Solar Fuel Production from Hydrogen Sulfide: An Upstream Energy Perspective

Hydrogen sulfide is readily available in vast quantities in the subsurface as a byproduct of industrial processes. Hydrogen evolution from H2S can transform this highly toxic gas into a source of green fuel. Compared to water splitting, H2S dissociation is thermodynamically more favorable. However, feasible industrial‐scale catalytic technologies are not developed yet. The recovery of valuable chemicals using carbon‐neutral photocatalytic processes can capitalize on abundant solar irradiation and advanced semiconductors. The challenge is developing photocatalysts that can efficiently operate over the long term in the harsh environment of subsurface and industry, while utilizing as much of the light source spectrum as possible and providing optimum adsorption/desorption abilities of hydrogen and sulfur‐containing intermediates. Meeting these requirements demands improved kinematic models of photocatalytic H2S decomposition to assess the effect of high temperatures, pressures, mixtures of hydrocarbons, produced water, and other contaminants. Metal sulfides‐based catalysts may be the key to H2S decomposition in the subsurface (e.g., oil and gas reservoirs) and wellbores, but first they need to be upscaled as bulk, robust, and recyclable materials. This review presents a guide for the development of the upstream energy production technology via photocatalytic H2S conversion.

DOI: 10.1002/aesr.202200201 Hydrogen sulfide is readily available in vast quantities in the subsurface as a byproduct of industrial processes. Hydrogen evolution from H 2 S can transform this highly toxic gas into a source of green fuel. Compared to water splitting, H 2 S dissociation is thermodynamically more favorable. However, feasible industrialscale catalytic technologies are not developed yet. The recovery of valuable chemicals using carbon-neutral photocatalytic processes can capitalize on abundant solar irradiation and advanced semiconductors. The challenge is developing photocatalysts that can efficiently operate over the long term in the harsh environment of subsurface and industry, while utilizing as much of the light source spectrum as possible and providing optimum adsorption/desorption abilities of hydrogen and sulfur-containing intermediates. Meeting these requirements demands improved kinematic models of photocatalytic H 2 S decomposition to assess the effect of high temperatures, pressures, mixtures of hydrocarbons, produced water, and other contaminants. Metal sulfides-based catalysts may be the key to H 2 S decomposition in the subsurface (e.g., oil and gas reservoirs) and wellbores, but first they need to be upscaled as bulk, robust, and recyclable materials. This review presents a guide for the development of the upstream energy production technology via photocatalytic H 2 S conversion.
electrolysis (H 2 S vs. H 2 O: 0.14 V vs. 1.23 V//33 kJ mol À1 vs. 237 kJ mol À1 ). [15,16] Since the sour wells' H 2 S content can be at least 3.5%, a significant amount of both hydrogen and sulfur can be generated via H 2 S decomposition. [17] Therefore, the challenge is to develop cost-and energy-efficient technology for hydrogen generation from H 2 S at an industrial scale. The possible solutions include, but are not limited to, thermochemical, electrochemical, photolytic, and photocatalytic processes. [1,[18][19][20][21][22] The thermal decomposition of H 2 S using concentrated solar irradiation has shown the potential to be %45% more costeffective than the Claus process due to H 2 generation. [23] The use of a catalyst could make the process even less energy demanding. For instance, such plasmonic metal-based photocatalysts as SiO 2 -supported gold nanoparticles can decrease the apparent activation energy barrier of H 2 S decomposition from 0.61 to 0.32 eV (for thermocatalysis and photocatalysis, respectively). Hot carriers, generated upon photoexcitation of the catalyst with visible light (λ = 525 nm), accelerate H─S bond cleavage. That, combined with photothermal heating effect observed for plasmonic nanoparticles, demonstrated 20 times higher H 2 S decomposition reaction rates than thermocatalysis. [24] Hence, the photocatalytic conversion of H 2 S can potentially be an economically profitable ecofriendly technology.
Photocatalysis is powered by the most abundant and renewable energy source, solar irradiation, used to drive pairs of chemical reactions. Photocatalysts are activated by photons of suitable wavelengths, generating charge carriers which are subsequently transferred to the reactants and simultaneously participate in coupled redox reactions. [25,26] Currently, photocatalytic H 2 S decomposition is at the technology readiness level (TRL) of 3, equivalent to the experimental proof-of-concept stage. [27] Hence, a semiconductor photocatalyst is needed for H 2 S decomposition into hydrogen and sulfur. These catalysts must be affordable, easily scalable, ecofriendly, photo and chemically stable, responsive to visible light, and catalytically efficient at upstream reaction conditions. H 2 S decomposition should be performed in the subsurface to reduce emissions. However, the development of catalysts is hampered by the harsh conditions inside reservoirs and wellbores (temperatures up to 200°C, pressures up to 1500 bar, and high concentrations of H 2 S). Consequently, the possibilities of H 2 S gas conversion under these conditions have not been addressed. Earlier reviews on liquid-and gas-phase H 2 S photoconversion considered the latest semiconducting materials and plausible reaction mechanisms at standard conditions. [11,28] However, semiconductors initially designed for other light-driven reactions (e. g., water splitting, CO 2 reduction, etc.) are rarely considered for H 2 S decomposition, mostly because it is presumed that photocorrosion and catalyst poisoning may inhibit their catalytic potential.
Here we discuss the use of inorganic semiconductor photocatalysts, primarily metal sulfides, from an industrial application viewpoint: state-of-the-art photocatalytically active materials, reported for various applications, are analyzed and assessed for use in in situ hydrogen production and elemental sulfur recovery from H 2 S gas.

Photocatalytic H 2 S Decomposition
Photocatalysts are standalone semiconducting materials or composites that produce electrical charges upon photoexcitation.
Following photogeneration, the electrical charges, namely, electrons and holes, can participate in reduction and oxidation (half-) reactions. The electrochemical potential of the charges driving the redox reaction is determined by the semiconductors' energy levels and their bandgap, that is, the valence band (VB) and the conduction band (CB) and the energy difference between them. Electron/hole (e À /h þ ) pairs can only be created if the energy of photons incident on the semiconductor is larger than or equivalent to its bandgap (hν ≥ E g ). The lifetimes of photogenerated charges in prototypical TiO 2 and CdS are on the order of nanoseconds. [29][30][31] During their lifetimes, electrons and holes should diffuse from the bulk of the semiconductor to its surface to participate in redox reactions; the latter can take milliseconds or even seconds. [29] Consequently, many charges recombine before reaching the reaction site. Hence, heterojunction engineering using semiconductors responsive to complementary spectral regions is advantageous in comparison to a single photoactive material. It enhances solar energy utilization and significantly increases the lifetimes of photogenerated charge carriers. The principle behind heterojunctions lies in the energy band alignment, which drives charge carriers to different sides of the junction. [32] A schematic representation of some types of junctions is shown in Figure 1.
