Regulating the Photoelectric Properties of ZnIn₂S₄: Structure, Interfaces, and Applications

1. Introduction

The transition toward low-carbon energy is taking place alongside a growing environmental burden [1,2,3]. Electricity, fuels and chemical products still depend heavily on energy-intensive processes, while the treatment of polluted water often requires additional chemicals and separation steps [4,5,6]. Solar energy can provide a cleaner source of driving force, but its practical value depends on whether light can be converted into useful chemical work with reasonable efficiency and selectivity [7,8,9]. This challenge is shared by several fields that are often discussed separately. Hydrogen production requires the storage of solar energy in a fuel. Carbon dioxide conversion seeks to use an abundant carbon source while avoiding uncontrolled products. Water treatment aims to destroy persistent molecules without transferring the problem to a secondary waste stream. At the same time, rapid monitoring of trace contaminants and biological markers requires photoactive materials that can generate stable and interpretable electrical signals. These tasks all depend on how photogenerated electrons and holes are created, separated and delivered to a surface or electrode. They also expose the limits of laboratory comparisons, because reaction media, product analysis and device configuration can change the apparent benefit of a material. The search for versatile visible-light absorbers therefore belongs to a wider effort to connect solar conversion, environmental remediation and chemical sensing [10,11].

Semiconductor photocatalysis links sunlight with reactions that would otherwise require chemical or electrical input [12]. The demands are different in each case. Hydrogen evolution needs electrons and accessible proton-reduction sites [13]. CO₂ conversion adds the problem of binding and activating a stable molecule while controlling product selectivity [14]. Water treatment is governed by the interaction of the catalyst with the pollutant and with the surrounding matrix [15]. A photoelectrochemical sensor has a further requirement [16]. Its electrical response must be reproducible for a defined analyte. Visible-light absorption helps in all of these systems, but it does not decide the result by itself. ZnIn₂S₄ is attractive because it can be obtained as nanosheets or hierarchical assemblies and can be connected to other semiconductors without changing the whole synthesis route. Early studies of layered ZnIn₂S₄, together with later hydrogen-evolution reviews, established the importance of its morphology and interfaces [17,18]. More recent work on nanosheets and hollow structures has shifted attention toward local coordination, wetting and carrier transport [19,20]. The reported rate is therefore only the starting point. The more useful question is why the same modification helps one reaction and has little effect in another.

A ZnIn₂S₄ powder and a ZnIn₂S₄ electrode do not experiences the same experiment. In a slurry, the important variables include exposed surface, mixing and contact with dissolved reactants. The product still has to be quantified, and the source of the electrons or holes must be accounted for. Once the material is deposited on an electrode, adhesion, film resistance, applied potential and electrolyte become part of the result. A larger photocurrent may make a better transducer, but the assay also needs analyte recognition, calibration, interference tests and sample recovery. Photoelectric sensing is therefore treated here as a separate application rather than another activity number. Focused reviews have already discussed ZnIn₂S₄ hydrogen generation and irradiation conditions [21,22]. The present review takes a narrower route. It links each structural change to the process it is expected to influence and then considers how well the available measurements support that link. Representative studies are used in place of a catalogue of every composite so that reaction conditions and qualifications are not lost.

ZnIn₂S₄ does not have a single mechanism that applies to every reaction. Synthesis may change phase, thickness, defect population and surface residues together, and the surface can continue to evolve under illumination. The discussion therefore treats charge-transfer labels as working assignments and considers them in relation to the measurements reported for each system.In this review, ZnIn₂S₄ is abbreviated as ZIS when the composition is unambiguous. The discussion first outlines its crystal structure, morphology, electronic properties and surface chemistry, followed by the main strategies used to regulate these features. The applications are grouped into hydrogen evolution, CO₂ reduction, environmental purification, selective organic transformations, H₂O₂ production and photoelectric sensing. Solid-state photodetectors are treated separately from PEC sensors because their performance is evaluated with different parameters. PEC, AQY and FTO denote photoelectrochemical, apparent quantum yield and fluorine-doped tin oxide, respectively. The terms type-II, Z-scheme and S-scheme are retained when used by the original authors, but these labels are not considered direct evidence of a charge-transfer pathway. Mechanistic assignments are therefore discussed together with the measurements supporting them and the stability of the material during operation. Studies of related semiconductors and functional materials have examined how vacancies, interfaces, composition and local coordination affect photo response. These reports provide context for the characterization principles discussed here, but they do not establish the same behavior in ZnIn₂S₄.

2. Structural and Electronic Basis of ZnIn₂S₄

Before discussing individual modification strategies, it is useful to relate the structure of ZnIn₂S₄ to the processes it supports. The crystal framework determines local coordination and available surfaces, while light absorption generates carriers that must reach reaction sites or be collected by an electrode. These links between structure, carrier behavior and photocatalytic or photoelectric output are summarized in Figure 1. The layered structure shown is representative; the phase and surface features of ZnIn₂S₄ depend on the synthesis conditions. The following section describes its crystal structures and synthesis-dependent surfaces [23,24,25,26,27].

2.1. Crystal Structure and Synthesis-Dependent Surfaces

The hexagonal layered form is commonly used in photocatalytic studies, although ZnIn₂S₄ also has other reported crystal structures. Phase labels should therefore come from diffraction and, where needed, local structural measurements of the actual sample [28]. A schematic crystal model cannot stand in for this verification. In layered materials, the arrangement of Zn, In and S across the sheet gives a different chemical environment to a basal plane and an edge [29,30]. The distinction matters because a large external surface area does not necessarily mean a large population of the sites that perform a given reaction [31,32]. The rhombohedral structure examined in a ZnIn₂S₄ phototransistor also illustrates why optoelectronic conclusions from one polymorph should not be transferred automatically to a typical hydrothermal nanosheet [33,34,35,36,37].

Hydrothermal and solvothermal routes often produce flower-like microspheres assembled from thin sheets [38,39]. The open assembly may expose edges and scatter light, although a dense core and overlapping sheets can hide part of the surface. Isolated nanosheets shorten the path from the bulk to the solution, but they may restack after drying and are not always easy to recover. A powder that disperses well in a stirred vial may therefore be a poor continuous coating. The useful morphology depends on the reaction medium and on the form in which the material will finally be used [40,41,42,43,44].

The synthesis route leaves chemical traces as well. Thiourea or thioacetamide, solvent choice and post-treatment can change sulfur content, surface oxide coverage and adsorbed residues. A small XPS shift may result from oxidation, charging or a change in coordination [45]. EDX is valuable for a compositional survey, but it cannot locate a dilute surface defect with confidence [46,47]. Claims about a particular facet, vacancy concentration or isolated site should therefore be supported by methods with the required spatial and chemical resolution. A comparison made from one precursor batch before and after a controlled treatment is especially useful because it reduces the number of variables that can explain a new rate [48,49,50,51].

