Biomass or bio-mess: tackling reproducibility in biomass-derived carbon electrocatalysts

Abstract

The pyrolysis of biomass is a promising route toward carbon electrodes, but its adoption in electrocatalysis is mostly limited to the oxygen reduction reaction. To prepare precise active sites for other electrocatalytic reactions in the oxygen, nitrogen, and carbon cycles, the complexity of both the biomass precursor and the pyrolysis process must be reigned in. We now report a two-stage strategy for stabilizing the synthesis of a reproducible electrocatalyst for a reaction requiring a precise active site, namely the hydrazine oxidation reaction. The strategy starts with a common yet variable biomass (ground coffee waste) and proceeds through (1) optimized activation by a range of methods, and (2) scalable introduction of Fe–N₄ centers. The result is a sustainable, highly active, and most importantly, reproducible, Fe–N–C electrocatalyst. This work should help the scientific and technological communities to realize the full potential of biomass as a source for carbon electrodes.

---

Introduction

Electrocatalysis is key to many global challenges related to energy conversion and storage, enabling a speedy exchange of electrical and chemical energy.¹ Carbon-based materials have emerged as electrocatalyst supports and electrode materials in a broad range of devices, such as fuel cells,^2,3 batteries,⁴ and supercapacitors.⁵ Carbon electrocatalysts offer high surface area,⁶ electronic conductivity,⁷ tuneable porous structures,⁸ and easy doping by heteroatoms such as N, O, S, B, or P.⁹ Most carbons can be prepared at remarkably low cost, compared to platinum group metal (PGM) catalysts,^10,11 since they can be made by pyrolysis and graphitization of biomass – a sustainable, near-infinitely abundant source of organic matter, ranging from spirulina algae¹² to human hair.¹³ But although biomass is an all-too-common carbon precursor in the scientific literature, it comes with a catch: reproducibility.¹⁴

The pyrolysis of biomass is a complex process applied to a complex starting material, creating an inherent challenge for electrocatalysis, where precise active sites are desired. First, any biomass is naturally doped with many elements, and is already pre-structured hierarchically into cells and tissues. Second, these elements and structures all scramble upon pyrolysis, yielding extremely complex 3D materials with many sites with catalytic potential. Even seemingly well-defined biomass such as rice can vary drastically, depending on biomass age,¹⁵ soil and climate of growth,¹⁶ and even between different organs of the plant.¹⁷ Thus, biomass-derived carbons are more often used as electrodes where such differences are not crucial. These include supercapacitors, where surface area is the chief (if not only) consideration;^18,19 or hard-carbon battery anodes, where high-temperature graphitization near-erases the difference in composition and structure.²⁰ In electrocatalysis, biomass precursors are most popular for the oxygen reduction reaction (ORR),^21–24 where many dopants (e.g., N, P, B) can all contribute to the activity.²⁵ Thus, in such reactions, a precise active site structure is not strictly necessary for good (if not outstanding) activity. In other electrocatalytic reactions, such as NO_x reduction,^26,27 N₂ reduction,^28,29 CO₂ reduction,^30,31 NH₃ oxidation,³² and N₂H₄ oxidation,^33,34 biomass precursors did not yield outstanding carbon electrocatalysts. Thus, the vast potential of biomass as a cheap carbon source remains insufficiently explored for most electrocatalytic reactions, where precise active sites are needed.

We now report a route towards reproducible multi-doped electrocatalysts for a key reaction in the nitrogen cycle, derived from a common biomass waste (ground coffee waste). We focus on bridging the gap between precursor complexity and active-site precision, by constructing the material in two steps: (1) a combined pyrolysis-activation, aimed at a limited homogenization of the doping and structure; and (2) doping with well-defined Fe–N₄ sites, recently discovered in our group on a model support³⁵ and here implemented firstly on a real-world, high surface area carbon. The procedure yields reproducible materials with outstanding electrocatalytic activity, demonstrated both on a lab scale and in a fuel cell.

As the model reaction we choose the hydrazine oxidation reaction (HzOR: N₂H₄ + 4OH⁻ ⇌ N₂ + 4H₂O + 4e⁻, E° = −0.33 V vs. RHE at pH 14).³⁶ Hydrazine hydrate (N₂H₄·H₂O) is an energy-dense alternative to hydrogen.³⁷ Direct hydrazine fuel cells (DHFCs) operate at room temperature, consuming liquid fuel and producing only clean nitrogen and water with no CO₂ (unlike alcohols).³⁸ They offer a theoretical electromotive force of 1.56 V, much higher than the 1.23 V for H₂ oxidation.³⁹ Moreover, hydrazine can be used as a sacrificial fuel for H₂ evolution.^40–42 However, hydrazine is more expensive than hydrogen or ethanol,⁴³ as are the electrocatalysts for its oxidation. Currently, the best available HzOR electrocatalysts are based on PGMs,⁴⁴ while more abundant transition metals like Ni and Co present geopolitical challenges regarding mining and extraction.⁴⁵

Iron- and nitrogen-co-doped carbons (FeNCs) are promising electrocatalysts for the HzOR.^36,46 They are based solely on earth-abundant elements, as Fe is the fourth most abundant element on earth. FeNCs can be prepared by joint pyrolysis of Fe, N, and C precursors,^47,48 by combined self-templating and transmetalation,^49–51 or by deposition of Fe-N_x sites on a carbon support.^52,53 The exact structure of the catalytic site is critical for its HzOR activity. For example, carbon edges⁵⁴ and N-heterodopants⁵⁵ are modestly active catalytic sites, but the introduction of a Fe–N_x moiety boosts activity dramatically.⁵⁵ The binding configuration of the atomically dispersed Fe–N_x moiety determines activity and selectivity,³⁵ as does nano-structuring of the electrocatalyst for optimal mass transfer.⁵⁶ Thus, there is a built-in gap between the variability and complexity of biomass, and the precise requirements of an HzOR electrocatalyst.

