Electrochemical valorization of H₂S in natural gas to sulfate under mild conditions

Abstract

H₂S capture and valorization from polluted natural gas offer environmental and resource recovery benefits, but current approaches produce moderate-value sulfur with intensive carbon footprint. Herein, we develop an electrochemical deep oxidation method that converts H₂S from polluted natural gas into value-added K₂SO₄ using in-situ cathodically generated H₂O₂. We first validate this concept using commercial H₂O₂ and then in-situ generated H₂O₂ in H-cell, revealing the importance of high H₂O₂ concentration for deep H₂S oxidation, especially sluggish S₂O₃²⁻-to-SO₃²⁻ conversion. We then showcase its application potential in 4-cm² and then 100-cm² flow reactor with high interfacial H₂O₂ concentration and large current, with the latter achieving H₂S removal (100,000 ppm to <15 ppm), >70% K₂SO₄ selectivity, and 100-h stable operation. Life-cycle assessment and techno-economic analysis confirm the strategy’s sustainability advantages and economic viability. We finally extend this method to produce a 1.4 wt% H₂SO₄ solution by modifying the flow reactor with a solid-electrolyte type.

Introduction

Hydrogen sulfide (H₂S) commonly exists in natural gas and crude oil reservoirs. It is a harmful gas, including threatening human health, endangering industrial processes (results in facility and pipeline corrosion during transportation, and catalyst deactivation during refinery process), and polluting environment through off-gases and wastewaters. In this context, H₂S capture and valorization (H₂S-CV) is of significant importance for protecting human, equipment, and environment from danger, as well as for producing valuable sulfur-involved chemicals.

As an important scenario, raw natural gas contains 1–10 vol% of H₂S (corresponding to 10,000–100,000 ppm, Supplementary Table 1), calling for efficient H₂S-CV technology. To date, H₂S impurities are first captured using amine absorption, separating CH₄ for downstream use, followed by distillation to regenerate H₂S gas (Fig. 1a). Subsequently, the pure H₂S gas is conventionally treated through the Claus process, in which 1/3 H₂S is firstly oxidized to sulfur dioxide (SO₂) by air in a combustion furnace, and then the remaining 2/3 H₂S reacts with SO₂ over Al₂O₃-based catalyst in Claus reactor to afford sulfur (S) as the final product^1,2. This method features high H₂S treatment capacity, accounting for more than 95% of total S production globally³. However, harsh reaction conditions (high temperature >900 °C) are necessary for the distillation and combustion step, resulting in a carbon-intensive process. In addition, complete reaction of H₂S and SO₂ cannot be achieved due to the thermodynamical equilibrium of the Claus reaction step⁴, resulting in the emissions of residual H₂S and SO₂ into atmosphere.

Fig. 1: Schematic illustration of conventional H₂S-CV technologies and this work.

a Conventional industrial H₂S treatment process. b Previously reported electrochemical anodic sulfion oxidation method. c Proposed strategy in this work: electrochemical H₂S oxidation via in-situ generated H₂O₂ at cathode.

In recent years, renewable electricity powered electrochemical reaction has emerged as a promising approach for sustainable pollutants capture and chemicals production⁵, which can be operated under ambient conditions with ideally low carbon emission. Regarding H₂S-CV, an electrochemical strategy based on anodic sulfion (S²⁻) oxidation reaction (ASOR, Fig. 1b) was previously reported, involving the oxidation of S²⁻ to S at the anode⁶. Furthermore, advanced modulation strategies including sulfurization^7,8, high-entropy alloying⁹, and carbon encapsulation^10,11, were developed to improve the stability and sulfophobicity of anode materials, thereby mitigating the issue of sulfur passivation. The introduction of redox pairs with superior reaction kinetic, such as I⁻/I₃⁻ (ref. ¹²) and Fe²⁺/Fe³⁺ (ref. ⁴), are also efficient strategies to mediate the oxidation of H₂S to S in the solution, by which sulfur deposition can be avoided. Additionally, the ASOR can couple with the hydrogen evolution reaction (HER) at the cathode to produce green hydrogen^13,14. Nevertheless, the produced S shows a moderate price (110 $ per ton S), and it is not the end product on market but rather the feedstock for other chemicals production via additional oxidation reactions. For instance, oxidation of S to more valuable sulfates (500 $ per ton K₂SO₄) accounts for more than 90% of S consumption¹⁵. Moreover, when treating same quantities of H₂S contaminants, the value-added from K₂SO₄ production remains much higher than S production, even after accounting for either the market price or production costs associated with KOH feedstock (Supplementary Fig. 1).

In this regard, it is attractive to develop an electrochemical strategy to directly transform H₂S (ideally, using natural gas as the feedstock) into sulfates via deep oxidation at ambient conditions. This concept has been explored in the aerobic oxidation of sulfides at the anode, with successful production of sulfates^16,17,18. However, sulfur deposition occurs at the same time, calling for advanced strategy to improve sulfide removal efficiency and sulfate selectivity. Typically, strong oxidant is required to achieve deep H₂S oxidation, such as Cl₂ or H₂O₂ (refs. ^19,20), but the widespread applications are restricted by the formation of stoichiometric chlorine-containing waste or the use of expensive oxidant (i.e., H₂O₂). Alternatively, electrocatalytically generated H₂O₂ via two-electron oxygen reduction reaction (2e⁻ ORR) serve as a promising oxidant, for it can be powered by renewable electricity and can be used on demand. Benefiting from the great progress of associated electrocatalysts design, especially the carbon-based one^21,22, the development of in-situ generated H₂O₂ for oxidation reactions to produce value-added chemicals have been witnessed in a handful cases, including epoxidation of olefins²³ and activation of N₂ and CH₄ (refs. ^24,25.). Owing to the high oxidation efficiency of in-situ generated H₂O₂ and the sustainable nature of H₂O₂ electrosynthesis, there is plenty room for further exploration for other scenarios, not limited to the few applications shown above. For instance, the treatment of pollutant H₂S is an important issue, and using in-situ generated H₂O₂ for its oxidation has not been reported. Therefore, we conceive to couple in-situ H₂O₂ generation at cathode with deep H₂S oxidation to produce sulfates.

In this work, we report an electrochemical H₂S oxidation reaction (EChem-SOR) for SO₄²⁻ production from H₂S-involved natural gas, via in-situ H₂S oxidation by H₂O₂ generated from cathodic 2e⁻ ORR (Fig. 1c). We first showed that H₂S can be deeply oxidized into sulfate ions (SO₄²⁻) by commercial H₂O₂ solution via chemical oxidation reaction (Chem-SOR), and recognized the importance of high-concentration H₂O₂ for deep H₂S oxidation. Then, we synthesized an oxidized carbon nanotube (OCNT) as an active 2e⁻ ORR electrocatalyst for H₂O₂ generation, and demonstrated the feasibility of EChem-SOR in H-cell. During H₂S conversion process, H₂S was initially captured by an alkaline medium to form S²⁻, followed by stepwise oxidations to eventually produce SO₄²⁻ by the in-situ generated H₂O₂ (in the form of HO₂⁻ in alkaline medium). We recognized that deep H₂S oxidation was limited by the low H₂O₂ concentration, owing to low current and diluted solution in H-cell, leaving thiosulfate ions (S₂O₃²⁻) an accumulated intermediate that is difficult to be further oxidized. To overcome this issue, we constructed a 4-cm² flow reactor for continuous EChem-SOR at high currents, enabling high-concentration H₂O₂ over the cathodic interface and in the bulk solution, reaching an over 90% H₂S_cap and 90% K₂SO₄ selectivity. Furthermore, we scaled up the flow reactor to 100 cm², showing the capacity to desulfurize real H₂S-involved natural gas with stable operation for 100 h and efficiently convert H₂S into K₂SO₄ at large-scale. By performing life-cycle assessment (LCA) and techno-economic analysis (TEA), we demonstrated that EChem-SOR strategy is potentially more sustainable for H₂S treatment and K₂SO₄ production compared with industrial methods, and shows profitability for K₂SO₄ production. Furthermore, the EChem-SOR system can be extended to produce H₂SO₄ solution in a modified solid-electrolyte flow reactor.