Precursor chemistry can determine how a sample behaves after it is placed in water. Residual sulfur-containing ligands may alter wetting, and a weakly bound oxide layer can evolve during the first minutes of illumination. These changes are easily missed when characterization is limited to dried powders. For repeated use, the initial suspension, recovered solid and reaction solution should be considered together, especially when the rate changes between cycles.

2.2. Light Absorption, Carriers and Surface Reactions

Visible-light response is one reason ZnIn₂S₄ is used so often for hydrogen evolution. After a photon is absorbed, the carriers still have to survive recombination, reach a surface and complete the two half-reactions [52,53]. Band edges and optical gaps are frequently treated as fixed material constants, although phase, stoichiometry, surface chemistry and the measurement method can move the reported values. A Tauc plot estimates optical absorption. It does not directly give the redox potentials of a working interface. Flat-band data are useful only when the reference electrode, pH and conversion procedure are stated, and calculated vacuum-level positions should not be mixed silently with electrochemical measurements.

The surface can be as important as the band structure. Proton reduction requires a suitable adsorption site and a route for transferring charge to hydrogen intermediates [54,55]. CO₂ reduction must activate a stable molecule while competing with hydrogen evolution. H₂O₂ production depends on a selective two-electron oxygen-reduction path and on avoiding peroxide decomposition. In water treatment, a strongly adsorbed pollutant can block the surface, while the active oxidant may come from dissolved oxygen rather than from direct hole transfer [56]. A defect that helps adsorption in one reaction can therefore trap carriers or accelerate an unwanted reaction in another.

A lower steady-state PL intensity, a smaller impedance arc or a larger photocurrent may support an argument for improved charge separation, but none of these signals has a single interpretation [57,58]. PL can fall when non-radiative traps increase, and impedance measured on an electrode does not directly give the turnover of a powder in suspension. Transient measurements, selective deposition, surface photovoltage and operando spectroscopy are more persuasive when they are measured under conditions close to those used for product formation. A mechanism is convincing only when the materials data and the chemical products point in the same direction.

2.3. Sulfide Stability and Working-State Evidence

The sulfur framework also creates a stability problem. If the intended oxidation reaction is slow, photogenerated holes may attack the sulfide. Sulfur loss, partial oxidation and dissolved Zn or In can then change the surface during a long run [59,60,61]. Reusing a recovered powder may not reveal the change, especially when every cycle starts with fresh solution. Continuous product monitoring and post-reaction composition are more informative than an unchanged final rate. The issue is especially serious in overall water splitting, where the sulfide must support water oxidation without a sacrificial hole scavenger [62]. The adjacent-site electronegativity study offers one example in which activity and stability were considered together [63].

Results from different ZnIn₂S₄ samples should first be read in light of the specimen that was tested. Suspended powders, supported films and isolated nanosheets expose different surfaces and receive different amounts of light. Their working interfaces are also unlike a dried powder prepared for XPS. Water, adsorbed reactants, a binder and the electrolyte double layer can change surface charge and access to sites. For this reason, a comparison between studies is meaningful only when the physical form, illuminated area and chemical surroundings are clear [64,65]. Optical and electrochemical measurements help interpret performance but cannot replace product analysis. Greater light absorption does not establish useful charge separation, while a larger photocurrent from a deposited film does not measure the quantum yield of a powder in suspension. Stability requires a similar distinction. Short reuse cycles show whether a catalyst can be handled repeatedly, whereas continuous operation can reveal slow drift and accumulation of dissolved or oxidized species [66,67]. If the surface reconstructs yet stays active, the reconstructed surface should enter the description of the working material.

The duration of a test matters because the chemistry may change during the test itself. A catalyst can retain its initial rate for several short cycles while slowly losing sulfur or releasing metal ions into the solution [68,69,70]. Conversely, a mild surface oxidation may lower the first rate but create a more stable working layer during later cycles. Time-resolved gas or liquid analysis, together with a final measurement of dissolved Zn, In and sulfur species, would help distinguish these situations. Such data are particularly valuable for overall water splitting and for sensor electrodes that are repeatedly switched between light and dark [30,71,72].

3. Structural Engineering Strategies

The following sections consider each modification in relation to the step it is intended to affect as shown in Figure 2. Thinner sheets may shorten the path to a surface, but can also restack or expose more edge sites. Vacancies can change adsorption and may become recombination centers after hydration or oxidation. Cocatalysts and junctions affect carrier collection only when their contact is close and remains stable during reaction. Because synthesis often changes several features together, the discussion considers material structure, reaction conditions and controls alongside the reported activity. For sensing electrodes, film integrity and receptor loading also affect the measured signal.

3.1. Thickness, Porosity and Hierarchical Assembly

Thin sheets expose more atoms to the solution and shorten the distance that carriers travel before reaching a surface. This is the main reason for using exfoliation and controlled sheet growth. Thickness, however, is not an isolated variable. Smaller sheets have a different edge-to-basal-plane ratio and may contain more defects or adsorbates [73,74,75,76]. They can also restack during drying. A meaningful comparison should therefore give the thickness distribution, exposed structure and reaction conditions rather than relying on the word ultrathin. Film continuity and electrical contact become equally important when the sheets are deposited on a substrate [77,78,79,80].

Porous assemblies offer a different balance. They allow light and reactants to reach internal sheets, while products can leave through the same network. Ding and co-workers compared hollow flowers made from defective ZnIn₂S₄ nanosheets with less favorable structures in hydrogen evolution [81]. Jia et al. later examined a porous, vacancy-containing preparation and reported a high donor-assisted H₂ rate. These studies support the value of an open architecture, but porosity and vacancy density changed together in the latter case. A BET area measured on a dry powder is also not the same as the wet, illuminated area in a reactor. Pore accessibility, settling and gas bubbles may all change the effective surface during operation.

The morphology selected for a specific application should follow its mass-transfer needs [82,83,84]. In a hydrogen slurry, a loose flower can maintain dispersion while presenting many accessible sheets. For CO₂ conversion, the adsorption of gas or dissolved bicarbonate and the local supply of proton donors’ matter. In pollutant removal, a porous material can concentrate an organic molecule before oxidation, making dark adsorption controls essential. In a PEC sensor, highly textured films can load more receptor but also increase nonspecific adsorption and background current. These distinctions argue against a single morphology ranking across all reported ZnIn₂S₄ applications [85,86,87,88,89].

3.2. Vacancies and Composition Control

Sulfur vacancies are often used to explain stronger absorption, altered charge distribution and enhanced reactant adsorption. They can change local coordination, but the experimental task is to show where the vacancies are and how many are present, then separate their effect from changes in morphology or oxidation state [90,91,92]. EPR, XPS and elemental analysis can support one another when appropriate controls are included. None of them, by itself, counts the vacancies on a working surface. Excessive defect density may instead create deep recombination centers. Activity should therefore be considered as a function of defect population rather than assumed to increase continuously. One study introduced Zn vacancies through sulfhydryl coordination and examined their influence on CO₂ photoreduction [93]. Metal vacancies present a related but distinct case because they can change CO₂ binding and local charge distribution. The assignment should remain consistent with the preparation chemistry and with the measurements made after reaction [94,95,96,97,98].