Experimental

Synthesis and activation of biomass-derived carbons

The materials were synthesized from waste coffee grounds collected from the Schulich Faculty of Chemistry cafeteria at Technion, Haifa, Israel. The coffee beans manufacturer is Nescafe, Nestle, Portugal, and the beans are a blend of Arabica and Robusta, medium roasting level. Nescafe states their coffee is sourced mainly from Brazil, Vietnam, Colombia, Indonesia, and Honduras.⁵⁷ Prior to pyrolysis, the biomass was dried at 50 °C for 72 h and ground using a planetary ball mill (Pulverisette 6, Fritsch). The biomass was chemically activated based on a procedure of Mullins et al.⁵⁸ at different pyrolysis temperatures: 700, 800, and 900 °C and using different potassium salts, based on materials used by Antonietti et al.,⁵⁹ namely KOH (Bio-Lab), K₂CO₃ (Merck Chemicals), and KH₂PO₄ (Spectrum Chemicals). All chemicals were used without further purification. Each salt was dissolved in 13.35 mL of DI water to obtain 1 M and stirred with 1 g of waste coffee grounds for 3 h at room temperature. The mixture was left at 120 °C overnight in a QSR box furnace in air. Upon evaporation, the obtained powder was manually ground and pyrolyzed in the tube furnace (Carbolite CTF) under argon atmosphere at either 700, 800, or 900 °C, at a heating rate of 3 °C min⁻¹ and dwell time of 1 h. In all cases, the resulting carbon powders were soaked in 150 mL of 1 M HCl overnight and then washed with 4 L of DI water until reaching neutral pH. The powders were dried at 50 °C in air. The resulting N-doped carbons are named according to the following convention: NC_coffee-A-T, where A is the activating agent (none/KOH/K₂CO₃/KH₂PO₄) and T is the pyrolysis temperature (700/800/900 °C).

Doping with FeN₄ centers to yield FeNC electrocatalysts

The twelve NC_coffee carbons were impregnated by atomically dispersed iron sites according to the ‘universal ligand’ method of Yang et al.,⁶⁰ where 13.8 mg of iron(II) acetate (95%, Acros Organics) and 100.3 mg of 1,10-phenanthroline monohydrate (99+%, Alfa Aesar) were mixed in ethanol (abs., Bio-Lab) for 20 min. The solution was mixed with 69.6 mg of carbon and left to evaporate at 70 °C. The black solid was heated under argon atmosphere to 70 °C/1 h, and then to 600 °C/2 h (heating rates 10 °C min⁻¹). The resulting FeNC materials are named according to the following convention: FeNC_coffee-A-T, where A is the activating agent (none/KOH/K₂CO₃/KH₂PO₄) and T is the pyrolysis temperature (700/800/900 °C).

Material characterization

High-resolution scanning electron microscopy (HR-SEM) was performed on a Zeiss-ultra+ at 1 keV and with an in-lens detector. N₂ adsorption–desorption isotherms were measured on a Micromeritics 3Flex instrument at 77 K, using vacuum-dried samples. The isotherms were analyzed using the two-parameter Brunauer–Emmett–Teller (BET) model for specific surface area (SSA), at P/P⁰ values of 0.01–0.15. Raman spectroscopy was performed on two devices: eight of the FeNC_coffee samples, pyrolyzed at 700 °C and 800 °C, were analyzed in a Horiba LabRam HR Evolution Raman microscope using a ×50 lens, 532 nm laser excitation wavelength, and 1800 grating and the remaining four FeNC_coffee samples, pyrolyzed at 900 °C, were analyzed in a WITec alpha300 Raman microscope using a ×100 lens, 532 nm laser excitation wavelength, and 300 grating. First-order Raman spectra were fitted iteratively with four Lorentzian components. The degree of graphitization is correlated to the intensity ratio between the D and G Raman peaks (I_D/I_G).⁶¹ Elemental analysis by inductively coupled plasma mass spectroscopy (ICP-MS) was performed by Mikrolab Kolbe (Germany). High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was performed on FEI Titan Cubed Themis G2 60-300, at an acceleration voltage of 200 kV. X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo Scientific ESCALAB QXi instrument. X-ray diffraction (XRD) measurements were performed on a Rikagu SmartLab 9 kW instrument. The fuel cell test was performed using a Greenlight fuel cell test station.

HzOR electrocatalysis

Electrocatalyst inks were prepared by sonicating a mixture of 1.0 mg carbon electrocatalyst powder, 0.50 mL dimethylformamide, and 20 μL Nafion® 5 wt% dispersion (Alfa Aesar) for 30 min. The ink was dropcast (10 μL) on a polished glassy carbon electrode (5 mm diameter) and dried at 50 °C. The catalyst loading was 0.02 mg or 0.1 mg cm⁻². Electrochemical measurements were performed by using a Biologic VSP multichannel potentiostat in 1 M KOH. A reversible hydrogen electrode (RHE, Pine) was used as a reference electrode, and a graphite rod as a counter electrode. Before measurements, the electrolyte was purged for 30 min by argon gas at 25.0 ± 0.1 °C and flowed continuously above the solution during the experiments. Cyclic voltammetry (CV) was recorded from 0.15 V to 0.8 V vs. RHE, at a scan rate of 10 mV s⁻¹. A positive-feedback automatic iR correction of 85% was used, with an alternating voltage amplitude of 10 mV. Before measurements, the electrode was cycled in the non-faradaic region to improve wetting (0.4 V to 0.7 V vs. RHE, 100 mV s⁻¹, 0.4 V to 0.7 V vs. RHE, 10 mV s⁻¹). For the hydrazine oxidation reaction, 20 mM hydrazine hydrate (99+%, Fisher Chemical) in 1 M KOH solution was used. For the cyanide poisoning tests, 20 mM KCN (97%, Alfa Aesar) were added to the electrolyte. For fuel cell testing, the coffee-derived catalysts were used as anodes (loadings of 2 and 4 mg cm⁻²), and CoFe catalysts as cathodes^22,62 (1 mg cm⁻² loading), separated by a Piperion (20 μm-thick) anion-exchange membrane. Gas diffusion electrodes were prepared in a manner according to procedures reported in previous publications.^63–65 Cell assembly and operating procedures were described in detail elsewhere.^66,67

Results and discussion

To create a uniform carbon scaffold, we examined several chemical activation processes. Such activations, whether physical or chemical, help control the pore formation and structural evolution of the high-variability biomass-derived carbons. In chemical activation, the precursor is impregnated with an activating solution; during heat treatment, the biomass undergoes the partially overlapping processes of pyrolysis, carbonization, and pore etching. Known activating agents include alkaline salts such as KOH,⁶⁸ acidic salts such as H₃PO₄,⁶⁹ or others such as ZnCl₂.^70,71 To gain control and understanding of the process, we employed three alkaline salts with a common counterion, namely KOH, K₂CO₃, and KH₂PO₄, and scanned the temperature range of 700–900 °C, where their reactions are meaningful.

