonation of Fe^III-H⁻ produced but rather HER is observed upon its reduction to Fe^II-H⁻. Thus, the pK_a3 value of Fe^III-H⁻ must be lower than 14.98, and/or the rate-determining step of HER by an iron(I) porphyrin must be the second protonation step, i.e., protonation of the Fe^III-H⁻ species. Alternatively, FeTPP shows a small catalytic current from the Fe^I state in the presence of DEA-H⁺, suggesting that the pK_a3 value of Fe^III-H⁻ is >11.42. Note that the fact that pK_a of the monoanionic Fe^I species is higher than that of the neutral Fe^III-H⁻ species, i.e., pK_a2 > pK_a3, is logical. Having established the basic pK_a values of the species involved in HER catalyzed by a simple FeTPP, the effect of pendant amine groups has been investigated and is now discussed.

The protonation of Fe^I to form Fe^III-H⁻ occurs in the presence of all acids (pK_a = 9.43–14.98) used here, and subsequent protonation of Fe^III-H⁻ (11.42 < pK_a3 < 14.98) to produce H₂ results in the catalytic HER current at the Fe^II/I potential. A feature of the catalytic behavior that had eluded past investigations that used strong acids is saturation of the HER current originating from the Fe^I species. Such a behavior is suggestive of a fast precatalytic equilibrium binding step between the Fe^III-H⁻ species and substrate, proton source, involved. This is supported by a plot of i_cat/i_p versus equivalents of acid used, which shows that saturation is achieved at a much lower concentration for acids having a lower pK_a (Figure 5A), i.e., better hydrogen-bond donors. This situation is characteristic of Michaelis–Menten-type enzyme kinetics, where a pre-equilibrium enzyme substrate binding step precedes the rate-determining step and saturation of the rate is observed at higher substrate concentration. Because the TPP ligand framework does not offer a distinct binding site for these anilinium ions from the acid source, it is likely that the Fe^III-H⁻ species associates with the anilinium ion of the corresponding conjugate acid in solution via intermolecular dihydrogen bonding interaction before catalysis occurs. Such interactions are weakened in the presence of electrolytes. Accordingly, saturation behavior of the HER current is lost when the concentration of the electrolyte is increased to 1 M (Figures 5B and S2).

Figure 5.

i_cat/i_p plot for the catalytic wave at the Fe^II/I redox potential with the addition of equivalents of four different acids for FeTPP in the presence of (A) 100 mM TBAP and (B) 1 M TBAP.

Effect of a Distal Base on the Electrochemical HER.

The porphyrins are grouped in two classes: (a) varying the pK_a value of the distal base; (b) varying the number of basic residues having the same pK_a.

a. Effect of the pK_a Value of the Distal Base.

The approximate pK_a value of the acidic forms of pyridine in FeL2 is 13.11, and that of the primary amine in FeL3 is 16.9 in acetonitrile.^62,63 In acetonitrile, the HER by FeL2 and FeL3 is probed in the presence of 5 equiv of [BH⁺]-OTs, where the pK_a value of the conjugate acid is varied from 9.43 to 14.98. Protonation of the ring (i.e., step 1 in Figure 3, right), which has pK_a < 9.43, is not observed in any of these cases for the acid sources having pK_a ≥ 9.43. The Fe^III/II CV for FeL2 moves anodically by 60 mV once the pK_a value of the acid is lowered below the pK_a value of the pendant pyridine (13.11), suggesting protonation of the pendant pyridine by the acid source in the solvent accompanying the reduction. Similarly, the Fe^III/II CV for FeL3 is shifted anodically by 60 mV even when Col-H⁺ is used as the acid source because the pK_a value of the pendant amine in FeL3 (16.9) is higher than that of Col-H⁺ (14.98). The protonation of Fe^II-H⁻ (pK_a > 14.98) resulting in a HER current at potentials below that of the Fe^II/I redox process is observed for all of these acids. The protonation of Fe^I (pK_a > 14.98), as is evident by a slight anodic shift and the irreversible nature of the Fe^II/I redox wave, is observed for all of the acids used here (Figure 6). In these cases, the HER current gradually increases in intensity and shifts anodically as the pK_a value of the acid source is lowered (Figure 6). The electrocatalytic hydrogen evolution by Fe^III-H⁻ protonation remarkably slows in FeTPP when using a weaker acid source like DEA-H⁺ (pK_a = 11.42), but significant catalytic current enhancement occurs in these two catalysts bearing pendant bases. The HER from Col-H⁺ results in a small catalytic current in FeL2 with a pendant pyridine but is clearly observed with the pendant alkyl amine in FeL3. Unlike FeTPP, the HER by Fe^I is observed with Col-H⁺ (pK_a = 14.98) even under high scan rates in FeL2 and FeL3 (Figure S3), suggesting that pK_a of Fe^III-H⁻ produced in these systems is higher than or similar to (due to hydrogen bonding with the pendant group) those of the pendant bases (pyridine in FeL2 and amine for FeL3) in these porphyrins. Similarly, pK_a enhancement by hydrogen bonding was observed in the same system for an axial −OOH ligand.³¹ Thus, the protonation of Fe^III-H⁻, which is completely inhibited in FeTPP, is enabled in the presence of pendant bases from Col-H⁺, allowing HER in FeL2 and FeL3. Note that a blank glassy carbon electrode does not show HER at these potentials with these acid sources in the absence of a catalyst under similar experimental conditions (Figure S7), but rather a HER current is observed at more negative potentials, as previously reported.⁷¹

