Enhanced and selective oxygen reduction by iron porphyrin with a biguanide residue in the second coordination sphere

Abstract

The role of 2nd sphere residues in catalytic processes that require multiple protons and multiple electrons is a growing area of interest. These 2nd sphere effects manifest through hydrogen bonding, proton transfer, and electrostatic interactions, resulting in very selective catalysis. Here, a biguanide functional group, which can form very strong hydrogen bonding with several water molecules, is installed in the 2nd sphere of an iron porphyrin. The resulting porphyrin, Fe–biguanide, selectively reduces O₂ to H₂O via 4e⁻/4H⁺ pathway in both aqueous and organic media, with rates much faster than those reported for any porphyrin to date. In the aqueous medium, the biguanide is protonated, whereas in the organic medium, it exists in neutral form. Several spectroscopic techniques confirmed the presence of hydrogen bonding with water molecules in the 2nd sphere. These results indicate that water molecules, held by 2nd sphere residues, can accelerate the kinetics of 4e⁻/4H⁺ reduction of O₂ as well as any other functional group.

Fe–biguanide catalyzes selective four-electron/four-proton reduction of O₂ to H₂O, exhibiting rates that surpass those reported for any porphyrin catalyst to date. Spectroscopic studies reveal that second-sphere hydrogen-bonding interactions with water significantly enhance O₂ reduction kinetics.

Introduction

Selective reduction of O₂ by 4e⁻/4H⁺ to H₂O is an active area of research, which is being pursued by several groups in the world.^1–5 While material-based catalysts can be robust, molecular catalysts offer an opportunity to introduce nuanced changes in the catalyst design to investigate the reaction mechanism in great detail, resulting in a better understanding of the chemical steps involved in determining the rate and selectivity of O₂ reduction, which helps to develop better catalysts for the purpose.^6–9 The 4e⁻/4H⁺ reduction of O₂ to H₂O can proceed through numerous combinations of electron transfer (ET), proton transfer (PT) and proton-coupled electron transfer (PCET) steps (Scheme 1, exemplified using iron porphyrin), resulting in multiple branching points in the reaction landscape that can be controlled through catalyst design. For example, the reduction of an Fe(iii)–O₂˙⁻ species, formed after O₂ binding to Fe(ii), to an Fe(iii)–OOH species can occur via either PCET or ETPT or hydrogen atom transfer (HAT) pathways. These pathways have all been observed in different synthetic and enzymatic systems and impose distinct requirements on their respective transition states.^10–15 Fundamental understanding of the mechanism of the O₂ reduction process allows better design of catalysts to navigate through this labyrinth of possible routes.

Scheme 1. Mechanistic possibilities of 4e⁻/4H⁺ O₂ reduction. Compound I is better described as an Fe(iv) = O unit bound to a porphyrin radical cation.

Iron porphyrins, analogues of the heme active site of cytochrome c oxidase—an enzyme that reduces O₂ to H₂O in mitochondria—have been investigated as potential O₂ reduction catalysts for several decades now. Elaborate models of cytochrome c oxidase have been developed and investigated by several groups. These models have been very useful in understanding the mechanism of O₂ reduction by the metallo-enzyme.^16–20 However, these models require elaborate synthesis and, thus, there has been a continuing effort to develop simpler iron porphyrins which can reduce O₂ to H₂O selectively. This will require a detailed understanding of the mechanistic landscape and tools to navigate it.

The O₂ reduction is probed electrochemically both in organic solvents and aqueous solutions. In organic solvents, where the porphyrins are readily soluble, acids are used as proton sources. In aqueous solutions, porphyrins are adsorbed on graphitic electrodes or on self-assembled monolayers on Au electrodes.^17,21–25 In organic solvents and under homogeneous electrochemical conditions, the rate-determining step of O₂ reduction is the protonation of the Fe(iii)–O₂˙⁻ species, and in the presence of strong acids, it generally results in 4e⁻/4H⁺ reduction.^2,26,27 In the case of sterically congested iron porphyrins, O₂ binding has been recently reported to be the rate-determining step.²⁸In situ spectro-electrochemistry indicates that in the presence of weak acids, the reduction of O₂ proceeds via an ECEC mechanism where the reduction of the iron to its Fe(ii) state initiates O₂ binding to result in Fe(iii)–O₂⁻, which is then reduced to an Fe(iii)–O₂²⁻ species, which is then protonated, resulting in the cleavage of the O–O bond.^8,15 In the case of simple porphyrins, a linear free energy-based correlation exists between the rate of O₂ reduction and the overpotential of O₂ reduction, in the presence of strong acids. However, this correlation breaks down when 2nd sphere residues that can participate in proton transfer are included in the catalyst.^29,30

In contrast, in situ mechanistic investigations of iron porphyrins have revealed that the rate and selectivity of O₂ reduction under heterogeneous conditions in aqueous medium are determined by the fate of an Fe(iii)–OOH species. The rate of O–O bond cleavage is the rate-determining step, and the site of protonation of this species determines the selectivity of O₂ reduction.³¹ Protonation of the proximal oxygen of the Fe(iii)–OOH species results in the release of H₂O₂, i.e. 2e⁻/2H⁺ O₂ reduction, while protonation of the distal oxygen results in O–O bond cleavage and 4e⁻/4H⁺ reduction.^7,32,33 However, the mechanism of electrochemical O₂ reduction varies substantially depending on the presence of 2nd sphere residues, which can participate in hydrogen bonding with or proton transfer to the active site. Using pendant groups like pyridine, imidazole and amines in the 2nd sphere of iron porphyrins, the rates of O₂ reduction could be enhanced by several orders of magnitude relative to previous reports on elaborate heme/Cu systems while keeping the selectivity for 4e⁻/4H⁺ reduction intact. These residues can hydrogen-bond to the Fe(iii)–OOH intermediate, preventing its hydrolysis, which would otherwise lead to the 2e⁻/2H⁺ reduced product H₂O₂, and can facilitate proton transfer to the distal oxygen, thereby facilitating heterolytic cleavage of the O–O bond.^{20–22,34–38}

An unresolved question is the potential role of solvent H₂O molecules in these reactions. Although in some cases, the 2nd sphere residues have been proposed to “hold” water molecules, their role/effect in catalysis has not been unambiguously established.³⁹ The presence of water clusters near the active site can, in principle, provide the same hydrogen bonding stability to intermediates formed during O₂ reduction and facilitate proton transfer to the active site. However, formation of such clusters near the active site is entropically unfavorable. Investigations show that the –NH termini of metformin, a demethylated biguanide, form strong hydrogen bonds with multiple water molecules in both its neutral and protonated states.^40,41 In this study, an iron porphyrin with a distal biguanide residue is synthesized, and its O₂ reduction is investigated in both organic and aqueous media. The data presented in this study unambiguously demonstrate that, indeed, biguanide residues in the 2nd sphere stabilize a cluster of water molecules, which in turn facilitate proton transfer to the reaction intermediates, resulting in rapid and selective 4e⁻/4H⁺ reduction of O₂ to H₂O in both organic and aqueous solvents.