The morphology variations were monitored using HR-SEM, with a clear difference between the non-activated controls and the chemically activated materials (Fig. 1). In the absence of an activating agent, the pyrolyzed carbons are chunky, irregular, and with glass-like fracture patterns, while the chemically activated carbons display relatively regular meso- and macroporosity. Pore formation is clearly enhanced by higher pyrolysis temperatures, with maximum porosity at 900 °C regardless of the activating agent. Samples activated by KH₂PO₄ show little porosity when heated at 700 °C or 800 °C, and significant mesoporosity at 900 °C (Fig. 1, NC_coffee-KH₂PO₄-900, inset). Activation with KOH yielded larger pores, around 300–800 nm in diameter in NC_coffee-KOH-900. With K₂CO₃ as an activating agent, pores across a range of diameters are observed (tens to hundreds of nanometres), especially in NC_coffee-K₂CO₃-900. In addition to the specific types of pores introduced, it is clear from the micrographs that chemical activation helps gain structural control over the complex and variable waste biomass precursor.

Fig. 1 Characteristic high-resolution scanning electron microscopy (HR-SEM) micrographs of the 12 NC_coffee samples. Different morphologies and porosity between the activated samples and the non-treated samples are visible. Inset: magnified view of NC_coffee-K₂CO₃-900 and NC_coffee-KH₂PO₄-900.

The qualitative structural information was complemented with quantitative analysis of the microporosity by N₂ gas sorption porosimetry (Fig. 2). The isotherms of the non-treated carbons are of type III, corresponding to a non-porous carbon, while the isotherms of biomass activated by KOH and K₂CO₃ are type IV for all three pyrolysis temperatures, indicating hierarchical porosity, with micro- and mesopores at different extents. In all carbons at low activation temperatures (700 °C), the only pores are micropores, as there is no evidence yet of the hysteresis and sloping that indicate mesoporosity (they could even be classified as type II). At pyrolysis temperatures above 800 °C, hysteresis becomes more pronounced and micropores grow. In fact, micropore activation by KH₂PO₄ begins only at 900 °C. The KOH-activated and K₂CO₃-activated carbons contain smaller micro- and mesopores while the KH₂PO₄-activated samples show bigger macropores according to the SEM micrographs (Fig. 1). Ultimately, all activated samples show hierarchical porosity, with the K₂CO₃ activation achieving the highest surface areas, at the lowest pyrolysis temperatures.

Fig. 2 N₂ adsorption–desorption isotherms of the 12 NC_coffee samples were measured on a Micromeritics 3Flex instrument at 77 K, using vacuum-dried samples.

The specific surface areas (SSAs) of the materials were calculated from the N₂ sorption isotherms using the Brunauer–Emmett–Teller model (Fig. 3). The non-treated carbons have ultra-low SSAs (<1 m² g⁻¹), as correlated with the type III isotherm. For activated samples, the trend is increasing SSA with increasing pyrolysis temperature, which is especially pronounced for K₂CO₃ (reaching 1800 m² g⁻¹ at 900 °C, after being near constant at 700 °C and 800 °C) and even more for KH₂PO₄: the 900 °C sample exhibiting a much higher surface area of over 1000 m² g⁻¹ compared to the samples at 700 °C and 800 °C, exhibiting a very low surface area (<70 m² g⁻¹).

Fig. 3 Brunauer–Emmett–Teller specific surface areas of the 12 NC_coffee carbons at 77 K, after vacuum-drying.

The trends in activation temperatures are repeated in the electrochemical surface area (ECSA) measurements (Fig. 4), performed after Fe impregnation (FeNC_coffee). The ECSAs were calculated from the double layer capacitance (C_dl), measured by voltammetric cycling at different scan rates (5, 10, 20, 50, 100 and 250 mV s⁻¹) in the non-faradaic potential window (0.4 to 0.5 V vs. RHE). The C_dl was divided by 40 mF cm⁻² typical surface capacitance to obtain the ECSA.⁷² The non-treated samples have a smaller ECSA compared to the chemically activated samples, while an increase in ECSA is seen for the FeNC_coffee-K₂CO₃ samples. In FeNC_coffee-KH₂PO₄, the 700 °C and 800 °C carbons show both low ECSAs and BET-SSAs while the 900 °C exhibit a high ECSA and BET-SSA.

Fig. 4 Electrochemical surface area measurements of the 12 FeNC_coffee, measured between 0.4 to 0.5 V vs. RHE, at scan rates between 5 and 250 mV s⁻¹.

The dramatic boost in SSA and formation of micro- and mesopores in the KH₂PO₄-activated carbons can be attributed to the highly endothermic reaction that occurs around 900 °C, which involves the decomposition of H₃PO₄ to P₂O₅ (eqn (2)) which can then oxidize the carbon (eqn (3)). Through a carbothermal reduction reaction, the etching of the carbon framework leads to the creation of more pores.^73,74 As for KOH, activation initially involves the melting of the KOH at 406 °C and reaction with the volatile material of the char produced from the biomass, producing H₂ and K₂CO₃ (eqn (4)) and further promoting porosity formation in the carbon. Chemical activation with K₂CO₃ occurs through reaction with surface N-motifs⁷⁵ at higher temperatures (>600 °C); and through the decomposition to K₂O and CO₂ (eqn (5)). The K₂O melts around 900 °C under N₂ and oxidizes the carbon,⁷⁶ while the hot CO₂ gas etches the pyrolysis char, broadening the pores (>700 °C).⁷⁷ Other carbonates have shown similar effects.^58,71 Furthermore, this set of redox reactions might also explain why the KOH and the K₂CO₃ chemically-activated carbon samples exhibit similar physical characteristics such as type IV isotherms, PSD and BET-SSA, as they produce similar chemical etching reactivity throughout the pyrolysis process. Overall, the SSA, ECSA, and SEM micrographs confirm that chemical activation succeeded in not only increasing the surface areas of the carbons but also generating hierarchical constant porous structures.

KH₂PO₄ + H₂O → H₃PO₄ + K⁺ + OH⁻ (1)

2H₃PO₄ → P₂O₅ + 3H₂O (2)

P₂O₅ + 5C → 2P + 5CO (3)

6KOH + 2C → 2K + 3H₂ + 2K₂CO₃ (4)

K₂CO₃ → K₂O + CO₂ (5)

Beyond porosity, graphitization is a key structural component of electrocatalyst activity, as graphitic domains conduct electrons from active sites to the external circuit. To gauge graphitization, Raman spectroscopy was used to characterize all twelve NC_coffee carbons (Fig. 5), and the peaks were deconvoluted.⁷⁸ The degree of graphitization is correlated to the intensity ratio between the disordered (D) and graphitic (G) bands (I_D/I_G, Fig. 6a), from which the characteristic length of graphitic domains in the a direction (L_a) can be calculated.⁶¹ All carbons exhibit with I_D/I_G > 1 and L_a between 9.7 and 17.6 nm. At 700 °C, only K₂CO₃ and KH₂PO₄ promote sufficient graphitization, lowering the I_D/I_G ratio significantly, while at 800 °C All four carbons present a similar degree of graphitization, with and without activating agents. Graphitization further increases with activation temperature, especially between 800 °C to 900 °C. This indicates the natural growth of graphitic domains.^79,80 The enhanced graphitization by K₂CO₃ and KH₂PO₄ indicates that the improved carbothermal reduction, due to melting of K₂CO₃ and decomposition of KH₂PO₄, is not only key for pore formation (and surface area), but also for promoting graphitization (and ultimately, electrical conductivity) in the carbon electrocatalysts.