Figure 6.

CV overlay of 1 mM FeL2, FeL3, and ⁶LFe in an acetonitrile solution for HER with a glassy carbon working electrode and a platinum wire counter electrode using 0.1 M TBAP as the electrolyte at a scan rate of 100 mV/s with respect to a Fc⁺/Fc⁰ couple using four different acid sources: (A) TsOH (pK_a = 8.64) associated with protonation steps 1–4 for all; (B) BA-H⁺ (pK_a = 9.43) associated with protonation steps 2–4 for all; (C) DIPA-H⁺ (pK_a = 10.62) associated with protonation steps 2–4 for all; (D) DEA-H⁺ (pK_a = 11.42) associated with protonation steps 2–4 (Figure 3, right) for FeL2 and FeL3 and processes 2 and 4 for ⁶LFe; (E) Col-H⁺ (pK_a = 14.98) associated with mainly processes 2 and 4 for all of the complexes except for FeL3, where the protonation step 3 is involved.

Both FeL2 (Figure 7A–D) and FeL3 (Figure 8A–D) show a clear saturation of the HER current catalyzed by the Fe^I state using the different acid sources. Controlled potential electrolysis experiments at −1.2 V (vs Ag/AgCl reference electrode) confirm the release of H₂ gas during a headspace gas analysis in gas chromatography (GC; Figure S4). Although the HER current saturates in FeL2 and FeL3 with increasing concentration of [BH⁺] like FeTPP, the saturation and activity is not lost for FeL2 in the presence of 1 M electrolyte (Figure S5). Thus, saturation of the HER current from the Fe^I state in these neutral iron(I) porphyrins having protonated bases does not involve a precatalytic bonding/association equilibrium of Fe^III-H⁻ with the external proton source [BH]⁺ but rather reflects an intramolecular proton transfer from the protonated pendant base to Fe^III-H⁻. Protonation of the pendant base in the presence of these conjugate acids is evident from the shift in the Fe^III/II CV depending on the pK_a value of the acid source and equivalents of the external acid source needed to protonate the pendant base and to reach saturation of the catalytic current at the Fe^II/I potential; unsurprisingly, increase with its pK_a (Figure S6). For example, 3 equiv of BA-H⁺ (pK_a = 9.43) is enough to protonate the pendant pyridine in FeL2 (pK_a = 13.11), while 5 equiv of DEA-H⁺ (pK_a = 11.42) is required to achieve the same (Figure S6).

Figure 7.

CV diagrams of 1 mM FeL2 in an acetonitrile solution with a glassy carbon working electrode using 0.1 M TBAP as the electrolyte for HER using four different acid sources: (A) BA-H⁺; (B) DIPA-H⁺; (C) DEA-H⁺; (D) Col-H⁺. The scan rate is 100 mV/s. Platinum wire is used as the counter electrode. All potentials are reported with respect to a Fc⁺/Fc⁰ couple using ferrocene as an internal standard.

Figure 8.

. CV diagrams of 1 mM FeL3 in an acetonitrile solution with a glassy carbon working electrode using 0.1 M TBAP as the electrolyte for HER using four different acid sources: (A) BA-H⁺; (B) DIPA-H⁺; (C) DEA-H⁺; (D) Col-H⁺. The scan rate is 100 mV/s. Platinum wire is used as the counter electrode. All potentials are reported with respect to a Fc⁺/Fc⁰ couple using ferrocene as an internal standard.

