Hierarchical Ternary Carbon Nanocomposite on Stainless Steel for Low−Overpotential Hydrogen Evolution in Alkaline Water Electrolysis ^†

Abstract

Alkaline water electrolysis (AWE) is a promising technology for sustainable hydrogen production, although its performance is limited by electrode overpotential, interfacial charge-transfer resistance, and limited electrochemically accessible surface area. In this study, a monolayer ternary carbon nanomaterial (CNM) composite comprising reduced graphene oxide (rGO), carbon nanotubes (CNTs), and Vulcan XC-72 was fabricated on stainless steel (SS) using a hybrid polyvinyl alcohol–polytetrafluoroethylene (PVA–PTFE) binder. Thermal treatment generated a porous conductive network that enhanced electrolyte accessibility and electron transport. Electrochemical characterization in 0.12 M NaOH showed that the CNM-modified electrode exhibited substantially higher current response and CV-derived double-layer capacitance (C_dl) of 62.61–78.51 mF/cm², compared with 3.43–3.74 mF/cm² for bare SS. Electrochemical fitting further showed markedly higher exchange-current density (i₀) parameters for the modified electrode, along with a reduced solution resistance (Rs) of ~2.1–2.2 Ω·cm² and a lower Rct. The oxyhydrogen (HHO) production rate reached 0.304 mL/min at 3.8 V, compared with 0.262 mL/min for bare SS at 4.0 V. Repeated HHO measurements showed ~2% variation (n = 3), indicating good reproducibility of the gas-production response. These results demonstrate that the rGO/CNT/XC-72 composite provides an effective and reproducible surface-engineering approach for enhancing electrochemical performance and HHO production in alkaline electrolysis systems.

1. Introduction

The global shift toward renewable energy has intensified demand for technologies for clean hydrogen production [1]. Among these, alkaline water electrolysis (AWE) stands out as a commercially mature and cost-effective approach, utilizing an alkaline electrolyte (typically KOH or NaOH) to drive the Hydrogen Evolution Reaction (HER) at the cathode and the Oxygen Evolution Reaction (OER) at the anode [2,3,4]. Despite its advantages, including the use of non-precious metal electrodes and operational robustness, AWE is fundamentally limited by high overpotentials, sluggish OER kinetics, and poor electrode durability [5,6]. Conventional electrode materials, such as bare graphite and stainless steel (SS), exhibit limited electroactive surface areas and inadequate charge-transfer properties, leading to elevated solution resistance (Rs), significant energy losses, and reduced hydrogen yields [5,7,8].

Surface engineering with carbon nanomaterials (CNMs) has emerged as a promising strategy to overcome these limitations [5,9,10]. Reduced graphene oxide (rGO) offers a two-dimensional conductive scaffold with abundant active sites; carbon nanotubes (CNTs) provide efficient one-dimensional electron transport pathways; and Vulcan carbon black (XC-72) enhances macroscopic conductivity as a conductive filler [5,11,12]. When combined into ternary composite architectures, these materials form synergistic, interconnected networks that simultaneously enhance electron transport, double-layer capacitance (C_dl), and electrolyte accessibility. The binder system further governs electrode performance; hybrid polyvinyl alcohol–polytetrafluoroethylene (PVA–PTFE) binders are particularly notable for their ability to generate hierarchical porous microstructures upon controlled thermal decomposition of PVA, significantly enhancing mass transport within the electrode layer [5,11,13].

However, the existing literature predominantly investigates single-component or binary CNM systems or relies on precious-metal catalysts that are cost-prohibitive at scale. Systematic studies on metal-free ternary rGO/CNT/XC-72 composites within practical AWE configurations, particularly examining the combined effect of PVA-induced hierarchical porosity on real gas production performance, remain limited. This study addresses that gap by fabricating and characterizing an optimized ternary CNM composite electrode on SS current collectors (CCs) using a PVA–PTFE binder, and evaluating its HER/OER activity via cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS), alongside direct oxyhydrogen (HHO) gas flow measurements. It is hypothesized that the synergistic CNM network, reinforced by thermally induced hierarchical porosity, will substantially reduce Rs and improve HHO production rates compared with bare SS electrode applications.