Fig. 5 Raman spectra of the NC_coffee carbons, after chemical activations at different temperatures. An example for a deconvolution is presented in the NC_coffee-K₂CO₃-900 sample.

Fig. 6 a) Raman intensity ratios between deconvoluted D and G bands for the NC_coffee carbons. (b) Inductively coupled plasma mass spectroscopy results for the 12 FeNC_coffee materials, showing the atomic ratio of Fe : N_x.

To determine the elemental composition of the catalysts, we used the following sampling strategy for inductively coupled plasma mass spectrometry (ICP-MS) analysis of the FeNC_coffee carbons (Fig. 6b). First, FeNC_coffee-KH₂PO₄-800 and FeNC_coffee-800 were tested to represent activated and non-activated carbons (respectively) for a survey analysis of C, H, N, and Fe (the most common elements in Fe–N–C catalysts), as well as S and P from the biomass,⁸¹ and K from the chemical activating agent. Based on the low content of H, S, and K in these samples, we did not test for these elements in the other carbons, nor for P when not using KH₂PO₄. The nitrogen content varies between 5–11 wt% across the carbons (Table S1†), which is typically more than enough for electrocatalysis, and for anchoring Fe sites. As the activation temperature increases, the nitrogen content either rises (KOH), falls (KH₂PO₄), or varies inconsistently (K₂CO₃ and non-treated). The iron content is >2 wt% for all activated samples, generally independent of pyrolysis temperature. The samples may also contain some Si and O from the quartz tube, as seen in XRD (Fig. 7). In addition, some of the samples contain a minute amount of iron oxide. The fact that it wasn't washed out by the acid wash, suggests that it is well encased in graphitic shells, and is assumed not to contribute to electrocatalysis.⁵⁵ Finally, the carbons show two broad peaks corresponding to amorphous carbon around 2θ = 24°, 44°.⁸²

Fig. 7 X-Ray diffractograms of 8 FeNC_coffee samples, smoothed. Peaks assigned to SiO₂ (JCPDS 01-076-0936), graphite (JCPDS 01-075-1621) and Fe₃O₄ (JCPDS 01-080-6402) are marked.

The electrocatalytic performance of the FeNC_coffee carbons was tested for hydrazine oxidation in 1 M KOH at 25 °C, normalized for either geometrical surface area (Fig. 8) or electrochemical surface area (ECSA, Fig. S1†). All catalysts have shown good activity, with onset potentials below 0.40 V vs. RHE, while some were outstanding: FeNC_coffee-K₂CO₃-800 oxidized hydrazine at E_onset = 0.26 V vs. RHE, earlier than nearly all Fe–N–C electrocatalysts of the HzOR.³⁶ The electrochemical performance can be analyzed along two aspects: by comparing different activation methods at the same pyrolysis temperature (Fig. 8a), or by monitoring the effect of temperature within each activating procedure (Fig. 8b). At 900 °C the K₂CO₃-activated carbon exhibited the earliest onset potential at 0.3 V vs. RHE and current around 1.4 mA cm⁻². Other electrocatalysts exhibited late onset potential beyond 0.3 V vs. RHE and currents below 1 mA cm⁻². For all activating agents, pyrolysis temperatures above 800 °C were needed to achieve currents above 1 mA cm⁻². Pyrolyzing the precursor at 900 °C gave the highest current densities and the earliest onset potentials, <0.3 V vs. RHE. Carbons activated by K₂CO₃ reached optimal activity already at 800 °C, while a higher temperature was needed for optimal activation with KH₂PO₄ or KOH. Moreover, K₂CO₃ activation yielded the highest current densities and earliest onset potentials than all other activating agents, let alone the non-treated samples.

Fig. 8 Cyclic voltammograms of hydrazine oxidation on FeNC_coffee catalysts, normalized to the geometrical electrode surface area (0.196 cm²), and arranged by (a) pyrolysis temperature, or (b) activating agent. Scan rate 10 mV s⁻¹, 20 mM N₂H₄, 1 M KOH.

Half-wave potentials (E_1/2), at which the current equals half of the limiting current,^83,84 are identical for all 4 samples at 700 °C (Fig. 9). The changes appear at 800 °C, with E_1/2 equals 0.3 V vs. RHE for the K₂CO₃ sample and at 900 °C, where both K₂CO₃ and KH₂PO₄ exhibit low E_1/2 of 0.36 V vs. RHE and 0.38 V vs. RHE, respectively. Although hydroxide ions have been found to compete with substrate binding (e.g. O₂) in alkaline media, this has not been studied yet in hydrazine oxidation;^85,86 our findings suggest that the reaction proceeds well on these materials. Overall, the onset potential of most of the chemically-activated carbon electrocatalysts correlated with the onset potential reported in the literature for hydrazine oxidation in alkaline media, at around 0.3 V vs. RHE.³⁶ All of the activity trends remain the same when correcting the samples for different electrochemical surface areas (Fig. S1†), rather than geometrical electrode areas (Fig. 8).

Fig. 9 Half-wave potentials for the HzOR on the FeNC_coffee catalysts (20 mM N₂H₄ in 1 M KOH).

The correlation between HzOR catalytic activity, and the efficiency of chemical activation (in terms of microporosity and SSA) suggests that micropores help accommodate the Fe–N₄ single-atom active sites in Fe–N–C biomass-derived carbon catalysts.⁸⁷ In contrast, electrocatalytic activity is not correlated with Fe content, which may reach saturation of the active Fe–N₄ sites, with the remainder going to inactive or encapsulated Fe-based inorganic particles.⁵⁵

Reproducibility study

Following an optimization of the waste coffee activation and pyrolysis process, and in order to establish the reproducibility of the full synthetic strategy of activation and active site embedding catalyst, we chose FeNC_coffee-K₂CO₃-800 as the model material, since activation with K₂CO₃ yielded catalysts with the earliest HzOR onset potentials and the highest current densities. Hence, we repeated the synthesis of this material 6 more times (dubbed S1–S6). The resulting catalysts had identical overall morphologies, with similar hierarchical porosity in the meso- and macropore range (Fig. 13). Their average BET-SSA is 470 ± 120 m² g⁻¹, and they all exhibit type IV isotherms (Fig. S3†). The iron content of the samples is similarly reproducible, with an average value of 2.4 ± 0.3% wt.

