Cobalt-based Metalloporphyrins As Efficient Electro-catalysts for Hydrogen Evolution From Acetic Acid and Water

Four molecular electrocatalysts based on cobalt complexes, CoT(X)PP (X = H (1), OH (2), CN (3), COOH (4)), were prepared from meso-tetra-p-X-phenylporphin (H2T(X)PP, X = H, OH, CN, COOH) by reaction with cobalt acetate to be used for electrolytic proton or water reduction. The electrochemical properties and the corresponding catalytic activities of these four catalysts were investigated by cyclic voltammetry. Controlled potential electrolysis with gas chromatography analysis confirmed that the turn-over frequencies (TOF) per mol of catalyst per hour were 42.4, 38.6, 55.5, and 70.1 mol H2 at an overpotential of 941.6 mV (in DMF) in the acetic acid solution containing catalyst. In neutral buffered aqueous solution (pH 7.0), these four molecular catalysts had TOF per mol of catalyst per hour of 352.53, 313.7, 473.4, and 714.6 mol H2, respectively, with an overpotential of 837.6 mV, indicating that complex 4 had better activity than complexes 1, 2, and 3. The Faraday efficiencies of complexes 1-4 were 99.1, 99.6, 100.4, and 99.0% at 72 h of consecutive reduction on a glassy carbon electrode, respectively. These results indicate that the electronic properties of the ligands play a crucial role in determining the catalytic activity of the cobalt complex and are consistent with the phenomenon that the catalytic activity of the benzene porphyrins is significantly increased in the presence of electron-withdrawing groups, and the CoT(COOH)PP is the most active catalyst.


Introduction
Finding alternative clean and renewable energy sources is crucial given the diminishing fossil fuels and the significant environmental pollution issues [1][2][3][4]. Due to its abundant supply, lack of pollution, and increased calorific value, hydrogen has drawn a lot of attention. Currently, the electro-and photocatalytic separation of water into hydrogen and oxygen is the focus of the study [5][6][7]. However, the development of effective catalysts with reduced overpotential, superior stability, and a high rate of turnover for the water reduction reaction is the primary problem for water decomposition [8,9]. Consequently, a substantial amount of investigational work has been dedicated to the exploitation of efficient catalysts of transition metal complexes based on cobalt [10,11,20], molybdenum [12], manganese [13][14][15][16], copper [17][18][19], nickel [20][21][22][23][24] and iron [25,26] for the reduction of protons or water to form H 2 . Macrocyclic complexes of cobalt have been broadly studied and reported as electrocatalysts for hydrogen production since the seminal research work of Sutin [27][28][29]. Cobalt porphyrins were also demonstrated to be a catalyst with high Faraday efficiency in homogeneous solutions [30][31][32][33]. Spiro and colleagues were the first to report that water-soluble cationic cobalt porphyrins could generate hydrogen by electrocatalysis [30], but this report displayed that hydrogen release occurred only at a very low turnover rate and high-potential acidity. Subsequently, a large number of related studies were reported. Nevertheless, the majority of reports utilized acid or organic solvent buffers in water to deliver the proton source. In this paper, we present an efficient bimolecular porphyrin cobalt catalyst for the reduction of water to hydrogen in phosphate buffer solutions. It possesses higher Faraday efficiency, turnover, and reversal rates compared to the first reported porphyrins [33]. Whereas, the donor type and electronic properties of the ligands perform a crucial role in determining the structure and reactivity of the corresponding metal complexes [27,34,35]. We are continuously studying different substituted ligands and their complexes with cobalt [36][37][38][39]. Our group has already investigated the electrocatalytic hydrogen production of two cobalt(II) complexes [40], designated CoT(4-Cl)PP and CoT(4-OMe)PP. In this context, we proceeded to investigate the electrocatalytic properties of the subsequent four cobalt complexes, CoT(X)PP (X = H (1), OH (2), CN (3), COOH (4)), and the implications of the four tetravalent ligand modifications on the catalytic properties of this four cobalt (II) complexes (Scheme 1).