Results

Chemical oxidation of H₂S using commercial H₂O₂

Prior to in-situ generated H₂O₂ via 2e⁻ ORR for H₂S oxidation, we first investigated if H₂S can be deeply oxidized into SO₄²⁻ by commercial H₂O₂ solution via Chem-SOR. As shown in the calculated Pourbaix diagram (Fig. 2a; see methods in Supplementary Note 1), H₂S and its dissolved forms (HS⁻ and S²⁻) can be converted into SO₄²⁻ within a wide pH window. The conversion process may involve an intermediate sulfur in solid phase at pH below 8.4, whereas its formation can be avoided under highly alkaline conditions, leading to direct HS⁻/S²⁻-to- SO₄²⁻ conversion. Additionally, the equilibrium potentials of H₂S/HS⁻/S²⁻-to- SO₄²⁻, O₂-to-OH⁻ and H₂O₂/HO₂⁻-to-OH⁻ increase sequentially within the full pH window. This indicates that H₂O₂ shows the thermodynamical ability for deep oxidation of H₂S to SO₄²⁻, and its oxidative ability exceeds that of O₂.

Fig. 2: Proof-of-concept for oxidizing H₂S using H₂O₂.

a Pourbaix diagram of SO₄²⁻ and S production from H₂S in its various forms (H₂S, HS⁻, or S²⁻) on the SHE (standard hydrogen electrode) scale. The dashed lines represent the reduction reactions of H₂O₂-to-H₂O and O₂-to-H₂O, respectively, suggesting the thermodynamical ability of H₂O₂ for deep H₂S oxidation, as well as stronger oxidizing capability than O₂. H₂S_cap conversion (left in b), productivity of various oxidation products (right in b), and SO₄²⁻ selectivity (c) at different H₂O₂ concentrations (0, 0.05, 0.1, 0.5, and 1 wt%) in 0.1 M KOH, with feeding 5% H₂S/Ar at a flow rate of 20 ml min⁻¹ for 2 h. The error bars represent the standard deviation of three independent measurements. Variation in the content of S²⁻, S₂O₃²⁻, SO₃²⁻, and SO₄²⁻ over time at H₂O₂ concentrations of 0.08 wt% (d), 0.4 wt% (e), and 1 wt% (f) in 0.1 M KOH. The amounts of NaHS and H₂O₂ were fixed at 1 mmol and 5 mmol, respectively. g Reaction pathway of H₂O₂-mediated H₂S stepwise oxidation reactions under alkaline conditions.

Guided by the fundamental analysis, we evaluated the effect of H₂O₂ concentration on the H₂S oxidation performance in an alkaline condition (0.1 M KOH), while the overall H₂O₂ amounts were increased accordingly. H₂S/Ar mixing gas containing 5 vol% H₂S was selected as the sulfur source (Ar was used as the diluted gas for safety concern, and H₂S concentration mimics that in natural gas reservoirs). Since H₂S was initially captured by alkaline solution and subsequently converted into products, H₂S conversion calculation was based on the captured amount, which was defined as “H₂S_cap conversion”. As shown in Fig. 2b, H₂S_cap conversion was below 20% at <0.1 wt% H₂O₂ concentration and nearly 100% at 1 wt% H₂O₂ concentration, indicating that the oxidation rate of H₂S can be significantly accelerated by increasing H₂O₂ concentration. The product distribution at different H₂O₂ concentrations (right in Fig. 2b) implies a stepwise oxidation of H₂S to S₂O₃²⁻, SO₃²⁻, and finally to SO₄²⁻. At low H₂O₂ concentrations (0.05 and 0.1 wt%), the primary products were S₂O₃²⁻ and SO₃²⁻, with only <25% SO₄²⁻ selectivity (Fig. 2c). As the H₂O₂ concentration increased to 0.5 wt%, SO₄²⁻ became the main oxidation product with >75% selectivity. When the H₂O₂ concentration further increased to 1 wt%, SO₄²⁻ became the sole liquid-phase product, while solid S was also observed. These results preliminarily demonstrate deep oxidation of H₂S to SO₄²⁻ is viable by using commercially available H₂O₂ solution, and facilitated by high H₂O₂ concentration.

To more quantitatively study the dependance of product distribution and reactivity of different intermediates (S²⁻, S₂O₃²⁻, SO₃²⁻) on H₂O₂ concentration. NaHS was utilized as the substrate to precisely control the introducing amount, considering that H₂S is transformed to HS⁻ and S²⁻ when it is captured in an alkaline solution. We fixed the input amounts of NaHS (1 mmol) and H₂O₂ (5 mmol), while changed H₂O₂ concentration. As shown in Fig. 2d–f, S₂O₃²⁻ was accumulated at lower H₂O₂ concentration, suggesting that S₂O₃²⁻ oxidation is more sluggish compared to S²⁻ and SO₃²⁻. As the H₂O₂ concentration increased, S²⁻ consumption, S₂O₃²⁻ conversion and SO₄²⁻ generation were greatly facilitated. Collectively, we proposed that H₂S undergoes a stepwise oxidation pathway to SO₄²⁻ in the presence of commercial H₂O₂ via Chem-SOR (Fig. 2g), and the reaction is limited by kinetically sluggish S₂O₃²⁻ oxidation. It is important to apply high-concentration H₂O₂ to facilitate deep H₂S oxidation. In addition, the maintenance of strong alkaline condition is also critical for the above transformation, since solid-state S was generated under more acidic conditions (Supplementary Figs. 2 and 3), which may be related to the disproportionation reaction of the formed S₂O₃²⁻ in acidic conditions²⁶.

EChem-SOR in H-cell

The above demonstration of H₂S-to-SO₄²⁻ conversion via Chem-SOR (that utilizes H₂O₂ as the oxidant) encourages us to develop an electrochemical strategy for H₂S-to-SO₄²⁻ conversion mediated by cathodically in-situ generated H₂O₂, the so-called EChem-SOR. The prerequisite for EChem-SOR is to prepare an efficient 2e⁻ ORR catalyst for H₂O₂ electrosynthesis. Based on previous work²⁷, we selected a well-established oxidized carbon nanotube (OCNT) as the electrocatalyst, by heat-treating pristine CNT with HNO₃. The high-resolution transmission electron microscopy (HR-TEM) images and X-ray diffraction (XRD) patterns (Supplementary Fig. 4a–e) showed that OCNT retains the tubular and crystalline structure of CNT. Both the results of elemental analysis (EA) and X-ray photoelectron spectroscopy (XPS) revealed a higher oxygen content in OCNT (Supplementary Fig. 4f, g), indicating the successful introduction of oxygen-containing functional groups (OFGs) through HNO₃ oxidation. Raman spectra results showed an increased I_D/I_G ratio for OCNT (Supplementary Fig. 5a), suggesting more defects or disordered structures were introduced into catalyst. Moreover, OCNT exhibited a higher ratio of C = O bond in high-resolution O 1 s XPS spectra, and a stronger signal of carboxyl groups (–COOH) in Fourier-transform infrared (FTIR) spectra (Supplementary Fig. 5b–d). These results suggest that more carboxyl groups are involved in the OFGs.

We then evaluated the 2e⁻ ORR performance of OCNT and CNT using a rotating ring-disk electrode (RRDE) setup in a three-electrode system. RRDE LSV curves showed that the disk current density, ring current, and calculated H₂O₂ selectivity of OCNT are all higher than that of CNT (Supplementary Fig. 6a), indicating a superior 2e⁻ ORR performance of OCNT. Moreover, the HNO₃ treatment time was optimized to obtain an optimal OCNT, achieving a maximum H₂O₂ accumulated concentration of 0.12 wt% within 2 h at 0.2 V vs RHE in H-cell (Supplementary Fig. 6b–d), which is comparable to previously reported oxidized CNT electrocatalysts^27,28,29. Furthermore, the in-situ FTIR spectra of OCNT and CNT at the open-circuit potential (OCP) and increasing potentials were collected (Supplementary Fig. 7). The results showed the peak intensity of OOH* intermediate gradually increases with the 2e⁻ ORR overpotential, and OCNT exhibited a higher relative peak intensity and increased quickly with the potential. These prove that the more favorable formation of OOH* intermediate in OCNT compared to CNT, contributing the enhanced 2e⁻ ORR performance.