Changing composition or introducing heteroatoms may affect the adsorption site, the oxidation half-reaction and the durability of a sulfide simultaneously. Oxygen doping combined with surface sulfur vacancies has been studied for visible-light overall water splitting [99]. The claim is more demanding than a donor-assisted H₂ rate because both water oxidation and hydrogen evolution need to be accounted for. Phosphorus filling of surface sites in few-layer ZnIn₂S₄ has also been investigated in sacrificial-agent-free hydrogen production [100]. These examples are best compared by what the added atom does locally, because doping and vacancy chemistry often occur together. A phosphorus atom occupying or interacting with a vacancy can alter the identity of the reactive site itself. Defect models need an operating-state qualification. A vacancy favorable in a DFT slab may be hydrated, oxygenated or occupied by a reaction intermediate in water. The lowest calculated adsorption energy for a chosen intermediate does not establish the full catalytic route. Isotopic labeling, in situ vibrational signals, controlled defect healing and measurements after extended illumination can help connect an idealized structural model to the real catalyst. These tests are especially important for CO₂ and H₂O₂ reactions, in which the adsorbed intermediate and competing reduction path determine the product.

3.3. Cocatalysts and Interfacial Active Sites

A cocatalyst may lower a surface-reaction barrier or collect carriers before they recombine. Noble metals are convenient electron collectors in hydrogen studies, yet loading, particle size and deposition route also change the number of exposed sites [101,102]. Non-noble components can be attractive for practical reasons, although a high powder rate is not a cost analysis. The useful controls include bare ZnIn₂S₄, the isolated cocatalyst and a loading series tested with the same light and donor. Dai et al. photo-deposited Ni(OH)₂ and nickel phosphide species on ZnIn₂S₄/ZnS [103]. The design assigns different interfacial functions to the two nickel-containing phases, but the final composition and distribution still need to be confirmed after deposition. A nominal precursor ratio cannot show which phase works during illumination. In a different route, an In-MOF precursor generated a ZnIn₂S₄ heterojunction for hydrogen formation without an added cocatalyst [104]. Its performance should be compared with the complete precursor-derived surface, under the solution and illumination used for the H₂ test [105,106,107,108,109].

At the atomic scale, an isolated metal site may alter local adsorption and improve atom utilization. Its assignment requires microscopy and coordination analysis that exclude small clusters. The site should also be examined after reaction, since an atom bound to the pristine lattice may migrate, coordinate to oxidized sulfur or aggregate under illumination.

3.4. Heterojunctions and Mechanistic Evidence

Contact with a second semiconductor can change absorption, local fields, adsorption and carrier collection, but the result depends strongly on how the two phases meet [102,110,111]. Zhang et al. reported a sheet-on-sheet ZnIn₂S₄/g-C₃N₄ structure for H₂ formation and related its performance to an intimate interface [112]. Later work introduced defects into both ZnIn₂S₄ and carbon nitride [113]. These systems cannot be reduced to one band diagram because their contact geometry and local chemistry are different. A physical mixture is a useful control for showing whether the interface itself matters. The label S-scheme requires further evidence. Type-II transfer, Z-scheme-like recombination and defect-mediated transfer can all produce a larger steady-state photocurrent. Work-function and light-dependent surface-potential measurements, spatially resolved deposition and selective radical tests can help separate these possibilities. When the surviving carriers are not located directly, the transfer path should remain a proposal rather than a settled fact [114,115,116,117,118].

Interfaces can also be designed for a particular chemical task. A MoO_3-x@ZnIn₂S₄ core-shell structure was studied for photothermal-assisted CO₂ conversion [119]. Here the thermal contribution and the charge-separation proposal both matters. A comparison without matching temperature would overstate the purely photocatalytic [120] advantage. Growth of ZnIn₂S₄ on amine-functionalized SiO₂ addressed CO₂ adsorption and nanosheet arrangement, showing that a support can contribute through reactant supply even if it does not form an electronically active junction. The appropriate comparison is therefore between complete reaction environments, with light flux, temperature, feed and support all specified. [121,122,123,124]

Synthesis rarely changes one feature alone. Vacancy formation can alter sheet thickness or wetting, while cocatalyst deposition can cover existing sites. Matched series that vary one parameter at a time are therefore more informative than a comparison of an optimized composite with pristine ZnIn₂S₄. A loading series can reveal the balance between interfacial contact and light shielding; defect studies can use controlled vacancy creation or healing to test whether activity follows the assigned site. Such controls are rarely perfect, but they help distinguish a cause from a structural change that accompanies synthesis.

For a heterojunction, work functions and surface potentials help establish the direction of charge redistribution, while transient measurements describe its timescale [125]. Their interpretation must be consistent with the product distribution. A proposed S-scheme should be presented as a working explanation when the evidence is indirect, especially if the partner semiconductor or the ZnIn₂S₄ surface changes during irradiation. Reporting the assumptions behind the model makes comparisons among junctions more useful than assigning every improved composite to the same pathway [126,127,128,129,130].

Reproducibility matters for both powders and electrodes. Independent batches can show whether a morphology or interface is controllable, while supported films should be assessed for loading, adhesion and electrical contact across replicate electrodes. A flower-like powder prepared in a small vial may have a size distribution that is difficult to retain when precursor concentration, mixing or heating history changes. Conversely, a film may appear uniform while its current varies from electrode to electrode. Reporting this spread identifies which preparation step limits transfer to repeatable reactions or sensing devices [131,132,133,134,135].

4. Photocatalytic Applications

Applications are grouped by measured output rather than by composite name. Donor-assisted H₂ evolution is separated from overall water splitting because a sacrificial donor supplies the oxidation half-reaction. CO₂ conversion is discussed with carbon-source controls and product balance in mind, while environmental purification is evaluated by distinguishing adsorption and parent-compound loss from mineralization. H₂O₂ production and selective organic reactions are considered through the products formed on both sides of the reaction where these were reported. Work on other semiconductor and catalyst systems has considered how morphology, porosity, heterojunctions and composition influence light-driven reactions. These examples are useful for framing materials design, although their activity values cannot be transferred directly to ZnIn₂S₄.

4.1. Hydrogen Evolution

Green hydrogen can be produced by renewable-powered water electrolysis or through solar-driven water splitting. Photocatalytic hydrogen production is of interest because semiconductor absorbers can use sunlight to drive chemical conversion directly, potentially avoiding separate photovoltaic and electrolyser units. Its practical performance, however, depends on light utilization, charge transfer, product separation and catalyst stability. Many studies use sacrificial donors to consume photogenerated holes; these demonstrate hydrogen evolution but do not establish overall water splitting or, on their own, a complete green-hydrogen process.

