用于光電催化水分解的氧化亞銅基光電陰極研究進展

摘要：隨著不可再生資源的消耗及環(huán)境污染日趨嚴重，開發(fā)環(huán)境友好、可再生的新能源受到廣泛關(guān)注。氫可通過燃料電池進行發(fā)電，被認為是理想的潔凈能源載體。耦合可再生能源，如光能、風能、海洋能等，進行光電水分解制氫是有效途徑之一。氧化亞銅（Cu2O）具有合適的能帶結(jié)構(gòu)、制備簡單、資源豐富，成為光電陰極半導體的研究熱點。然而，Cu2O光電陰極面臨光生電荷復(fù)合較快、光腐蝕嚴重等挑戰(zhàn)，導致其光電效率低、穩(wěn)定性差。本綜述首先簡要介紹光電水分解制氫原理以及Cu2O的能帶結(jié)構(gòu)適配性，其次總結(jié)氧化亞銅的制備方法；重點概述提高Cu2O光電效率和穩(wěn)定性的策略，包括形成氧化亞銅-n型半導體p-n結(jié)、添加助催化劑、引入空穴傳輸層等；結(jié)合近年來表征技術(shù)的發(fā)展，介紹先進的光電陰極表征手段；最后，對光電陰極未來的研究方向進行展望。

關(guān)鍵詞：光電催化；氧化亞銅；析氫反應(yīng)；光陰極；水分解

中圖分類號：O646

Advances in Cu2O-based Photocathodes for Photoelectrochemical Water Splitting

Abstract： Owing to the growing consumption of non-renewableresources and increased environmental pollution， significant attentionhas been directed toward developing renewable and environmentallyfriendly energy sources. Hydrogen has emerged as a clean energycarrier and is considered an ideal chemical for power generation via fuelcells. Using renewable energy to power hydrogen production is anattractive prospect， and hydrogen production throughphotoelectrochemical water splitting is considered a promising area ofinterest; consequently， significant research is being conducted onrationally designed photoelectrodes. Generally， a photocathode forhydrogen evolution must have a conduction band that is more negativethan the reduction potential of hydrogen. Numerous photocathode materials have been developed based on this premise;these include p-Si， InP， and GaN. Compared with other photocathode materials， Cu-based compounds are advantageousowing to their low preparation costs and diverse chemical states. For example， Cu2O is a non-toxic p-type metal oxidesemiconductor material with an appropriate band structure for water splitting and a direct band gap of 1.9–2.2 eV.Furthermore， the production of Cu2O is facile， and the required materials are abundant; thus， it has attracted significantinterest as a material for photocathodes. However， Cu2O suffers from rapid recombination of photogenerated carriers andsevere photo-corrosion， leading to unsatisfactory efficiency and poor stability. Intrinsically， the poor photo-stability of Cu2Ocan be attributed to the location of the redox potential of Cu2O within its bandgap， owing to which photoelectrons tend topreferentially reduce Cu2O to Cu rather than reduce water to reduction. Therefore， Cu2O itself is not an ideal hydrogenevolution catalyst. Thus， co-catalysts are necessary to improve its hydrogen evolution activity and photostability. In additionto co-catalysts， combining Cu2O with tailored n-type semiconductors to generate built-in electric fields of p–n junctions hasattracted extensive attention owing to its ability of increasing the separation of photogenerated carriers. Similarly， applyinga hole transfer layer on the substrate can promote photocarrier separation. Furthermore， considering that water isindispensable for Cu2O reduction， one effective approach to improve the stability of Cu2O is the addition of aprotective/passivation layer to isolate Cu2O from water in aqueous electrolytes. In this review， we provide a brief overviewof the mechanism of photoelectrochemical water splitting and the band structure of Cu2O; preparation methods of Cu2Ophotocathodes; strategies to improve the efficiency and stability of Cu2O photocathodes， including the construction of p–njunctions， integration with co-catalysts， and modifications using hole transport layers; advanced photoelectrochemicalcharacterization techniques; and a discussion regarding the direction of future photocathode research.

Key Words： Photoelectrocatalysis; Cuprous oxide; Hydrogen evolution reaction; Photocathode; Water splitting

1 Introduction

Presently， with the rapid development of industrialization， thedemand of energy is more and more increasing. Worse of all， theenergy used now mainly comes from coal， oil and othertraditional non-renewable fossil fuels， and the low utilizationefficiency also leads to environmental pollution and energyshortage. In order to solve these problems， exploring anddeveloping new clean and renewable energy has become animportant task. Hydrogen as a clean energy carrier， is believedto be an ideal chemical to generate power via fuel cells. Toproduce hydrogen from water splitting， integrated withrenewable energies from solar， wind， ocean and so on， is veryattractive. Especially， considering the considerable solar energyon earth 1， it is a very attractive to produce hydrogen from watersplitting via a photoelectrochemical process.