Based on the high activity of OCNT for H₂O₂ electrosynthesis, we first investigated the feasibility of EChem-SOR route in a H-cell. The electrochemical setup contained two tandem parts (Fig. 3a and Supplementary Fig. 8), involving a H₂S-capture tank (on the left) and a S²⁻-oxidation H-cell (on the right). For the H₂S-capture unit, 5% H₂S/Ar mixing gas was introduced into a catholyte reservoir containing 0.1 M KOH, where H₂S was captured by the alkaline solution and dissociated to S²⁻ ions via acid-base reaction. The dissociated S²⁻ solution was then pumped into the cathodic chamber of H-cell. It is important to note that the H₂S capture unit is important to the following S²⁻-oxidation unit (that is, electrolysis unit) to obtain a stable current-time (i-t) curve of 2e⁻ ORR (Supplementary Fig. 9). If H₂S/Ar was directly fed into H-cell without a H₂S-capture tank, the current continuously decayed (Supplementary Fig. 10), owing to Ar expelled the dissolved O₂ over the two-phase interface (between electrode and electrolyte) that hampered 2e⁻ ORR. For the S²⁻-oxidation unit, the dissolved oxygen underwent 2e⁻ ORR at the electrode-electrolyte interface, with the generated H₂O₂ participating in S²⁻ oxidation. The electrode was fabricated by depositing OCNT electrocatalyst on a gas diffusion layer (OCNT/GDL) with an area of 1 cm². Meanwhile, oxygen evolution reaction (OER) took place at a Pt foil counter electrode. We expected that the influent S²⁻ ions would undergo deep oxidation process by H₂O₂ generated at cathode. The oxidation products (S₂O₃²⁻, SO₃²⁻ and SO₄²⁻) and unreacted S²⁻ were then recirculated to the tank for further oxidation. As a result, H₂S capture and conversion were operated in a batch manner.

Fig. 3: Deep H₂S oxidation via in-situ generated H₂O₂ at cathode in H-cell.

a Schematic illustration of the electrochemical setup for EChem-SOR in H-cell separated by a bipolar membrane (BPM). H₂S_cap conversion (b), productivity of oxidation products (c), SO₄²⁻ selectivity (d), and e⁻-to-sulfate efficiency (e) for CNT and OCNT without electrolysis or at different applied potentials, with feeding 5% H₂S/Ar at a flow rate of 20 ml min⁻¹ for 2 h. The error bars represent the standard deviation of three independent measurements.

As shown in Fig. 3b, c, H₂S oxidation took place with low reactivity in the absence of an applied potential, which can be attributed to the limited oxidative capability of O₂ for deep H₂S oxidation, as discussed in the Pourbaix diagram (Fig. 2a). By contrast, when cathodic potential was applied, H₂S_cap conversion and total productivity of three oxidation products (S₂O₃²⁻, SO₃²⁻ and SO₄²⁻) were largely promoted, demonstrating the feasibility of using in-situ generated H₂O₂ as the oxidant for H₂S-to-SO₄²⁻ conversion. In addition, OCNT exceeded CNT by showing higher H₂S_cap conversion and total productivity at various potentials. These results demonstrate that achieving higher activity for H₂O₂ electrosynthesis (by applying a more negative potential or utilizing a more active 2e⁻ ORR electrocatalyst) is important for promoting H₂S conversion.

Nevertheless, we recognized that the efficiency of deep H₂S oxidation in H-cell requires substantial improvement, considering that H₂S_cap conversion was less than 45% (Fig. 3b) and SO₄²⁻ selectivity was less than 25% (Fig. 3d) for both CNT and OCNT. In addition, an e⁻-to-sulfate efficiency (%) metric was defined to evaluate the utilization efficiency of electron for SO₄²⁻ production. The calculation results showed that e⁻-to-sulfate efficiency in H-cell (for both CNT and OCNT) are less than 25% (Fig. 3e). We speculated that H₂O₂ concentration in H-cell may be insufficient, considering the limited current and H₂O₂ dilution in the bulk solution, calling for more advanced reaction system for EChem-SOR.

EChem-SOR in a 4-cm² flow reactor

Based on H₂O₂-concentration dependence of H₂S oxidation (in Chem-SOR) and the limited deep H₂S oxidation in H-cell (in EChem-SOR), we deduce that the efficiency of EChem-SOR system in valorization of H₂S to SO₄²⁻ is significantly reliant on the enhancing H₂O₂ concentration. Therefore, we constructed a 4-cm² flow reactor for EChem-SOR system (Supplementary Fig. 11), aiming at enhancing H₂O₂ concentration to further improve H₂S oxidation performance.

The schematic of the flow reactor (Fig. 4a), the enlarged view of cathode chamber (Fig. 4b) and all the possible chemical and electrochemical reactions (Fig. 4c) are shown. The H₂S-to-SO₄²⁻ pathway is depicted below: (1) H₂S/Ar mixing gas (20 ml min⁻¹) and O₂ (100 ml min⁻¹) was together fed into the cathode side (Fig. 4b), migrated through GDL and reached the three-phase interface (between electrode, electrolyte, and gas). (2) O₂ immediately underwent 2e⁻ ORR over OCNT electrocatalyst (Eq. 1 in Fig. 4c), with the generation of high concentration of HO₂⁻ and OH⁻ locally. Due to gaseous O₂ was fed in this flow reactor, presumably showing higher efficiency than the use of water-dissolved O₂ in aforementioned H-cell, the interference of co-feeding Ar can be neglected. (3) H₂S was captured by the OH⁻ generated at the interface and in the circulating electrolyte (1 M KOH), and was dissociated into S²⁻ ions (Eq. 2). (4) Subsequently, the S²⁻ ions were oxidized by in-situ generated HO₂⁻ to form SO₄²⁻ (Eq. 3), which then flowed out as a K₂SO₄ solution. (5) Additionally, water dissociation reaction (Eq. 4) on bipolar membrane (BPM) and subsequent protonation reaction (Eq. 5) at cathodic side can prevent the S²⁻ ions and oxidation products from migrating to anodic side, thus sulfur loss can be avoided (Supplementary Fig. 12). As a result, H₂S capture and conversion can be operated in a continuous manner.

Fig. 4: Engineering strategies to improve deep H₂S oxidation performance in a 4-cm² flow reactor.

a Schematic illustration of the electrochemical setup for EChem-SOR in flow reactor separated by a BPM. The cathode is OCNT/GDL electrode with an electrode area of 4 cm² and a catalyst loading of 1 mg cm⁻². The anode is a Ti-based phase supported IrRu oxide (commercial IrRuO_x/Ti) electrode. b Enlarged structural diagram and reaction pathway on the cathodic side. c Summary of electrochemical and chemical reactions occurring on the cathodic side and BPM. H₂S_cap conversion (d), productivity of oxidation products (e), SO₄²⁻ selectivity (f), and e⁻-to-sulfate efficiency (g) for OCNT without electrolysis or at different applied currents, with feeding 5% H₂S/Ar at a flow rate of 20 ml min⁻¹ for 2 h. The error bars represent the standard deviation of three independent measurements.

The performance of H₂O₂ electrosynthesis over OCNT in flow reactor was first evaluated by only feeding O₂ gas. The results showed that a high H₂O₂ FE over 90% was obtained under absolute current between 100 and 400 mA, with H₂O₂ productivity and accumulated concentration increasing with current (Supplementary Fig. 13a–c). Compared with preceding H-cell, the flow reactor delivered significantly higher H₂O₂ concentration over the same OCNT electrocatalyst (Supplementary Fig. 13d), with the maximum H₂O₂ concentration increasing from 0.12 wt% (in H-cell) to 0.98 wt% (in flow reactor). This performance is comparable to that reported in other electrochemical systems under similar experimental conditions (Supplementary Table 1), indicating a good H₂O₂ synthesis performance of the flow reactor system. Moreover, even under identical electrolysis conditions (current density: 50 mA cm⁻², electrolyte volume: 50 ml, electrolysis time: 2 h), the flow reactor exhibited higher H₂O₂ concentration compared to the H-cell (Entry 1 and 2 in Supplementary Table 2). This enhancement correlates with both the enlarged electrode area and improved H₂O₂ FE in the flow reactor. Notably, the effect of FE on H₂O₂ concentration becomes more pronounced at elevated current densities (Entry 3 and 4 in Supplementary Table 2). This originates from the efficient O₂ mass transport in flow reactor configuration, which enables maintenance of high H₂O₂ FE even at increased current densities.

Subsequently, the performance of in-situ H₂O₂ generation coupled with H₂S oxidation was studied (Fig. 4d–g). In the absence of an applied current, the captured H₂S can still be partially oxidized by O₂, with over 70% being converted to S₂O₃²⁻ (Fig. 4d, e). Upon applying a constant current, H₂S_cap conversion as well as SO₄²⁻ productivity and selectivity significantly increased. With continuous increase in applied current, the H₂S-to-SO₄²⁻ conversion was further enhanced, achieving a H₂S_cap conversion of 90%, a SO₄²⁻ selectivity of 69%, and a maximum e⁻-to-sulfate efficiency of 80% under 200 mA. Moreover, compared to the H-cell, the flow reactor exhibited enhanced deep oxidation performance under the same constant-current density condition (Supplementary Fig. 14), which can be attributed to the higher H₂O₂ concentration in the bulk solution and at the cathode interface within the flow reactor system.