2. Materials and Methods

2.1. Materials and Instrumentation

The experimental setup utilized a high-purity SS sheet (316L) as the CC. The CNMs employed for the electrode coating included CNT powder (99% purity), rGO powder (99% purity), and XC-72 as the CCF. The binding matrix was formulated using Teflon™ PTFE dispersion, grade 30 (60% dispersion) (Fuel Cell Store online store, USA) and PVA powder (98% purity). Analytical-grade NaOH (99.98% purity) served as the electrolyte for the AWE process. For the fabrication and operation of the HHO generator, transparent plexiglass (acrylic) sheets and polyvinyl chloride (PVC) end-caps were used to construct the cell housing. The preparation of the ternary CNM composite involved using an ultrasonicator (Faithful Instruments, Huanghua, China) for homogeneous dispersion and a tube furnace (Zhengzhou Protech Technology, Zhengzhou City, Henan Province, China) under a nitrogen (N₂, 99% purity) atmosphere for thermal treatment. The composite was applied to the SS CC via an airbrush system equipped with a dedicated compressor. Electrochemical measurements, including CV and EIS, were performed using a CS310 electrochemical workstation (Wuhan Corrtest Instruments, Wuhan, China).

2.2. HHO System Design and Specifications

The HHO system was based on a dry-cell-type generator with a bubbler configuration as previously described [3,5]. Performance improvements were achieved by replacing conventional electrodes with surface-engineered electrodes incorporating CNMs, while maintaining the original system geometry, as shown in

Table 1. HHO system specifications.

Parameter	Unit	Specification
Operating Voltage (DC)	V	1.3–2.5
Working Current Range	mA	1–250/cell
Volume of HHO Generator	cm3	2.4 per cell
Electrolyte Solution Molarity	M	0.12
Volume of Gas Bubbler	cm3	157
Electrolyte Flowrate	mL/min	10–15

(Source: Author’s compilation).

.

A 0.12 M NaOH electrolyte was selected based on the operating characteristics established in Siriwardena et al.’s previous study of a dry-cell HHO system, in which NaOH provided favorable gas-production performance, and increasing the electrolyte concentration within the investigated range improved ionic conductivity and gas evolution [3]. The 0.12 M concentration was therefore adopted as a practical operating condition, given that it reflects the concentration dependence of HHO production, as shown in the HHO production graph for the present electrode-comparison study, rather than as a universal optimum electrolyte concentration.

2.3. Current Collector (CC) Preparation

SS CCs were mechanically polished using 800-grit followed by 1000-grit sandpaper to obtain a uniform surface finish. The polished CCs were ultrasonically cleaned in acetone for 5 min to remove surface contaminants, preheated at 50–60 °C to ensure complete drying, and stored in a dust-free environment before spray coating. Then, the coating slurry was prepared using a hybrid polyvinyl alcohol–polytetrafluoroethylene (PVA–PTFE) binder with reduced graphene oxide (rGO), carbon nanotubes (CNTs), and Vulcan XC-72 conductive carbon black. The CNM ratios were maintained at rGO/CNT (0.30/0.80) and XC-72 (0.50), while 1 wt% sodium lauryl sulfate (SLS) was used as the dispersing agent. The CNM and conductive carbon filler (CCF) values are normalized, dimensionless ratios that represent the relative positions of the investigated minimum and maximum levels. PVA was dispersed in distilled water by magnetic stirring (350–500 rpm, 30 min) followed by ultrasonication (40 kHz, 20 min, 40 °C). Separately, CNTs, rGO, and XC-72 were manually ground and dispersed in the SLS solution [5,11]. The carbon dispersion was then combined with the PVA solution under high-shear stirring (400–500 rpm, 30 min) and further sonicated (40 kHz, 20 min). PTFE dispersion was subsequently added dropwise under continuous stirring, and the final slurry was mixed overnight at 500 rpm with a solids-to-liquid ratio of 30:70. Slurry quality was evaluated using a tape peel test by applying a thin coating onto a glass substrate, allowing it to dry, and removing Scotch adhesive tape to verify coating adhesion and confirm the suitability of the formulation for spray coating.