Since the discovery of photo-induced water splitting on theTiO2 electrode in 1972 by Fujishima and Honda 2， photo-drivenwater splitting has become a research hotspot around the world.After decades of research， three configurations for solar-drivenwater splitting have been proposed， i.e. photo-catalysis （PC），photoelectrocatalysis （PEC） and photovoltaic coupled withelectrolysis （PV-E） 3， as shown in Fig. 1. They all have their ownadvantages and disadvantages. Most of the PC systems usesemiconductor powder as photocatalysts， which means theoxidation and reduction of water proceed at the same catalystparticles. The advantage of such a configuration includes asimple reaction system， convenient preparation ofphotocatalysts， and low cost. However， the shortage is that theefficiency is severely limited by the fast recombination ofphotogenerated electrons and holes in the transmission process，which requires highly efficient semiconducting materials. Forexample， a shorter charge transfer path is necessary to promotethe diffusion of photogenerated carriers to the interface and toparticipate in water splitting 4. PEC and PV-E systems divideoverall water splitting reaction into cathodic reaction and anodicreaction by applying an external potential bias. Although thesystems become more complex and the cost is relatively higher，the external electric field makes the carrier separating moreeffectively. PV-E systems have been demonstrated the highestefficiency of solar to hydrogen production in a broad sense， butthe cost of hydrogen production by this configuration is ratherhigh， so is not believed as the most suitable route for solarhydrogen production 5. According to the economic evaluation，the cost of hydrogen prepared by PEC is between $1.60 and$10.40 kg?1， which is quite competitive with traditional fossilfuels 6， which is very promising.

PEC cells are composed of photocathode， photoanode andelectrolyte. At present， the design and improvement of PEC cellsmainly start from cathodes and anodes. Semiconductors can bedivided into n-type and p-type according to the types of carriers（or crystal defects）. The n-type semiconductors with electrons asmajor carriers are usually adopted as photoanodes， and p-typesemiconductors with holes as major carriers are mainly used forphotocathodes. Common photoanode materials include binary orternary oxides composed of Ti-based， W-based， Fe-based andother metals （such as TiO2， WO3， Fe2O3， ZnFe2O4， etc.）. Thecommon photocathode materials include Cu-based metal oxides，p-type silicon， III-V main group compounds （such as GaN， InP），etc. This review focuses on discussing the photoelectrochemicaldecomposition of water to produce hydrogen， so only theresearch progress on photocathodes is included. Thesesemiconductors have many exciting advantages， such as narrowbandgaps， fast photocharge generation/separation efficiencies，etc.， as shown in Fig. 2.

However， there are still significant barriers for practicalapplication. One of the barriers is the cost of production， giventhe expensive materials and complex manufacturing processesinvolved. Furthermore， there are also concerns about the safetyand toxicity of certain semiconductor materials such as galliumarsenide （GaAs）， which can pose health risks to workersinvolved in their production. Thus， Cu2O stands out fromnumerous photo electrode materials， which is cheap， friendly tonature， and has suitable band edges compared to water redoxlevels. It could absorb wide range solar radiation and itstheoretical photocurrent could reach ?14.7 mA·cm?2. However，it is not easy to apply it with some limitations， and what needs tobe addressed the most is its extreme photo corrosion. Given thatRyu et al. 8 has provided a comprehensive review on the researchstatus of Cu2O photocathode prior to 2019， this review will focuson recent progress on the improvement strategies and advancedcharacterization techniques of Cu2O.

2 Principle of photoelectrochemical watersplitting and characteristics of Cu2O

2.1 Principle of photoelectrochemical water splitting

It is well known that once a semiconductor is irradiated withlight of certain energy （hν）， electrons at valence band （VB） willbe excited and transfer to conduction band （CB）， leaving a holeat VB. The photogenerated electron and hole will be separatedand diffuse to the surface of semiconductors， driven by electricpotential （built-in electric field or applied bias）， to participate inreduction and oxidation reactions， respectively. The specificreaction equation is displayed as follows：

Cathode： 4H+ + 4e? → 2H2

Anode： 2H2O + 4h+ → O2 + 4H+

Compared with PC， the PEC system divides the reaction sitesinto photocathode and photocathode to separate the redox sitesand the oxidation sites in space， which could avoid the reduction（or oxidation） of products and increase productivity. Moreover，a bias voltage can be easily applied to the system to assist inseparating of electrons and holes. Especially， some photosemiconductorsthat could not decompose water inthermodynamics， can also photoelectrochemically split water byapplying a bias voltage.