To verify the interfacial H₂O₂ concentration effect, two different H₂S oxidation experiments were conducted (Supplementary Fig. 15a, b): (1) EChem-SOR method (that is integrated oxidation): Simultaneous 2e⁻ ORR electrolysis and H₂S oxidation by introducing O₂ and H₂S into the flow reactor during electrolysis; (2) Electrochemical-chemical tandem method: H₂O₂ is first electrochemically generated via 2e⁻ ORR, followed by H₂S oxidation using the resulting homogeneous H₂O₂ solution without applying electrolysis. The results show that the EChem-SOR method achieves superior H₂S oxidation performance compared to the tandem method under all tested currents (Supplementary Fig. 15c). To eliminate potential H₂O₂ decomposition effects in tandem method, we also conducted additional short-duration tests, which confirmed the similar observed trend (Supplementary Fig. 15d). Moreover, the interfacial concentration advantage becomes more pronounced at lower flow rate of electrolyte (Supplementary Fig. 16). Furthermore, the COMSOL model were performed to visualize this concentration gradient (Supplementary Figs. 17 and 18), revealing that the H₂O₂ concentration near the electrode surface is much higher than in the bulk solution. Moreover, under elevated currents and reduced electrolyte flow rates, the interfacial H₂O₂ concentration was further enhanced (Supplementary Fig. 18f, l), which is consistent with the experimental observations. Specifically, the H₂O₂ concentration can reach up to 1.04 wt% within 10 μm from the electrode at 200 mA and 10 ml min⁻¹ of electrolyte, which is more than 86 times higher than the bulk concentration (within 200 μm). These findings prove that a pronounced interfacial H₂O₂ concentration effect in flow reactor, making this system more suitable for scaled-up in-situ H₂O₂ applications.

Valorization of H₂S in natural gas to K₂SO₄

The above promising results (over 90% H₂S_cap conversion and 90% SO₄²⁻ selectivity) of EChem-SOR in 4-cm² flow reactor encourages us to further evaluate its performance in desulfurization from H₂S-involved natural gas (H₂S/CH₄), with the designed processing shown in Fig. 5a. To prevent potential operation risks and CH₄ oxidation reaction induced by H₂O₂, as well as to separate CH₄ from H₂S for downstream utilization, the H₂S/CH₄ gas was first fed into a gas−liquid separator containing 1 M KOH solution, where H₂S was captured in the form of S²⁻ and the low-soluble CH₄ was expelled upward. Benefiting from the high capture rate of base solutions for acid gas^30,31, H₂S can be almost completely absorbed by KOH solution (Supplementary Fig. 19).

Fig. 5: EChem-SOR for valorization of H₂S in natural gas using a 100-cm² flow reactor.

a Schematic illustration of continuous H₂S capture and K₂SO₄ production from a H₂S-involved (100,000 ppm) natural gas via EChem-SOR system that enables nearly complete capture H₂S and highly efficient K₂SO₄ production. The system comprises a simulated natural gas, a flow reactor, a gas-liquid separator filled with 1 M KOH, and a crystallizer. b A 100-h stability test of EChem-SOR system using 100-cm² flow reactor at 5 A and 300 ml min⁻¹ of 10 vol% H₂S/CH₄, with refreshing the catholyte daily for maintaining the alkaline environment to capture H₂S. H₂S_cap conversion (left in c), H₂S concentration in tail gas (right in c), K₂SO₄ selectivity (left in d), e⁻-to-sulfate efficiency (left in d), and accumulated K₂SO₄ mass (right in d) detected at different time points.

With the feed gas change from 5% H₂S/Ar to 5% H₂S/CH₄, the H₂S oxidation performance at the same current (400 mA) was maintained (Supplementary Fig. 20). And no CH₄ oxidation products were detected in catholyte. These results indicate the generated H₂O₂ in EChem-SOR system is incapable of activating CH₄, which is different from the recent reported systems where CH₄ oxidation took place, which is owing to the utilization of acid conditions and the generation of oxygen radical species in their studies^25,32. On the other hands, the effect of trace amounts of CO₂ (1000 ppm), another prevalent impurity gas in raw natural gas, was also considered (Supplementary Fig. 21). The results indicate the system’s operation and H₂S treatment performance were largely unaffected, while the introduced CO₂ is almost completely converted into carbonate ions under alkaline environment of the system.

Moreover, the continuous H₂S oxidation was preliminarily conducted using 4-cm² flow reactor at 400 mA, showing that the EChem-SOR system maintained stable operation over a 40-hprolonged test, with cell voltage remained stable between 2.34 and 2.82 V (Supplementary Fig. 22a). Simultaneously, the products were analysis at intervals, achieving the stable H₂S_cap conversion (88–93%), SO₄²⁻ selectivity (88–93%), and e⁻-to-sulfate efficiency (55–61%) throughout the duration of the experiment. The post-stability characterization confirmed the well-preserved structure of both the cathode (OCNT/GDL) and anode (IrRuO_x/Ti) after continuous electrolysis (Supplementary Fig. 22b, e–g), with no notable transformation observed. And the post-OCNT/GDL retained >85% of its initial current density at 2 V while showing negligible decay (<3%) in H₂O₂ production performance (Supplementary Fig. 22c, d). This performance demonstrated the applicability of EChem-SOR system in H₂S valorization.

Furthermore, to evaluate the performance of EChem-SOR system in application scenarios with other typical H₂S concentration levels (Supplementary Table 2), the H₂S concentration in feed gas was adjusted to 10,000 ppm and 100,000 ppm, respectively, while a flow rate of 20 ml min⁻¹ was maintained. The results indicate that the system maintains high H₂S_cap conversion (>85%) and K₂SO₄ selectivity (>80%, Supplementary Fig. 23a), demonstrating its broad applicability. Furthermore, by increasing the H₂S influx with higher H₂S flow rate (>20 ml min⁻¹, Supplementary Fig. 23b), K₂SO₄ production was promoted with higher e⁻-to-sulfate efficiency. However, both H₂S_cap conversion and K₂SO₄ selectivity showed gradual decreases, indicating that there were more unconverted S²⁻ and accumulated S₂O₃²⁻. This may be attributed to a mismatch between the limited generation rate of H₂O₂ and the required high oxidation rate of H₂S under increased H₂S influx, resulting in a decline in H₂S processing capacity.

EChem-SOR in a 100-cm² flow reactor

To increase the H₂S processing capability (high ratio and high flow rate of H₂S), the flow reactor was further scaled up by increasing electrode working area from 4 cm² to 100 cm² (Supplementary Fig. 24), with the aim of increasing H₂O₂ productivity to match high influx of H₂S. Compared to the preceding 4-cm² flow reactor, the 100-cm² one showed obvious advantages for treating high flow rate of H₂S/CH₄ (>100 ml min⁻¹; Supplementary Fig. 25). At a flow rate of 100 ml min⁻¹, a superior performance was achieved at 3 A, with high H₂S_cap conversion (87%), K₂SO₄ productivity (1.7 g h⁻¹), K₂SO₄ selectivity (73%), e⁻-to-sulfate efficiency (76%). At even higher flow rate of 300 ml min⁻¹, the current is optimized to be 5 A, leading to high H₂S_cap conversion (93%), K₂SO₄ productivity (4.5 g h⁻¹), K₂SO₄ selectivity (72%), and e⁻-to-sulfate efficiency (84%). These results reveal that a balance between the H₂O₂ production and H₂S oxidation is achievable during scaling up (with higher electrode area) by balancing the applied current and flow rate of H₂S/CH₄.

We then evaluated stability of EChem-SOR system in the 100-cm² flow reactor by employing continuous operation at 5 A for 100 h, with H₂S/CH₄ feeding at 300 ml min⁻¹. During the operation, the cell voltage remained stable (between 2.52 and 2.94 V; Fig. 5b). According to the liquid products and tail gas analysis result at intervals (Fig. 5c, d), the reaction system showed efficient H₂S capture (H₂S concentration decrease from 100,000 ppm to <15 ppm, H₂S_cap conversion: >90%) and valorization (K₂SO₄ selectivity: 70%, e⁻-to-sulfate efficiency: 80%) abilities. At the end of the 100-h test, an accumulation of 397 g K₂SO₄ was produced, as confirmed by the XRD measurement result (Supplementary Fig. 26).