Physical Measurements
Infrared spectra in the range of 400-4000 cm −1 were recorded on a Nicolet Avatar 360 FT-IR spectrophotometer with samples in the form of KBr pellets. The absorption spectra of the samples in DMF solution were measured with a Shimadzu UV-2550 double-beam spectrophotometer in the vertical incident light range of 200-800 nm. The ESI-MS spectra of the four complexes were measured on an AB SCIEX, API3200.

Electrochemical Methods
Cyclic voltammograms (CVs) were as well obtained on a CHI-660E electrochemical analyzer under N 2 using a threeelectrode cell with a glassy carbon electrode as the operating electrode, saturated Ag/AgNO 3 or Ag/AgCl electrode as the reference electrode, as well as a platinum wire as the auxiliary electrode. In organic media, a ferrocene/ferrocene (1+) couple was used as an internal standard and 0.10 M [(n-Bu) 4 N]ClO 4 was utilized as a supporting electrolyte. Controlled potential electrolysis (CPE) in aqueous media was undertaken using a hermetically sealed glass double chamber cell separated by a glass frit. The working compartment was instrumented with a glassy carbon plate and an Ag/ AgCl reference electrode. The auxiliary compartment was installed with a platinum gauze electrode. The working compartment and the auxiliary compartment contained 40 mL of phosphate buffer. The CVs were documented after the addition of cobalt complexes. After electrolysis, 0.50 mL of the headspace sample was withdrawn and exchanged with 0.50 mL of CH 4 . The headspace sample was injected into a gas chromatograph (GC). The GC experiments were executed using an Agilent Technologies 7890A gas chromatograph.

Characterizations
In the FT-IR spectra of complexes 1-4 (Fig. S1), the N−H stretching vibrational bands of 1 and 3 are located at 3309 and 3313 cm −1 , respectively, while the N−H stretching vibrational bands of 2 and 4 are covered by their −OH at 3430 cm −1 and −COOH stretching vibrational bands at 3436 cm −1 , respectively. The N−H bending vibrational bands were located at 966, 962, 966, and 960 cm −1. whereas when the free base porphyrins are converted to cobalt (II) complexes, the vibrational bands of N−H disappear and a new vibrational band appearance near 1000 cm −1 , which is the characteristic band of N−Co bond, signifying that the hydrogen atoms on the nitrogen atoms are substituted by metal ions to form metalloporphyrins [43].
The UV-Vis spectra of complexes 1-4 dissolved in DMF are demonstrated in Fig. S2. The intense absorption peak near 412-421 nm belongs to the Soret band and the weak absorption at 500-700 nm to the Q band, which accounts for the π→π* leap in the porphyrin ring (a 2u -e g * leap and a 1u -e g *). While, the four characteristic peaks near 515, 550, 590 and 648 nm in the Q-band of the free base porphyrins were reduced to one peak near 530 nm, confirming the formation of the cobalt(II) complex [43].
The ESI-MS spectra of complexes 1-4 in CH 3 OH solution exhibited in Fig