Hydrogen evolution is the most developed application of ZnIn₂S₄, but the terminology used for it still matters. Many studies use triethanolamine, lactic acid or sulfide/sulfite as a sacrificial donor, sometimes together with a deposited cocatalyst. The donor supplies the oxidation half-reaction, so the product is photocatalytic H₂ evolution rather than overall water splitting. Rates can be compared only when the lamp spectrum and intensity, catalyst mass, reactor geometry, donor concentration and normalization are known. An AQY at a defined wavelength is helpful, although the calculation still depends on whether incident or absorbed photons are used. The morphology studies illustrates the point. Ding et al. used hollow flowers assembled from defective ZnIn₂S₄ sheets, while Jia et al. reported 61,506 ± 4746 μmol g^-1 h^-1 of H₂ and an apparent quantum efficiency of 56.53% at 420 ± 15 nm for a porous, vacancy-containing material [81]. The second result was obtained with a donor and should be read together with its lamp, catalyst loading and solution. It is not an overall-splitting rate. Both reports suggest that open surface architecture can matter without changing the bulk band gap, but BET area or PL intensity alone does not tell how many sites receive light in a stirred suspension. Action spectra, transient measurements and comparable catalyst loadings are needed for a closer comparison. Sedimentation can further reduce the illuminated fraction of a powder during the experiment [136,137,138,139,140,141].

A junction adds another possibility. The sheet-on-sheet g-C₃N₄ example placed two light-responsive solids in extensive contact. Defect-containing ZnIn₂S_4-x/g-C₃N_4-x [113] added local electronic states to that interfacial picture. Both reports may show higher donor-assisted H₂ output than their component controls, but the explanation need not be identical. An interfacial potential can change carrier direction while a defect may change the proton-binding surface. Loading experiments are particularly revealing. A thin partner layer can improve contact, whereas too much of it can shade ZnIn₂S₄ or block the surface. This often produces an optimum rather than a continuous increase with composite content. Cocatalysts address the surface step more directly. In Dai et al.’s Ni(OH)₂/Ni_xP-modified ZnIn₂S₄/ZnS system [103], two nickel-containing components were deliberately placed on a sulfide framework. To interpret such a result, one must know whether nickel phases collect electrons, assist water or donor oxidation, change wetting, or serve more than one function. The in-situ route itself can create a ZnS-rich interface and alter the underlying ZnIn₂S₄, making a bare ZnIn₂S₄ comparison only the first of several necessary controls. For the MOF-derived junction [104], the preparation route may supply an intimate interface without a separate cocatalyst deposition. Its value is best judged against the complete precursor-derived surface, not merely the presence or absence of a noble metal [142,143,144,145,146,147,148].

A much smaller set of studies tackles pure-water overall splitting. The oxygen-doped and vacancy-containing ZnIn₂S₄ reported in [99] and the green hollow gZIS (Figure 3) reported by Chong et al.[19] explicitly address this more difficult task. Xin et al. linked adjacent-site electronegativity differences to activity and stability in overall splitting [149]. Such claims demand H₂ and O₂ quantification, a plausible stoichiometric relationship, exclusion of sacrificial sulfur oxidation, and evidence that the solid remains chemically intact. Ideally, a gas balance is accompanied by isotope tests or dissolved-species analysis where source assignment is uncertain. A daytime stability run is meaningful, but it should be interpreted together with whether rates and gas ratios remain steady throughout the run. The distinction between H₂ evolution and overall splitting also changes the outlook for scale-up. A high mass-normalized rate from a few milligrams of powder under a concentrated lamp does not specify production per illuminated area in an outdoor reactor. Sacrificial donors add material and separation costs. On the other hand, an overall-splitting material with lower initial rate may be more important scientifically if it suppresses photocorrosion and closes the oxidation balance. Reports should therefore provide both mass- and area-based outputs where possible, light input, gas verification and continuous operating duration. Those details let readers evaluate the chemistry without making a premature commercial claim [150,151,152,153,154].

Powder rates do not directly predict performance in a photoelectrode. A ZnIn₂S₄ slurry exposes both sides of a sheet to liquid, whereas immobilization creates a preferred contact with the substrate and electrolyte. Electrical connection may collect carriers before they react at the exposed surface, while a binder or protective layer can improve film integrity but limit proton or water access. Cocatalyst distributions optimized for suspended sheets may therefore need redesign for a supported device. Reports should give active mass and illuminated area, and state applied bias and collection method where relevant. Reactor geometry also affects the illuminated fraction of a slurry; a flow reactor adds residence time, pressure drop and product-removal effects. These variables can change the apparent rate even when catalyst loading is unchanged [155,156,157]. Related studies outside ZnIn₂S₄ have considered hydrogen evolution, water splitting and oxygen evolution across photocatalytic and electrocatalytic materials. The different reaction conditions and charge-balance requirements limit direct performance comparisons [158,159,160,161].

4.2. CO₂ Photoreduction

The conversion of CO₂ into useful chemicals and fuels is being investigated as part of broader efforts in carbon capture and utilization. Depending on the energy source and reaction system, this conversion can proceed through electrochemical, thermochemical, biological or light-driven routes, yielding products such as carbon monoxide, formate and hydrocarbons. Each route has distinct requirements for energy input, reaction conditions and product separation. Solar-driven conversion is of particular interest because it can couple light absorption with chemical transformation, although its practical performance remains dependent on efficient charge utilization and suitable reaction pathways. Within this area, semiconductor photocatalysts provide a means of using absorbed photons to drive CO₂ reduction

CO₂ photoreduction is difficult to assess because activation, proton delivery and product desorption are coupled to competing H₂ evolution. At low conversion, trace carbon contamination can distort apparent product selectivity, and different studies report products that require different numbers of electrons. ZnIn₂S₄ has been examined both as an electron-rich component and as a surface whose metal sites or vacancies may bind intermediates. Product identity, carbon source and the contribution of the support therefore need to be considered together.

CO₂ conversion is more demanding than proton reduction. The molecule is difficult to activate, several proton and electron pathways compete, and trace carbon contamination can produce misleading signals when conversion is low. ZnIn₂S₄ has been used both as an electron-rich component and as a surface whose metal sites or vacancies bind intermediates. The Zn-vacancy study based on sulfhydryl coordination emphasized local adsorption chemistry [93]. Nanosheets grown on amine-functionalized SiO₂ emphasized the supply of CO₂ near the sulfide [120]. These are different solutions to the same reaction problem. A reactive site has little value if CO₂ does not reach it, while stronger adsorption does not guarantee selective reduction. Other reports introduced a photothermal MoO_3-x@ZnIn₂S₄ composite [119], a 2D/2D sulfide and polymeric junction, or a hollow TAPB-COF@ZnIn₂S₄ structure [162]. In each case, the interface can change adsorption, water access, light penetration and temperature at the same time. The isolated components, a physical mixture and a dark experiment at the same temperature are therefore important controls [163,164,165,166,167,168].