To sum up， the PEC system can be seen as the coupling ofphotocatalysis system and electrocatalysis system. It combinesthe advantages of both configurations， and overcomes someshortages of photocatalysis system in some aspects （such asbroadening the applied range of various photocatalyst materials，reducing the contact of oxidation and reduction products toimprove the yields）， and also takes full advantages of solarradiation as a new energy.

2.2 Energy band structure of Cu2O

Among a wide range of semiconductor materials， copperbasedcompounds have become one of the most competitivephotoelectrochemical electrode materials due to their abundancein reserves， simple preparation and environmentally friendlycharacteristics. Among them， Cu2O has attracted the mostattention. First of all， the band gap of Cu2O and its conductionband position play an important role. The width of the band gapdetermines the utilization efficiency of lights， and the conductionband position determines whether the semiconductor is able tosplit water to hydrogen. The band gap of Cu2O is 2.2 eV， whichcan absorb light in the wavelength range from 300 to 620 nm，accounting for nearly 50% of the visible sunlight 9. In addition，the maximum photocurrent density of Cu2O can reach ?14.7mA·cm?2 in theory， according to Formula 1 10， and the maximumconversion efficiency of solar energy to hydrogen can reach 18%in theory， according to Formula 2 and 3 10. These merits make ita very excellent candidate as the photocathode material.

where RH2 is H2 production per unit time （mmol·s?1）; ΔGr isstandard Gibbs free energy for water formation （237 kJ·mol?1）;I0"is the power density of incident light （100 mW·cm?2）; S isilluminated area.

where e0 is elementary charge （1.6 × 10?19 C）; λ（Eg） is band-edgeabsorption wavelength of semiconductor （nm）; IPCE or EQE isexternal quantum efficiency; Nλ is number of incidentwavelength λ monochromatic photons; jph is photo currentdensity （mA·cm?2）; I（λ） is the power density of wavelength λmonochromatic light.

Although Cu2O has a broad application prospect in the PECsystem， some disadvantages limit its application. Among thesedefects， the most serious and concerned is thephotoelectrochemical stability of Cu2O in aqueous solution， for the redox potential of Cu2O （+ 0.47/+ 0.6 V vs. normal hydrogenelectrode， abbreviated as NHE） is located within its band gap， asshown in Fig. 3 11. When electrons and holes are generated andaccumulate in the bulk of Cu2O， Cu2O would be photo-corrodedextremely easily 12. In addition， one of the main factors affectingthe PEC performance is the width of the space charge layer （WSC，related with the diffusion length of carriers）. In order to improvelight absorption， the thickness of Cu2O should be controlled atleast 1 μm. However， limited by the narrow width of the spacecharge layer， the number of carriers would decrease dramaticallywhen the thickness is thicker than 200 nm. So， it is not easy toreach the maximum photocurrent density for Cu2O. For a dopingsemiconductor， the thickness of the space charge layer could beobtained according to Formula 4 13.

where ε is dielectric constant; ε0 is vacuum dielectric constant;Nd is doping concentration; E is the applied potential， Efb is theflat band potential.

3 Preparation techniques and synthesis"routes of Cu2O

3.1 Electrodeposition

Electrodeposition （ED） has become one of the mostcommonly used methods for preparing Cu2O， considering theadvantages in economics and efficiency. Electrodeposition couldbe proceeded on a large and various conductive substrate undermild conditions. Most importantly， Cu2O with differentmorphology and properties can be obtained by simply changingthe deposition conditions （deposition potential or current，composition of electrolyte， pH， temperature， etc.）， which isconvenient to investigate and improve the performance of Cu2O.