Collectively, by a series of engineering strategies (reactor design and gas feeding manner), we have demonstrated the capacity and stability of EChem-SOR system for effectively capturing H₂S from natural gas and producing K₂SO₄ at large scale. As summarized in Table 1, the initial H-cell in Fig.3a featured a working electrode area of 1 cm² and operated in batch manner. Due to the limited O₂ mass transfer to the two-phase interface (electrode−electrolyte), the achievable current was relatively low (56 mA), resulting in a low H₂O₂ productivity and moderate H₂S oxidation performance (H₂S_cap conversion: 45%, K₂SO₄ selectivity: 21%, e⁻-to-sulfate efficiency: 20%). To construct an efficient three-phase interface (electrode−electrolyte−gas), a 4-cm² flow reactor with continuous operation manner and increased electrode area was employed in Fig.4a, enabling the formation of high concentration of H₂O₂ at the cathodic interface, thereby achieving higher H₂S oxidation performance (H₂S_cap conversion: 98%, K₂SO₄ selectivity: 90%, e⁻-to-sulfate efficiency: 61%). Furthermore, for treating simulated natural gas with higher H₂S influx, the flow reactor was scaled up to 100 cm² and operated at higher current in Fig. 5a. Consequently, a superior H₂S treatment capacity (flux up to 300 ml min⁻¹), the maintained H₂S oxidation performance (H₂S_cap conversion: 93%, K₂SO₄ selectivity: 72%, e⁻-to-sulfate efficiency: 84%), and larger-scale K₂SO₄ production (4.5 g h⁻¹) were achieved. Even with normalized H₂S influx per unit area (3 ml min⁻¹ cm⁻²), the 100-cm² flow reactor performed comparably to the 4-cm² flow reactor and outperformed the H-cell (Supplementary Fig. 27). Additionally, long-term operation stability of 100 h was also achieved. These achievements highlight the versatility and potential of EChem-SOR system using flow reactor for future applications in natural gas desulfurization.

Table 1 Key metrics comparison of three reactors in EChem-SOR system

Sustainability and economic evaluation of EChem-SOR

Based on the potential application of the developed EChem-SOR system in H₂S capture and K₂SO₄ production, we performed a life-cycle assessment (LCA) to evaluate its sustainability (Supplementary Note 2). The model was illustrated (Supplementary Fig. 28), and the base case was according to the performance in 100-cm² flow reactor, with the calculated various environmental implications being listed (Supplementary Table 3). The result showed that global warming potential (GWP) was the most notable environmental matric that was influenced by EChem-SOR system, which was contributed by electrochemical reaction, separation and H₂S production processes. Specifically, the GWP of electrochemical reaction mainly originated from the electricity consumption thus is affected by different electricity sources (Supplementary Fig. 29a, b). The GWP of separation process is from the evaporation of water in crystallizer and evaporator, thus applying KOH electrolyte with higher concentration would reduce associated energy and thus GWP (Supplementary Fig. 29c, d). Moreover, the GWP contributed by H₂S in the based case was assumed from its production process, whereas it can be reduced to zero in actual implementation by using waste H₂S from natural gas (as demonstrated in the above study). Although the GWP values of base case during K₂SO₄ production and H₂S removal are higher than those of industrial processes (shown by the dotted line in Fig. 6a, b and discussed in Supplementary Note 2), a more competitive target case with progressively reduced GWP values can be expected by a series of optimizations (Fig. 6a, b). The optimizations include utilizing renewable electricity powered by wind, increasing KOH concentration (to reduce water evaporation energy) and changing H₂S resource to H₂S-involved natural gas (H₂S production process and associated energy would be excluded). As a result, the target case delivers a GWP value of 0.4 kg CO₂e per kg K₂SO₄ and 2.7 kg CO₂e per kg H₂S, indicating that EChem-SOR system was potentially more sustainable than industrial K₂SO₄ production and H₂S removal processes.

Fig. 6: Sustainability and economic evaluation of EChem-SOR system.

Roadmap to reducing GWP for K₂SO₄ production (a) and H₂S removal (b) in the base case through progressive optimization of carbon emission-relevant parameters. c Sensitivity analysis of various single-variables (current density, e⁻-to-sulfate efficiency, electricity price and KOH price) on the production cost per kg K₂SO₄. The base point of relative change was chosen at a current density of 50 mA cm⁻², e⁻-to-sulfate efficiency of 84%, electricity price of 0.03 $ kWh⁻¹, KOH price of 0.38 $ kg⁻¹. d A contour map for the total K₂SO₄ production costs in EChem-SOR system with respect to the different current densities and e⁻-to-sulfate efficiencies. The dotted line represents the breakeven line at current market price of 0.5 $ per kg K₂SO₄. The direction of the arrow represents the profitable region. All other parameters were fixed as the base point parameters in (c).

Subsequently, we conducted a techno-economic analysis (TEA) to assessed the economic viability of EChem-SOR system for K₂SO₄ production (Supplementary Note 3). The K₂SO₄ production costs include capital, operating, and material costs, in which current density, e⁻-to-sulfate efficiency, electricity price, and KOH price are the main factors (Supplementary Fig. 30). A single-variable sensitivity analysis of the above four factors on production costs were performed (Fig. 6c). The results showed that a reduction in current density and e⁻-to-sulfate efficiency, and substantial increases in electricity and KOH prices may lead to unprofitability. Furthermore, we calculated the total K₂SO₄ production cost at variable current density and e⁻-to-sulfate efficiency (Fig. 6d). The results revealed that higher current density coupled with improved e⁻-to-sulfate efficiency is vital for profitability of EChem-SOR system, since the H₂O₂ production rate and concentration are improved accordingly, thereby deep H₂S oxidation to K₂SO₄ can be facilitated. Specifically, based on the high performance of EChem-SOR system in 100-cm² flow reactor (with an e⁻-to-sulfate efficiency of 84% at 50 mA cm⁻²), we envisage a profit margin (with total production cost of 0.44 $ and market price of 0.5 $ per kg K₂SO₄) for K₂SO₄ production. Furthermore, we compared the economic viability of the EChem-SOR route with the current ASOR route, which can be coupled with cathodic H₂ production (Supplementary Note 4). The results show that although the ASOR system achieves a lower S and H₂ production cost (0.31 $ per kg S) compared to the EChem-SOR route, the AOR system reaches the profitability region only under higher current density and S Faradaic efficiency (Supplementary Fig. 31). This demonstrates that the promising profitability potential of the EChem-SOR system.

Extending EChem-SOR for H₂SO₄ solution production

Besides the above promising results for K₂SO₄ production from H₂S by using in-situ generated H₂O₂, we conceived to expand the application of EChem-SOR system to produce other SO₄²⁻-involved valuable chemicals. Benefiting of the progress in solid-electrolyte (SE) flow reactor^5,30,33, we sought to produce sulfuric acid (H₂SO₄) solution under electrolyte-free conditions, by introducing solid electrolyte into the middle layer and feeding pure H₂O as a circulating solution in both middle SE layer and anodic chamber (Supplementary Fig. 32a). Notably, due to the weak absorption capacity of pure H₂O for H₂S, we are able to feed H₂S/CH₄ gas along with the circulated H₂O into the middle SE layer (Supplementary Fig. 32b), which is conducive to H₂S capture by OH⁻ that was generated from 2e⁻ ORR. The captured H₂S were then oxidized to SO₄²⁻ by HO₂⁻ that migrated from the cathodic interface, followed by combination with H⁺ that migrated from the anodic side to eventually yield outflowing H₂SO₄ solution (Fig. 7a).

Fig. 7: Extending EChem-SOR system for H₂SO₄ solution production using a 100-cm² SE flow reactor.

a Schematic illustration of the reaction pathway in SE flow reactor for H₂SO₄ production. The SE flow reactor comprises an OCNT/DGL cathode, IrRuO_x/Ti anode, anion exchange membrane (AEM), proton exchange membrane (PEM), and middle SE layer filled with H⁺-conducting ion exchange resin. b H₂SO₄ production and e⁻-to-sulfate efficiency of EChem-SOR system at 5 A with different H₂S/CH₄ flow rate. c Long-term stability test at 5 A and 60 ml min⁻¹ of 10% H₂S/CH₄. After every 10-h electrolysis (50 h in total), the SE layer was chemically rinsed by Na₂SO₃ solution to remove deposited S without the need to disassemble the reactor (for 2 h, without applying a current). The right axis represents the corresponding H₂SO₄ concentration and e⁻-to-sulfate efficiency in each electrolysis period.