Results

Synthesis

The pendant biguanide group was introduced into the iron porphyrin by reacting (o-aminophenyl)triphenylporphyrin (H₂MAPP) with cyanoguanidine in the presence of AlCl₃, a Lewis acid, in THF solvent under reflux conditions (Scheme 2). The resulting solution was then neutralized with ammonium hydroxide, yielding pure H₂–biguanide after further workup and column chromatography. Hexagonal purple crystals were obtained by slow evaporation of the ether in DCM (Fig. 1B). H₂–biguanide was crystallized in the C2/c space group. The crystal structure reveals a cluster promoted by intermolecular hydrogen bonding interactions between the polar N–H groups of four H₂–biguanide molecules. These polar biguanide moieties are encapsulated by the planar porphyrin rings, forming a cage-like arrangement (Fig. 1A). To evaluate the effect of the chelating biguanide moiety on the oxygen reduction reaction, a control porphyrin H₂–triazine was synthesized. This involved blocking the biguanide residue of H₂–biguanide by reacting it with ethyl acetate (Scheme 2) in the presence of sodium methoxide as a deprotonating agent, resulting in a 1,3,5-triazine derivative. H₂–triazine was further purified using column chromatography. Both ligands were metalated with FeBr₂ salt and collidine as a base in THF solvent under an inert atmosphere, followed by aerobic workup with 4 N HCl, resulting in the formation of chloride-bound iron complexes in both cases. Characterization data for the ligands and the iron porphyrins (Fig. S1–S11 and Table S1) are provided in the SI.

Scheme 2. Synthetic procedure adapted to access the complexes investigated.

Fig. 1. X-ray crystallographic structures of biguanide. (A) Ellipsoid plot of the cluster showing four biguanide macrocycles with coordinated chloride ligands and associated solvent molecules. (B) Simplified molecular view highlighting the porphyrin core and the appended biguanide functionality. Hydrogen atoms are omitted for clarity.

Oxygen reduction in organic solvent

The electrochemical properties of the Fe–biguanide porphyrin were probed in acetonitrile solvent. Under a nitrogen atmosphere, cyclic voltammetry (CV) revealed a chemically reversible Fe(iii/ii) redox process at −0.65 V versus Fc⁺/Fc⁰ (Fig. 2A, green trace), which is similar to that of other iron tetraphenylporphyrin derivatives.^20,35 Upon the addition of 5% (v/v) water, this couple shifted anodically by 80 mV, appearing at −0.57 V versus Fc⁺/Fc⁰ (Fig. 2A, blue trace). Such an anodic displacement is indicative of microsolvation of the iron site when water is added to the organic solution. This is made possible by the 2nd sphere biguanide residue, which can hydrogen bond with several water molecules and thereby form a water cluster in the vicinity of the iron.³⁹ The presence of a microsolvation environment around the iron is further evidenced by the absence of comparable shifts in Fe(iii/ii) E⁰ in the Fe–triazine control complex (Fig. 2B), in which the biguanide group is blocked and thus incapable of engaging in strong hydrogen bonding with the added water. This result highlights a secondary coordination sphere effect, whereby the distal functionality, biguanide, tunes redox energetics by as much as 80 mV through microsolvation.

Fig. 2. (A) Cyclic voltammograms of Fe–biguanide recorded in CH₃CN under N₂ atmosphere, in the absence and presence of H₂O, and under O₂ atmosphere with added H₂O. (B) Identical measurements for Fe–triazine, which serves as the control complex. (C) Effect of adding a strong acid (HDMF) to Fe–biguanide solutions, under both N₂ and O₂ atmospheres. (D) Results of parallel experiments with Fe–triazine in the presence of HDMF.

When the Fe–biguanide solution was saturated with O₂, the Fe(iii/ii) couple became electrochemically irreversible. This behavior is characteristic of oxygen coordination to Fe(ii) and subsequent electron transfer to generate an Fe(iii)–O₂˙⁻ superoxo intermediate. In the presence of 5% water, a well-defined catalytic wave emerged at −0.80 V versus Fc⁺/Fc⁰, following a precatalytic feature at −0.7 V (Fig. 2A, red trace). The separation between the two processes suggests an EC mechanism, in which the electron transfer to Fe–O₂˙⁻ is followed by a chemical step, specifically, protonation of the Fe–O₂²⁻ (peroxide) species. The Fe–triazine did not display any substantial catalytic current under the same conditions in the presence of added water (Fig. 2B).

Iron porphyrins without 2nd sphere assistance typically require strong acids (e.g., trifluoroacetic acid or perchloric acid) to protonate the Fe(iii)–O₂˙⁻ intermediate in aprotic media. Water is generally too weak a proton source to participate in this step. Previously, Anxolabéhère-Mallart and co-workers demonstrated that in the absence of strong acids, the O₂ reduction proceeds through electrochemical conversion of Fe(iii)–O₂˙⁻ to an Fe(iii)–O₂²⁻ intermediate, with subsequent protonation by weak acids like (CF₃)₃C–OH.¹⁵ Against this backdrop, the observation that an iron porphyrin can catalyze O₂ reduction with water as the sole proton source in an organic solvent is unprecedented. This unique reactivity suggests that the biguanide substituent, which will stay neutral in an organic solvent, not only pre-organizes water molecules around the porphyrin periphery but may also effectively tune the pK_a of the reactants involved, thereby rendering proton transfer to the bound oxygen intermediates through an extended hydrogen-bonding network. Such organization of water molecules could potentially stabilize protonated water clusters or “caged hydronium” motifs resembling Zundel- or Stoyanov-like structures,⁴² which is consistent with the observed reactivity trends. Density functional theory was used to examine the spatial arrangement of the appended biguanide group relative to the iron center and the axial ligand environment. The optimized structure supports the proposed secondary coordination sphere interaction, showing intramolecular hydrogen bonding between the axial Fe–OH ligand and the appended biguanide –NH group, with an H–O⋯H–N distance of 1.79 Å. These results support the feasibility of the proposed hydrogen-bonding interaction (Fig. S12). However, accurate modelling in the presence of explicit water clusters and a hydrogen-bonding network would require extensive computational treatment, which is beyond the scope of the present work.