2.4. Coating and Heat Treatment Process

The prepared slurry was spray-coated onto the SS CCs using an airbrush operated at 0.82 bar. The coating was applied in two successive layers, with each layer air-dried at 25 °C before the next was applied. A final drying step was performed at temperatures below 40 °C, resulting in a total coating thickness of ~10 µm (5 µm per layer). The reported loading corresponds to the total dry CNM mass, comprising rGO, CNTs, and Vulcan XC-72, per unit geometric area, and does not include the PVA–PTFE binder. It was determined gravimetrically from the difference between the mass of the coated electrode after drying and heat treatment and that of the corresponding bare SS current collector (CC). The geometric coated area (A_geo) was 1.00 cm² per side of each exposed electrode. For three independently prepared electrodes, the deposited CNM loading was 80.0 ± 1.0 mg/cm² (n = 3), with a relative standard deviation (RSD) of 1.2% (Equation (1)). The reported loading refers to the dry CNM component and excludes the SS CC.

$$L_{coating}={\displaystyle \frac{m_{coated}-m_{bare}}{A_{geo}}}$$

(1)

where ($L_{coating}$) is the coating in mg/cm², ($m_{coated}$) is the mass of the dried/heat-treated coated electrode in mg, ($m_{bare}$) is the mass of the corresponding bare SS electrode in mg, and ($A_{geo}$) is the geometric coated area in cm².

After that, the coated electrodes were subsequently subjected to thermal treatment under a nitrogen atmosphere. The furnace chamber was initially purged with N₂ at 20 mL/min for 5 min to remove residual oxygen, after which the flow rate was maintained at 3 mL/min throughout the heating process. The PVA–PTFE hybrid binder was heat-treated according to the sintering profile shown in

Table 2. PVA–PTFE heat treatment parameters.

Step	Temperature Range (°C)	Heating Rate (°C/min)	Hold Time (min)
1	27 → 85	5	15
2	85 → 230	5	30
3	230 → 327	5	15
4	327 → 360	5	60
5	Cooling to ambient	5	240

(Source: Author’s compilation).

.

It enabled controlled partial decomposition of PVA to generate porosity while promoting PTFE melting and recrystallization, thereby enhancing coating integrity and adhesion.

2.5. Electrochemical Characterization

Electrochemical measurements were performed in a three-electrode configuration, with the surface-engineered SS electrode as the working electrode, a graphite rod as the counter electrode, and an Ag/AgCl reference electrode in 3 M KCl. The electrochemical workstation was configured using a reference-electrode potential of 0.197 V, and the potentials in the raw CV data were recorded relative to the Ag/AgCl reference electrode. The electrolyte pH was measured as 13.12 at 25 °C. For consistency with the alkaline electrocatalysis literature, the measured potentials were converted to the reversible hydrogen electrode (RHE) scale at 25 °C, giving a conversion offset of +0.973 V [14,15]. All potentials presented in the revised CV figures are therefore reported versus RHE, according to Equation (2).

$$E_{RHE}=E_{Ag/AgCl}+0.197+0.05916\times pH$$

(2)

All electrochemical measurements were performed in five independent replicates (n = 5). The experimental variability of the electrochemical parameters was ~2.8%, and the corresponding values are reported with the associated experimental variation. CV measurements scanned two potential windows: −1.12 V to +0.65 V (bare electrode) and −1.4 V to +0.85 V (modified electrode) vs. Ag/AgCl. CV measurements were performed at scan rates (SRs) of 50 and 150 mV s over five consecutive cycles to evaluate the electrochemical response and cycle-to-cycle reproducibility. Furthermore, no separate electrochemical activation procedure was performed before CV measurements. The electrodes were subjected directly to the prescribed CV protocol, and five consecutive CV cycles were recorded under the same experimental conditions. The recorded cycles were used to evaluate the electrodes’ electrochemical response and reproducibility. The capacitance values obtained from the CV electrochemical analysis were treated as the C_dl and normalized to the geometric electrode area. The C_dl analysis follows the general electrochemical interpretation of capacitive current response reported in recent alkaline electrocatalysis studies [15]. Five independent measurements were performed for each electrode condition (n = 5), yielding an experimental variation of ~2.8%. The values were 3.43 and 3.74 mF/cm² for bare SS and 62.61 and 78.51 mF/cm² for the CNM-modified SS electrode at 50 and 150 mV/s, respectively. The substantial increase was interpreted as evidence of an enlarged electrochemically accessible interfacial area following CNM modification.