We first examined the H₂SO₄ production performance at 5 A and different H₂S/CH₄ flow rates. As the flow rate increased, the H₂SO₄ production and e⁻-to-sulfate efficiency improved but with reducing trend (Fig. 7b). This suggests the amount of H₂S captured by OH⁻ slowly reached its maximum under this condition. Notably, S₂O₃²⁻ was not detected in all the tests, even at lower currents but only SO₄²⁻ (Supplementary Fig. 33a, b). Under such conditions, we expected S₂O₃²⁻ might be accumulated. Note that the middle circulation solution became acidic and sulfur deposition was observed in the middle SE layer (Supplementary Fig. 33c, d). Therefore, we speculated that the intermediate S₂O₃²⁻ was not only oxidized to SO₄²⁻, but also underwent a disproportionation reaction in the presence of H⁺ with the formation of sulfur (Fig. 7a). To address the potential impact of sulfur deposition on system operation, we rinsed the middle SE chamber of the reactor with Na₂SO₃ solution (without applying a current) every 10 h of electrolysis (Supplementary Fig. 34). This approach utilizes the comproportionation reaction between Na₂SO₃ and sulfur to generate Na₂S₂O₃ (SO₃²⁻+ S → S₂O₃²⁻), effectively removing deposited sulfur, which extends the system’s operating time without the need to disassemble the reactor. Note that the recovered S₂O₃²⁻ was not included in the final H₂SO₄ calculation. As a result, the EChem-SOR system achieved stable operation for over 50 h in total at 5 A and 60 ml min⁻¹ of 10% H₂S/CH₄ (Fig. 7c), with efficient H₂SO₄ production (1.4 wt% concentration and 30% e⁻-to-sulfate efficiency), demonstrating its potential for long-term H₂SO₄ solution production.

Discussion

Overall, we developed an electrochemical method for simultaneous removal of high-concentration H₂S from natural gas and production of value-added K₂SO₄ or H₂SO₄, via coupling cathodic in-situ H₂O₂ generation and deep H₂S oxidation.

We performed a progressive study to demonstrate the above concept: (1) Showing the ability of deep H₂S oxidation to SO₄²⁻ using commercially available H₂O₂ via Chem-SOR; (2) Preliminarily demonstrating the feasibility of EChem-SOR in a H-cell; (3) Designing a 4-cm² flow reactor with three-phase interface to generate interfacial high-concentration H₂O₂, showing over 90% H₂S_cap conversion and 90% K₂SO₄ selectivity; (4) Scaling up the flow reactor to 100 cm² with stable IrRuO_x/Ti anode, achieving efficient H₂S capture (from 100,000 to below 15 ppm), large-scale K₂SO₄ production (72% selectivity and high productivity of 4.5 g h⁻¹) and good stability (100 h); (5) Demonstrating the sustainability (in terms of environmental implications and carbon emission intensity) and profitability (in terms of K₂SO₄ production) of EChem-SOR system; (6) Extending the product of H₂S conversion to H₂SO₄ in a solid-electrolyte-based flow reactor, with the stable production of 1.4 wt% H₂SO₄ solution in each electrolysis cycle.

We envision that this system may show broad opportunity for the treatment of other sulfur-containing industrial waste gases and waters. Moreover, more application scenarios for in-situ H₂O₂ utilization are expected to expand in the future. Through development of highly efficient catalysts, innovative reactor design and optimization of reaction systems, more sustainable and economic route for sulfur-containing wastes valorization are expected.

Methods

Chemicals

All chemicals were used as purchased without further purification. Deionized water was used throughout the experiments. Potassium hydroxide (KOH, 99.99%) and sodium hydroxide (NaOH, 97%) were purchased from Macklin Reagent Co. Ltd (Shanghai, China). Hydrogen peroxide (H₂O₂, 30 wt%), concentrated nitric acid (HNO₃, 68 wt%), concentrated sulfuric acid (H₂SO₄, 98 wt%) were TongGuang Fine Chemicals Co. Ltd (Beijing, China). Sodium sulfate (Na₂SO₄, 99%), sodium sulfite (Na₂SO₃, 98%), sodium thiosulfate (Na₂S₂O₃, 99%), sodium hydrosulfide (NaHS, 70%), cerium sulfate (Ce(SO₄)₂, 99%), N, N-dimethyl-p-phenylenediamine dihydrochloride (C₈H₁₂N₂·2HCl, 96%) and ammonium iron sulfate dodecahydrate (NH₄Fe(SO₄)₂·12H₂O, 99.95%) were purchased from Aladdin Reagent Co. Ltd (Shanghai, China). Carbon nanotube (CNT, 99%) and Nafion solution (5 wt%) were purchased from Sigma-Aldrich Chemical Reagent Co. Ltd (Shanghai, China). Binary mixture of hydrogen sulfide and argon (5% H₂S/Ar), binary mixture of hydrogen sulfide and methane (5% H₂S/CH₄ and 10% H₂S/CH₄) and oxygen (O₂, 99.999%) were purchased Air Liquide Co. Ltd (Tianjin, China). Gas diffusion layer (GDL, YLS-30T), bipolar membrane (FBM-PK), anion exchange membrane (FAA-3-50) and proton exchange membrane (Nafion 117) were purchased from Shengernuo Co. Ltd. (Suzhou, China). Solid electrolyte (Amberlite IR-120, hydrogen form) was purchased from Alfa Aesar Co. Ltd. (Shanghai, China).

Synthesis of OCNT

OCNT was prepared by heat treating pristine CNT with HNO₃ solution. Typically, 0.1 g of pristine CNT was added to 100 ml of 13 M HNO₃ solution. The mixed solution was heated to 80 °C in an oil bath and stirred for 12 h under reflux condensation. After cooling to room temperature, the solution was centrifuged and washed with water until the supernatant became neutral. The precipitate was then dried in an oven at 60 °C overnight to obtain OCNT electrocatalyst. HNO₃-6h and HNO₃-24h were prepared following the same procedure as OCNT, except that the heat treatment time were changed to 6 and 24 h, respectively.

Preparation of electrodes

To maintain consistency between the tests conducted in H-cell and flow cell, and to avoid introducing additional variables that could affect the electrochemical synthesis of H₂O₂, the same substrate (i.e., GDL) was utilized to support OCNT. For preparation of OCNT/GDL electrode with different working area (1, 4, and 100 cm²), typically, a catalyst ink with a concentration of 2 mg ml⁻¹ was first prepared through mixing electrocatalyst, ethanol, and Nafion solution. The ink was sonicated for 1 h to obtain a homogeneous mixture and then spray-coated onto the GDL using an ultrasonic spraying system (SCP101, Shengernuo Co. Ltd., Suzhou, China). After drying, the OCNT/GDL electrode with a catalyst loading of 1 mg cm⁻² was obtained. The RRDE-loaded electrocatalysts were prepared by dropping 13 µl of catalyst ink with a concentration of 2 mg ml⁻¹ onto the disk electrode (0.247 cm²) of RRDE, and then drying at room temperature, resulting in a catalyst loading of 0.1 mg cm⁻².

Materials characterization

Scanning electron microscopy (SEM) images were obtained by ApreoC microscopy. High-resolution transmission electron microscopy (HR-TEM) images were obtained by FEI Tecnai F20 microscopy. X-ray diffraction (XRD) patterns were recorded on Bruker D8-Focus using Cu Kα radiation in the range from 10° to 60° at 40 kV and 40 mA. Raman spectroscopy was recorded on the confocal Raman microscopy (HORIBA LabRAM HR Evolution) with signals excited by 532 nm (50 mW) laser. Elemental analysis (EA) was performed on the Unicube Elemental Analyzer from Elementar, Germany. X-ray photoelectron spectroscopy (XPS) spectra were performed on on ESCALAB Xi+ photoelectron spectrometer. Fourier transform infrared spectroscopy (FT-IR) spectra were recorded on Vertex 70 FT-IR Spectrometer. In-situ FTIR spectra were collected on a TENSOR II FT-IR (NETZSCH GERAETEBAU GMBH). The experiments were conducted in 0.1 M KOH solution with O₂ introduced into the custom built three-electrode PTFE cell, with OCNT/GDL or CNT/GDL coated ZnSe prism as the working electrode, Pt wire as the counter electrode and Hg/HgO as the reference electrode. Each spectrum was collected with an acquisition time of 100 s at a potential interval of 100 mV from OCP or 0.9 V vs RHE.