The catalytic performances of Fe–biguanide and Fe–triazine were benchmarked in the presence of HDMF (protonated DMF) as the proton donor (Fig. 2C and 2D). Under acidic conditions, the Fe(iii/ii) redox process, probed under N₂ atmosphere, shifts to −0.49 V vs. Fc/Fc⁺, which is anodically shifted by 160 mV with respect to its value in the absence of acid. This is likely the result of protonation of the biguanide residue in the 2nd sphere, which results in a positively charged electrostatic potential near the iron. In the presence of HDMF, an O₂ reduction current is observed which peaks at Fe(iii/ii) potential (Fig. 2C). Under acidic conditions, oxygen reduction follows a CE pathway in an organic solvent, in which the Fe(iii)–O₂˙⁻ superoxide intermediate undergoes protonation prior to electron transfer such that the protonation is the rate-determining step and all subsequent steps are relatively rapid.¹¹ As a result, the Fe(iii)–O₂˙⁻ superoxide species is the only intermediate observed under these conditions. The anodic shift of the Fe(iii/ii) reduction potential in the presence of acid, thus, automatically lowers the overpotential of O₂ reduction. The kinetic parameters, including k_obs, TOF_max, and onset values obtained by foot-of-the-wave analysis (FOWA), are summarized in Table 1 (Fig. S13).

Table 1. Comparative catalytic parameters for oxygen reduction by FeTPP, Fe–biguanide, Fe–triazine, Fe-DB, Fe-OB, FeL2, and Fe-TPPMIm in CH₃CN with HDMF and TsOH as the proton source.

2nd Sphere	k obs (102 M−1 s−1)	TOFmax (s−1)	Overpotential (V)	Proton source	Reference
	3.09	2.50	1.38	HDMF	This work
	85.1	68.1	1.14	HDMF	This work
	6.00	4.80	1.27	HDMF	This work
	12.8	10.3	1.08	TsOH	35
	15.4	12.3	1.01	TsOH	35
	7.8	6.25	1.05	TsOH	35
	1.99	1.59	1.20	TsOH	35
	16	13	1.10	TSOH	20

The rate of the oxygen reduction reaction (ORR) catalyzed by Fe–biguanide under these conditions is 8510 M⁻¹ s⁻¹. The rate for FeTPP, which does not have any 2nd sphere residues, under the same conditions is 309 M⁻¹ s⁻¹. The rate for Fe–triazine is 600 M⁻¹ s⁻¹, which is similar to that for FeTPP. The data clearly establish Fe–biguanide as the superior catalyst, having both lower overpotential and significantly higher catalytic turnover relative to FeTPP and Fe–triazine. Importantly, Fe–biguanide reaches TOF values nearly 14-fold greater than that of Fe–triazine at a lower overpotential, illustrating how the second-sphere biguanide residue favorably affects both the kinetics and thermodynamics of ORR (Table 1). In fact, the Fe–biguanide surpasses the kinetic rates of similar iron porphyrins with pyridine (FeL2, Table 1), dibenzylamine (FeDB, Table 1) and imidazole (FeMIm, Table 1). Needless to say, the higher O₂ reduction rate at lower overpotential contradicts the expected trend based on a scaling relationship proposed in the past.³⁰ This is a characteristic of iron porphyrins with proton transfer residues in the 2nd sphere.³⁵

EPR measurements of Fe(iii)–biguanide showed a characteristic high-spin Fe(iii) signal (at g = 6), indicating the absence of µ-oxo dimer formation, as µ-oxo Fe(iii) porphyrins are typically EPR-silent due to antiferromagnetic coupling (Fig. S14A). Additionally, controlled potential electrolysis (CPE) under O₂-saturated conditions showed less than 10% catalyst degradation after 1 h. Both pre- and post-electrolysis UV-visible spectra displayed a Soret band at 416 nm, characteristic of monomeric Fe(iii) porphyrins, whereas µ-oxo species typically exhibit a Soret band at around 410 nm. The high ORR catalytic rate of Fe–biguanide (8510 M⁻¹ s⁻¹) further supports the absence of µ-oxo formation under electrochemical conditions (Fig. S14B and C).

Microsolvation environment of Fe–biguanide

A key question that arises from the reactivity exhibited by Fe–biguanide is how O₂ reduction in organic solvents can proceed using a weak acid source such as water, when all other iron porphyrins require stronger acid sources. Several complementary spectroscopic studies were employed to probe the interaction of the biguanide group in the 2nd sphere with water in the medium. ¹H NMR spectroscopy of the free H₂–biguanide–porphyrin ligand in DMSO-d₆ shows that the ¹H resonance of the dissolved water (750 mM, chosen arbitrarily) is at 3.392 ppm, whereas it is found at 3.355 ppm in DMSO-d₆. The water content in DMSO-d₆ was quantified using the –OMe (9H's) resonance of 10 mM 1,3,5-trimethoxybenzene as an internal standard (Fig. 3). The observed downfield shift of the water signal in the presence of H₂–biguanide supports hydrogen bonding interactions, which weaken the O–H bond. Upon increasing the water content from 2 M to 15 M in neat DMSO-d₆ and in DMSO-d₆ that contains H₂–biguanide, the water resonance shifts progressively downfield (Fig. S15–S18). The shift in DMSO-d₆ arises from enhanced water–water as well as water–DMSO hydrogen-bonding interactions. However, in the presence of H₂–biguanide, the water signal is always shifted ∼0.1 ppm further downfield compared to neat DMSO-d₆, indicating additional interaction between water and the biguanide moiety. To confirm whether this interaction originates from the biguanide unit itself, the experiment was repeated using phenylbiguanide (lacking the porphyrin core), which produced a comparable downfield shift of the water resonance to that observed with H₂–biguanide (Fig. S15–S18).