EIS was performed to investigate the charge-transfer resistance (Rct) and interfacial electrochemical properties of the electrodes under conditions relevant to the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). Measurements were conducted over a frequency range of 100 kHz to 0.1 Hz using a sinusoidal perturbation amplitude of 10 mV, with the applied DC bias corresponding to the selected potential of interest. The resulting Nyquist plots were analyzed by fitting to appropriate equivalent circuit models (ECMs). As illustrated in Figure 1A, the ECM for the bare electrode consisted of the Rs, Rct, a constant phase element (CPE) representing the non-ideal C_dl arising from surface heterogeneity, and a Warburg impedance (Ws) accounting for diffusion-controlled mass transport. In contrast, the modified electrodes were fitted using the ECM shown in Figure 1B, comprising Rs, two Rct components (Rct1 and Rct2), and two CPEs (CPE1 and CPE2), reflecting multiple electrochemical interfaces and heterogeneous charge-transfer processes associated with the surface modification.

All electrochemical measurements were performed in five independent replicates (n = 5). The measured electrochemical parameters exhibited an overall experimental variation of ~2.8%.

2.6. HHO Production Rate Measurement

The HHO system was fabricated with surface-modified and bare electrodes in two operations, operated at 2–3 V DC, with electrolyte circulation at 10–12 mL/min via a peristaltic pump, under standard conditions (25 °C, 1 atm). HHO output was routed through the bubbler unit and a gas-capturing assembly to a flow meter for production quantification, as shown in Figure 2. To evaluate the stability and reproducibility of HHO production, repeated measurements were performed under identical operating conditions using the CNM-modified and bare SS electrode configurations. Each condition was tested in three independent measurements (n = 3), and the observed variation in HHO production was ~2% of RSD. The mean HHO production rate and the corresponding experimental variation were used to compare the electrode configurations.

3. Results and Discussion

3.1. Electrochemical Characterization

3.1.1. CV Analysis

The electrochemical behavior of the bare and CNM-modified SS electrodes was evaluated by CV at SRs of 50 and 150 mV/s to assess the effect of surface modification on charge-storage and electron-transfer characteristics. The bare SS electrode (Figure 3A) exhibited quasi-rectangular voltammograms with weak redox features, indicating that electrical C_dl primarily governed the current response with limited Faradaic contribution. Increasing the SR from 50 to 150 mV/s increased the current response, accompanied by broader peak separation, suggesting increased polarization and relatively slow charge-transfer kinetics arising from the limited number of electroactive sites on the unmodified SS surface. The CNM-modified SS electrodes were prepared with a mean CNM loading of 80.0 ± 1.0 mg/cm² (n = 3), providing a consistent coated surface for electrochemical comparison with bare SS.

In contrast, the CNM-modified SS electrode (Figure 3B) exhibited substantially higher anodic and cathodic current densities and a much larger enclosed CV area at both SRs, confirming enhanced electrochemical activity after surface modification. The anodic Ip increased from 3.12 to 13.81 mA/cm². The CV-derived increased from 62.61 to 78.51 mF/cm² between 50 and 150 mV/s, while the corresponding values for bare SS were only 3.43 and 3.74 mF/cm², respectively. Thus, the CNM-modified electrode exhibited ~18.2− and ~21.0-fold higher capacitance than bare SS at 50 and 150 mV/s, respectively. The substantially larger modified electrode is consistent with the formation of a hierarchical conductive network comprising interconnected rGO sheets, CNT pathways, and XC-72 particles, thereby increasing the electrolyte-accessible interfacial area compared with the relatively smooth bare SS surface.

The exchange current density (i₀) obtained from the electrochemical fitting also increased markedly following CNM modification, as shown in Figure 4. At 50 mV s, it increased from 4.71 × 10⁻⁵ to 3.26 × 10⁻³ mA/cm², corresponding to a ~69-fold increase. At 150 mV/s, it increased from 1.83 × 10⁻⁴ to 1.76 × 10⁻² mA/cm², corresponding to an ~96-fold increase. The substantially lower cathodic Tafel slope of the modified electrode further indicates more favorable interfacial charge-transfer kinetics than those of bare SS.