In the study from Chen et al., hexagonal and cubic ZnIn₂S₄ nanosheets were successfully prepared via liquid ultrasonic exfoliation and lamellar hybrid-based synthetic strategies, respectively, as shown in Figure 4 [18]. The photocatalytic performance of these two ZnIn₂S₄ nanosheet phases was evaluated through CO₂ photoreduction reactions for methyl formate production, with methanol acting as both the carbon dioxide absorbent and reducing agent. Photocatalytic tests confirmed that hexagonal ZnIn₂S₄ possesses better catalytic activity than cubic ZnIn₂S₄. In addition, both nanosheet samples exhibit remarkably higher reaction reactivity than hydrothermally synthesized ZnIn₂S₄ microspheres. Density functional theory calculations were further implemented to explore the inherent electronic structural characteristics of hexagonal and cubic ZnIn₂S₄ phases [169,170,171,172,173].

A CO₂ product rate is most informative when reported with selectivity, absolute carbon conversion and competing H₂ production. Isotope-labelled CO₂ and carbon-containing blanks are particularly important at low conversion; for the Zn-vacancy system, these controls would help test whether the proposed local coordination change accounts for the observed product distribution [174,175,176,177].Product distributions may change during illumination as adsorption or the catalyst surface evolves. A shift from CO to CH₄ after the first hour, for example, may reflect reconstruction rather than a fixed selectivity of the as-prepared solid. When water or a donor is present, the source of carbon and hydrogen should be established. Reporting electron-equivalent product yields alongside H₂ would help compare studies in which products consume different numbers of electrons. For vacancy-containing catalysts, time-resolved products can also show whether the proposed adsorption site remains active or is modified during the reaction. These measurements are particularly relevant when a support changes local CO₂ supply or when surface composition changes under illumination [178,179,180,181,182].

The support studies illustrate distinct ways to alter the environment around ZnIn₂S₄. Amine-functionalized silica provides potential CO₂-binding groups while also controlling nanosheet arrangement. In a porous COF shell, adsorption capacity has to be considered together with permeability and product diffusion. Oxygen-deficient MoO_3-x introduces a temperature-dependent contribution to the core-shell system, so its thermal and electronic effects should be separated. Their relative roles can be compared only when feed, light flux and temperature are matched. A support-only control and a physical mixture can help determine whether the benefit comes from local CO₂ supply, an electronic junction or simply a change in dispersion [183,184,185,186,187]. Work on CO₂ conversion and capture spans photocatalytic reduction, photothermal conversion and adsorption. These approaches address different stages of carbon utilization, so adsorption capacity should not be treated as evidence of photoreduction activity.

4.3. Environmental Purification

The discharge of industrial chemicals, pharmaceuticals, pesticides and other persistent pollutants has increased the demand for water-treatment methods that can operate under diverse environmental conditions. Conventional approaches, including adsorption, biological treatment and membrane separation, are useful in many settings, but they may transfer contaminants to another phase or require additional treatment when pollutants are poorly biodegradable. Advanced oxidation processes offer an alternative by converting contaminants into less harmful products, although their energy and chemical requirements vary with the treatment system.

Water purification requires evidence that contaminant concentration and hazard have both decreased. ZnIn₂S₄ composites have been tested against dyes, antibiotics and inorganic species, but removal can mean adsorption, loss of the parent signal, transformation or mineralization. For dyes, sensitization and adsorption can change the spectrum without extensive chemical destruction. Antibiotic studies are more informative when calibrated chromatography follows the parent compound and major products, with total organic carbon or toxicity measurements used where available. The composition of the water also matters because pH, dissolved oxygen and background organic matter can change adsorption and reactive-species chemistry [188,189,190,191,192].

ZnIn₂S₄ composites have been tested against dyes, antibiotics and inorganic contaminants, but removal can mean several different things, shown in Figure 5. It may be dark adsorption, disappearance of the parent molecule, transformation into smaller products or actual mineralization. A fading dye is weak evidence of chemical destruction because sensitization and adsorption can change its spectrum. Antibiotic studies are more informative when they report calibrated concentration data, products and, where possible, toxicity. Dark adsorption and light-only controls should come before a kinetic interpretation. A 2D/0D ZnIn₂S₄/ZnO structure and a hierarchical ZnIn₂S₄/C₃N₄ composite were both tested with tetracycline [193,194]. Their similar endpoint does not mean that their interfaces work in the same way. pH, initial concentration, lamp spectrum, catalyst loading and dissolved oxygen should be aligned before the rates are compared. A pseudo-first-order constant measured over one concentration interval is not an intrinsic surface constant. If one composite adsorbs more tetracycline before illumination, its higher apparent removal does not isolate the role of charge separation [195,196,197,198,199,200,201].

La-doped ZnIn₂S₄ was examined for oxygen activation and antibiotic degradation in water [202]. The proposed change to oxygen chemistry is a useful alternative to a discussion centered entirely on hole oxidation. If superoxide or another oxygen-derived species contributes, oxygen availability in the reactor and scavenger controls become important. Scavengers themselves can change adsorption, pH and interfacial kinetics, so the assignment is stronger when spin-trapping or another independent method supports it. A ZnIn₂S₄-decorated MIL-88B(Fe) system combined Cr(VI) reduction with tetracycline removal [203]. During simultaneous treatment, electrons used for metal reduction and holes used for organic oxidation may complement one another, but the final chromium species and transformed organics still need to be quantified. The two targets have different analytical endpoints. A lower Cr(VI) concentration does not establish recovery of harmless Cr(III), and a declining antibiotic peak may be replaced by products with different toxicity. The iron-containing framework can also contribute adsorption or secondary oxidative chemistry. Solution composition therefore matters when the contribution of ZnIn₂S₄ is discussed. This paired application is interesting because the reductive and oxidative sides can be followed separately. Real wastewater presents a more difficult test. Chloride, bicarbonate and dissolved natural organic matter may compete for oxidants or occupy surface sites, and suspended matter can shorten the optical path. A catalyst that works in distilled water at a chosen pH may behave differently in buffered, saline or colored water. Stability testing should include Zn and In release as well as repeated pollutant removal. Mineralization, intermediate toxicity, separation of the spent solid and performance at realistic concentrations deserve more weight than the near-complete disappearance of one parent peak [204,205,206,207,208,209].

The analytical endpoint can change the apparent ranking of water-treatment materials. A sample collected after tetracycline conversion may show substantial parent-compound loss while retaining aromatic fragments with biological activity. Chromatography can identify major transformation products, whereas total organic carbon provides a separate measure of mineralization; neither follows from the disappearance of the parent peak. In the ZnIn₂S₄-decorated MIL-88B(Fe) system, Cr(VI) reduction and tetracycline removal occur together, but the final chromium species and organic products should both be quantified. The iron-containing framework may contribute through adsorption or secondary oxidative chemistry. Matrix composition, catalyst recovery and Zn/In release also affect whether a result obtained in clean water is relevant to wastewater treatment [210,211,212,213,214]. Studies on environmental treatment and resource recovery have used photocatalytic, adsorptive and hybrid approaches. Their reported targets and endpoints differ, and pollutant removal or capture alone does not establish mineralization or selective conversion. [215,216,217,218,219,220] Related studies have addressed aqueous contaminants through adsorption, biomass-derived materials, photocatalytic degradation, and hybrid treatment. The targets include arsenic, nutrients, dyes, antibiotics, and oily wastes, but the reported endpoints differ. These examples provide environmental-treatment context; removal or adsorption should not be equated with mineralization or selective conversion. [221,222,223,224,225]

4.4. H₂O₂ Production and Selective Organic Transformations

Solar-driven chemical conversion can be directed toward products that serve as fuels, oxidants or chemical feedstocks. H₂O₂ is of interest because it is used in water treatment, chemical synthesis and disinfection, while its conventional production and transport can involve energy-intensive processes and concentrated solutions. Photocatalytic routes may allow H₂O₂ to be generated on site from water and oxygen, although product accumulation depends on controlling both its formation and subsequent decomposition. Light-driven organic transformations address a related but distinct objective: converting readily available substrates into specific products under comparatively mild conditions. In these reactions, selectivity is as important as conversion, since competing pathways can consume the substrate or oxidize the desired product. ZnIn₂S₄ has been examined in both areas, with material modification used to adjust charge-carrier behavior and surface reactivity for the target reaction (Figure 6).