Fig. 3. ¹H NMR spectra of DMSO-d₆ + 750 mM H₂O (blue spectrum) and H₂–biguanide + 750 mM H₂O (orange spectrum).

The presence of water in the vicinity of the porphyrin can be evidenced more directly by using TCSPC (time-correlated single-photon counting). Local water molecules quench the excited state rapidly, leading to shorter excited-state lifetimes. Fluorescence studies of Zn–biguanide (a fluorescent analogue of Fe–biguanide) revealed that its emission properties are highly sensitive to the presence of water in the medium. In dry DMSO, Zn–biguanide displayed a sharp emission peak at 452 nm with a lifetime of 3.21 ns (Fig. 4A) whereas no emission peak at 452 nm was observed for Zn–triazine (Fig. 4B). Incremental water addition progressively quenched the emission, and at 5% water, the lifetime decreased to 1.54 ns, representing a nearly 50% reduction (Fig. 4C) of the excited-state lifetime. No such effect was observed for Zn–triazine, confirming that the biguanidinium groups are directly responsible (Fig. 4D) for holding water molecules in the vicinity of the active site, as has been indicated by the ¹H NMR results for samples in DMSO-d₆. A satisfactory fit was confirmed when the weighted residuals were randomly distributed around zero (Fig. S19).

Fig. 4. (A) Emission spectra of Zn–biguanide in pure acetonitrile and in the presence of 1% and 5% H₂O, showing progressive quenching of fluorescence intensity with increasing water fraction. (B) Corresponding spectra for Zn–triazine under identical conditions, displaying only minimal changes in intensity upon water addition. (C) Time-resolved fluorescence decay profiles of Zn–biguanide with increasing water content, demonstrating a marked decrease in excited-state lifetime as the H₂O percentage increases, consistent with enhanced non-radiative deactivation. Instrument response function (IRF) represents the temporal response of the detection system to an instantaneous excitation pulse and is used for accurate fluorescence lifetime analysis. (D) Decay profiles for Zn–triazine in acetonitrile with 1% and 5% H₂O, showing only slight modulation of lifetime relative to that of Zn–biguanide.

Heterogeneous O₂ reduction reaction

The Fe–biguanide complex was immobilized on edge-plane graphite (EPG) electrodes and on self-assembled monolayer (SAM)-modified Au electrodes via drop-casting, and their electrochemical responses were recorded in pH 7 phosphate buffer solutions. At this pH, the biguanide residue should be protonated as the pK_a of phenylbiguanide is ∼11.0.^40,43 In the absence of O₂, the complex exhibited an electrochemically reversible Fe(iii/ii) redox couple, with half-wave potentials (E_1/2) of −320 mV vs. Ag/AgCl on octanethiol (C₈SH) SAM and −328 mV on EPG (Fig. 5A). Upon exposure to O₂, the reversible Fe(iii/ii) feature is replaced by a diffusion-limited catalytic wave corresponding to O₂ reduction, with peak potential similar to the Fe(iii/ii) E_1/2 on EPG (Fig. 5A) as well as on C₈SH-SAM (Fig. S20). Rotating disk electrode (RDE) measurements revealed that the catalytic current increased proportionally with rotation rate, indicating mass transport-controlled kinetics (Fig. 5B). Koutecký–Levich (K–L) analysis yielded a linear relationship between 1/I_cat and 1/ω^1/2, with a slope consistent with a four-electron reduction of O₂ to water (Fig. 5C). From the K–L intercept, the second-order rate constant (k_cat) was determined to be 2.5 × 10⁷ M⁻¹ s⁻¹, approximately an order of magnitude higher than that for the Fe–porphyrin bearing triazine (Fig. S21) or guanidine substituents,²² implying that the definite role of the biguanide 2nd sphere functionality in facilitating O₂ reduction extends to aqueous medium as well (Table 2).

Fig. 5. (A) Cyclic voltammograms of Fe–biguanide recorded under N₂ and O₂ atmospheres in pH 7 phosphate buffer on EPG, showing the reversible Fe(iii/ii) couple under N₂ and the appearance of a catalytic wave in O₂. (B) Linear sweep voltammograms with a RDE at various rotation rates, displaying diffusion-limited catalytic currents consistent with multi-electron oxygen reduction. (C) K–L analysis of the RDE data at different potentials, yielding linear fits that confirm first-order dependence on O₂ concentration and establish the number of electrons transferred per O₂ molecule, consistent with a 4e⁻/4H⁺ pathway to H₂O. (D) RRDE experiment showing the disk current corresponding to O₂ reduction and the simultaneous ring current for peroxide detection, indicating high selectivity for complete 4e⁻ reduction with only minor H₂O₂ release.

Table 2. Kinetics and selectivity of oxygen reduction by Fe–biguanide under heterogeneous conditions.

Catalyst		Kinetics of ORR	PROS analysis (%)
		k cat (×106 M−1 s−1)	EPG	C8SH-SAM	C16SH-SAM
Fe–biguanide	H2O	25.02 ± 0.32	5.6	9.0	11.7
	D2O	8.30 ± 0.22	2.4	3.0	4.6
	H/D	3.0	2.3	3.0	2.5