3.1.2. EIS Analysis

EIS was performed to evaluate the influence of the rGO/CNT/XC-72 conductive carbon network on the interfacial charge-transfer characteristics of SS electrodes under HER and OER conditions. Nyquist plots exhibited depressed semicircles followed by diffusion tails, indicating coupled charge transfer and mass transport processes. For the bare SS electrode (Figure 5A), the Rs was 2.90 Ω·cm² (HER) and 3.84 Ω·cm² (OER). The HER Rct was relatively low (11.21 Ω·cm²), whereas the OER Rct increased by several orders of magnitude (5.74 × 10⁸ Ω·cm²), reflecting sluggish OER associated with passive oxide film formation. The corresponding increase in Ws from 465 to 14,513 Ω·cm² further indicates greater diffusion limitations during OER. The CPE exponent (n ≈ 0.78–0.79) suggests a non-ideal capacitive interface due to surface heterogeneity. Following CNM modification (Figure 5B), Rs decreased to ~2.1–2.2 Ω·cm², confirming improved electrical contact through the conductive carbon network. The modified electrodes exhibited two Rcts (Rct1 and Rct2), corresponding to the CNM coating and SS CC, with values of 3.88 and 6.53 Ω·cm² (HER) and 3.28 and 12.67 Ω·cm² (OER). Higher CPE exponents (0.88–0.95) indicate a more homogeneous and stable electrode/electrolyte interface. These findings demonstrate that the synergistic interaction of rGO sheets, CNT conductive pathways, and XC-72 nanoparticles forms an interconnected conductive framework that enhances interfacial stability and facilitates efficient electron transport.

3.2. HHO Production Rate Measurement

The HHO production rates of the bare and CNM-modified SS electrodes were evaluated at applied voltages of 1.6–4.0 V (Figure 6). Below 1.8 V, the measured HHO output remained at or below the flow-meter resolution limit (≤0.005 mL/min). At 1.8 V, the modified electrode produced 0.032 ± 0.0006 mL/min, compared with 0.005 ± 0.0001 mL/min for the bare electrode. At 2.2 V, the production rate increased to 0.105 ± 0.0021 mL/min for the modified electrode, ~3−fold higher than that of the bare electrode (0.036 ± 0.0007 mL/min). At 3.0 V, the modified electrode reached 0.238 ± 0.0048 mL/min, compared with 0.138 ± 0.0028 mL/min for the bare electrode. At higher voltages, the modified electrode approached a plateau, producing 0.285 ± 0.0057, 0.304 ± 0.0061, and 0.302 ± 0.0060 mL/min at 3.6, 3.8, and 4.0 V, respectively. The maximum HHO production rate of 0.304 mL/min was therefore obtained at 3.8 V, compared with 0.262 mL/min for the bare electrode at 4.0 V. The higher HHO production at lower applied voltage is consistent with the enhanced electrochemical response and reduced Rct of the CNM-modified electrode. To assess reproducibility, HHO production measurements were repeated three times under identical operating conditions (n = 3), with a variation of ~2%. The relatively low variation indicates good repeatability of the measured HHO production performance.

4. Conclusions

This study demonstrates that a single-layer ternary CNM composite electrode can effectively enhance the electrochemical performance of alkaline water electrolysis. The synergistic combination of rGO, CNTs, and XC-72 formed a conductive porous network that improved electrolyte accessibility and electron transport, compared with bare SS, the CNM-modified electrode exhibited substantially higher CV-derived double-layer capacitance (C_dl) and enhanced electrochemical i₀ parameters, along with reduced Rs and Rct. At SRs of 50 and 150 mV/s, the modified electrode exhibited ~18–21-fold higher (C_dl) and ~69–96-fold higher fitted (i₀), respectively, than bare SS. EIS analysis further indicated a more homogeneous electrode/electrolyte interface and improved interfacial charge-transfer characteristics following CNM modification. Repeated HHO production measurements (n = 3) showed ~2% experimental variation, indicating good reproducibility. The modified electrode achieved a maximum HHO production rate of 0.304 mL/min⁻¹ at 3.8 V, compared with 0.262 mL/min for bare SS at 4.0 V, demonstrating higher HHO production at a lower applied voltage. Overall, the results demonstrate that the rGO/CNT/XC-72 composite provides an effective and reproducible surface-engineering strategy for improving electrode performance and HHO production in alkaline electrolysis systems.