To drive a redox reaction, that is, to be photocatalytically active, the semiconductor's energy band structure should be precisely tuned in accordance with the energetics of the desired reaction. The semiconductor's CB minimum must be more negative than the half-cell electrochemical potential of the reduction halfreaction. At the same time, its VB maximum must be more positive than the half-cell electrochemical potential of the oxidation half-reaction (vs. normal hydrogen electrode (NHE), pH 0). In the case of H 2 S conversion, the following requirements apply: The predicted thermodynamic equilibrium potential difference for simultaneous H 2 evolution and S-recovery from solution is 0.14 V (vs. NHE, pH 0). [33] However, kinetic barriers associated with both half-reactions necessitate an additional overpotential (η red and η ox ) which must be provided by the photocatalyst in each half-reaction ( Figure 1). The overpotential required for a particular half-reaction is determined by how efficiently the photocatalyst's surface-active sites catalyze the said reaction. Adding a cocatalyst, typically metal or metal oxide nanoparticles, can often lower the overpotential by reducing the reaction activation energy and promoting its kinetics. [34] Furthermore, Schottky junctions created between a semiconductor and a cocatalyst are beneficial for charge separation. [34,35] In addition, photocorrosion resistance requires the semiconductor's reductive decomposition potential and the CB minimum to be more negative than the reduction half-reaction potential, while its oxidative decomposition potential and VB maximum to be more positive than the oxidation half-reaction potential (vs. NHE). [36] All these considerations widen the bandgap of the semiconductor. However, most solar irradiation is present in the visible and NIR spectral regions, precisely λ = 400-1000 nm ( Figure 2). [37] Consequently, photocatalysts must be tailored to absorb as much of the solar spectrum as possible while considering the aforementioned constraints. Specifically, semiconductors responsive to visible and NIR irradiation, with a bandgap narrower than 2.5 eV, are preferred in photocatalysis.
The photocatalyst's performance is typically evaluated by measuring the H 2 evolution reaction (HER) rate (mmol g À1 h À1 ), solar-to-hydrogen conversion efficiency (STH, %), quantum efficiency (QE, %), and apparent quantum yield (AQY, %). [38] However, it is challenging to compare HER rates directly across reported results due to the gamut of reaction conditions (including the light source, catalyst amount, temperature, type of sacrificial agents, etc.).
The rate of the H 2 evolution reaction is determined by the Gibbs free energy of hydrogen adsorption (ΔG H* ) on the catalyst's surface. This concept is typically implemented in electrocatalysis but should also apply to the photocatalyst surface. The relationship between experimentally measured exchange current densities of some cocatalysts and ΔG H* calculated via density functional theory (DFT) is shown in Figure 3. [39,40] An optimum catalyst should bind reaction intermediates neither too strongly nor too weakly. The closer the ΔG H* value is to zero, the easier it is for hydrogen to desorb from the catalyst's surface.
Several strategies have been developed to modify a photocatalyst's surface to tune ΔG H* . For example, DFT calculations demonstrated that ΔG H* of In 2 S 3 doped with Mn is 0.74 eV, while ΔG H* of pure In 2 S 3 is À4.29 eV. [41] Likewise, DFT calculations of the sulfur adsorption energy (E ads ) revealed that sulfur desorption from Cu-doped β-In 2 S 3 ((In x Cu 1Àx ) 2 S 3 ) is more favorable (À0.37 eV) than from an individual β-In 2 S 3 (À2.01 eV). [42] Hence, Mn doping increases the HER rate by rapid adsorption/desorption of H intermediates. It also improves solar energy utilization by narrowing the bandgap of the solid-solution Mn 2x In 2(1Àx) S 3 from 1.82 to 1.58 eV. [41] At the same time, Cu doping broadens the spectral response of a catalyst and allows Figure 1. Schematic energy level diagram of a) single semiconductor, and different types of heterojunctions for simultaneous solar fuel generation and S-recovery: b) Schottky junction between a semiconductor and a metal particle adsorbed on its surface, c) type I junction with a straddling bandgap, and d) type II junction with a staggered bandgap between two semiconductors. Solid horizontal lines represent the reduction potentials of the half reactions at pH 0. All the electrochemical potentials are shown with reference to the absolute electrochemical scale (0 V vs. NHE = À4.44 V vs. vac.). Figure 2. Solar irradiance spectra (red) and corresponding photon flux curve (blue) at AM1.5 as functions of light wavelength and material bandgap. Reproduced with permission. [37] Copyright 2017, Wiley VCH. Reproduced with permission. [40] Copyright 2017, The American Association for the Advancement of Science.
www.advancedsciencenews.com www.advenergysustres.com efficient S desorption, solving the challenge of catalyst surface passivation by elemental sulfur. [42][43][44] Apart from tuning ΔG H* of a photocatalyst, anion or cation doping can introduce trapping sites for the photogenerated charges and inhibit their recombination. This technique also broadens the spectral response of the semiconductor due to the formation of an intermediate band. [45,46] Overall, an ideal photocatalyst for upstream applications must fulfill essential requirements: 1) operate at in situ conditions; 2) exhibit low-to-none photocorrosion; 3) be responsive in wide spectral range (i.e., narrow bandgap); 4) be infinitely recyclable; and 5) provide optimum adsorption/desorption abilities of hydrogen and sulfur-containing intermediates. Practical and industrial applications should also optimize the photocatalytic process concerning various operating parameters, such as H 2 S flowrate, concentration, catalyst dosage, and light irradiation.

Liquid-Phase H 2 S Dissociation
In the upstream energy production, H 2 S may be dissolved in liquids (oils or water) or mixed with gas flows. In situ conversion of H 2 S requires the development of methods and photocatalysts for each set of conditions. The research is mainly focused on the photocatalytic conversion of H 2 S in aqueous solutions driven by the need to minimize the hazardous effects of H 2 S rather than commercial imperatives. However, the potential reforming of H 2 S dissolved in liquid hydrocarbons and produced water remains an open challenge.
In the liquid phase, H 2 S decomposition occurs via dissociation in a basic aqueous solution (e. g., NaOH, ethanolamine, etc.) rather than via a photochemical process (Equation (3)). Subsequently, the target product, H 2 , is obtained via proton (H þ ) reduction with photogenetared electrons (Equation (4) and (5)), regardless of whether H þ comes from the dissociated H 2 S or water. Aqueous solutions of Na 2 S/Na 2 SO 3 are routinely used as the reaction medium. These carbon-neutral solutions are promising for sustainable energy technologies compared to other common organic hole scavengers, for instance, alcohols, amines, and lactic or ascorbic acids, among others. H 2 S dissociation can proceed via the following steps. [47] Dissociation∶ The minimum required potential for HS À oxidation to polysulfides S n 2À is only %0.3 V (vs. reversible hydrogen electrode [RHE], pH 14, dissolved sulfur concentration 0.5 м) ( Figure 4). Hence, they are common intermediates in the photocatalytic process. [47] The generation of elemental sulfur and S n 2À species (Equation (6) and (7)) presents a significant issue for H 2 S decomposition. They form a yellowish solution absorbing some irradiation or even passivating the catalyst's surface. Thus, Na 2 SO 3 acts as a scavenger consuming the intermediates that poison the catalyst (Equation (8) and (9)). Most studies claim the formation of thiosulphate S 2 O 3 2À anions during the photoreaction. However, it is unclear whether byproducts, such as dithionate S 2 O 6 2À , sulfate SO 4 2À , and other polysulfides S n 2À , are also present in the system. [48] While S 2 O 3 2À and SO 4 2À are valuable products used in clinical applications and as fertilizers in agriculture, their production as a mixture might not be economically viable. That being the case, precise control over the reaction conditions is required for the selective and efficient recovery of valuable sulfur-containing chemicals. In general, the photocatalysts for liquid-phase H 2 S dissociation can be divided into two groups: first, non-metal-sulfidebased and second, metal sulfide (MS)-based materials. These two categories will be discussed in more detail in the next subsections.