Selectivity for 4e⁻ reduction and PROS analysis

O₂ reduction can proceed via either a 2e⁻/2H⁺ pathway, producing H₂O₂ or a 4e⁻/4H⁺ pathway, producing H₂O. The RDE data indicate that the O₂ reduction involves the delivery of 4e⁻ (Fig. 5C, n = 4). Selectivity was further evaluated using rotating ring-disk electrode (RRDE) experiments, which detect partially reduced oxygen species (PROS) released from the working electrode and oxidized at the surrounding Pt ring (Fig. 5D).⁴⁴ PROS during O₂ reduction in aqueous medium result from the hydrolysis of the Fe(iii)–OOH intermediate.⁴⁵ On EPG, Fe–biguanide generated only ∼6% PROS, corresponding to 94% selectivity for complete 4e⁻/4H⁺ reduction (Table 2), consistent with the RDE data. The negligible release of H₂O₂ indicates stabilization of the Fe(iii)–OOH species against hydrolysis by the 2nd sphere biguanide residue by hydrogen bonding. This stability is tested better under slow ET rates. Au electrodes functionalized with SAMs of varying alkyl chain lengths (C₈SH and hexadecanethiol (C₁₆SH)) were employed. As the chain length increases, electron tunneling slows down substantially, providing more time for partially reduced species to diffuse before full reduction (∼10³ s⁻¹ in C₈SH, ∼4–6 s⁻¹ C₁₆SH and >10⁵ s⁻¹ in EPG).^46–49 Even under these slow ET conditions, PROS formation increased only modestly to 9.0% on C₈SH-SAM and 11.7% on C₁₆SH-SAM (Table 2), indicating that Fe–biguanide maintains high selectivity for complete O₂ reduction across a broad range of ET rates.

Kinetic isotope effect

The rate of O₂ reduction on the EPG electrode decreased to 8.3 × 10⁶ M⁻¹ s⁻¹ in D₂O, yielding a kinetic isotope effect (KIE) of ∼3.0 (Table 2). This indicates that a PT or a PCET is involved in the rate-determining step.²⁷ Additionally, the KIE on PROS indicates that the step responsible for the generation of PROS involves protonation or PCET. Past mechanistic investigations on iron porphyrins using surface-enhanced resonance Raman spectroscopy (SERRS) coupled to rotating disc electrochemistry have indicated that PROS is generated via the protonation of an Fe(iii)–OOH intermediate species (pK_a ∼7) generated during O₂ reduction to release H₂O₂.⁴⁵ Thus, the KIE in PROS generation is consistent with the protonation of Fe(iii)–OOH to release H₂O₂. The KIE on the rate of ORR recorded for a thiolate-bound iron porphyrin without a distal superstructure was 18 owing to the lack of proton transfer pathways to the catalytically active site.²⁷ The low KIE of 3.0 in Fe–biguanide is consistent with a well-organized proton transfer channel during ORR.

Identification of catalytic intermediates

The SERRS-RDE technique was developed to identify the reactive intermediates formed during catalytic ORR activity on heterogeneous electrodes using a home-built set-up combining in situ SERRS and rotating disc electrochemistry.^50,51 Oxygen reduction is a multistep process involving delivery of 4e⁻ and 4H⁺ to O₂ during the catalytic cycle. Under steady-state conditions, different oxygen-bound intermediate species accumulate depending on the rate-determining step of the ORR cycle, i.e., the intermediate with the slowest decay rate accumulates under steady state. If these intermediates can be identified, direct characterization of the key intermediate species that dominate the rate and selectivity of the overall process can be attempted. When excited into the Soret region, iron porphyrin complexes exhibit intense Raman bands characteristic of the metal ligand as well as oxidation and spin states.^52–54 In particular, the v₄ and the v₂ marker bands represent the oxidation state, spin state and coordination number of the iron with high fidelity. The Fe–biguanide complex immobilized over a C₈SH-modified Ag electrode shows the presence of Feⁱⁱⁱ HS only marked by v₄ and v₂ at 1365 cm⁻¹ and 1555 cm⁻¹, respectively (Fig. 6B), when the electrode is held at 0 V in a pH 7 buffer, indicating a high-spin Fe(iii) resting state of the iron porphyrin. In the low-energy region of the SERRS data, a band at 582 cm⁻¹ is observed (Fig. 7A), which loses intensity upon reduction of Feⁱⁱ, which arises from Fe–O stretching of the hydroxide ligand of the Feⁱⁱⁱ–biguanide complex in aqueous solutions.⁵⁵

Fig. 6. (A) SERRS of Fe–biguanide immobilized on C₈SH-modified Ag electrode recorded under air-saturated pH 7 buffer at different applied potentials. (B) Deconvolution of the ν₄ (left) and ν₂ (right) regions reveals the coexistence of Fe(iii) HS and Fe(ii) HS states, along with the emergence of new bands at 1371 cm⁻¹ and 1570 cm⁻¹, assigned to ferryl (Fe(iv) = O) species accumulated under catalytic turnover.

Fig. 7. (A) Low-frequency region of the SERRS-RDE spectra showing the formation of the Fe(iv) = O species, evidenced by a band at 785 cm⁻¹ upon applying potentials from 0 to −0.5 V. (B) The band shifts to 754 cm⁻¹ when ¹⁸O₂ is used instead of ¹⁶O₂, confirming the isotopic substitution of the oxo species.

In the absence of O₂, applying a potential below the Fe^iii/iiE_1/2 results in conversion of the resting high-spin Fe(iii) to high-spin Fe(ii), indicated by a downshift of v₄ and v₂ to 1344 cm⁻¹ and 1543 cm⁻¹, respectively (Fig. S22). Under catalytic conditions in air-saturated buffer, these Feⁱⁱ HS marker bands are replaced by broad bands indicating the presence of a mixture of species (Fig. 6A). The deconvolution of the v₄ and v₂ peaks shows a rise of v₄ and v₂ bands at 1371 cm⁻¹ and 1570 cm⁻¹, respectively, along with residual Feⁱⁱⁱ HS and Feⁱⁱ HS species with ν₄ and ν₂ vibrations at 1362, 1555 cm⁻¹ and 1346, 1539 cm⁻¹, respectively. These high-energy marker bands (1371, 1570 cm⁻¹) indicate the accumulation of ferryl (Fe^iv = O) species during ORR.^21,56 As the potential was lowered from −0.3 V to −0.5 V, these 1370 cm⁻¹ and 1571 cm⁻¹ bands gradually gained intensity, indicating that this species is indeed formed during the course of ORR.