Electrochemical 2e⁻ ORR measurements and H₂O₂ production

The 2e⁻ ORR performance of electrocatalysts were measured using a rotating ring-disk electrode (RRDE) in a three-electrode system. The RRDE-loaded electrocatalyst, graphitic rod and Hg/HgO electrode were used as the working electrode, counter electrode and reference electrode, respectively. All the potentials measured against Hg/HgO (E_Hg/HgO) were converted to the reversible hydrogen electrode (E_RHE) scale in this work using E_RHE = E_Hg/HgO + 0.0591 × pH +  0.098 V. Unless otherwise specified, all curves are reported without iR compensation. The solution resistance was determined by electrochemical impedance spectroscopy technique at open circuit potential in a frequency range from 105 to 1 Hz with a perturbation of 5 mV. The linear scan voltammetry (LSV) curve was measured from 1 to 0.2 V vs RHE at 10 mV s⁻¹ and 1600 rpm in O₂-saturated electrolyte, while the ring potential of RRDE was fixed at 1.23 V vs RHE. The H₂O₂ partial current density (j_H2O2) and selectivity (%) were calculated as follows:

$${j}_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}=\frac{{i}_{{{\rm{ring}}}}}{N\times {S}_{{{\rm{disk}}}}}$$

(1)

$${{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}{{\rm{selectivity}}}\left(\%\right)=\frac{200\times {i}_{{{\rm{ring}}}}/N}{{i}_{{{\rm{disk}}}}+{i}_{{{\rm{ring}}}}/N}$$

(2)

where i_ring is the measured ring current, i_disk is the measured disk current, S_disk is the area of disk electrode (0.247 cm²), and N is collection efficiency (37%).

The electrochemical H₂O₂ production in H-cell was performed using chronoamperometry with a CS1350 electrochemical analyzer (CS Instruments, Inc.,Wuhan). OCNT/GDL (1 cm²), Pt foil and Hg/HgO electrode were used as the working electrode, counter electrode and reference electrode, respectively. The working and counter chambers were both filled with 0.1 M KOH solution and separated by a BPM. 100 ml min⁻¹ of O₂ was continuously bubbled into the working chamber. The electrochemical H₂O₂ production in flow reactor was performed using chronopotentiometry. OCNT/GDL (4 cm²) and IrRuO_x/Ti (4 cm²) electrode were used as the cathode and anode, respectively. The cathodic and anodic chamber were separated by a BPM. 1 M KOH solution was circulated in both chambers at a rate of 10 ml min⁻¹.

The H₂O₂ concentration was measured by Ce(SO₄)₂ titration method based on the colorimetric reaction between Ce⁴⁺ and H₂O₂. After electrolysis, 50 µl of catholyte was mixed with 10 ml 0.5 mM Ce⁴⁺ solution. The Ce⁴⁺ concentration can be detected by ultraviolet-visible spectrophotometer (UV-Vis, UV-3600, Shimadzu) at 318 nm (Supplementary Fig. 35). The H₂O₂ FE, productivity and accumulated concentration can be calculated according the following equations:

$${{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}\,{{\rm{FE}}}\,\left(\%\right)=\frac{{N}_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}\times F\times {C}_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}\times V}{Q}$$

(3)

$${{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}\,{{\rm{productivity}}}\left({{\rm{mmol}}}{{{\rm{h}}}}^{-1}\right)=\frac{{C}_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}\times V}{t}$$

(4)

$${{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}\,{{\rm{concentration}}}\,\left({{\rm{wt}}}\%\right)=\frac{100\times {C}_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}\times {M}_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}}{{\rho }_{{{{\rm{H}}}}_{2}{{{\rm{O}}}}_{2}}}$$

(5)

where N_H2O2 is the electron transfer number of 2e⁻ ORR, F is the Faraday constant, C_H2O2 is the measured concentration of H₂O₂, V is the volume of catholyte, Q is total charge through electrolysis, t is the electrolysis time, M_H2O2 is the molar mass of H₂O₂, and ρ_H2O2 is the density of H₂O₂ solution at room temperature.

Chemical oxidation of H₂S to SO₄ ²⁻

The chemical H₂S oxidation reaction was conducted by using prepared H₂O₂ solution to oxidize the captured H₂S. Typically, 50 ml of 0.1 M KOH solution was first prepared for the capture of H₂S gas, a certain volume (0–1.5 ml) of commercial H₂O₂ solution (30 wt%) was then added to the KOH solution to obtain defined-concentration H₂O₂ solutions (0–1 wt%). Subsequently, 5% H₂S/Ar mixing gas was fed into the above solution with a flow rate of 20 ml min⁻¹, and the tail gas was absorbed using a 3 M NaOH solution. After aeration for 2 h, the unreacted S²⁻ in the solution were detected through the methylene blue spectrophotometric method using UV-Vis spectrophotometer, and the liquid oxidation products (S₂O₃²⁻, SO₃²⁻ and SO₄²⁻) of H₂S were analyzed through an ion chromatography (IC, ICS-5000, Thermo Fisher Scientific) equipped with a conductivity detector (CD). The quantitative methods are provided in the later section. The solid oxidation product (S) of H₂S was collected by centrifuging and drying the post-reaction solution, and its content was determined by weighing.

Electrochemical oxidation of H₂S to K₂SO₄

The electrochemical H₂S oxidation to produce K₂SO₄ was carried out by utilizing in-situ generated H₂O₂ at cathode to oxidize the captured H₂S in KOH electrolyte. The electrolysis in H-cell and flow reactor were performed with a CS1350 electrochemical analyzer (CS Instruments, Inc., Wuhan). For the experiments in H-cell, 50 ml of 0.1 M KOH solution was circulated between the H₂S-capture tank and the cathode chamber at a flow rate of 10 ml min⁻¹. The anode chamber contained 30 ml of 0.1 M KOH solution, separated from the cathode chamber by a BPM. OCNT/GDL (1 cm²), Pt foil and Hg/HgO electrode were used as the working electrode, counter electrode and reference electrode, respectively. During electrolysis, the chronoamperometry test was performed at a constant potential. Meanwhile, O₂ was fed into the cathode chamber and 5% H₂S/Ar was fed into the H₂S-capture tank, with the tail gas being absorbed by a 3 M NaOH solution. After electrolysis, the unreacted S²⁻ and oxidation products were detected using UV-Vis and ion chromatography, respectively.

For the experiments performed in a 4-cm² flow reactor, OCNT/GDL (4 cm²) and commercial IrRuO_x/Ti (4 cm²) electrode were used as the cathode and anode, respectively. The cathode and anode chamber were separated by a BPM, with 50 ml of 1 M KOH solution circulating on each side at a flow rate 10 ml min⁻¹. During electrolysis, the chronopotentiometry test was performed at a constant current. O₂ and H₂S/Ar were together fed into the cathode chamber, part of the mixing gas passed through the GDL to participate in the reaction, while the remaining gas was recirculated into the KOH circulation solution for residual H₂S capture. After electrolysis, the unreacted S²⁻ and oxidation products were detected using UV-Vis and ion chromatography, respectively.

For the experiments performed in a 100-cm² flow reactor, a commercial IrRuO_x/Ti electrode was used as the anode and the electrode working area was enlarged to 100 cm². To simulate the treatment of H₂S-involved natural gas using EChem-SOR system, 5% H₂S/CH₄ or 10% H₂S/CH₄ was as the gas source to replace 5% H₂S/Ar. For high flux H₂S/CH₄ (>100 ml min⁻¹), H₂S/CH₄ was first fed into a 500 ml of 1 M KOH solution for H₂S capture and then along with the circulation solution for subsequent reactions in cathode chamber.

For the long-term operation of EChem-SOR system in flow reactor, the electrolysis process followed the aforementioned flow reactor procedure. Additionally, the cathodic KOH circulation solution was refreshed daily for maintaining the alkaline environment to capture H₂S. An additional gas bag was connected to the end of KOH reservoir to collect the tail gas, and the H₂S concentration in tail gas was measured via a gas chromatograph (GC, SOE−131, Shimadzu) equipped with a flame photometric detector (FPD). After the long-term operation, the catholyte was collected and K₂SO₄ product was obtained through rotary evaporation and crystallization.

Electrochemical oxidation of H₂S to H₂SO₄ in solid-electrolyte flow reactor

The electrochemical H₂S oxidation to produce H₂SO₄ solution was carried out in a SE flow reactor. The SE flow reactor includes a 100 cm² OCNT/DGL cathode, 100 cm² commercial IrRuO_x/Ti anode, anion exchange membrane (AEM), proton exchange membrane (PEM), and middle SE layer filled with H⁺-conducting ion exchange resin. The middle SE layer and the anode chamber were circulated by 500 ml of deionized water at a flow rate of 1 ml min⁻¹ and 10 ml min⁻¹, respectively. During electrolysis, the chronopotentiometry test was performed at a constant current. O₂ was fed into the cathode chamber. 10% H₂S/CH₄ was first mixed with the circulating water, both were then circulated in the middle layer.