Cyclic Voltammetry Studies
To explore the electrochemical properties of four molecular catalysts, CVs were measured in DMF with 0.10 M [(n-Bu) 4 Fig. 1a, the CV of 0.89 mM complex 1 exhibits two reversible peaks at −0.85 V and −1.97 V verse Ag/AgNO 3 , respectively. For comparison, CV of H 2 T(4-H)PP was measured in a similar condition, H 2 T(4-H)PP exhibits one reversible peak at −0.78 V verse Ag/AgNO 3 . The first reduction wave of 1 at −0.85 V, being close to the first reduction of the free ligand, is assigned to H 2 T(4-H)PP. The second redox peak can be assigned to Co II /Co I couple. From Fig. 1b, complex 2 also shows two reversible peaks at −0.85 V and −2.07 V verse Ag/AgNO 3 , respectively. H 2 T(4-OH)PP exhibits one reversible peak at −0.86 V verse Ag/AgNO 3 . The first reduction wave of 2 at −0.85 V, being close to the first reduction of the free ligand, is assigned to H 2 T(4-OH)PP. The second redox peak can be assigned to Co II /Co I couple. From Fig. 1c, complex 3 also shows two reversible peaks at −0.78 V and −1.90 V verse Ag/ AgNO 3 , respectively. H 2 T(4-CN)PP exhibits one reversible peak at −0.91 V verse Ag/AgNO 3 . The first reduction wave of 3 at −0.78 V, being close to the first reduction of the free ligand, is assigned to H 2 T(4-CN)PP. The second redox peak can be assigned to Co II /Co I couple. From Fig. 1d, complex 4 also shows two reversible peaks at −1.34 V and −2.05 V verse Ag/AgNO 3 , respectively. H 2 T(4-COOH)PP exhibits one reversible peak at −1.04 V verse Ag/AgNO 3 . The first reduction wave of 4 at −1.34 V, being close to the first reduction of the free ligand, is assigned to H 2 T(4-COOH)PP. The second redox peak can be assigned to Co II /Co I couple. The Potential of Co II /Co I is shifted by 100, 70, and 80 mV correspondingly since H at the para-position of the benzene ring is superseded by hydroxyl, cyano, and carboxyl groups, respectively.

Catalytic Hydrogen Evolution From Acetic Acid in DMF
The performance of hydrogen evolution for acetic acid in DMF solution was investigated by cyclic voltammetry test, as depicted in Fig. 2a-e. As depicted in Fig. 2a, the significant voltammetric current emerges at −1.99 V with the addition of different concentrations of acetic acid (from 0.0 to 5.56 mM) and significantly increases with the proton concentration. The results indicated that 1 electrocatalytic hydrogen evolution required the reduction of Co(II) to Co(I) and the protonation reaction. Interestingly, the catalytic wave onset of Ag/AgNO 3 remained almost constant at about −1.75 V as the acetic acid concentration proceeded to increase from 0.0 to 5.56 mM (Fig. 2a). complexe 2 increased significantly with the change of acetic acid concentration (from 0.00 to 6.30 mM) around −2.04 V, and the initiation point of the catalytic peak remained almost around −2.04 V compared to Ag/AgNO 3 (Fig. 2b). Along with the change in acetic acid concentration (from 0.00 to 7.78 mM), complexe 3 exhibited a significant increase around −1.82 V, and the catalytic peak onset remained almost at −1.49 V compared to Ag/AgNO 3 (Fig. 2c). A systematic peak also emerges at −2.03 V with varying acetic acid concentration (from 0.00 to 6.30 mM) compared to the previous complexes of complex 4 (Fig. 2d). Following the above observations, only the Co II /Co I couple was engaged in the proton reduction. As the acetic acid concentration increased, the Co II / Co I reduction peak current gradually enhanced. All the complexes exhibited a significant positive shift in the starting reduction potential, diminishing the energy required for electron transfer.
In accordance with the existing literature [44][45][46] and the mentioned analytical work, the potential mechanism of acid-mediated catalytic cycling for hydrogen production is described in Scheme 3. The single-electron reduction of [Co (II) T(X)PP] provides a [Co (I) T(X)PP] − species. The addition of protons produces Co (III) -H, a potentially highly reactive intermediate [43]. Further reduction of Co (III) -H generates H 2 reconstitutes the initiating complex [CoT(X)PP]. To develop a further assessment of the catalytic performance of complexes 1-4, bulk electrolysis was carried out in DMF solution using glassy carbon plate electrodes in acetic acid at mutually variable applicative potentials. Fig. 3a-d illustrate the aggregate overall electrolytic charge of complexes 1-4 in the existence of acid. While the applied potential was −1.45 V, the maximum charges reached 57.8 mC, 45.8 mC, 52.3 mC and 71.8 mC in 2 min. The CPE experiment without complex 1-4 at the same potential has a charge of a mere 8.7 mC (Fig. 4), indicating that these four complexes in the presence of such conditions are effective in hydrogen production indeed. Assuming that each of the catalyst molecules is exclusively distributed on the electrode surface and that each electron is used for the reduction of protons, complexes 1, 2, 3 and 4 provide 42.4, 38.6, 55.5 and 70.1 moles of hydrogen (Eqs. S1-S4) per mole of catalyst in the 941.6 mV OP in accordance with equations (1) [47] and (2) [48]. It is remarkably that complexes 1, 3 and 4 are significantly more active than complex 2, which is in agreement with the apparently higher catalytic activity in the presence of the electronwithdrawing group relative to the activity of the electrondonating group. In the presence of hydrogen, hydroxyl or carboxyl radicals on the porphyrin benzene ring, mesotetra-p-carboxyphenylporphyrin is a superior active catalyst. Increasing the Co II /Co I redox potential, the decrease in the hydrogen-producing activity of the cobalt complexes.
(1) TOF = ΔC∕(F × n 1 × n 2 × t) Overpotential = Applied potential − E ⊙ Note that ΔC is the charge from the catalyst solution during CPE minus the charge from solution without catalyst during CPE, F is Faraday's constant, n 1 is the number of moles of electrons required to generate one mole of H 2 , n 2 is the number of moles of catalyst in solution, and t is the duration of electrolysis.