Nonsulfide-Based Semiconductors
Previous reviews on photocatalytic HER via H 2 S dissociation discussed using oxides of transition metals with d 0 and d 10 electron configuration alone or in combination with other cocatalysts (e. g., noble metals, RuO 2 , MoS 2 , etc.). However, their solar irradiation response is often limited to the UV spectral range due to their wide bandgap. Therefore, junction engineering between multiple semiconductors and doping strategies is essential for optimal solar energy use. [1,10,15,20] Quantum dots (QDs) have also shown great potential as stateof-the-art photocatalysts due to their easily tunable bandgaps, precisely controlled composition and morphology, efficient exciton generation, and charge separation. [49] Recently, Zn-decorated InP QDs (E g = 1.35 eV) have been suggested as a promising nontoxic alternative to the conventional CdSe, CdTe, and CdS QDs. A steady HER rate of 7.60 mmol g À1 h À1 was observed for 10 h. The presence of Zn on the surface of QDs increased the  surface's zeta-potential, thereby weakening the electrostatic repulsion between the photocatalyst's surface and HS À , as well as S 2À anions in the solution, consequently, accelerating their photooxidation. Additionally, the redshifted absorption spectrum of Zn-InP advanced the visible light activity of the material. [50] Nonmetallic semiconductors (e.g., graphitic carbon nitride (g-C 3 N 4 ), carbon nanotubes (CNTs), and graphene oxide (GO), among others) have come to limelight due to their favorable band structures, good conductivity, high photoluminescence, chemical stability, and facile preparation. [51][52][53] Coupling them with other semiconductors broadens the spectral response and prevents exciton recombination. For example, g-C 3 N 4 porous nanosheets were modified with S, N-codoped carbon dots (CDs), which acted as photosensitizers under 460 nm light irradiation. The S,N-CDs/g-C 3 N 4 composite loaded with 1 wt% Pt cocatalyst demonstrated steady H 2 production over 10 h with an AQY of 3.47% at 460 nm. [54] MnO 2 -activated carbon (AC) nanocomposites (E VB = 0.4 eV, E CB = À2.0 eV, E g = 2.41 eV vs. vac.) were suggested for H 2 S dissociation. The highly porous AC provided an increased amount of catalytically active sites. Unfortunately, these sites were blocked by S-species at sulfite concentrations as low as 0.3 м. Additionally, a decrease in HER performance was observed after one cycle due to 30% weight loss during the catalyst recovery process. [55] Similarly, Oladipo et al. observed reduced H 2 production when using N-doped TiO 2 /graphene composite decorated with Pt cocatalyst. Sulfur-containing species were adsorbed on the catalyst surface at HS À and S 2À concentrations of 0.4 and 0.284 м, respectively, restricting the catalyst exposure to irradiation. [56] Moreover, TiO 2 is known to release oxygen atoms from its lattice under UV irradiation forming Ti 3þ -O v states. [57][58][59] Due to this self-reduction, S-intermediates adsorbed on the catalyst surface can react with the free oxygen forming the toxic byproduct SO 2 (Equation (10) and (11)).
Schematic representation of sulfur dioxide gas formation∶ Catalyst surface passivation∶ MO solid þ H 2 S gas !MS solid þ H 2 O gas (12) The extreme sensitivity of these photocatalysts to increased S-species concentrations in solution is a major bottleneck of the aqueous process. In case elemental sulfur is produced during the reaction, it poisons the catalyst by blocking its active sites. At the same time, the catalyst can be partially occluded from the incident light due to the yellow polysulfide solution. The Na 2 S/Na 2 SO 3 solution is a common scavenger for the photogenerated holes and polysulfide anions. However, once most of the scavenger is consumed, the holes cannot be extracted efficiently anymore, thus causing self-oxidation of the material.
Furthermore, for some metal oxides, transformation into sulfides in an H 2 S environment is a thermodynamically favorable process even at standard conditions (e.g., ZnO þ H 2 S ! ZnS þ H 2 O, approx: ΔG 0 rxn À67 kJ mol À1 , at 298.15 K and 1 atm). Considering the highly reactive nature of H 2 S, it is evident that at in situ conditions, the catalysts' degradation caused by surface passivation with a metal sulfide layer or even a complete transformation into a sulfide is bound to happen (Equation (12)). [60] Hence, applying previously reported nonsulfide-based materials for the in situ H 2 S photoconversion is impractical, while metal-sulfide-based semiconductors seem more appropriate.

Metal Sulfide-Based Semiconductors
MS-based photoactive materials are typically made of metal cations with a d 10 electron configuration. The CB of these catalysts consists of d and sp hybrid orbitals of the metal, and the VB consists of 3p orbitals of sulfur. [61] This structure explains their narrow bandgap (%0.26 eV ≤ E g ≤ 2.8 eV), which supports a wide response to the solar spectrum, including the visible and NIR range, when compared to prototypical UV light-active metal oxides, such as commercially available TiO 2 (E g (anatase) = 3.2 eV, E g (rutile) = 3.0 eV). [62][63][64][65] MS-based photocatalysts are often synthesized via solvothermal or hydrothermal methods using metal salts, such as nitrates, chlorides, or acetates. The solvothermal method typically employs ethanol or pyridine as solvents, whereas for the hydrothermal process the solvent is water. Thioacetamide C 2 H 5 NS or thiourea SC(NH 2 ) 2 are widely used as a sulfur source in both techniques. Wang et al. demonstrated that CdS-based semiconductors hydrothermalized from thioacetamide have a broader light absorption range, weaker photoluminescence, and larger surface area than those made from L-cysteine or thiourea. [66] Despite the need for elevated temperatures and high-pressure conditions, both procedures could be scaled up easily for industrial applications. [61,67] Nevertheless, a low-cost and large-scale synthetic procedure of these materials with a well-defined structure still needs further research and development.