Examining the lower energy region of the SERRS data helps identify metal–oxygen stretching frequencies that can be confirmed using labeled ¹⁸O₂. In air-saturated pH 7 buffer, a new band at 785 cm⁻¹ increases in intensity as the potential of the working electrode is lowered (Fig. 7A). When the buffer is saturated with ¹⁸O₂, this band shifts to 754 cm⁻¹ (Fig. 7B). The observed isotopic shift of 31 cm⁻¹ with ¹⁸O substitution is in good agreement with the calculated value for a diatomic Fe–O oscillator for an Fe^iv = O species, confirming its accumulation on the electrode during catalytic O₂ reduction. This implies that reduction of Fe^iv = O is the rate-determining step in catalytic O₂ reduction. Note that the accumulation of the Fe^iv = O species during catalysis indicates that the reduction of Fe^iv = O to Feⁱⁱⁱ–OH is a PCET process, which is consistent with the KIE observed.

The rate of ORR and its selectivity for 4e⁻/4H⁺ of Fe–biguanide is the highest among all iron porphyrin complexes reported in a pH 7 buffered solution. These porphyrins differ in their rate-determining step, and it is difficult to compare rates for iron porphyrins having different rate-determining steps (Table 3). The only clear comparison can be with Fe-MARG, which has a guanidine group instead of a biguanide group in Fe-biguanide. Both of these show the accumulation of Fe(iv) = O species during catalytic steady state as indicated by the SERRS-RDE data. The guanidine group should also be protonated at pH 7 and, in that way, is similar to biguanide. However, the guanidine groups will not be able to hold on to as many water molecules in the distal cavity as biguanide as the latter has more tethering points for hydrogen bonding with water molecules.⁵⁷ In organic solvents, Fe-MARG does not show ORR in the presence of H₂O in the solvent, unlike Fe–biguanide (Fig. S23). This again reflects the inability of guanidine to stabilize H₂O molecules in the 2nd sphere. This is the reason for more facile ORR in Fe–biguanide relative to Fe-MARG in organic as well as aqueous media.

Table 3. The ORR kinetics and selectivity of Fe–biguanide and some related iron porphyrins in a pH 7 buffered solution. An extensive list is included in Table S2.

2nd sphere	K ORR (×106 M−1 s−1)	PROS	KIE	Rate-determining-step species	Reference
	25.02 ± 0.32	5.6	3	Fe(iv) = O	This work
	18.0	5	—	Fe(iii)–OOH	34
	8.34	2.2	—	Fe(iii)–OOH	20
	22.0	6.1 ± 0.3	10.6	Fe(iii)–O2˙−	35
	7.26 ± 0.3	2.3 ± 0.2	1.9	Fe(iii)–OOH	35

Conclusion

The Fe–biguanide porphyrin represents a rare and mechanistically distinct case where water molecules can mediate the ORR in a non-aqueous solvent. The key enabling factor is the distal biguanide group, which holds water molecules in the 2nd sphere, modulates the redox potential through microsolvation, and enables catalytic O₂ reduction at relatively low overpotentials, with TOFs exceeding those of conventional iron porphyrins by more than an order of magnitude and low H/D KIE. This work further underscores the importance of secondary-sphere effects in molecular electrocatalysis, where substituents can create local solvation environments that fundamentally alter proton availability and catalytic pathways. The results point toward broader opportunities in catalyst design where weak acids, such as water in organic solvents, can be rendered chemically active through tailored second-sphere interactions as demonstrated here.

Experimental section

Materials

All reagents were of the highest grade commercially available. Cyanoguanidine and AlCl₃ were purchased from Sisco Research Laboratories (SRL). 2,4,6-Collidine, FeBr₂, Zn(OAc)₂, chloroform-d₃, tetrabutylammonium perchlorate (TBAP) and ferrocene (Fc) were purchased from Sigma-Aldrich. Sodium chloride and ammonium chloride were purchased from Spectrochem Ltd Unless otherwise mentioned, all reactions were performed at room temperature. Column chromatography was performed with silica gel (mesh size: 100−200) purchased from Spectrochem Ltd Neutral alumina was purchased from Merck. Tetrahydrofuran (THF), ethylacetate and acetonitrile (CH₃CN) were purchased from FINAR Ltd THF was used after drying over Na metal with the addition of benzophenone indicator, and CH₃CN was dried over CaCl₂ and CaH₂.

Instrumental details

All electrochemical experiments were performed using CH instruments (CH 700 E electrochemical analyser). Glassy carbon, platinum, and standard double-junction Ag/AgCl reference electrode (standard double-junction silver/silver chloride filled with 4 M KNO₃ solution) were purchased from Pine Instruments. All the NMR spectra were recorded with a Bruker DPX-300, Bruker DPX-400, or DPX-500 spectrometer at room temperature. The mass spectra were recorded with a QTOF Micro YA263 instrument. Steady-state absorption spectra were measured using an Evolution 201 (ThermoFisher) UV-visible spectrophotometer. The slit width was kept at 2 nm, and the experiments were carried out at room temperature. Fluorescence measurements were carried out using a Fluoromax 4 (HORIBA Scientific) at room temperature. The slit width was fixed at 2 nm. The lifetime of complexes in different conditions was estimated by measuring the emission decays of the samples using the TCSPC setup of Deltaflex (HORIBA). A laser source of 405 nm with a full width at half-maximum (FWHM) of 900 ps was used for the TCSPC experiments. The emission decays were collected at the magic angle polarization and detected using a photomultiplier tube. EzTime software was used to fit the decay traces. All electrochemical experiments were conducted using CH Instruments (model CHI700E, CHI720D Electrochemical Analyzer). The bipotentiostat, reference electrodes, and Teflon plate material evaluating cell (ALS Japan) were purchased from CH Instruments. The RRDE setup from Pine Research Instrumentation (E6 series Change Disk tips with AFE6M rotor) was used to obtain RRDE data. SERRS-RDE data were collected using a Trivista 555 spectrograph (Princeton Instruments) and 413.1 nm excitation from a Kr⁺ laser (Coherent Innova Saber SBRCDBW-K), rotating the electrode at 200 rpm unless noted otherwise. The region of interest in the Raman spectrum was calibrated against naphthalene resonance Raman spectra during each instance of data collection. Single-crystal X-ray data were collected using a Bruker D8VENTURE Microfocus diffractometer equipped with a PHOTON II detector, with Mo Kα radiation (λ = 0.710 73 Å), controlled by the APEX3 software package. 87 Raw data were integrated and corrected for Lorentz and polarization effects using the Bruker APEX III program suite. Absorption corrections were performed using SADABS. Space groups were assigned by analysis of metric symmetry and systematic absences (determined by XPREP) and were further checked by PLATON 88,89 for additional symmetry. All of the structures were solved by direct methods and were refined against all data in the reported 2θ ranges by full-matrix least-squares on F2 with the SHELXL program suite using the OLEX 2 91 interface. Hydrogen atoms at idealized positions were included during the final refinements of each structure. The OLEX 2 interface was used for structure visualization, analysis of bond distances, angles, and drawing ORTEP plots.