The electrochemical tests confirm the reproducibility of the catalyst synthesis, with one outlier (Fig. 14a). The E_onset and E_1/2 are highly repeatable (0.27 ± 0.01 and 0.41 ± 0.03 V vs. RHE, respectively) and the peak current density varies between 0.9–1.35 mA cm⁻². The outlier (S3) has a slightly belated onset potential of 0.32 V vs. RHE, and a lower peak current of 0.6 mA cm⁻²; we hypothesize that inhomogeneity during the activation step is the reason for the lower SSA of this sample, and the resulting lower efficiency of Fe embedding. The reproducibility of the electrode preparation was further examined by depositing the FeNC_coffee-K₂CO₃-800 ink on 3 different electrodes (Fig. 14b). The electrochemical reproducibility is thus further validated, with identical onset potentials of 0.25 V vs. RHE and highly similar peak current densities of 1.6 mA cm⁻². To conclude, the reproducibility of the synthetic strategy, combining activation and active site embedding, was confirmed by the example of FeNC_coffee-K₂CO₃-800 catalysts. These materials showed highly similar onset potentials, half-wave potentials, and peak current densities, arising from similar physical parameters such as hierarchical porosity and surface area.

Following the optimization process of the catalysts, samples S1, S2, S4–S6 were mixed to create a sample named FeNC_coffee-K₂CO₃-800-mixed. The catalyst's iron and nitrogen oxidation states, as well as the nitrogen chemical environments were determined using X-ray photoelectron spectroscopy (XPS) (Fig. 10 and S2†). Deconvoluted XPS in the N 1s region, showing pyridinic (N_pyri), metal-bound (Fe–N), pyrrolic (N_pyrro) and graphitic (N_g) nitrogens at 398.7 eV, 399.5 eV, 400.6 eV and 401.8 eV, respectively.^88–90 In the Fe 2p region, deconvoluted XPS shows Fe²⁺ and Fe³⁺ oxidation states, which are typically hard to assign to specific bonds.^91–93 The high dispersion of Fe–N_x sites was supported by high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), combined with elemental mapping (Fig. 11(a–d)). Iron sites, from atomic to clusters (0.5–2 nm) are well dispersed in several regions of the material. While the presence of atomic Fe–N_x sites is thus corroborated, the exact contribution of the larger clusters (which could be direct or by modulation⁹⁴) is out of the scope of this paper. To further test the contribution of atomic sites to the activity, we performed cyanide testing, known to poison such sites (Fig. 12).^95,96 Indeed, the CN⁻-poisoned catalyst shows a dramatically lower current density, testifying to the existence of Fe–N_x sites.

Fig. 10 Deconvoluted XPS of FeNC_coffee-K₂CO₃-800-mixed catalyst in the (a) Fe 2p region and (b) N 1s region.

Fig. 11 (a–c) HAADF-STEM micrographs of the FeNC_coffee-K₂CO₃–800-mixed catalyst and d) mapping of iron.

Fig. 12 Cyanide poisoning of Fe–N_x sites with the FeNC_coffee-K₂CO₃-800-mixed catalyst. Scan rate 10 mV s⁻¹, 20 mM N₂H₄, 1 M KOH, 20 mM KCN.

Fig. 13 Characteristic SEM micrographs of the 6 repeated syntheses of the six FeNC_coffee-K₂CO₃-800 catalysts (dubbed S1–S6).

Fig. 14 Electrochemical reproducibility of FeNC_coffee-K₂CO₃-800 electrocatalysts for the HzOR. (a) Cyclic voltammetry of 6 separate syntheses (S1–S6 samples). (b) Three electrodes (E1–E3) with the same ink of FeNC_coffee-K₂CO₃-800. Scan rate 10 mV s⁻¹, 20 mM N₂H₄, 1 M KOH.

The electrocatalytic performance towards hydrazine oxidation of FeNC_coffee-K₂CO₃-800-mixed was examined in a PGM-free direct hydrazine fuel cell (Fig. S4†). Two cells were prepared using this catalyst in the anode, coupled with CoFe catalysts^22,62 (1 mg cm⁻² loading) in the cathode. Polarization curves for anode catalyst loadings of 2 mg cm⁻² and 4 mg cm⁻² were acquired (Fig. 15). Open circuit voltage (OCV) for the 2 mg cm⁻² loading cell is at 0.68 V at 80 °C and it increased to 0.75 V at 80 °C with higher loading. These promising results show the potential of these catalysts for practical applications such as direct hydrazine fuel cell. It must be noted though that these are initial results, with non-optimized cells. Mass transport limitations, changes to the ionomer/catalyst interface, and variations in catalyst properties pose challenges in the rapid translation of the intrinsic catalytic activity observed in lab-scale rotating disk electrode (RDE) to a membrane electrode assembly (MEA). Thus, extensive optimization at the electrode preparation stage is required to maximize the performance.^97,98

Fig. 15 Electrocatalytic activity of the FeNC_coffee-K₂CO₃-800-mixed catalyst in a direct hydrazine fuel cell.

Conclusions

To help realize the full potential of biomass as a source for precise and reproducible carbon electrodes, we developed a two-stage process for converting a readily available, if variable, biomass (waste coffee grounds) to an active and reproducible electrocatalyst. The chemical activation step allowed homogenization of structural differences between the biomass particles, without erasing them entirely. This process was optimized in terms of reactant (KOH, K₂CO₃, or KH₂PO₄) and pyrolysis temperature (700–900 °C), yielding optimal surface area, porosity hierarchy, and energy for a combination of K₂CO₃ at 800 °C. In the second step, Fe–N₄ sites were introduced by the impregnation of Fe salts and phenanthroline ligands, yielding a precise Fe–N–C material with outstanding activity towards the hydrazine oxidation reaction, comparable to the best Fe–N–C electrocatalysts in the literature.⁹⁹ The best electrocatalyst (onset potential 0.26 V vs. RHE) could be synthesized again with outstanding reproducibility, in terms of electrochemical activity and material properties. This material was further utilized in a direct hydrazine fuel cell, expanding the scope of PGM-free-based DHFCs. Overall, we expect that this work will stress the importance of tackling reproducibility in biomass-derived materials for electrocatalysis.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