The results presented above clearly demonstrate the advantage of having a pendant base in iron porphyrins during HER, where these bases allow the protonation and/or proton translocation to its Fe^III-H⁻ center, resulting in higher HER currents at the Fe^II/I potential than FeTPP for the same acid source. Given that the pendant base is protonated by external acid sources having pK_a values of less than that of the pendant base and considering the fact that the rate-determining step is H⁺ transfer to the Fe^III-H⁻ species, these results may imply that the pendant bases facilitate the electrcatalytic HER via enhancement of the protonation rate of the Fe^III-H⁻ species produced, and the fact that this HER current becomes independent of the external conjugate acid concentration, further indicates that this protonation step is likely intramolecular. Further insight into this is obtained from the catalytic rates presented in a later section.

b. Effect of the Number of Bases on the HER.

The number of pendant bases was varied from one pyridine to three pyridines for complexes FeL2 and ⁶LFe, respectively (Figure 1). The CV data for the FeL2, FeL3, and ⁶LFe complexes with 1 mM catalyst concentration were recorded in an acetonitrile solution using the same set of acids (pK_a = 8.64 − 14.98). The CV of these catalysts without any acid showed the Fe^III/II, Fe^II/I, and Fe^I/0 redox processes (peak separation of 75 ± 10 mV) having potentials similar to FeTPP, indicating that the incorporation of these bases in the porphyrin ligand does not alter the electronic structure of the iron center. Note that the CV response of the ⁶LFe complex is inherently complicated because of the counteranion dissociation and coordination of one of the free pyridines to the Fe^III and Fe^II states, likely that previously established.^72,73 The large potential shifts and multiple catalytic waves of Fe^III/II and Fe^II/I redox couples in the presence of BA-H⁺ and DIPA-H⁺ can be attributed to the protonation of pyridine (Figure S8). Because there are three pyridines, there is the possibility of multiple protonations in ⁶LFe. This is the case for BA-H⁺ and DIPA-H⁺ (Figure 6, green and red lines) where the shifts in the Fe^III/II and Fe^II/I processes are more than that observed for DEA-H⁺ and Col-H⁺ (Figures 6, blue and yellow lines, and S8). Furthermore, for DEA-H⁺ and Col-H⁺, the shifts become saturated upon using 1 equiv of acid, suggesting that in these two cases only one proton binds and the binding of this proton is more favorable than in FeL2, which has a single pendant pyridine but requires more than 1 equiv of acid that has pK_a > 10, suggesting that the three pyridines in ⁶LFe form a “proton sponge” with a very high first pK_a. For these acid sources, the catalytic current of HER increases because of protonation of the Fe^III-H⁻ species; i.e., the Fe^I state can catalyze HER similar to that in FeL2 and FeL3. The magnitude of the HER current from the Fe^III-H⁻ species is smaller in ⁶LFe relative to FeL2. In fact, while FeL2 shows a catalytic current at Fe^II/I redox potential from DEA-H⁺ (pK_a = 11.42), ⁶LFe having three pyridines does not show such a current enhancement from the Fe^I state and the CV diagram of the Fe^II/I redox wave becomes irreversible owing to the protonation and formation of a Fe^III-H⁻ species (Figure S8). Thus, ⁶LFe, which features a greater number of pendant bases relative to FeL2, shows a weaker catalytic current for HER. This is because the three pyridines together can act as proton sponge, causing the intramolecular protonation of Fe^III-H⁻ to be slow (see DFT Calculations).

Rate Determination Using the Foot-of-the-Wave Analysis (FOWA).

The catalytic rates of HER from the Fe^I state can be determined either by using the FOWA of the kinetic region developed by Costentin and Savéant at which the HER current depends on the substrate concentration (acid)⁷⁴ or by analyzing i_cat/i_p at acid concentrations where the scan-rate-independent HER current plateaus.^75–77 In our cases, the electrochemical processes deviate from the ideal S-shaped behavior, and the catalytic rates are analyzed under purely kinetic conditions at the foot of the wave in the absence of any side phenomena. Here, we have evaluated the reaction rates for the heterolytic reaction pathway (ECCE mechanism) from the slope of the linear fit of i/i_p versus [1 + exp(F/RT)(E − E_cat/2)]⁻¹ (Figures S9–S13) using eq 1 by which the FOWA expression can be generally represented.^78,79

$$i/i_{\text{p}}=2.24\text{n}{\left[(RT/F\nu )k{C_{\text{A}}}^0\right]}^{1/2}/[1+\text{exp}(F/RT)\left(E-E_{\text{cat}/2}\right)]$$ (1)

where n = number of electrons transferred from the electrode, T is the temperature, R is the gas constant, i_p = CV peak current of one electron-transfer process, i = catalytic peak current, k is the second-order rate constant of the rate-determining step, F is the Faraday constant, C_A⁰ is the acid concentration, kC_A⁰ = k_obs is the first-order rate, ν is the scan rate, and E − E_cat/2 is the difference between the applied potential (E) and the potential of the half-peak catalytic current (E_cat/2) in the presence of a substrate.