Homogeneous electrochemical experiments

Cyclic voltammograms were recorded with a potentiostat (CHI 700 E or CHI720 D) using glassy carbon as a working electrode. All CV data were collected in a sealed electrochemical cell. Five milliliters of 0.5 mM compound solution in ACN/THF (9 : 1) was taken in the presence of 100 mM TBAP as the supporting electrolyte. The glassy carbon electrode was taken as the working electrode, standard double-junction silver/silver chloride as the reference electrode, and Pt as the counter electrode. Ferrocene (Fc) was used as an internal reference, and the potential scale was normalized with respect to the potential of the Fc^+/0 couple. The solution was deaerated by bubbling N₂ or O₂ before recording the CV measurements as required. All the electrochemical experiments were performed at room temperature.

Heterogeneous electrochemical experiments

All electrochemical experiments were conducted in pH 7 buffer (unless otherwise mentioned) containing 100 mM Na₂HPO₄·2H₂O and 100 mM KPF₆ (supporting electrolyte), using a Pt wire as the counter electrode and a double-layer Ag/AgCl electrode as the reference electrode. All electrochemical data were plotted following the polarographic convention. A 3-neck glass cell fitted with PTFE stoppers from Pine Research was used for all electrochemical experiments unless otherwise mentioned. Separation of the WE compartment from the CE compartment did not result in any difference in the data output. All experiments were conducted at 25 °C. Applied voltages were corrected for IR compensation by running the IR compensation test in the electrochemical workstation. For anaerobic experiments, the electrolyte was purged with UHP-grade argon gas at a 10 mL min⁻¹ flow rate. The starting point of the scan was fixed at 0.2 V, and the initial direction of the scan was cathodic. Electrochemical stability and purity of the catalyst at the electrode surface during electrochemical experiments were verified by taking CV for several cycles, where no visible changes were encountered in the Fe^iii/ii potential and current density.

Attachment of catalysts on SAM-modified Ag and Au electrodes

The SAM-modified Au and roughened Ag disks were taken out of the depositing solutions, rinsed with ethanol and a copious amount of water, and then mounted on the RRDE setup. The solutions of the catalysts were prepared in chloroform, followed by drop-casting onto the electrode for 30 minutes to ensure complete loading. After each respective loading, the surfaces were thoroughly rinsed with chloroform, ethanol, and triply distilled water before carrying out electrochemical or SERRS/SERRS-RDE experiments.

UV-visible absorption data collection

Absorption spectra of the pure complexes were collected in ACN : THF (9 : 1) (2 mL; 5 µM). The background was corrected before the experiments, using an identical amount of solvent mixture.

¹H NMR data collection

¹H NMR spectra were recorded with a 300 MHz NMR spectrometer. A 60 mM solution of the ligand was prepared in 400 µL of DMSO-d₆ containing a 10 mM internal standard, followed by the addition of H₂O. Two scans were collected to obtain spectra with clarity, and the change in chemical shift value of the 2H's of H₂O was monitored after each addition.

Fluorescence data collection

The samples were excited with the respective excitation wavelength as obtained from absorption data. The emission spectra of the pure complexes were collected in ACN : THF (9 : 1) (2 mL; 5 µM).

TCSPC data collection

The lifetime experiments of the pure complexes were conducted in ACN : THF (9 : 1) (2 mL; 5 µM). The decay traces were fitted using EzTime software and the quality of the fitting was judged by weighted residuals (close to 1). All the traces were described by a single exponential function. The quality of fitting was evaluated from the weighted residual distribution and the autocorrelation function of the weighted residuals. A satisfactory fit was confirmed when the weighted residuals were randomly distributed around zero, the autocorrelation function became close to the zero line, and the reduced chi-square value was close to unity.

Synthetic details

H₂MAPP (1)

MAPP was synthesized following the reported protocol from Collman and co-workers.⁵⁸ ESI-MS (positive-ion mode, CH₂Cl₂): 630.2630 (100%, [M + H]⁺) (Fig. S1). ¹H NMR (CDCl₃): δ (ppm) −2.66 (bs, 2H), 3.36 (s, 2H), 6.8–8.3 (m, 19H), 8.6–9.1 (m, 8H) (Figure S2).

H₂–biguanide (2)

MAPP (1) (1 eq., 400 mg, 0.634 mmol) was dissolved in 40 mL of dry THF. To this solution, dicyandiamide (1.5 eq., 80 mg, 0.952 mmol) and AlCl₃ (2 eq., 170 mg, 1.268 mmol) were added under reflux conditions while maintaining an inert atmosphere. After 1 day, 4 eq. of dicyanodiamide and 3 eq. of AlCl₃ were added, the reaction mixture was refluxed for another 2 days. The organic layer in DCM was washed with saturated NH₄Cl and NaCl solution, collected and dried over anhydrous Na₂SO₄. The solvent was evaporated under reduced pressure and was purified through column chromatography using neutral alumina and 30% methanol-DCM solvent mixture. Hexagonal purple crystals were obtained by slow vapour diffusion of the ether in DCM (Fig. 1b). Yield (410 mg, 90.5%). ¹H NMR (300 MHz, DMSO-d₆) δ 8.99 (d, J = 4.8 Hz, 2H), 8.89–8.74 (m, 6H), 8.34–8.19 (m, 6H), 7.96 (dd, J = 7.5, 1.5 Hz, 1H), 7.82 (dq, J = 9.5, 5.3 Hz, 9H), 7.76–7.65 (m, 1H), 7.42 (d, J = 8.1 Hz, 1H), 7.31 (t, J = 7.4 Hz, 1H), 5.10 (s, 2H), −2.80 (s, 2H) (Fig. S3). ¹³C NMR (101 MHz, DMSO-d₆) {¹H} δ 158.55, 156.87, 141.91, 141.85, 137.13, 135.43, 135.10, 134.75, 134.63, 130.80, 129.98, 128.92, 128.47, 128.24, 127.44, 127.17, 122.95, 120.04, 119.99 (Fig. S4). ESI-MS (positive ion mode, CH₂Cl₂): 714 ([M]⁺) (Fig. S5). UV-visible (THF λ/nm = 414, 512, 546, 587, 647) (Fig. S6).