STA performed synthesis, experiments and characterizations (with SB). SMZ performed fuel cell experiments, supervised by DRD. STA wrote the original manuscript. DE supervised the project and edited the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank Dr. Thierry K. Slot and Dr. Tomer Burshtein for fruitful discussions, Dr. Inna Zeltser (Technion) for N₂ gas sorption porosimetry, Dr. Rachel Edrei and Assaf Licht (Technion) for Raman measurements, Dr. Anton Twersky, Dr. Maria Koifman-Khristosov and Dr. Yael Etinger-Geller (Technion) for XRD measurements, Dr. Yaron Kauffmann (Technion) for HAADF-STEM measurements and Dr. Pini Shekhter (Tel Aviv University) for XPS experiments. We thank Prof. Kaido Tammeveski (University of Tartu) for providing the cathode catalyst samples for the fuel cell. This work was partially funded by the Nancy & Stephen Grand Technion Energy Program (GTEP), and by the Ministry of National Infrastructure, Energy and Water Resources of Israel through grants no. 2032808 (222-11-055) and no. 2033260 (222-11-064).

References

1. C. Hu, R. Paul, Q. Dai and L. Dai, Chem. Soc. Rev., 2021, 50, 11785–11843.
2. K. Jiao, J. Xuan, Q. Du, Z. Bao, B. Xie, B. Wang, Y. Zhao, L. Fan, H. Wang, Z. Hou, S. Huo, N. P. Brandon, Y. Yin and M. D. Guiver, Nature, 2021, 595, 361–369.
3. R. K. Singh, J. C. Douglin, V. Kumar, P. Tereshchuk, P. G. Santori, E. B. Ferreira, G. Jerkiewicz, P. J. Ferreira, A. Natan, F. Jaouen and D. R. Dekel, Appl. Catal., B, 2024, 357, 124319.
4. J. F. Baumgärtner, K. V. Kravchyk and M. V. Kovalenko, Adv. Energy Mater., 2024, 202400499.
5. M. R. Benzigar, V. D. B. C. Dasireddy, X. Guan, T. Wu and G. Liu, Adv. Funct. Mater., 2020, 30, 2002993.
6. L. Dai, Y. Xue, L. Qu, H.-J. Choi and J.-B. Baek, Chem. Rev., 2015, 115, 4823–4892.
7. H. Fei, J. Dong, D. Chen, T. Hu, X. Duan, I. Shakir, Y. Huang and X. Duan, Chem. Soc. Rev., 2019, 48, 5207–5241.
8. W. Tian, H. Zhang, X. Duan, H. Sun, G. Shao and S. Wang, Adv. Funct. Mater., 2020, 30, 1909265.
9. K. Gao, B. Wang, L. Tao, B. V. Cunning, Z. Zhang, S. Wang, R. S. Ruoff and L. Qu, Adv. Mater., 2019, 31, 1805121.
10. M. Shao, Q. Chang, J.-P. Dodelet and R. Chenitz, Chem. Rev., 2016, 116, 3594–3657.
11. M. Shao, J. Power Sources, 2011, 196, 2433–2444.
12. M. Sevilla, C. Falco, M.-M. Titirici and A. B. Fuertes, RSC Adv., 2012, 2, 12792–12797.
13. W. Qian, F. Sun, Y. Xu, L. Qiu, C. Liu, S. Wang and F. Yan, Energy Environ. Sci., 2013, 7, 379–386.
14. S. Tabac and D. Eisenberg, Curr. Opin. Electrochem., 2021, 25, 100638.
15. N. Zeba, T. D. Berry, K. Panke-Buisse and T. Whitman, PLoS One, 2022, 17, e0265663.
16. N. Hussain, M. Ahmed, S. Duangpan, T. Hussain and J. Taweekun, Sustainability, 2021, 13, 10449.
17. Y. Zhang, S. Liu, X. Zheng, X. Wang, Y. Xu, H. Tang, F. Kang, Q.-H. Yang and J. Luo, Adv. Funct. Mater., 2017, 27, 1604687.
18. Y. Gao, S. Zheng, H. Fu, J. Ma, X. Xu, L. Guan, H. Wu and Z.-S. Wu, Carbon, 2020, 168, 701–709.
19. A. R. Selvaraj, A. Muthusamy, Inho-Cho, H.-J. Kim, K. Senthil and K. Prabakar, Carbon, 2021, 174, 463–474.
20. H. Jäger and W. Frohs, Industrial Carbon and Graphite Materials: Raw Materials, Production and Applications, Wiley-VCH, 2021.
21. U. Sajjad, A. Sarapuu, J. C. Douglin, A. Kikas, A. Treshchalov, M. Käärik, J. Kozlova, J. Aruväli, J. Leis, V. Kisand, K. Kukli, D. R. Dekel and K. Tammeveski, ACS Catal., 2024, 14, 9224–9234.
22. K. Kisand, A. Sarapuu, J. C. Douglin, A. Kikas, A. Treshchalov, M. Käärik, H.-M. Piirsoo, P. Paiste, J. Aruväli, J. Leis, V. Kisand, A. Tamm, D. R. Dekel and K. Tammeveski, ACS Catal., 2022, 12, 14050–14061.
23. R. Cheng, J. Yang, M. Jiang, A. Dong, M. Guo, J. Zhang, B. Sun and C. Fu, J. Electrochem. Soc., 2020, 167, 110552.
24. C. Jiao, Z. Xu, J. Shao, Y. Xia, J. Tseng, G. Ren, N. Zhang, P. Liu, C. Liu, G. Li, S. Chen, S. Chen and H.-L. Wang, Adv. Funct. Mater., 2023, 33, 2213897.
25. J. Quílez-Bermejo, E. Morallón and D. Cazorla-Amorós, Carbon, 2020, 165, 434–454.
26. J. Wang, J. Liang, P. Liu, Z. Yan, L. Cui, L. Yue, L. Zhang, Y. Ren, T. Li, Y. Luo, Q. Liu, X.-E. Zhao, N. Li, B. Tang, Y. Liu, S. Gao, A. M. Asiri, H. Hao, R. Gao and X. Sun, J. Mater. Chem. A, 2022, 10, 2842–2848.
27. T. Xie, X. Li, J. Li, J. Chen, S. Sun, Y. Luo, Q. Liu, D. Zhao, C. Xu, L. Xie and X. Sun, Inorg. Chem., 2022, 61, 14195–14200.