The first-order rates (kC_A⁰) of HER for all of these complexes are determined at 5 equiv of acid concentration with respect to the catalyst concentration (Table 1), where the pendant bases are protonated (see above). The rate of HER for FeTPP drops from 4.9 ± 0.5 to 1.2 ± 0.3 s⁻¹ when the pK_a value of the acid source is increased from 9.43 in BA-H⁺ to 11.42 in DEA-H⁺. This is due to stronger N–H bonds in acids having higher pK_a, which would result in higher activation energies for the rate-determining protonation step by the acid [BH⁺]-OTs⁻, leading to H₂ release. The rates of HER are, in general, estimated to be higher for iron porphyrins with pendant bases like FeL2 and FeL3 relative to FeTPP for any acid source including Col-H⁺, where the rate for FeTPP is not measurable. Overall, these pendant bases help to accelerate the catalytic current for HER compared to that for FeTPP. For structurally analogous complexes FeL2 and FeL3 (the distance between the proton bearing a base and iron is the same), the rate is always higher in FeL2, which has a lower pK_a (13.11) for the pendant group than does FeL3 as long as the pK_a value of the acid source is less than that of the pendant base; i.e., the pendant base is protonated. This is due to the stronger N–H bond strength of the protonated amine with higher pK_a in FeL3 (pK_a = 16.9), which makes it a weaker proton donor relative to the protonated pyridine in FeL2 (pK_a = 13.11). When Col-H⁺ is used as the acid source, the rate of FeL3 (higher pK_a than Col-H⁺) is greater than that of FeL2 (lower pK_a than Col-H⁺) because the pendant base is protonated in the former and only partially (with ΔpK_a ~ 1.6 of pendant bases Py-H⁺ and Col-H⁺) in the latter. The CV experiments suggest that the rate-determining step for HER is the protonation of Fe^III-H⁻ species. Normally, FOWA yields the rate of the chemical step after the redox event, leading to the formation of an active catalyst (EC), i.e., Fe^I + H⁺ → Fe^III-H⁻. However, if the redox event is followed by two consecutive chemical steps (ECC), FOWA should reflect the rate of the second step if that is lower than the rate of the first chemical step, which is the case here (i.e., Fe^III-H⁻ + H⁺ → Fe^III + H₂).

Table 1.

First-Order Reaction Rate kC_A⁰ (s⁻¹)

acid (pKa)	FeTPP	FeL2 (13.11)	FeL3 (16.9)	6LFe (13.11)
BA-H+ (9.42)	4.9 ± 0.5	10.4 ± 0.6	5.9 ± 0.7	1.3 ± 0.4
DIPA-H+ (10.62)	2.7 ± 0.3	9.4 ± 0.5	7.6 ± 0.4	1.8 ± 0.3
DEA-H+ (11.42)	1.2 ± 0.3	4.9 ± 0.5	3.9 ± 0.3	0.5 ± 0.4
Col-H+ (14.98)	ND	0.8 ± 0.2	1.1 ± 0.4	ND

The relationship between ΔG (Gibbs free energy of the reaction) and ΔG^⧧ (Gibbs free energy of activation) describing the transition state (TS) can be approximated as

$$\Delta G^{⧧}=\alpha \Delta G+C$$ (2)

where α is the extent of conversion from the reactant to the product and C is the constant.