The attractive features of MS materials include, but are not limited to, suitable and easily tunable bandgap, abundant photocatalytically active sites, and high porosity. Compared to metal oxides, MS-based semiconductors have lower redox potentials and longer lifetimes of excited states. [61] Their major drawbacks include 1) tendency to agglomerate; 2) fast recombination rate of photogenerated electron/hole pairs; and 3) photocorrosion due to oxidation of lattice S 2À ions by photogenerated holes to sulfur or SO 4 2À (depending on the presence of oxygen in the system), resulting in catalyst degradation. [67,68] Yet, MS-based semiconductors are promising candidates for H 2 S decomposition since they offer chemical stability in H 2 S environments and broad solar response. [1,28] Single and cocatalyst-decorated semiconductors exhibit stateof-the-art HER performance, yet their photostability in solution is limited. Heterojunction formation is preferred for enhanced charge separation and advanced light harvesting performance. For instance, individually, Bi 2 S 3 is known for its high photoluminescence intensity (i.e., highly radiative recombination rate of photogenerated charges), whereas MnS has a low absorption coefficient. Lashgari and Ghanimati showed that S atom cosharing in n-Bi 2 S 3 /p-MnS junctions provided a better p-n contact between the two semiconductors. [69] That in turn improved charge separation. Thus, the p-n junction (E g = 2.18 eV) should enhance the photocatalytic activity of the material in comparison to the individual components. However, the HER efficiency was not reported. The authors suggested elemental sulfur extraction from an alkaline solution (pH 11) via titration of S 2 2À with 1 м HCl: sulfur particles precipitated from the solution at pH 5. Yet, this approach appears to be impractical for large-scale sulfur production.
Another example of MnS-based photocatalyst is MnS/In 2 S 3 -MoS 2 . It is a double-interface composite with exceptional catalytically active sites, 12 h photostability, and AQY of 72% (400 nm). [70] The crystal structure of the material remained intact after recycling. MoS 2 served as a cocatalytic electron sink, promoted charge separation and migration, and decreased the overpotential of MnS/In 2 S 3 from À0.47 to À0.41 V (vs. RHE). On the one hand, charge transfer may be facilitated by a larger overpotential. On the other hand, the energy difference can also contribute to energy losses, as other processes can occur (e.g., heat generation). A large overpotential provides more driving force but does not necessarily maximize the catalytic efficiency. Therefore, the key is to strive for a compromise between energy efficiency and sufficient driving force, that is, an appropriate energy band alignment.
ZnS was also reported to have a promising H 2 S conversion efficiency. However, it suffers from serious photocorrosion. [28] Element doping appears to be the most used technique for photocatalyst stabilization. It can introduce donor or acceptor states in the bandgap of a semiconductor and, consequently, change the redox potential of photogenerated charge carriers and shrink its bandgap. For instance, as a consequence of ZnS doping with electronegative nitrogen atoms, the VB of ZnS consists not only of sulfur 3p states but also of nitrogen 2p states. Therefore, the VB maximum is upshifted from À6.58 to À5.58 eV (vs. vac.) and the material's bandgap is reduced from 3.67 to 2.67 eV, while the CB is intact. As a result, the photostability of N-doped ZnS reached 12 h as compared to 5 h for pristine ZnS. [71] Similarly, ZnS with incorporated iron, an abundant metal commonly used in industry for H 2 S treatment, has a reduced bandgap of 2.04 eV (from 3.5 eV), turning it into a visible light-active semiconductor with low toxicity. The catalyst's activity was found to be stable over 9 h. It also showed a higher HER rate and the presence of HS À species in the solution for pH < 9, while under more basic conditions, pH 9-11, it showed lower H 2 production and sulfide S 2À , as well as bisulfite HSO 3 À anion formation. [72] CdS is another well-known photocatalyst, which suffers from photocorrosion due to the deep VB. Ca 2þ -modified CdS colloidal nanocrystals (E g % 2.5 eV) were tested for H 2 S dissociation. The photocatalyst maintained its activity for only 10 h with HER rate of 56 mmol g À1 h À1 under LED light illumination (460 nm, 3 W). Importantly, S 2 O 3 2À was the only product of the oxidation halfreaction. Selective thiosulfate formation prevented the catalyst's poisoning by other S-species, thus improving its photocatalytic activity. [48] Dan et al. studied the effect of the crystal structure and morphology on the energy band position of CdS-InS solid solutions. Compared to hexagonal nanosheets, cubic Cd x In 1Àx S (E g = 2.19 eV) with regular particle morphology and large surface area demonstrated a better HER rate of 16.35 mmol g À1 h À1 , AQY of 26.7% (420 nm), and 15 h photostability. The deeper VB position (1.52 eV vs. vac.) promoted the oxidation of sulfur deposited on the surface to S 2 O 3 2À , preventing the catalyst surface passivation. Simultaneously, the more positive E CB of À0.67 V helped to avoid unnecessary energy losses while providing enough overpotential for H 2 evolution. [73] The formation of more complex heterostructures with various morphologies, such as CdIn 2 S 4 stacked nanosheets loaded with 10% MoS 2 , Co 9 S 8 nanocages decorated with CdIn 2 S 4 nanospheres, or ZnIn 2 S 4 nanosheets with CdIn 2 S 4 octahedrons, did not dramatically improve the resistance to self-oxidation compared to the single-catalyst Cd x In 1Àx S. [74][75][76] The same research group also proposed a dual-solid-solution composite, Mn x Cd 1Àx S/Cd y Mn 1Ày S. It showed H 2 S dissociation with HER rate of 113 mmol g À1 h À1 , selective formation of S 2 O 3 2À , and a remarkable AQY of 77% (440 nm). [77] The heterostructure increased the light-harvesting ability and spatial separation of charges while facilitating charge transfer to the reaction sites. Compared to individual MnS and CdS, the heterostructure showed improved visible light absorption (up to 500 nm) and an increased charge carrier density. Additionally, an increased concentration of H 2 S in the solution reduced its pH from basic (12.77) to near-neutral (7.99), facilitating the selective formation of thiosulfate anions. Thus, the dual-solid-solution Mn x Cd 1Àx S/ Cd y Mn 1Ày S is a photocatalyst with one of the best performances in H 2 S dissociation. Nonetheless, further evaluation of its stability, recyclability, and scalability is still required to determine its full potential as a catalyst.
Overall, photocatalysts based on CdS, MnS, and their solid solutions with In 2 S 3 have demonstrated a significant potential for H 2 evolution via liquid-phase H 2 S dissociation. However, their reported photocatalytic activity lasts only a few tens of hours. This short photocatalytic lifespan could be attributed to the rapid catalyst photocorrosion and the hazardous nature of H 2 S, which limit the catalyst evaluation over an extended time. The role and advantage of S-species in the oxidation half-reaction remains unresolved. The recovery of elemental sulfur or valuable S-containing chemicals from an aqueous system might be challenging due to the formation of a complex mixture of oxidation reaction products. The composition of that mixture and the quantitative yield of the oxidation products are rarely determined since they are mainly used as a sacrificial agent rather than as a precursor to high-value chemicals. Since developing low-waste ecofriendly technologies is one of the primary goals of modern industry, at the very least, Na 2 S/Na 2 SO 3 should be recovered and recycled or better be converted into valuable chemical products.
The photocatalytic materials used for H 2 S conversion and discussed in this section are summarized in Table 1.