H₂O₂ production and selective organic conversion use ZnIn₂S₄ differently from pollutant degradation. Here photogenerated charges are directed toward retaining a chosen product, and reactions on both sides can have value. Peroxide synthesis requires oxygen activation while avoiding H₂O₂ decomposition. In paired organic conversion, the oxidation product is collected rather than treated as a sacrificial by-product. The two topics are grouped because both require product selectivity and a clearer account of the complementary half-reaction, as shown in Figure 5 [226,227,228].

H₂O₂ provides a different target for oxygen reduction. A two-electron route can form peroxide, but the newly formed H₂O₂ may also be consumed by photogenerated charges or by the catalyst surface. The measured accumulation is therefore the difference between formation and loss. ZnIn₂S₄ nanosheets with tunable dual vacancies were studied for sacrificial-agent-free H₂O₂ production [229]. This is distinct from systems in which an alcohol supplies the oxidation half-reaction. The oxygen source, solution conditions and peroxide stability during illumination all affect the result. Oxygen-free controls and a time-dependent concentration profile are needed before the dual-vacancy composition can be linked to net formation. A surface that binds oxygen more strongly may also decompose peroxide more quickly. The optimum defect population should therefore be understood as a balance between the two processes, not as proof that more vacancies are always better. Radical trapping can suggest an oxygen-reduction route, but it does not establish two-electron selectivity or close the peroxide mass balance. Isotope tests, electron balances and independent decomposition experiments provide stronger support. Analytical methods should also be checked for interference from sulfide, residual thiourea or a colored catalyst suspension [230,231,232,233,234,235,236].

The two benzyl alcohol studies use related chemistry but arrive at it through different material designs. In one, separately introduced cocatalytic components modify the two redox sides of ZnIn₂S₄ nanoflowers [237,238]. In the other, an ultrathin nonporous host is intended to use both electrons and holes within one architecture. A fair comparison requires benzaldehyde selectivity at the same substrate conversion, rather than relying on the hydrogen rate alone. It also helps to know whether dissolved oxygen was excluded, since oxygen can consume electrons and alter alcohol oxidation. Both preparations coupled H₂ evolution with benzaldehyde formation. The oxidation product is therefore part of the value of the reaction rather than a disposable donor product. Alcohol conversion, benzaldehyde selectivity and H₂ yield should be reported together, along with the fate of the hydrogen and oxygen atoms. Further oxidation to benzoic acid or coupling byproducts may occur, so a high benzaldehyde signal at low conversion is only one part of the assessment [239,240,241,242,243,244].

Wu et al. examined C–H activation and coupling over zinc-indium sulfides and proposed a concerted proton-coupled electron-transfer route from experiments and calculations [245]. The study is useful because it addresses the oxidative elementary step instead of assigning every change to generic hole consumption. Cocatalyst-modified monolayer ZnIn₂S₄ has also been used to steer amine transformations [246]. In such reactions, rapid charge extraction does not guarantee selectivity if an intermediate remains on the surface or reacts again. Substrate conversion, desired product and side products should therefore be followed under the stated atmosphere and illumination. Both products in paired redox chemistry need to be reported. H₂ generation with alcohol oxidation does not imply overall water splitting, but it can be a legitimate route to two useful products. Conversely, a selective organic product obtained with an externally supplied electron acceptor has an input that must be acknowledged. Benzyl alcohol is a simple model with benzaldehyde as a defined oxidation product, whereas an amine coupling reaction may involve two adsorbed reactants, sequential bond formation and several competing products. For the first case, the link between alcohol consumption, benzaldehyde and H₂ can be followed directly. For the second, the principal carbon- and nitrogen-containing products and their selectivity at a stated conversion are more informative than substrate disappearance alone. The atmosphere and the identity of the complementary electron acceptor also matter. These details show whether a modified ZnIn₂S₄ surface improves a particular bond-forming step or merely accelerates loss of the starting compound [247,248,249,250,251]. Related literature on peroxide formation and selective chemical transformations covers photocatalytic and other catalytic systems. The studies are cited as application context, while their reaction pathways and performance remain specific to each material and protocol [252,253,254]. Studies on other photocatalysts have examined H2O2 formation from water and air, including donor–acceptor frameworks and heterojunction systems. They offer a useful comparison for product accumulation and charge-transfer analysis, although the reaction conditions and product accounting remain material-specific.

4.5. Photocatalytic Ammonia Synthesis

Ammonia is produced mainly through the Haber–Bosch process, which requires elevated temperature and pressure and is closely tied to centralized production and hydrogen supply. Photocatalytic nitrogen reduction has been investigated as a possible route to ammonia formation under light irradiation and relatively mild conditions (Figure 7). In this process, the catalyst must facilitate N₂ adsorption and reduction while competing with proton reduction, which can lower the selectivity toward ammonia. Reported ammonia formation rates also require careful interpretation because trace nitrogen-containing impurities may contribute to the measured product.

Early ZnIn₂S₄-based studies mainly used semiconductor junctions to improve visible-light-driven nitrogen reduction. Zhang et al. combined one-dimensional In₂O₃ nanorods with two-dimensional ZnIn₂S₄ nanosheets [255]. The authors attributed the improved response to oxygen vacancies in In₂O₃ and the proposed step-scheme charge-transfer pathway. Under simulated sunlight, the optimized material containing 50 wt% In₂O₃ produced 18.1 ± 0.77 mg L^-1 of ammonia after 2 h. This value is reported as a solution concentration rather than a catalyst-mass-normalized rate, so it should not be compared directly with rates from other reactor configurations. [256,257,258]

A different approach was reported by She et al., who coupled flower-like, spiky ZnIn₂S₄ with molecular cobalt phthalocyanine [259]. The composite generated ammonia at 156.59 μmol g^-1 h^-1 under visible-light irradiation without a sacrificial hole scavenger. The authors associated the result with improved light harvesting and charge separation after immobilizing the molecular catalyst on the semiconductor. This design shifts part of the catalytic function from the ZnIn₂S₄ surface to a molecular component, although the reported activity remains specific to the study’s illumination and reactor conditions [260,261].