Catalytic Hydrogen Evolution From Aqueous Buffer
We conducted further investigations on the electrochemical properties of these four complexes in buffered aqueous solutions at pH 3.5-7.0, the pH range associated with catalytic water reduction. As indicated in Fig. 5a, in the absence of complex 1, the catalytic current is not obvious until the potential is up to −1.50 V versus Ag/AgCl. On the addition of complex 1, the catalytic current emerges at around −1.37 V relative to Ag/AgCl, indicating that the addition of complex 1 can reduce the potential. Furthermore, the current intensity increased significantly as the concentration of complex 1 increased from 0.72 to 1.67 μM, where the peak current is up to 23.6 μA and the initial reduction potential of Ag/AgCl is positively shifted by 130 mV. It is apparent from Fig. 5b that the initiation of the catalytic current is influenced by the pH of the solution, and the imposed potential decreases with increasing pH, demonstrating the involvement of protons in the initial stage of electrochemical catalysis.
Among the same conditions, as illustrated in Figs. 6a, S4a, and S5a, while no catalyst 2-4 was an addition, the catalytic current was not evident until the potential reached −1.50 V for Ag/AgCl. When catalyst 2-4 was added, the catalytic current increased significantly, demonstrating that the addition of complex 2-4 can diminish the potential as well. As the concentration of complex 2-4 increased from 0.07 to 1.55 μM, 0.05 to 0.86 μM, and 0.45 to 3.02 μM, respectively, the current intensity was significantly increased and the peak currents were up to 14.3, 24.6, and 18.7 μA, as the initial reduction potentials of Ag/AgCl were positively shifted by 144, 128, and 168 mV, respectively. From Figs. 6b, S4b, and S5b, it is clear that The starting potential of the catalytic current is regulated by the pH of the solution, and the applied potential decreases with increasing pH, which confirms the involvement of protons in the initial stage of electrochemical catalysis.
The results indicated that the addition of the complex can reduce the potential current when the pH is determined, as well as the current intensity increases with the concentration, and there is a saturation concentration. At a constant catalyst concentration, along with the decrease of pH, the proton in the system increases, and the peak current gradually increases, with a positive shift of the starting reduction potential amount.
To investigate further the electrocatalytic activity of complexes 1-4 in aqueous media, the bulk electrolysis of complexes 1-4 with an aqueous solution of 0.25 M buffer solution (pH 7.0) was evaluated at variable potentials using a glassy carbon electrode in a double-chamber cell. As can be observed in Fig. S6e, in the absence of complex 1, the maximum charge for 2 min of electrolysis was  (Fig. S6a), accompanied by the appearance of a large number of bubbles, which was confirmed by gas chromatography as H 2 (Fig. S7a). An electrolysis time of 1 h yielded about 3.62 mL of H 2 with a Faraday efficiency of 99.13% of H 2 (Fig. S7b). Following Eqs. (1) and (3) [49], we calculate the TOF of the catalyst as 352.53 mol of hydrogen per mol of catalyst per hour at a maximum overpotential of 836.7 mV (Eq. S5), where As it can be seen in Fig. S6b, at an applied potential of −1.45 Vs Ag/AgCl, the maximum charge generated by electrolysis for 2 min verse was 258.1 mC, while the generated gas was verified by gas chromatography to be hydrogen (Fig. S8a), and the electrolysis cycle for 1 h produced 3.24 mL of hydrogen with a Faraday efficiency of 99.6% (Fig. S8b).  