Gas-Phase H 2 S Decomposition
In gas mixtures, it is possible to filter or target H 2 S precisely. Nevertheless, the photodecomposition of H 2 S in the gas phase remains unexplored due to the challenges of manipulating the gas with high toxicity, corrosiveness, and flammability. However, from a practical perspective, gas-phase photoconversion is desirable for several reasons: 1) H 2 S exists already as a highly pressurized gas in the reservoir; 2) the two desired products of H 2 S decomposition, namely, H 2 gas and liquid or solid S n 0 , can be easily separated; 3) end products of a photocatalytic process are often environmentally benign and therefore, do not require further processing or complex disposal; and 4) no solvent or sacrificial agent (e.g., NaOH, TEOA, etc.) are required for the reaction, thus making it a cost-effective process. A typical hydrocarbon gas mixture consists mainly of H 2 S and methane (CH 4 ). The latter likely would not participate in a photocatalytic reaction because the bond dissociation energy of C-H in methane (439 kJ mol À1 ) is higher than the S-H in hydrogen sulfide (376 kJ mol À1 ). Hydrogen-enriched methane can be used to reduce the fuel's carbon content. Alternatively, the gas mixture can be separated to get pure hydrogen fuel.
The predicted minimum equilibrium potential difference for H 2 S gas decomposition into H 2 and S is 0.17 V (vs. NHE). Various examples of photocatalysts evaluated for gaseous H 2 S conversion can be found in the earlier reviews. [1,10,28] Most of the reported semiconductors for H 2 S treatment are based on metal oxides, particularly on the well-known TiO 2 . [78,79] However, as discussed earlier, metal oxides may not provide sufficient chemical stability in H 2 S-rich environments under the harsh upstream reaction conditions considered here.
Among the studied materials, MS-containing catalyst (CdS þ ZnS)/Fe 2 O 3 demonstrated one of the highest conversions of 91% in 10 ppm H 2 S, yet a meager H 2 evolution rate. Sulfate formation was observed after 30 min of irradiation, which subsequently formed crystalline sulfur, passivating the catalyst's surface and leading to its deactivation after only 60 min. Moreover, H 2 production decreased from 93 to 0.04 μmol as H 2 S concentration increased from 10 to 125 ppm, respectively, which might be attributed to the formation of iron sulfide. [80] The catalyst's deactivation at high concentrations of H 2 S, as a consequence of its surface passivation with S-containing products, is a common issue for most catalysts. [1] Elemental sulfur  [75] www.advancedsciencenews.com www.advenergysustres.com can block photocatalytically active sites and reduce the visible light absorption of the photocatalyst, thereby dramatically reducing the HER rate. The sulfur occlusion is nocuous when the photoreaction occurs at room temperature and atmospheric pressure. However, in situ H 2 S decomposition at a wellhead or subsurface can be performed at temperatures above 116°C, where elemental sulfur exists in its liquid form (T MP 115.21°C). Therefore, surface passivation can be circumvented, and catalyst recovery and recycling may be feasible. Further optimization of MS-based semiconductors toward improved resistance to self-oxidation and chemical stability in H 2 S-rich environments is of utmost importance.

Metal Sulfides: From H 2 O to H 2 S Dissociation
As mentioned earlier, H 2 evolution from water splitting is thermodynamically less favored than that from H 2 S dissociation (ΔG 0 H 2 O þ 237 kJ mol À1 vs. ΔG 0 H 2 S þ 33 kJ mol À1 ). [15,16] That makes H 2 evolution from H 2 S an attractive energy-efficient process for industrial applications. The minimum thermodynamic potential and minimum theoretical semiconductor bandgap required for water splitting is 1.23 eV. [36] However, the sluggish kinetics of H þ reduction and O 2 evolution require significant overpotentials, leading to a bandgap of at least 1.70 eV to drive the reaction. [49] In contrast, for H 2 S splitting, the minimal theoretical potential is 0.14 eV (H 2 S/S 0 , vs. NHE, pH 0). [33] Thus, the latest MS-based photocatalysts applied for water splitting are worth considering in H 2 S photoconversion. These photocatalysts exhibit high HER rates, quantum efficiencies, and activity in the vis-NIR spectral region. In this part of the review, photostability was not an evaluation metric because the reaction conditions in most reported studies are dramatically different from those of the upstream H 2 S decomposition. Furthermore, the overpotential provided by photocatalysts for water splitting would be unnecessarily large for H 2 S dissociation. Hence, the available catalysts could be modified further (e. g., doped) in order to upshift their VB, thereby boosting the visible light response and improving photostability.
Many examples of semiconductors for H 2 O splitting, including MS-based ones, can be found in previous reviews. In general, CdSbased and ZnIn 2 S 4 heterostructures appear to have the most promising characteristics for in situ H 2 S gas decomposition. [61,[81][82][83]

CdS-Based Photocatalysts
The photocatalytic activity and photostability of a semiconductor depend on its crystal structure, morphology, and thickness. Ultrathin (1 nm) Cd 4 S 5 nanosheets with MS/metal/MS sandwich structures and unsaturated surface S À anions were shown to have high electrical conductivity, upshifted VB and CB (1.41 and À0.91 V vs NHE, E g = 2.32 eV), and 50 h resistance to photocorrosion. Interestingly, structural degradation of thicker nanosheets was found already after 20 h of illumination. [84] MoS 2 is a well-known cocatalyst with good adsorption/desorption properties for H intermediates (ΔG H* close to 0), which promotes H 2 generation. [40] On its own, MoS 2 demonstrates poor HER activity, mainly due to the reduced activity of catalytic sites associated with nanostructure aggregation and poor electronic conductivity of its basal plane. Therefore, various strategies, including phase transformations, vacancy engineering, and morphology modifications, have been suggested to improve MoS 2 photocatalytic activity. [85] Heterostructure engineering using MoS 2 and CdS is a common approach to attain HER rates on the order of tens of mmol g À1 h À1 . [86][87][88][89] For example, 1T-metallic phase MoS 2 nanosheets with excellent metallic conductivity and many photocatalytically active basal and edge planes inhibited CdS photocorrosion due to the improved charge separation and transfer to the reactants. Furthermore, it upshifted the VB and CB of the material (E g = 2.34 eV) as compared to bare CdS (E g = 2.42 eV), which is favorable for HER and the desired S recovery. [89] An exceptional HER rate of 196 mmol g À1 h À1 combined with 60 h stability was achieved using CdS nanorods decorated with B-doped MoS 2 nanosheets. The multilayered structure of B-doped MoS 2 provided many catalytically active sites. Additionally, boron doping enhanced the separation and transfer of photogenerated charges. [88] Decorating a semiconductor with QDs, for example, MoS 2 QD, is another strategy to improve H 2 evolution efficiency. MoS 2 QDs demonstrate unique electronic and optical properties. Their size determines the number of exposed edges, namely, S atoms, per surface area, meaning that more H atoms can adsorb onto the catalyst's surface and thus participate in the reduction reaction. MoS 2 QD also possess appealing upconversion properties, broadening the spectral response of a catalyst. For example, Bi 2 S 3 decorated with 0.14 wt% MoS 2 QD was reported as an efficient (HER rate 17.7 mmol g À1 h À1 ), relatively photostable (24 h) vis-NIR active HER photocatalyst. [90] Similar to MoS 2 , WS 2 possesses cocatalytic HER activity comparable to Pt and noble metals. [91] Due to the thermocatalytic effect of 1T metallic WS 2 and an advanced interband charge transfer at the WS 2 -CdS interface, it was possible to reach 81.2 mmol g À1 h À1 under simulated sunlight. Additionally, the S vacancies of WS 2 facilitated H* adsorption (ΔG H* = 