Sulfur-containing compounds quantification

The concentration of S²⁻ ions in the solution can be determined by the methylene blue spectrophotometric method^34,35. Typically, after electrolysis, 30 μl of catholyte was added to a sample tube containing 10 ml of 0.25 M NaOH solution, leading to all sulfide ions (HS⁻) dissociate into S²⁻ in a strongly alkaline environment (pH > 13). Next, 5 ml of 0.01 M C₈H₁₂N₂·2HCl solution in 4 M H₂SO₄ was slowly added along the wall of the sample tube. After capping and shaking, 0.5 ml of 0.2 M NH₄Fe(SO₄)₂·12H₂O solution in 0.2 M H₂SO₄ was slowly added along the tube wall. Then, 20 mL of deoxygenated deionized water was added, and the mixture was shaken well. After standing for a moment, the solution turned blue. Finally, the absorbance of the test solution was measured at a wavelength of 665 nm using UV-Vis spectrophotometer. The concentration of S²⁻ in the solution can be determined according to the standard curve. To obtain the standard curve, 10 sets of NaHS solutions in 0.1 M KOH with known concentrations (ranging from 0.3 to 16.7 mM) were first prepared. The test solutions were then prepared according to the above procedure, and the standard curve was plotted by linearly fitting the absorbance values of the test solutions at a wavelength of 665 nm (Supplementary Fig. 36).

The liquid oxidation products (S₂O₃²⁻, SO₃²⁻ and SO₄²⁻) of H₂S were detected using ion chromatography equipped with a conductivity detector and AS11-HC column. The mobile phase was 100 mM NaOH solution with a flow rate of 0.38 ml min⁻¹. Typically, after electrolysis, 20 µl of catholyte was collected and diluted with 1000 µl of deionized water, and then 10 µl of the diluted solution was subjected to ion chromatography analysis. The concentrations of three oxidation products can be determined according to the calibration curves of known concentration samples (Supplementary Figs. 37–39).

Calculation of productivity, H₂S_cap conversion and e⁻-to-sulfate efficiency

The productivity of various oxidation products was calculated according the following equations:

$${{{\rm{SO}}}}_{4}^{2-}{{\rm{productivity}}}\left({{\rm{mmol}}}{{{\rm{h}}}}^{-1}\right)=\frac{{C}_{{{{\rm{SO}}}}_{4}^{2-}}\times V}{t}$$

(6)

$${{{\rm{SO}}}}_{3}^{2-}{{\rm{productivity}}}\left({{\rm{mmol}}}{{{\rm{h}}}}^{-1}\right)=\frac{{C}_{{{{\rm{SO}}}}_{3}^{2-}}\times V}{t}$$

(7)

$${{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}{{\rm{productivity}}}\left({{\rm{mmol}}}{{{\rm{h}}}}^{-1}\right)=\frac{{C}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}\times V}{t}$$

(8)

$${{\rm{S\; productivity}}}\left({{\rm{mmol}}}{{{\rm{h}}}}^{-1}\right)=\frac{{m}_{{{\rm{S}}}}}{{M}_{{{\rm{S}}}}\times t}$$

(9)

where C_SO42−, C_SO32− and C_S2O32− are the concentration of SO₄²⁻, SO₃²⁻ and S₂O₃²⁻ measured under electrolysis, respectively. m_S is the mass of S obtained by weighing, M_S is the molar mass of S, V is the volume of catholyte, t is the electrolysis time.

The H₂S_cap conversion (%) was calculated according the following equation:

$$ {{{\rm{H}}}}_{2}{{{\rm{S}}}}_{{{\rm{cap}}}} {{\rm{conversion}}}\, \left(\%\right)=\\ 100-\frac{100\times {C}_{{{{\rm{S}}}}^{2-}}\times V}{\left({C}_{{{{\rm{SO}}}}_{4}^{2-}}+{C}_{{{{\rm{SO}}}}_{3}^{2-}}+{C}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}+{C}_{{{{\rm{S}}}}^{2-}}\right)\times V+{m}_{{{\rm{S}}}}/{M}_{{{\rm{S}}}}}$$

(10)

where C_S2− is the concentration of S²⁻ measured by the methylene blue spectrophotometric method.

To describe the electron utilization efficiency for SO₄²⁻ production, we define the e⁻-to-sulfate efficiency (%) as follows:

$$ {{{\rm{e}}}}^{-}-{{\rm{to}}}-{{\rm{sulfate\; efficiency}}}\,\left(\%\right)=\\ \frac{\left({C}_{{{{\rm{SO}}}}_{4}^{2-}}-{{{\rm{C}}}}_{{{{\rm{SO}}}}_{4}^{2-}}^{0}\right)\times V\times F\times \left({N}_{{{{\rm{S}}}}^{2-}}\times {R}_{{{{\rm{S}}}}^{2-}}+{N}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}\times {R}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}+{N}_{{{{\rm{SO}}}}_{3}^{2-}}\times {R}_{{{{\rm{SO}}}}_{3}^{2-}}\right)}{Q}$$

(11)

$${R}_{{{{\rm{S}}}}^{2-}}=\frac{{C}_{{{{\rm{S}}}}^{2-}}^{0}}{{C}_{{{{\rm{SO}}}}_{4}^{2-}}^{0}+{C}_{{{{\rm{SO}}}}_{3}^{2-}}^{0}+{C}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}^{0}+{C}_{{{{\rm{S}}}}^{2-}}^{0}}$$

(12)

$${R}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}=\frac{{{{\rm{C}}}}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}^{0}}{{C}_{{{{\rm{SO}}}}_{4}^{2-}}^{0}+{C}_{{{{\rm{SO}}}}_{3}^{2-}}^{0}+{{{\rm{C}}}}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}^{0}+{C}_{{{{\rm{S}}}}^{2-}}^{0}}$$

(13)

$${R}_{{{{\rm{SO}}}}_{3}^{2-}}=\frac{{{{\rm{C}}}}_{{{{\rm{SO}}}}_{3}^{2-}}^{0}}{{C}_{{{{\rm{SO}}}}_{4}^{2-}}^{0}+{C}_{{{{\rm{SO}}}}_{3}^{2-}}^{0}+{C}_{{{{\rm{S}}}}_{2}{{{\rm{O}}}}_{3}^{2-}}^{0}+{C}_{{{{\rm{S}}}}^{2-}}^{0}}$$

(14)

where C_SO42−⁰, C_SO32−⁰, C_S2O32−⁰, and C_S2−⁰ is the concentration of SO₄²⁻, SO₃²⁻, S₂O₃²⁻, and S²⁻, respectively, which is measured under the condition of only O₂ feeding, without performing electrolysis. N_S2− (8) is the electron transfer number from S²⁻ to SO₄²⁻, N_S2O32− (4) is the electron transfer number from S₂O₃²⁻ to SO₄²⁻, N_SO32− (2) is the electron transfer number from SO₃²⁻ to SO₄²⁻. R_S2O32− is the ratio of S₂O₃²⁻ among all sulfur species under the condition of only O₂ feeding, without performing electrolysis. R_SO32− is the ratio of SO₃²⁻ among all sulfur species under the condition of only O₂ feeding, without performing electrolysis.

It is important to note that the concentrations of the three oxidation products (SO₄²⁻, SO₃²⁻ and S₂O₃²⁻) produced from O₂-assisted H₂S oxidation are dynamically changing during the electrolysis process, leading to the contribution of O₂ to the H₂S oxidation reaction cannot be precisely excluded. Therefore, the calculated e⁻-to-sulfate efficiency based on the equation may deviate from the actual situation. Nevertheless, this still allows for an objective comparison of electron utilization efficiency for different reactors.

COMSOL simulation of H₂O₂ concentration distribution

The COMSOL simulation was conducted by coupling electrochemistry and fluid flow modules. A two-dimensional axisymmetric model was adopted to simulate the flow cell, with dimensions matching the experimental setup (electrode width: 20 mm, channel height: 1.5 mm). Various currents were applied to the working electrode, while the inlet electrolyte flow rate was set to 0.5–10 ml min⁻¹ with fully developed laminar flow. Initial H₂O₂ concentration was set to zero throughout the domain. The numerical model solves the following governing equations while accounting for both diffusion and convection effects:

$$\nabla .\,{j}_{i}+\mu .\nabla {c}_{i}={R}_{i}+{S}_{i}$$

(15)

Where j_i is the applied current, c_i is the H₂O₂ concentration, R_i and S_i are the source items, μ is the velocity field, that is obtained by solving the Navier-Stokes equation.