Cu2O can be prepared by both cathodic reduction and anodicoxidation electrodeposition process. The cathodic reduction ismainly achieved by reducing the Cu2+ to Cu+ in the electrolyte，and then depositing it on the cathode under the action of electricfield and OH?. The cathodic electrodeposition of Cu2O can betraced back to 1983. Tench and Warren 14 summarized theprevious experience of forming metal oxide and hydroxide filmson metal electrodes by electrochemical means， and developedthe technology of electrodepositing copper oxide and hydroxidefilms from aqueous solution. It is believed that the reduction anddeposition of Cu2+ into Cu2O can be divided into two steps. First，Cu2+ gets electrons to be reduced to Cu+ and reacted with OH?in aqueous solution to form CuOH， which then lost watermolecule to form Cu2O. The specific steps are as follows：

Cu2+" e- + OH-→ CuOH

2CuOH → Cu2O + H2O

Anodic oxidation is mainly through applying positivepotentials on copper electrode to oxidize metal copper to Cu2+，which then combines with OH? to form Cu（OH）2 and depositedon electrode 15. Finally， Cu2O is obtained after high-temperatureannealing.

Synthesis of Cu2O via electrodeposition is popular andextensively adopted， not only because of its convenience andlow cost， but also because it can gain Cu2O with diversemorphologies and orientations by just simply changing pH andtemperature 16， which will be discussed in detail later.

3.2 Thermal oxidation

As a traditional preparation process， thermal oxidation iswidely used， and researchers have prepared countless materialsby this process. Aveline and Bonilla 17 prepared Cu2O by thermaloxidation and improved the optical characteristics throughchanging the stoichiometries. It is known that one-valent copperis more stable at high temperatures. In their experiments， eventhe oxidation time was as short as 3 min， the copper sheet wascompletely oxidized into Cu2O as well. In addition， changing theoxidation time or annealing time could gain Cu2O with differentoptical performance. In a word， if one wants to prepare Cu2O bythermal oxidation， high temperature is necessary.

3.3 Sputtering

As a kind of physical vapor deposition （PVD） film preparationtechnology， sputtering methods include direct current （DC）sputtering， alternating current （AC） sputtering， reactivesputtering and magnetron sputtering. These methods arecommonly used for film preparation， and are also suitable forpreparation of Cu2O films. During the process of forming films，the sputtering voltage plays an important role， and it affectsresistivity of film directly 18.

However， there are only a few research on Cu2O via thismethod and the results are not satisfactory， but this cannotdampen the researchers’ morale. As shown in Fig. 4， Qin et al. 19 applied the magnetron sputtering method to prepare a Cu2Ophotocathode. By controlling the O2 partial pressure andsubstrate temperature， a high-quality Cu2O photocathode wasobtained， which was better performed than a conventional EDCu2O（electrodeposition）. With precisely controlling thepreparation parameters， the defects in the Cu2O film could bediminished， and the carrier lifetime， mobility and electric fieldintensity could by thereby improved.

3.4 Chemical vapor deposition

Chemical vapor deposition （CVD） is a method to produce thinfilms by heating gaseous precursor compounds and allowingthem react chemically on some substrates. One major drawbackis， only a few special copper compounds （such as copperacetylacetonate 20 and copper hexafluoro acetylacetonate 21） areeasily vaporized， and could be precursors for the CVD process.Furthermore， choosing MgO as a substrate is beneficial for thegrowth of cubic Cu2O， which is with good PEC performance.Besides MgO， to create a p-n junction， an n-type material suchas TiO2 22 could also be used as an extra layer. What’s more，during Cu2O deposition， the properties of Cu2O thin films （suchas preferred orientation and carriers mobility） are resulted by thehumidity of the carrier gas on the texture 23. As its uncomplicatedand comprehensible deposition process and mechanism， CVD ispreferred by many researchers. Thus， CVD is a potential methodto prepare Cu2O photocathodes.。

3.5 Chemical bath deposition

Chemical bath deposition has attracted extensive attentions，due to its numerous advantages， such as low cost， mild reactionconditions and facile preparation of electrodes with a large area.Similar to those described in the ED process， Cu2O could begained by whether Cu0 oxidation 24 or Cu2+ reduction 25. Sincechemical reaction is involved， there are many parameters， suchas surfactants， precursor concentrations and impregnationduration， etc.， could influence the final products and thus thePEC performance. By changing surfactants， it could formdifferent nanostructures. As shown in Fig. 5， ethylenediamineand ethanolamine are in favor to forming nanorods 26， whilen-butylamine is conducive to the formation of hollownanospheres 27. In addition， the concentration of solution and theimmersion time also determine the nanostructure 28.

Besides above methods， there are some other means notmentioned here， such as hydrothermal methods and spraypyrolysis， could also fabricate Cu2O with high quality. For all themethods to fabricate Cu2O photocathodes， each method has itsadvantages and shortage. When choosing one method， it isnecessary to consider both from performance and also the costs.