28. D. Feng, L. Zhou, C. Liu, H. Li, Y. Zhang, Y. Yao and T. Ma, Electrochim. Acta, 2024, 484, 144096.
29. H. Huang, L. Xia, R. Cao, Z. Niu, H. Chen, Q. Liu, T. Li, X. Shi, A. M. Asiri and X. Sun, Chem. – Eur. J., 2019, 25, 1914–1917.
30. S. Zhang, X. Zhang, S. Zhang, J. Zhang, G. Li, Y. He, J. Shao, S. Zhang, H. Yang and H. Chen, Carbon Capture Sci. Technol., 2023, 9, 100135.
31. C. Quan, Y. Zhou, J. Wang, C. Wu and N. Gao, J. CO2 Util., 2023, 68, 102373.
32. J. Xu, X. Tang, M. Song and M. Tang, J. Biobased Mater. Bioenergy, 2014, 8, 444–448.
33. K. Koh, Y. Meng, X. Huang, X. Zou, M. Chhowalla and T. Asefa, Chem. Commun., 2016, 52, 13588–13591.
34. A. L. Cazetta, L. Spessato, S. A. R. Melo, K. C. Bedin, T. Zhang, T. Asefa, T. L. Silva and V. C. Almeida, Int. J. Hydrogen Energy, 2020, 45, 9669–9682.
35. I. Offen-Polak, S. N. Reddy, T. Y. Burshtein, S. M. Zahan, S. Xiang, Y. Shahaf, C. Studnik, L. Ni, M. U. Delgado-Jaime, U. Kramm, D. R. Dekel, C. Vogt, A. I. Frenkel, I. Grinberg and D. Eisenberg, J. Am. Chem. Soc., 2025 DOI:10.1021/jacs.5c08450.
36. T. Y. Burshtein, Y. Yasman, L. Muñoz-Moene, J. H. Zagal and D. Eisenberg, ACS Catal., 2024, 14, 2264–2283.
37. G. L. Soloveichik, Beilstein J. Nanotechnol., 2014, 5, 1399–1418.
38. C. W. Anson and S. S. Stahl, Chem. Rev., 2020, 120, 3749–3786.
39. L. J. M. J. Blomen and M. N. Mugerwa, Fuel Cell Systems, Springer Science & Business Media, 1994.
40. Z. Li, W. Wang, Q. Qian, Y. Zhu, Y. Feng, Y. Zhang, H. Zhang, M. Cheng and G. Zhang, eScience, 2022, 2, 416–427.
41. Y. Zhu, Y. Chen, Y. Feng, X. Meng, J. Xia and G. Zhang, Adv. Mater., 2024, 36, 2401694.
42. Q. Qian, Y. Zhu, N. Ahmad, Y. Feng, H. Zhang, M. Cheng, H. Liu, C. Xiao, G. Zhang and Y. Xie, Adv. Mater., 2024, 36, 2306108.
43. A. Badreldin, E. Youssef, A. Djire, A. Abdala and A. Abdel-Wahab, Cell Rep. Phys. Sci., 2023, 4, 101427.
44. A. Serov and C. Kwak, Appl. Catal., B, 2010, 98, 1–9.
45. A. Månberger and B. Johansson, Energy Strat. Rev., 2019, 26, 100394.
46. K. Ojha, E. M. Farber, T. Y. Burshtein and D. Eisenberg, Angew. Chem., Int. Ed., 2018, 57, 17168–17172.
47. G. Wu, K. L. More, C. M. Johnston and P. Zelenay, Science, 2011, 332, 443–447.
48. G. Wu, Z. Chen, K. Artyushkova, F. H. Garzon and P. Zelenay, ECS Trans., 2008, 16, 159.
49. B. Koyuturk, E. M. Farber, F. E. Wagner, T.-P. Fellinger and D. Eisenberg, J. Mater. Chem. A, 2022, 10, 19859–19867.
50. A. Mehmood, J. Pampel, G. Ali, H. Y. Ha, F. Ruiz-Zepeda and T.-P. Fellinger, Adv. Energy Mater., 2018, 8, 1701771.
51. D. Menga, F. Ruiz-Zepeda, L. Moriau, M. Šala, F. Wagner, B. Koyutürk, M. Bele, U. Petek, N. Hodnik, M. Gaberšček and T.-P. Fellinger, Adv. Energy Mater., 2019, 9, 1902412.
52. L. Jiao, J. Li, L. L. Richard, Q. Sun, T. Stracensky, E. Liu, M. T. Sougrati, Z. Zhao, F. Yang, S. Zhong, H. Xu, S. Mukerjee, Y. Huang, D. A. Cullen, J. H. Park, M. Ferrandon, D. J. Myers, F. Jaouen and Q. Jia, Nat. Mater., 2021, 20, 1385–1391.
53. Y. Li, X. Liu, L. Zheng, J. Shang, X. Wan, R. Hu, X. Guo, S. Hong and J. Shui, J. Mater. Chem. A, 2019, 7, 26147–26153.
54. T. Y. Burshtein, K. Tamakuwala, M. Sananis, I. Grinberg, N. R. Samala and D. Eisenberg, Phys. Chem. Chem. Phys., 2022, 24, 9897–9903.
55. T. Y. Burshtein, D. Aias, J. Wang, M. Sananis, E. M. Farber, O. M. Gazit, I. Grinberg and D. Eisenberg, Phys. Chem. Chem. Phys., 2021, 23, 26674–26679.
56. M. Gao, L. Wang, Y. Yang, Y. Sun, X. Zhao and Y. Wan, ACS Catal., 2023, 13, 4060–4090.
57. Origins & Manufacturing FAQs | Nescafé | UK & IE, https://www.nescafe.com/gb/faq/coffee-manufacturing, (accessed 13 February 2024).
58. J. E. Eichler, J. N. Burrow, N. Katyal, G. Henkelman and C. B. Mullins, J. Mater. Chem. A, 2023, 11, 12811–12826.
59. X. Liu and M. Antonietti, Carbon, 2014, 69, 460–466.
60. H. Yang, L. Shang, Q. Zhang, R. Shi, G. I. N. Waterhouse, L. Gu and T. Zhang, Nat. Commun., 2019, 10, 4585.
61. L. G. Cançado, K. Takai, T. Enoki, M. Endo, Y. A. Kim, H. Mizusaki, A. Jorio, L. N. Coelho, R. Magalhães-Paniago and M. A. Pimenta, Appl. Phys. Lett., 2006, 88, 163106.
62. J. Lilloja, E. Kibena-Põldsepp, A. Sarapuu, J. C. Douglin, M. Käärik, J. Kozlova, P. Paiste, A. Kikas, J. Aruväli, J. Leis, V. Sammelselg, D. R. Dekel and K. Tammeveski, ACS Catal., 2021, 11, 1920–1931.