0.85 eV) and captured electrons from CdS. [92] Zhong et al. took advantage of the CdS photocorrosion process to make the semiconductor more resistant to self-oxidation. Under irradiation, large nanoparticles of CdS embedded into WS 2 nanosheets would dissolve and then, upon reaching a specific concentration in solution, recrystallize into smaller CdS particles, which are favorable for HER. Even though the HER rate was moderate at 0.37 mmol g À1 h À1 , the photostability reached record 200 h. [93] NiS has also shown cocatalytic activity in HER. Compact stacking between a cocatalyst NiS x , well-dispersed CdS nanoparticles, and a highly conductive porous Ni foam was shown to accelerate electron transfer from CdS to the reaction site. As a result, the composite maintained 78% activity after 30 h of illumination with 6.2 mmol g À1 h À1 overall HER rate. [94] At the same time, CdS/NiS combined with Ti 3 C 2 MXene, a 2D metal carbide with superior metallic conductivity, intense redox activity, and many hydrophilic sites valuable for interactions with water, exhibited similar photostability of 30 h and QE of 40.1% at 420 nm. H 2 evolution with Ti 3 C 2 as a cocatalyst is thermodynamically favorable since ΔG H* was calculated to be as little as 0.00283 eV (at its optimal H* coverage θ = ½) for O-terminated Ti 3 C 2 MXene, which is even lower than the one for Pt, the most efficient HER cocatalyst. [95] NiS was suggested as a promising single HER photocatalyst as well. It was shown that the highly crystalline monophase α-NiS has a slightly higher catalytic activity than β-NiS in H 2 generation from Na 2 S/Na 2 SO 3 aqueous solution (13.41 vs. 12.73 mmol g À1 h À1 ). The reasons are the enhanced α-NiS absorption, better suited VB and CB alignment (1.66 and À0.17 V vs. NHE, E g = 1.83 eV) for H þ reduction, and the presence of defects in its structure, which act as photocatalytic reaction sites and nonradiative recombination centers. [96] Additionally, F À anion doping of Ni 3 S 2 was shown to strengthen the binding between H atoms and the material's surface, as well as accelerate water dissociation, thus improving H 2 production. [97] CdS-ZnS solid solutions are an emerging ternary metal sulfide with good photocatalytic activity, stability, and bandgap tunability through the Cd/Zn ratio. Ni-doped Zn 0.5 Cd 0.5 S demonstrated an advanced visible light activity compared to Mndoped Zn 0.5 Cd 0.5 S, which was attributed to its narrower bandgap of 1.64 eV (vs 2.37 eV). [98,99] Both materials are synthesized solvothermally in ethanol and have similar photostability over four cycles (16 h total), while Ni doping appears to improve the photocatalyst's activity. Yet, there are many factors which can affect the photocatalyst's performance and should be considered in order to unravel the reasons for enhanced HER rates of Ni/Zn 0.5 Cd 0.5 S. They include material's absorption coefficient, electron-hole recombination rate, AQY, STH efficiency, the resistance of the system to charge transfer, and presence of energy loss channels.
Another way to promote the light-harvesting ability of a semiconductor is to introduce intermediate-energy levels by oxygen anion doping, which act as electron/hole traps. [100,101] O-doped Cd x Zn 1Àx S solid solutions demonstrated a higher photocurrent density than pristine Cd x Zn 1Àx S. This finding suggests faster charge carrier separation and thus, enhanced photocatalytic activity of 50 h. [101] Similarly, iodide doping was shown to modulate the band structure, improve visible light activity, and enhance the absorption of semiconductors. [102][103][104] Metal-organic frameworks (MOFs) are rarely applied in photocatalytic reactions due to their poor stability and electronic conductivity. However, semiconductors obtained from MOF templates maintain some of the initial structural properties of MOFs (porosity, ion diffusion channels, etc.) while having enhanced electrical conductivity and monodispersed metal centers. The latter can increase the number of photocatalytically active sites, which leads to improved material performance. Moreover, a wide choice of templates provided by the incredible versatility of MOFs allows solid-solution fabrication with the desired bandgap. [105,106] For example, a type-I heterojunction between MoS 2 and Zn 0.5 Cd 0.5 S showed HER rates of 23.8 mmol g À1 h À1 , AQY 20.16% at 420 nm, and stability over 25 h under visible light irradiation (420-780 nm, E g = 1.80 and 2.51 eV respectively). [107] Even though MOF formation is relatively straightforward once the conditions are optimized, the addition of this extra synthetic step should be justified by significantly improved performance (e.g., a dramatic increase in HER rate, AQY, etc.). Yet, the characteristics of 5 wt% MoS 2 /Zn 0.5 Cd 0.5 S are comparable with other photocatalysts synthesized by more straightforward hydro-or solvothermal procedures. The MOF-templated synthetic approach is promising, but further research is needed to assert its potential in industrial applications.
As an alternative, the formation of type-II homojunctions (twinned) of zinc blend/wurtzite Cd 0.5 Zn 0.5 S combined with MoS 2 nanosheets with S vacancies resulted in an advanced HER of 69.25 mmol g À1 h À1 and AQY of 55.2% (420 nm). In this case, MoS 2 nanosheets provided the composite with many catalytically active sites due to the increase in structural defects. Additionally, they improved the light-harvesting ability of the heterostructure (E g = 1.56 eV, E VB = 1.44 V, E CB = À0.12 V vs. RHE) as compared to the single Cd 0.5 Zn 0.5 S (E g = 2.61 eV, E VB = 2.17 V, E CB = À0.44 V vs. RHE). The twinning structure of Cd 0.5 Zn 0.5 S prevented charge recombination in bulk; meanwhile, the nanosheets' good conductivity and reduced charge transfer distance ensured the efficient transfer of electrons and holes from Cd 0.5 Zn 0.5 S to MoS 2 catalytically active sites, initiating HER. [108]

ZnIn 2 S 4 -Based Photocatalysts
Zinc indium sulfide, ZnIn 2 S 4 (ZIS), is another promising photocatalyst. Its relatively narrow bandgap (E g % 2.0-2.9 eV) makes this semiconductor active in the visible light spectrum. It can be synthesized under mild conditions via simple, environmentally friendly, cost-effective hydrothermal methods from zinc chloride, indium nitrate, and thioacetamide. ZIS is less toxic than CdS, which is crucial for material handling in mass-scale applications. The hexagonal crystal structure of ZIS is the most thermodynamically and chemically stable phase and it is typically used as a visible and NIR photocatalyst. The absorption edge of ZIS at 550 nm can be further tuned via combination with other semiconductors, for example, modified with Ni as a cocatalyst, MoS 2 QDs, or in dual-Schottky junctions with UV-NIR-active Co 9 S 8 (E g = 1.30 eV) and PbS (E g = 1.52 eV). [109][110][111] The broadened ZIS spectral response and enhanced light absorption inhibit charge recombination and reduce charge transfer resistance. Additionally, the photocatalyst's stability reached 100 h and the HER rate increased by more than ten times compared to bulk ZIS. [110] NiCo 2 S 4 as cocatalyst provided an enhanced local electric field at the interface with ZnIn 2 S 4 (Schottky junction formation). That, combined with the photothermal effect, contributed to an enhanced charge transfer ability from ZIS to NiCo 2 S 4 co-catalyst, H─OH bond activation of adsorbed water, and decreased reaction activation energy. [112] As discussed earlier, element doping can improve catalyst performance by facilitating charge separation and narrowing the bandgap of a semiconductor. [67] Valuable insights into ZIS cation doping were published by Wang et al. By substituting only 0.5% of the Zn atoms with Cu, the ZIS crystal structure was maintained, and the Cu dopant acted as an electron trap instead of a recombination center. On the other hand, excess doping resulted in crystal distortion, higher resistance to charge transfer, and a VB upshift. [113] The latter could potentially lead to a limited electron/hole separation.