More recently, Cheng et al. anchored sulfurized MXene quantum dots on ZnIn₂S₄ through Ti–S bonds [262]. In aqueous suspension under a 300 W xenon lamp, the composite reached an ammonia-production rate of 360.5 μmol g^-1 h^-1 without an added cocatalyst or sacrificial reagent. The study included dark and Ar controls and used ¹⁵N₂ labeling to examine the nitrogen source. Spectroscopic measurements and calculations were used to relate the Ti–S-linked interface to carrier extraction and N₂ adsorption. The result illustrates how an interface may influence both charge transport and the surface steps involved in nitrogen reduction, rather than acting only as a charge-separation junction [263,264].

These reports show several ways in which ZnIn₂S₄ has been incorporated into ammonia-producing photocatalysts, from oxide and sulfide junctions to molecular and quantum-dot components. Their reported rates should be read alongside the nitrogen source, illumination, reactor geometry, sacrificial reagents and analytical protocol, as shown in Figure 5. This is particularly important because the products are often measured at low concentrations, while the activity units and control procedures differ among studies. Further work would benefit from consistent isotope-based source verification and reporting that allows rates to be compared on a catalyst-mass and photon-input basis. The broader materials literature also includes studies on catalyst preparation, energy storage, porous frameworks and electrochemical conversion. They are cited here as examples of adjacent materials research, rather than as direct evidence for ZnIn₂S₄ photocatalysis or sensing. Nitrogen conversion has also been studied through photocatalytic nitrogen fixation and electrochemical nitrate or nitrogen reduction. Because the nitrogen source and reaction route differ, their yields require separate interpretation.

4.6. Comparing Reaction Performance

Rates across these applications cannot be placed on a single scale. For hydrogen evolution, donor identity, catalyst loading, photon flux and normalization affect the reported rate; area-based output becomes important for supported systems. CO₂ studies require absolute product amounts, a carbon-source control and H₂ competition, while pollutant removal should be separated from dark adsorption and mineralization. H₂O₂ accumulation is a net value that includes decomposition. Selective organic reactions require substrate conversion and product selectivity, preferably alongside the paired redox product. Sensing performance is governed by calibration, detection limit, matrix effects and reproducibility rather than a catalytic rate. Meaningful comparisons therefore report reaction-specific conditions, photon input, product distribution and stability instead of combining unlike data in one ranking.

5. ZnIn₂S₄ in Photoelectric Sensing

Photoelectric sensing combines optical excitation with electrical readout and includes photoelectrochemical (PEC) sensors and solid-state photodetectors, as shown in Figure 8. PEC platforms have been used in bioanalysis and clinical diagnostics for detecting biomarkers, nucleic acids and pathogens, as well as in environmental and food monitoring of metal ions, pesticides, antibiotics and other contaminants. Photodetectors are used to measure light intensity or wavelength in applications such as imaging and optical communication. Unlike conventional electrochemical methods, PEC sensors use illumination as an independently controlled excitation source. Modulating the light can help distinguish the photocurrent from the dark background, and some device designs operate at low or zero bias. Their electrical output can also be read by compact circuitry, which may simplify integration into portable systems compared with colorimetric or fluorescence assays that require optical signal detection [265,266,267,268,269,270].

In a PEC assay, the photoactive layer generates charge carriers under illumination, while the analyte changes their transfer through molecular recognition, a surface reaction, interfacial blocking or a mediator. The photocurrent may therefore increase or decrease, depending on the assay design. These features do not guarantee high selectivity or sensitivity: performance also depends on receptor specificity, electrolyte composition, matrix effects, electrode architecture and calibration. For ZnIn₂S₄-based sensors, the relevant question is how its light absorption, carrier transport and interface with the recognition layer contribute to the measured signal under the stated assay conditions [271,272,273,274,275]. Outside PEC sensing, optical methods used in food and environmental analysis include SERS, fluorescence, colorimetric assays, hyperspectral measurements, and smartphone-assisted readout. Their targets range from pesticide and toxin residues to metal ions, pathogens, and freshness-related signals. These studies illustrate how recognition and signal generation can be coupled in different ways; their analytical figures should not be compared directly with those of a ZnIn₂S₄ PEC electrode. [276,277,278,279,280,281,282]

5.1. From Photocatalyst to Sensing Electrode

ZnIn₂S₄ can serve as the photoactive layer in a PEC sensor (the left of Figure 6). Illumination produces an anodic or cathodic current, and analyte binding, surface reaction, blocking or a mediator changes that signal. A larger background current may improve signal-to-noise, but does not by itself lower the detection limit. Calibration slope, baseline variation, selectivity and electrode stability remain relevant, and the direction of the response should be tied to the proposed recognition or reaction step [283,284,285,286].

An early ZnIn₂S₄/CdS junction was used for indirect dopamine detection [287]. The junction supplied the photoresponse, while dopamine altered the measured current. In another study, PEDOT was combined with flower-like ZnIn₂S₄ microspheres to construct a carcinoembryonic-antigen immunosensor and related photoanodic and photocathodic assays [288]. The selected PEDOT/ZnIn₂S₄ film produced a much larger anodic photocurrent than the comparison film. That improvement strengthens the transducer, but it does not replace antibody controls or tests for nonspecific adsorption [289,290,291,292,293,294,295,296].

ZnIn₂S₄/Ag₂CO₃ was developed as a PEC biosensing platform for neuron-specific enolase [297]. The report gave a range from 50 fg mL^-1 to 200 ng mL^-1 and a detection limit of 4.86 fg mL^-1 under its stated assay conditions. The proposed junction and associated Ag species increased the photocurrent compared with the individual components. Several effects may contribute to this result, including absorption, charge separation, electron mediation and immunorecognition. The proposed Z-scheme pathway cannot replace those separate controls. Calibration in a simple electrolyte should also be distinguished from validation in serum or another clinical matrix. A ZnIn₂S₄@ReS₂/AgInS₂ photoactive assembly was coupled to an aptamer for kanamycin detection [298]. A CdS-nanorod/ZnIn₂S₄-nanosheet 1D/2D junction was used in a tobramycin bioassay [299]. These cases illustrate how a heterointerface supplies charge collection while the recognition layer determines analyte specificity. The two functions should be tested independently. A larger photocurrent after junction formation says little about whether the aptamer or enzyme retains its function after immobilization. Selectivity experiments should include structurally related antibiotics and the real matrix in which an assay is intended to operate. A low instrumental detection limit in buffer is a starting point, not proof of field-ready antibiotic monitoring [300,301,302,303,304,305,306,307].