63. J. Xue, J. C. Douglin, K. Yassin, T. Huang, H. Jiang, J. Zhang, Y. Yin, D. R. Dekel and M. D. Guiver, Joule, 2024, 8, 1457–1477.
64. J. C. Douglin, A. Sekar, R. K. Singh, Z. Chen, J. Li and D. R. Dekel, Small, 2024, 20, 2307497.
65. J. C. Douglin, K. Vijaya Sankar, A. L. G. Biancolli, E. I. Santiago, Y. Tsur and D. R. Dekel, ChemSusChem, 2023, 16, e202301080.
66. J. C. Douglin, J. A. Zamora Zeledón, M. E. Kreider, R. K. Singh, M. B. Stevens, T. F. Jaramillo and D. R. Dekel, Nat. Energy, 2023, 8, 1262–1272.
67. K. Yassin, J. C. Douglin, I. G. Rasin, P. G. Santori, B. Eriksson, N. Bibent, F. Jaouen, S. Brandon and D. R. Dekel, Energy Convers. Manage., 2022, 270, 116203.
68. D. Lozano-Castelló, J. M. Calo, D. Cazorla-Amorós and A. Linares-Solano, Carbon, 2007, 45, 2529–2536.
69. P. Feng, J. Li, H. Wang and Z. Xu, ACS Omega, 2020, 5, 24064–24072.
70. Z. Heidarinejad, M. H. Dehghani, M. Heidari, G. Javedan, I. Ali and M. Sillanpää, Environ. Chem. Lett., 2020, 18, 393–415.
71. Y. Urabe, T. Ishikura and K. Kaneko, J. Colloid Interface Sci., 2008, 319, 381–383.
72. T. Y. Burshtein, I. Agami, M. Sananis, C. E. Diesendruck and D. Eisenberg, Small, 2021, 17, 2100712.
73. Y. Shi, G. Liu, L. Wang and H. Zhang, RSC Adv., 2019, 9, 17841–17851.
74. W. Lei, Y.-P. Deng, G. Li, Z. P. Cano, X. Wang, D. Luo, Y. Liu, D. Wang and Z. Chen, ACS Catal., 2018, 8, 2464–2472.
75. J. E. Eichler, H. Leonard, E. K. Yang, L. A. Smith, S. N. Lauro, J. N. Burrow, R. P. P. L. Ribeiro and C. B. Mullins, Adv. Funct. Mater., 2024, 2410171.
76. R. L. Lehman, J. S. Gentry and N. G. Glumac, Thermochim. Acta, 1998, 316, 1–9.
77. J. M. Illingworth, B. Rand and P. T. Williams, Fuel Process. Technol., 2022, 235, 107348.
78. D. Eisenberg, P. Prinsen, N. J. Geels, W. Stroek, N. Yan, B. Hua, J.-L. Luo and G. Rothenberg, RSC Adv., 2016, 6, 80398–80407.
79. A. Gomez-Martin, Z. Schnepp and J. Ramirez-Rico, Chem. Mater., 2021, 33, 3087–3097.
80. M. Kakunuri, S. Kali and C. S. Sharma, J. Anal. Appl. Pyrolysis, 2016, 117, 317–324.
81. D. Rossetto and S. S. Mansy, Front. Cell Dev. Biol., 2022, 864830.
82. M. D. Hossain, Q. Zhang, T. Cheng, W. A. Goddard and Z. Luo, Carbon, 2021, 183, 940–947.
83. S. Zago, M. Bartoli, M. Muhyuddin, G. M. Vanacore, P. Jagdale, A. Tagliaferro, C. Santoro and S. Specchia, Electrochim. Acta, 2022, 412, 140128.
84. S. Das, S. Ghosh, T. Kuila, N. C. Murmu and A. Kundu, Biomass, 2022, 2, 155–177.
85. J. H. Zagal and M. T. M. Koper, Angew. Chem., Int. Ed., 2016, 55, 14510–14521.
86. M. Mazzucato and C. Durante, Electrochim. Acta, 2021, 139105.
87. T. Asset and P. Atanassov, Joule, 2020, 4, 33–44.
88. T. Marshall-Roth, N. J. Libretto, A. T. Wrobel, K. J. Anderton, M. L. Pegis, N. D. Ricke, T. V. Voorhis, J. T. Miller and Y. Surendranath, Nat. Commun., 2020, 11, 5283.
89. H. Peng, Z. Mo, S. Liao, H. Liang, L. Yang, F. Luo, H. Song, Y. Zhong and B. Zhang, Sci. Rep., 2013, 3, 1765.
90. K. Artyushkova, I. Matanovic, B. Halevi and P. Atanassov, J. Phys. Chem. C, 2017, 121, 2836–2843.
91. X. Yi, H. Yang, X. Yang, X. Li, C. Yan, J. Zhang, L. Chen, J. Dong, J. Qin, G. Zhang, J. Wang, W. Li, Z. Zhou, G. Wu and X. Li, Adv. Funct. Mater., 2024, 34, 2309728.
92. X. Jiang, J. Chen, F. Lyu, C. Cheng, Q. Zhong, X. Wang, A. Mahsud, L. Zhang and Q. Zhang, J. Energy Chem., 2021, 59, 482–491.
93. R. Gokhale, Y. Chen, A. Serov, K. Artyushkova and P. Atanassov, Electrochem. Commun., 2016, 72, 140–143.
94. K. Strickland, E. Miner, Q. Jia, U. Tylus, N. Ramaswamy, W. Liang, M.-T. Sougrati, F. Jaouen and S. Mukerjee, Nat. Commun., 2015, 6, 7343.
95. M. W. Chung, G. Chon, H. Kim, F. Jaouen and C. H. Choi, ChemElectroChem, 2018, 5, 1880–1885.
96. I. Salton, K. Ioffe, T. Y. Burshtein, E. M. Farber, N. M. Seraphim, N. Segal and D. Eisenberg, Mater. Adv., 2022, 3, 7937–7945.
97. T. Soboleva, X. Zhao, K. Malek, Z. Xie, T. Navessin and S. Holdcroft, ACS Appl. Mater. Interfaces, 2010, 2, 375–384.
98. B. Britton and S. Holdcroft, J. Electrochem. Soc., 2016, 163, F353.
99. Y. Zheng, F. He, M. Chen, J. Zhang, G. Hu, D. Ma, J. Guo, H. Fan, W. Li and X. Hu, ACS Appl. Mater. Interfaces, 2020, 12, 38183–38191.

---

Footnote

† Electronic supplementary information (ESI) available: Additional material and electrochemical characterizations. See DOI: https://doi.org/10.1039/d4cy00991f

---