2D monolayers of ZIS were doped with 0.61 mol% of Ag via cation exchange reaction. As a result, S nanoholes were distributed across the entire Ag-ZnIn 2 S 4 nanosheets. Dual defect engineering narrowed the material's bandgap from 2.38 to 2.13 eV (E VB = 1.61 V, E CB = À0.52 V), increasing the density of states (DOS) around VB and CB and therefore improving the charge carrier mobility. The photocatalyst exhibited high stability over seven cycles (total of 56 h) of irradiation and HER efficiency of 7.3 mmol g À1 h À1 with Na 2 S/Na 2 SO 3 as a sacrificial agent. [114] www.advancedsciencenews.com www.advenergysustres.com Adv. Energy Sustainability Res. 2023, 4, 2200201 The introduction of anion (sulfur) vacancies is commonly used to regulate the local ion coordination and to promote visible light absorption of a semiconductor. [67,68,115] However, integration of anion vacancies into the photocatalyst's lattice is beneficial as long as the reaction is performed in an aqueous solution.
After the initial H 2 S decomposition, S-intermediates might prefer to interact with other S-species in solution rather than bind to a metal cation of the catalyst and occupy the vacancies. However, suppose the reaction is carried out in the gaseous phase under elevated H 2 S partial pressures and high temperatures. In that case, S atoms adsorbed on the catalyst surface are more likely to bind with metal cations and occupy the vacancies. [116] That could lead to a rapid decrease in charge trap states and consequently, charge recombination and deactivation of the catalyst. Further comprehensive information on ZIS preparation techniques, morphologies, and ZIS-based photocatalysts can be found in the detailed review by Song. [117] In brief, the critical problems of ZIS-based photocatalysts are as follows: 1) the effect of defects on its photocatalytic activity should be investigated; 2) the large-scale production of ZIS should be developed; 3) more easily recoverable materials (not powders) should be developed; and 4) more efforts should be made to put ZIS into an application for HER.

CdS and ZIS Potential for H 2 S Decomposition
Photocatalysts based on CdS, its solid solution with ZnS, and ZIS combined with other semiconductors and cocatalysts have demonstrated superior performance in photocatalytic liquid-phase H 2 evolution. These metal sulfides allowed to achieve drastically improved HER rates of tens and even a hundred mmol g À1 h À1 , along with elevated AQY, prolonged photocatalytic activity, and stability up to 100 h. [87,88,92,93,108,110,111] CdS, ZnS, and ZIS are easy to synthesize via typical hydro-or solvothermal procedures. Notably, the high toxicity of CdS (health hazard class 3) has to be considered prior to any large-scale production and widespread use. Overall, the reported advanced catalytic performance makes these materials excellent candidates for state-of-the-art photocatalysts for in situ solar fuel production from gaseous H 2 S.
Nevertheless, photocorrosion remains an open challenge in CdS-based and ZIS photocatalysts. Many H 2 evolution reactions are performed at 4-15°C to avoid material overheating and degradation by light. However, reactions under reservoir or field conditions and increased gas pressure (high H 2 S concentration) could provide enough S-containing species to be oxidized by photogenerated holes. As a result, most charges should be consumed by the desired photocatalytic reaction rather than wasted on the catalyst's self-oxidation process. Efficient charge separation and transfer from the bulk to the catalyst's active sites is essential for this process. The junction formation between these semiconductors and low-cost, abundant cocatalysts, for example, Ni, NiS, and MoS 2 , improves the charge separation and thus, the resistance to photocorrosion, broadens their absorption range, and increases the number of catalytically active sites.
Additionally, to avoid surface passivation, an ideal catalyst should have optimum adsorption/desorption of hydrogen and sulfur-containing intermediates. Hence, the in-silico catalyst design supported by DFT calculations of key parameters, such as Gibbs free adsorption energies of reaction intermediates (ΔG H* ! 0), could be a promising way forward. [33,39,118,119] Furthermore, most research has been focused on the morphology of photocatalytic nanostructures. The slightest changes in those parameters can significantly affect the amount of available catalytically active sites, consequently affecting the HER efficiency. Ideally, bulk MS-based semiconductors or thin films should be designed for the task, as nanostructures may not be stable under harsh in situ reaction conditions. However, suppose any specific morphology and nanostructure is deemed suitable, then encapsulating the photocatalyst into, for instance, semiconducting graphene shells could preserve its integrity. [120] The presented MS-based photocatalysts for H 2 O splitting are summarized in Table 2.

Other Metal Sulfides
Photocatalytic CO 2 reduction, degradation of environmental pollutants, and N 2 fixation have received growing attention in academia and industry. H 2 S photocatalysts could be derived from these applications provided their energy-level alignment is suitable for H 2 S decomposition, they demonstrate activity in the Vis-NIR spectrum, and maintain chemical and photostability.
For instance, bismuth sulfide-based catalysts offer excellent visible light absorption (E g % 1.3-1.8 eV vs vac). These materials can act as sensitizers in heterojunctions to broaden the semiconductors' light-harvesting range. [121] The inclusion of Bi 2 S 3 nanorods (E g = 1.38 eV) boosted the light-harvesting ability of the semiconductor Zn 0.5 Cd 0.5 S (E g = 2.47 eV). The photocatalyst demonstrated 25.71% photothermal conversion efficiency and good photothermal stability, promoting the thermocatalytic effect in photocatalysis. The photothermal effect increases the temperature in the catalyst's microenvironment, consequently promoting (temperature-activated) charge carrier transfer. At the same time, the type-II heterojunction architecture suppressed the recombination of photogenerated charges. It increased the photocurrent density by hot electron injection from the CB of Bi 2 S 3 to the CB of Zn 0.5 Cd 0.5 S. [122] Bi 2 S 3 nanorods (E g = 2.47 eV) were used to degrade Congo red dye under 500 nm irradiation. According to the researchers, the orthorhombic crystalline structure of the nanorods enhanced the photodegradation efficiency by decreasing the charge recombination rate. [123] However, bulk Bi 2 S 3 or other Bi 2 S 3 nanorod structures might be more attractive options since they can provide more efficient solar energy utilization due to their bandgap of %1.33 eV and thus, its activity in the NIR spectral region (λ ≤ 900 nm). [124][125][126] Bismuth sulfur iodate BiSI exhibits n-type conductivity and a bandgap of %1.38 eV, enabling efficient solar spectrum utilization. Yet, it suffers from fast charge recombination. The performance of the semiconductor was improved by introducing S vacancies and O-doping. O-doping changes the surface characteristics, making it less hydrophobic and more positively charged, while S vacancies inhibit charge recombination. Unfortunately, the material demonstrated photocatalytic activity only for four cycles of 30 min each. [127] Reduced photoluminescence intensity was observed for 30 wt % MoS 2 /BiSI heterojunction. This result indicates more efficient  [108] www.advancedsciencenews.com www.advenergysustres.com electron/hole separation compared to bulk BiSI, which should increase its photocatalytic activity. [128] Combining MoS 2 as a cocatalyst with the stable visible light-responsive BiSI makes this heterostructure appealing for H 2 S photocatalysis. However, a deeper understanding of its photoredox properties is still needed to evaluate its potential for further development and application. Numerous nonsulfide Bi-based photocatalysts have been reported over the past years. [121,129] However, they may be unsuitable for H 2 S splitting due to their insufficient chemical stability in H 2 S-rich environments found upstream. On the other hand, bismuth sulfides discussed above offer activity across a wide range of the solar spectrum. Nonetheless, the application of Bi 2 S 3 and BiSI photocatalysts appears to be relatively unexplored. For instance, their energy band structure is often unclear, as well as the precise causes of the rapid deactivation of these semiconductors and their full potential for hydrogen evolution. These significant gaps in understanding bismuth sulfide photoredox behavior make them attractive for academic research rather than for industrial applications.