Fig. S6c, the maximum charge production for 2 min of electrolysis was 521.2 mC when the applied potential was −1.45 V versus Ag/AgCl, while the generated gas was identified by gas chromatography as hydrogen (Fig. S9a), the electrolytic cycle for 1 h generated 5.11 mL of hydrogen with a Faraday efficiency of 100.4% (Fig. S9b). Based on Eqs. (1) and (3), complex 3 yields a TOF of 473.4 mol of hydrogen per mol of catalyst per hour and an OP of 837.6 mV (Eq. S7). It is noted from Fig. S6d that the maximum charge generated by electrolysis for 2 min was 741.7 mC when the applied potential was −1.45 V versus Ag/AgCl, whereas the generated gas was identified by gas chromatography as hydrogen gas, as observed in Fig. S10a. In the 1 h electrolysis cycle 8.56 mL of hydrogen was produced with a Faraday efficiency of 99.0% (Fig. S10b). Under Eqs.
(1) and (3), the TOF of complex 4 is 714.6 mol of hydrogen per mol of catalyst per hour and the OP is 837.6 mV (Eq. S8). It is clear from the above experimental data that the complexes examined in this paper have excellent catalytic properties. The catalytic performance of complex 4 was significantly superior to the other three complexes, which may be due to the enhanced electron absorption ability of complex 4 than the tetradentate ligand of complex 3. The catalytic activity of complex 1 was higher than that of complex 2, which may be since the tetradentate ligand of complex 2 is an electron-donating group, and the electron-donating group decreased the activity of the complex. In these results, it is suggested that the presence of ligands is a crucial structural feature of proton or water reduction catalysis and has a substantial influential effect on its catalytic performance.
It is well known that the catalytic efficiency of different electrocatalytic systems varies with different experimental conditions, which makes it difficult to compare the hydrogen evolution performance of the reported electrocatalytic systems based on cobalt porphyrins under the same conditions. Our group has studied the catalytic properties of cobalt complexes CoT(4-Cl)PP and CoT(4-OMe)PP. On this basis, we have carried out a more systematic study, under the same test conditions to compare, the results show that: 4 has higher activity than the reported tetra-p-Cl cobalt porphyrin complex (TOF = 632.9 in neutral phosphate buffer solution) [40]. In a neutral phosphate buffer solution, the TOF of complex 4 was equal to 714.6 mol, and the applied potential was 1.45 V versus Ag/AgCl ( Table 1). As presented in Fig. S11, the four catalysts had a powerful charge accumulation capacity in 0.25 M buffer solution (pH = 7) under the action of −1.40 V against Ag/AgCl with no remarkable loss of activity in 72 h.

Conclusions
We have successfully synthesized four electrocatalysts based on meso-tetra-p-X-phenylporphin cobalt complexes. The potentially higher the Co (II)/(I) redox potential the quieter the hydrogen production activity of the cobalt complexes. Electrochemical investigations demonstrated that the four complexes could generate hydrogen not only from acetic acid but also in buffer solutions without acetic acid. The electrochemical studies indicated that the four complexes had excellent catalytic properties for hydrogen production, and their catalytic activities (TOF) were in descending order: 4 (714.6) → 3 (473.4) → 1 (352.5) → 2 (313.7). The results are concordant with the results that the catalytic activity is significantly enhanced with the presence of electron-withdrawing groups on the phenol counterpart of the ligand.

Conflict of Interests
The authors declare no competing interests. Ref [40] Ref [40] This work  This work  This work  This work