The two antibiotic cases should therefore be compared at the assay level. Their calibration limits include incubation, recognition chemistry and signal amplification, rather than the ZnIn₂S₄ interface alone. Time to result, selectivity against related antibiotics and recovery from the intended matrix would be more useful for comparison than the photocurrent alone. Inorganic-ion sensing follows a different route. ZnIn₂S₄ microspheres grown on WO₃ nanosheets on FTO were used for PEC Cu²⁺ detection [308]. At 0 V versus SCE in pH 7 phosphate buffer containing 0.100 M L-ascorbic acid, Cu²⁺ lowered the anodic photocurrent through formation of Cu_xS at the surface. The reported logarithmic range was 5 nM to 100 μM, with a detection limit of 1.2 nM and recovery tests in river water. The ascorbic-acid medium is part of the assay and should be stated with the analytical figures. Competing ions, pH and complexation may alter the surface reaction, while fouling or sulfide corrosion must be excluded before the signal is assigned to Cu²⁺ [309,310,311,312,313].

Another self-powered design paired a ZnIn₂S₄@CdS photoanode with a Bi₂WO₆ photocathode for BACE1 detection [299]. An enzyme-cleavable probe enabled zero-bias operation, with a reported range of 1 fg mL^-1 to 500 pg mL^-1 and a detection limit of 0.30 fg mL^-1. The numerical limit belongs to the complete two-electrode architecture, including peptide cleavage and probe deposition, rather than to ZnIn₂S₄ alone. Serum spike recoveries were reported, but reproducibility across independently prepared devices and performance in heterogeneous clinical samples remain important tests [314,315,316,317,318].

Because Cu²⁺ detection relies on surface conversion to Cu_xS, regeneration and recovery of the initial current determine whether the electrode is reusable or better treated as disposable. Differences in microsphere coverage, WO₃ thickness or electrical contact with FTO may widen the spread in blank currents across electrodes. For river-water measurements, filtration, dilution and added ascorbic acid should be included in the assay protocol because each can change the concentration or speciation presented to the electrode. These details are part of analytical performance, not only experimental procedure [319,320,321,322].

Photoelectrochemical and electrochemical assays have also been developed for pesticides, antibiotics, toxins, and metal ions in food or water. Several use aptamers, immunorecognition, quantum-dot labels, microfluidic formats, or self-powered heterojunctions. Their performance depends on the complete assay and sample treatment, not solely on the photoactive material, so these reports are cited here as broader sensor-design context. [323,324,325,326,327,328]

5.2. Photodetectors and the Limits of Cross-Device Comparison

A photodetector responds to light itself and should not be compared directly with an analyte sensor (the right of Figure 6). Its relevant metrics include responsivity, detectivity, dark current, response time, wavelength and applied bias. Zhen et al. studied rhombohedral ZnIn₂S₄ nanosheets in a gated phototransistor and reported a dark current of 7 pA at 5 V, a specific detectivity of 1.8 × 10¹⁴ Jones, and sub-millisecond rise and decay times. These values describe a solid-state device, not a molecular detection limit in solution. They also depend on device area, noise estimate, wavelength, optical power and gate setting. ZnIn₂S₄ can generate a light-to-current response in both PEC electrodes and phototransistors, but the transport medium, junction, noise and meaning of the output are different. The two device classes can be discussed in one review only when their metrics remain separate [329,330,331,332,333].

For biomarker assays, the calibration range should be considered in relation to concentrations expected in the intended sample, not only the lowest detectable level. Antibody or aptamer immobilization can itself alter interfacial resistance, and later blocking or label-deposition steps may change the current before analyte binding is measured. In the self-powered BACE1 design, the enzyme-cleavable probe adds another step between target recognition and photocurrent. Measuring each layer or reaction stage separately helps locate the origin of the signal. A zero-bias response is useful, but does not remove the need to report illumination, incubation time and performance in a protein-containing matrix [334,335,336,337].

5.3. Analytical Validation and Durability

A PEC comparison should include the bare substrate, ZnIn₂S₄ layer and completed junction or receptor surface under the same illumination and potential. Reports should define the signal, baseline drift, regeneration protocol and detection-limit calculation. Measurement time is also relevant when comparing assays; a very low limit after lengthy incubation may not be practical for rapid screening. Signal-on and signal-off formats should be evaluated with appropriate blank and recognition controls [338,339,340,341].

Durability depends on the intended use. Repeated illumination may oxidize ZnIn₂S₄ or loosen a film, while storage can affect a bioreceptor. Useful measures include the number of full assay cycles, regeneration conditions, drift in blank current and calibration slope, and post-assay composition. Dissolved Zn and In measurements can complement microscopy or XPS after use. Storage experiments should state temperature and atmosphere. For disposable sensors, fabrication yield and batch-to-batch variation may matter more than regeneration [342,343,344].

Validation in a real sample requires more than spike recovery in a simple matrix. Serum proteins may foul the electrode, and natural-water ions or organic matter can compete for surface sites. Recovery tests should state sample preparation; agreement with an independent method on native samples provides stronger evidence. When clinical markers are discussed, their concentrations in the intended setting should be compared with the sensor’s validated working range. For multilayer assays, omission controls can identify whether recognition, label deposition or interfacial blocking produces the signal. Enzyme-based assays should also check whether illumination or applied bias affects receptor activity [345,346,347,348,349]. Optical and electrochemical sensing studies in other material systems illustrate the range of readout schemes used for chemical and biological targets. These reports provide general analytical context; their detection limits are not directly comparable with those of ZnIn₂S₄ devices [350,351,352,353]. Analytical studies further show the importance of matrix handling, sample pretreatment, and the choice of reference method. This is relevant when comparing sensor limits reported in buffer with measurements in food or environmental samples. The added examples include peroxide assays, microfluidic analyses, biosensor reviews, and combined detection–degradation platforms [354,355,356].

6. Conclusions and Outlook

ZnIn₂S₄ has been studied for donor-assisted hydrogen evolution, overall water splitting, CO₂ conversion, pollutant treatment, selective synthesis and photoelectric devices. This range reflects the flexibility of its structure and surface chemistry, but also complicates comparison: synthesis can alter several material features at once, and reaction tests use different light sources, donors and normalization methods. The central question is which step changed, under what conditions, and whether the difference persists during operation.

Future studies should establish phase and defect identity in the material actually tested and follow relevant surface changes during operation. Controlled defect healing or creation, paired with product analysis, can strengthen causal assignments. For heterojunctions, measurements that distinguish competing charge-transfer pathways are more informative than a band diagram alone. Application-specific reporting remains necessary. Overall water splitting requires a verified gas balance and evidence against sulfide oxidation; CO₂ reduction needs carbon-source controls and product balance; environmental and H₂O₂ studies should distinguish transformation from mineralization and peroxide formation from loss. In sensing, analytical validation must remain separate from photocatalytic activity, and photodetector metrics should not be compared with molecular detection limits.

The next step is to connect material-level measurements with repeatable reaction and device performance. For supported ZnIn₂S₄, film preparation, operating conditions and stability should accompany the initial photocurrent or catalytic rate. Powder activity with a sacrificial donor may not translate directly to an electrode used in a different electrolyte, while a sensing response depends on the receptor and sample-preparation steps as well as the semiconductor. Reporting these links will make the literature easier to compare without treating a high rate or low detection limit as a complete measure of performance.