Fe–biguanide (3)

H₂–biguanide (2) (200 mg, 0.280 mmol, 1 eq.) was dissolved in 20 mL of dry, degassed THF. To this solution, 2,4,6-collidine (74 µL, 0.560 mmol, 2 eq.) was added to deprotonate the pyrrolic protons, followed by the addition of FeBr₂ (240 mg, 1.12 mmol, 4 eq.). The resulting mixture was stirred under an inert atmosphere for 48 hours. After completion, the reaction was quenched with water. The organic layer in DCM was washed with brine solution and 4 N HCl, collected and dried over anhydrous Na₂SO₄. The solvent was evaporated under reduced pressure. The product was purified through column chromatography using neutral alumina and 40% methanol-DCM solvent mixture. Yield (235 mg, 91%). Anal. Calcd for C₄₆H₃₄ClFe N₉·2CH₃OH·CH₂Cl₂: C = 61.81%, H = 4.55%, Cl = 11.17%, Fe = 5.87%, N = 13.24%, O = 3.36%. Found: C = 60.62%, H = 4.88%, Cl = 11.58%, Fe = 5.67%, N = 13.75%, O = 3.24%. ESI-MS (positive ion mode, CH₂Cl): 768 ([M]⁺) (Fig. S7). UV-visible (THF λ/nm = 415, 510, 583, 645, 690) (Fig. S6).

Zn–biguanide (4)

H₂–biguanide (2) (1 eq., 100 mg, 0.140 mmol) was dissolved in 15 mL of dry THF. To this solution, Zn(OAc)₂ (4 eq., 123 mg, 0.560 mg) was added and stirred for 2 hours. No external base was added as the counter anion of Zn(ii) salt used, i.e. acetate, can deprotonate the pyrrolic protons. The organic layer in DCM was washed with a brine solution, collected and dried over anhydrous Na₂SO₄. The solvent was evaporated under reduced pressure. The product was purified through column chromatography using neutral alumina and 30% methanol-DCM solvent mixture. Yield (115 mg, 90%). ESI-MS (positive ion mode, CH₂Cl₂): 776 ([M + H]⁺) (Fig. S8).

H₂–triazine (5)

H₂–biguanide (2) (1 eq., 100 mg, 0.140 mmol) was dissolved in 5 mL methanol and 20 mL ethylacetate. To it NaOMe (10 eq., 76 mg, 1.40 mmol) dissolved in methanol was added and kept in refluxing condition for 2 days. The organic layer in DCM was washed with NH₄Cl and NaCl, collected and dried over anhydrous Na₂SO₄. The solvent was evaporated under reduced pressure and was purified through column chromatography using a mesh size of 100–200 silica gel and 3% methanol-DCM solvent mixture. Yield (120 mg, 89%). ESI-MS (positive ion mode, CH₂Cl₂): 738 ([M]⁺) (Fig. S9). UV-visible (THF λ/nm = 414, 511, 546, 587, 647) (Fig. S6).

Fe–triazine (6)

H₂–triazine (5) (100 mg, 0.136 mmol, 1 eq.) was dissolved in 20 mL of dry, degassed THF. To this solution, 2,4,6-collidine (36 µL, 0.271 mmol, 2 eq.) was added to deprotonate the pyrrolic protons, followed by the addition of FeBr₂ (160 mg, 0.544 mmol, 4 eq.). The resulting mixture was stirred under an inert atmosphere for 48 hours. After completion, the reaction was quenched with water. The organic layer in DCM was washed with brine solution and 4 N HCl, collected and dried over anhydrous Na₂SO₄. The solvent was evaporated under reduced pressure. The product was purified through column chromatography using mesh size 100–200 silica gel and 10% methanol-DCM solvent mixture. Yield (110 mg, 87.3%). Anal. Calcd for C₄₈H₃₅ClFeN₉·CH₃OH·2CH₂Cl₂: C = 59.41%, H = 4.20%, Cl = 17.19%, Fe = 5.42%, N = 12.23%, O = 1.55%. Found: C = 60.01%, H = 4.35%, Cl = 16.86%, Fe = 5.51%, N = 12.74%, O = 1.74%. ESI-MS (positive ion mode, CH₂Cl₂): 791 ([M]⁺) (Fig. S10). UV-visible (THF λ/nm = 415, 510, 583, 647, 691) (Fig. S6).

Zn–triazine (7)

H₂–biguanide (2) (1 eq., 50 mg, 0.063 mmol) was dissolved in 15 mL of dry THF. To this solution, Zn(OAc)₂ (4 eq., 55 mg, 0.252 mg) was added and stirred for 2 hours. No external base was added as the counter anion of Zn(ii) salt used, i.e. acetate, can deprotonate the pyrrolic protons. The organic layer in DCM was washed with a brine solution, collected and dried over anhydrous Na₂SO₄. The solvent was evaporated under reduced pressure. The product was purified through column chromatography using mesh size 100–200 silica gel and 5% methanol-DCM solvent mixture. Yield (55 mg, 88.14%). ESI-MS (positive ion mode, CH₂Cl₂): 800 ([M + H]⁺) (Fig. S11).

Author contributions

Suman Patra: conceptualization, investigation, methodology, formal analysis, data curation, visualization, writing. Soumili Ghosh: investigation, methodology, formal analysis, data curation, visualization, writing. Soumya Samanta: methodology, data curation, writing. Nisha Maurya: investigation, methodology, data curation. Somnath Das: methodology. Abhishek Dey: conceptualization, funding acquisition, project administration, resources, supervision, and writing.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

CCDC 2540588 contains the supplementary crystallographic data for this paper.⁵⁹

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information: additional UV-visible, HRMS, ¹H-NMR, electrochemistry data and rR data. See DOI: https://doi.org/10.1039/d6sc03084j.