The rate constant (k) and equilibrium constant (K) are related to ΔG^⧧ and ΔG, respectively.

$$k=k_{\text{B}}T/h\text{exp}\left(-\Delta G^{⧧}/RT\right)$$ (3)

$$K=\text{exp}(-\Delta G/RT)$$ (4)

When eqs 3 and 4 are combined in eq 2 and its logarithmic form is taken, the expression becomes

$$\text{ln}k=\alpha \text{ln}K+C^{\prime }$$ (5)

Again, pK_a = −ln K.

$$\text{ln}k=-\alpha \text{p}{K}_{\text{a}}+C^{\prime }$$ (6)

For the set of iron porphyrin complexes used here, FeTPP (Figure 9A, green), FeL2 (Figure 9A, violet), and FeL3 (Figure 9A, red), the log of the rate constant varies linearly with pK_a of the acid sources. The slopes obtained from the linear free energy plots are different for these complexes using the same conjugate acid sources having different pK_a values. The slopes of the lines are 0.35, 0.20, and 0.11 for FeTPP, FeL2, and FeL3, respectively. The slope of the plot of log(k) versus pK_a is indicative of the nature of TS involved in the protonation of a Fe^III-H⁻ species according to the previously established assumption where a smaller slope indicates a reactant-like early TS;⁸⁰ i.e., the extent of proton transfer is less to the Fe^III-H⁻ center. The magnitude of the slope is lowest for FeL3, where pK_a of the pendant base is the highest (intramolecular proton transfer irrespective of the external acid sources used) and FeTPP, without pendant groups (intermolecular proton transfer), exhibits the highest slope (green trace, Figure 9A,B). Although, qualitatively, the generalization of the pK_a dependence of log(rate) seems reasonable, neither FeL2 nor FeL3 shows rates that are independent of pK_a of the external acids to suggest purely intramolecular control of the rate-determining Fe^III-H⁻ protonation step. To limit any association of the external acid with FeL2 and FeL3 akin to FeTPP, the HER rates were obtained for FeL2 (slope 0.20 in Figure 9A) using 1 M TBAP (Figure S13 and Table S1). The resulting log(k) versus pK_a plot for FeL2 approaches linearity (slope 0.01; brown trace, Figure 9B), indicating mostly intramolecular proton transfer to Fe^III-H⁻ in the rate-determining step; i.e., the rate of HER becomes almost independent of pK_a of the external proton source. Density functional theory (DFT) calculations are used to understand proton transfer to Fe^III-H⁻ from the protonated pendant group to obtain insight into the difference in the intramolecular H–H bond formation step between FeL2, FeL3, and ⁶LFe.

Figure 9.

Plot of log(rate) versus pK_a of the external acid source for FeTPP (green), FeL2 (violet), and FeL3 (red) in the presence of 100 mM TBAP (A) and the same for FeTPP (green) and FeL2 (brown) with 100 mM TBAP and 1 M TBAP, respectively (B).

DFT Calculations.

Geometry-optimized DFT calculations are used to investigate the interaction between the protonated distal base and Fe^III-H⁻. Optimized structures show that the Fe^III-H⁻ species in both FeL2 and FeL3 show strong hydrogen-bonding interaction between the protonated distal base and hydride ligand (Figure 10A–B) with H---H distances of 1.39 and 1.30 Å, respectively. This interaction is also referred to as a dihydrogen bonding.^55,81 The energy difference between the hydrogen-bonded structure and the structure where the hydrogen-bonding BH⁺ group is rotated away is 8–17 kcal/mol in the gas phase (Figure S14). A similar energy for dihydrogen bonding was estimated for the active site of the mononuclear active site of mononuclear Fe-hydrogenase.⁸¹ The optimized structure of the hydrogen-bonded Fe^III-H⁻ in a singly protonated ⁶LFe molecule reveals that the dihydrogen-bonding distance is 3.35 Å (Figure 10C) relative to 1.39 Å (Figure 10A) in FeL2. In fact, the proton appears to be nicely lodged in the proton sponge created by the three pyridines. Logically, this dihydrogen-bonding interaction is weak and is unlikely to contribute to proton transfer to the Fe^III-H⁻ species for the catalytic HER. Consistently, ⁶LFe with three pyridines is much slower toward the catalytic activity from the Fe^I state than that of FeL2 with only one pyridine (Table 1).

Figure 10.

DFT-optimized structures of (A) Fe^IIIL2-H⁻, (B) Fe^IIIL3-H⁻, and (C) ⁶LFe^III-H⁻ with protonated distal bases.