Conclusion and Outlook
The latest state-of-the-art photocatalytic semiconducting materials, mainly metal sulfides, were reviewed, and their potential for upstream H 2 S conversion into hydrogen fuel and elemental sulfur was assessed. Numerous MS-based photocatalysts with improved performance have recently been developed for H 2 evolution via water splitting. Even though H 2 S dissociation is thermodynamically a much less energy-demanding process, photocatalytic H 2 S conversion into valuable chemicals, primarily via gas phase reaction, remains largely unexplored. The reason is the high toxicity of H 2 S and the specificity of the industrial process, which require the design of a reactor and a catalyst that are both physically and chemically stable in the harsh H 2 S-rich environment.
Due to their thermodynamic and chemical properties, metal oxide-based semiconductors may be unsuitable for H 2 S decomposition in the upstream energy industry. Alternatively, other well-known photocatalysts, such as CdS, ZnS, MnS, and their solid solutions, as well as ZnIn 2 S 4 , demonstrated promising HER efficiencies while being physically and chemically relatively stable in H 2 S atmosphere. Consequently, studying in situ H 2 S photoconversion at the surface and subsurface using these materials is a promising path for developing environmentally friendly industry-relevant photocatalytic technologies for solar fuel production and sulfur recovery. However, several challenges still limit the practical application of these materials. First, catalyst surface passivation by elemental sulfur, and thus rapid catalyst poisoning, is one of the main issues in photocatalytic H 2 S decomposition. In this regard, the reservoir and wellbore conditions could be advantageous to product separation and catalyst recovery.
Second, the susceptibility to photocorrosion reduces the photocatalyst's activity to a few tens of hours, which is modest from an industrial perspective. Hence, most liquid-phase systems depend on a sacrificial agent that facilitates the transfer of photogenerated holes from the semiconductor before the lattice S 2À ions can be oxidized. The photodegradation of a semiconductor in the gas phase remains unresolved; the role of S n 2À species in this problem remains an open question. Heterojunction engineering using multiple MS-based semiconductors is a promising technique to promote charge separation and prevent the catalyst's self-oxidation. However, the heterostructure design and precise H 2 S decomposition mechanism still lack fundamental understanding and require further systematic studies, including computational analysis. Additionally, anion (O, N, B, I) and cation (Ni, Cu) doping can facilitate charge separation and improve photostability by introducing charge carrier trap states. Importantly, the dopant should be chosen carefully, and its concentration should be optimized to boost the photocatalyst's performance effectively. For instance, the incorporation of dopants with large ionic radii can lead to lattice distortion, and thus, loss of semiconductor's crystallinity, while an excessive amount of a dopant can induce the bandgap broadening, therefore, limiting the spectral response of a semiconductor. [130,131] Third, HER efficiency can be improved by photocatalyst modification with cocatalysts. For instance, MoS 2 or WS 2 have demonstrated similar cocatalytic activity as noble metals. They lower the reaction's kinetic barriers, while reducing the cost of the material as compared to noble metals. Notably, an ideal photocatalyst should also harvest as much solar spectrum as possible. In order to maximize the spectral response of CdS-or ZnS-based semiconductors, they can be combined with emerging NIRactive photosensitizers such as Bi 2 S 3 or BiSI. Nevertheless, their photophysical properties still need to be explored.
The nanostructure of photocatalysts often determines their performance. It dictates the number of photocatalytically active sites and their morphology. Additionally, most reported materials are obtained as powders, which are challenging to recover and recycle. However, industrial applications necessitate the development of bulk, robust, and recyclable photocatalyst with sustained photocatalytic activity and physicochemical stability at high  [99] Mn-doped Zn 0.5 Cd 0.5 S Solvothermal 300 W Xe lamp (λ ≥ 420 nm) 0.6 -2.37 0.5 м Na 2 S, 0.5 м Na 2 SO 3 16 h stability, 5 o C 2022 [98] O-doped MoS 2 / Mn 0.5 Cd 0.5 S Hydrothermal, ultrasonic 300 W Xe lamp (λ ≥ 420 nm) 84.3 46.9 (420) 1.58/2.25 0.5 м Na 2 S, 0.5 м Na 2 SO 3 16 h stability 2020 [100] www.advancedsciencenews.com www.advenergysustres.com temperatures and pressures. Therefore, upscaling photocatalyst production and photoreactor design should be investigated further. To ensure sustained high-yield solar fuel production and sulfur recovery, the effect of in situ reaction conditions, such as high temperatures, pressures, H 2 S concentration, etc., on the catalyst performance should be studied during the transition from laboratory-scale reactions to macroscale testing. The photocatalytic H 2 S conversion under wellhead or wellborelike conditions has not been reported thus far. Another challenge emerges from delivering high-power photonic energy to upstream conditions. In recent work, kilowatt-range coherent photonic power has been successfully delivered to relevant upstream conditions through the use of fiber optics. [132][133][134][135] Other recent developments have also coupled incoherent concentrated solar light to fiber optics for diverse applications including photocatalysis. [136,137] That being said, upstream H 2 S to hydrogen conversion is an emerging field and requires thorough systematic theoretical and experimental studies before implementing it into industry-relevant processes.
To conclude, academia and industry should look at H 2 S gas beyond being an industrial nuisance and instead as a potential valuable precursor for green H 2 production and sulfur recovery. This is by no means a trivial task; nevertheless, it is both economically and environmentally critical to develop carbon-neutral catalytic processes powered by sustainable and renewable energy sources. The state-of-the-art materials, their current limitations, and possible ways forward toward an educated catalyst design presented in this review should provide guidance for future work in the upstream photocatalytic H 2 S decomposition.