SUMMARY

The pK_a values of the different possible protonation events in HER by iron porphyrins are estimated using proton sources with varying pK_a values. Protonation of the Fe^III-H⁻ species by an external proton source is likely to be the rate-determining step of the HER catalyzed by FeTPP. The role of pendant bases in HER are investigated using a series of synthetic iron porphyrins where both the pK_a and number of bases are systematically varied. The results show that the rate-determining step of HER is intramolecular proton transfer from the protonated pendant nitrogen base to a Fe^III-H⁻ species. Having pendant residues covalently attached to the porphyrin facilitates this rate-determining protonation step by providing a local source of proton mimicking the desired roles played by such residues in naturally occurring enzyme active sites. The rate of proton transfer is, however, lower for pendant bases with high pK_a because of stronger N–H bonds, resulting in more reactant like TS. Importantly, increasing the number of bases results in a lowering of the catalytic rate due to a competing proton sponge effect.

EXPERIMENTAL DETAILS

Materials.

All reagents were of the highest grade commercially available and were used without further purification unless mentioned. The reagents for the synthesis and other chemicals like ferrocene, p-toluenesulfonic acid (TsOH), bromoaniline, N,N-diethylaniline, 2,6-diisopropylaniline, collidine, and tetrabutylammonium perchlorate (TBAP, [Bu₄N][ClO₄)]) were purchased from Sigma-Aldrich. The nitrogenous bases were freshly distilled and used after purification whenever it was required, and acetonitrile was dried using the Pure Process Technology solvent system. All of the bases and solvent were degassed using a freeze–pump–thaw technique and stored in an inert-atmosphere glovebox.

Electrochemical Measurement.

All electrochemical experiments were performed using a CH Instruments model CHI710D bipotentiostat electrochemical analyzer. A platinum wire electrode was used as a counter electrode. The measurements were made against a silver wire reference electrode using ferrocene as an internal standard. Anaerobic experiments were performed inside a glovebox under a nitrogen atmosphere. The glassy carbon electrode was used as a working electrode, which was freshly polished to remove all contamination before every use. The polished electrode was then rinsed with water and the solvent used for electrochemical measurements. The electrodes used for homogeneous electrochemistry were purchased from CH Instruments.

Synthesis.

The catalysts (FeL2 and FeL3) were synthesized in our laboratory following an earlier reported procedure.^{31 6}LFe was synthesized in Prof. Karlin’s laboratory following a previous literature report.⁸² These catalysts are pictorially represented in Figure 1.

CV.

Homogeneous CV experiments were carried out in an acetonitrile solution of a 1 mM/0.5 mM catalyst with 100 mM TBAP (supporting electrolyte) in an electrochemical cell. The scan rate was 100 mV/s in most cases unless mentioned otherwise. An internal standard ferrocence was dissolved in the electrolytic solution, and the potentials are reported with respect to a Fc⁺/Fc⁰ couple. A stock solution of ferrocene was prepared, which was added to the solution of the complex before electrochemical measurements. Then two stock solutions of equal strength (250 mM/500 mM) were prepared for the base (B) used and the acid (TsOH). In order to perform the catalysis with different acids, 10 equiv (with respect to the catalyst) of the base to the solution and then increasing equivalents of acid was gradually added up to 10 equiv so that in situ generated conjugate acid [BH⁺]OTs⁻ behaves as the required acid source. The direct reduction of anilinium is sufficiently negative on glassy carbon such that it does not interfere in the analyzed region.

Bulk Electrolysis and Hydrogen Detection by GC.

Hydrogen evolution was confirmed by controlled potential electrolysis at −1.2 V versus Ag/AgCl in acetonitrile at room temperature. Headspace gas analysis was performed by GC fitted with a thermal conductivity detector using helium as the carrier gas in a sealed electrochemical cell containing a 0.5 mM solution of FeL2, 50 equiv of DEAH⁺ as a substrate, a glassy carbon working electrode with a large surface area (1 cm²), a platinum counter electrode, and a Ag/AgCl/1 M KCl reference electrode. Here, we performed electrolysis at a fixed potential for a certain period under the conditions mentioned above and obtained a substantial amount of H₂, which was confirmed by GC.

DFT Calculations.

DFT calculations of the complexes were carried out using the BP86 functional with unrestricted formalism.^83–86 For these complexes, a split basis set (6–311g* on the iron atom and 6–31g* on carbon, hydrogen, oxygen, and nitrogen atoms) was used. The frequencies were calculated on the optimized structures. An energy minimum is confirmed by performing frequency calculations on the fully optimized structure using the same basis set as that used for optimization to ensure no imaginary mode is present. The final energy calculations were performed using the 6–311+g* basis set on all atoms in the PCM model using acetonitrile as a solvent and a convergence criterion of 10⁻¹⁰ hartree.^87,88 The energies reported are the total free energies computed using zero-point energies and entropy (298 K).