4 Strategies to improve efficiency andstability of Cu2O photocathodes

In this section， we focus on giving a review on strategies toimprove efficiency and stability of Cu2O photocathodes，including bandgap engineering to enhance light adsorption ofCu2O， surface/substrate modification to improve charge transferability， introduction of protecting layers to increase stability andso on. It should be noted that some of strategies to improve theefficiency also enhance the stability by reducing thephotocharges participating in photocorrosion of Cu2O. Ingeneral， the first four strategies below mainly focus onimproving efficiency， and the last one mainly focuses onenhancing stability.。

4.1 Bandgap engineering

Although the bandgap of Cu2O （2.2 eV） is narrower thanmany other semiconductors， but there is still nearly 50% of thesolar energy not utilized even by assuming the maximumadsorption. So， reducing the bandgap of Cu2O is necessary toenhance the PEC performance and it has been proved that thebandgap could be regulated by strategies， such as annealingprocess 32， doping heteroatoms 33， applying strains 34 and so on.

For the heteroatom doping strategy， the size of doping ionaffects the bandgap of Cu2O directly. Nolan and Elliott 35 havestudied the effects of different doping ions and analyzed thepossible reasons. They concluded that when the size of dopant issmaller than Cu+， the bandgap is narrower and vice versa.Besides bandgap， the doping ion also influences the transparencyof the Cu2O film， which is the determinant of light absorption.Tseng et al. 36 doped Ag in Cu2O thin film， and studied the effectsof Ag contents and deposition temperatures on the property ofCu2O thin film. With increasing the Ag contents （which wascontrolled by changing the current power applied to Ag targets），the transparency decreased but also the resistivity， which isprobably due to the formation of Ag phase. So， it is necessary todetermine an appropriate doping concentration.

By changing the annealing temperature of Cu2O， Wang etal. 32 found that the defect band tail of Cu2O would be removedwith temperature increasing， which caused the widening of Cu2Obandgap.

Visibile et al. 34 made use of density functional theory （DFT）modeling to explore the influence of compressive and tensilestrains on the bandgap. Their results indicated that the bandgapof Cu2O would be narrowed under tensile strain and be widenedto a maximum under moderate compressive strain and decreaseagain under extreme compressive strain.

From the above experimental and theoretical investigations， itis clear that several strategies could be effective to engineer thebandgap of Cu2O， and thereby the light adsorption capability.

4.2 Surface modification

Surface modification is often adopted to provide catalytic sitesfor reactions， or to improve the separation of photogeneratedcharges by introducing a surface layer for selectively transferringwhether electrons or holes， or forming a p-n junction with abuild-in electronic field， and so on. Most of all， through surfacemodification， the photo-corrosion of Cu2O could be improvedgreatly， due to removing off photogenerated charges from Cu2Ofastly.

Noble metals， such as Pt and Pd， with excellent hydrogenevolution activity， are undoubtedly ideal co-catalysts. Chen etal. 37 deposited Cu2O on an Au substrate and introduced Pt as aco-catalyst on the Cu2O surface. Pt could accept photoelectronsfrom Cu2O and allow HER proceeding efficiently; meanwhilethe Au substrate could enhance the absorption of light. As aresult， the recombination of photogenerated carriers is alleviatedand a photocurrent density as high as 3.55 mA·cm?2 is obtained，which is 4.63 times higher than pure Cu2O.

Apart from noble metals as co-catalysts， some non-noblemetal compounds， such as transition metal dichalcogenides（TMDs）， could also be co-catalysts for HER. Li et al. 38 tookWS2 as both catalyst for HER and electronic transport layer（ETL）， in which only electrons could be transferred but no holes，and the photocurrent density of the Au/Cu2O/TiO2/WS2 cathodecould reach 10 mA·cm?2 at ?0.33 V （vs. RHE） in 0.1 mol·L?1PBS buffer. It was also discussed that the high photocurrentdensity partly gives credit to the p-n junction between Cu2O andTiO2， which facilitate the electron transfer in-betweens.

Different from electron transport layer， the role of p-n junctionis to promote the electron transfer through forming a built-inelectric field to drive the separation of electrons-hole pairs. Topromote charge separation， a staggered type II band offset isexpected， as shown in Fig. 6. TiO2 （bandgap = 3.2 eV） iscommonly used to form a p-n type II junction with Cu2O， and aPt/TiO2/ZnO：Al/Cu2O/Au photocathode could reach a high HERphotocurrent of 7.6 mA·cm?2 11. Besides TiO2， some TMDs arealso potential materials. Shinde et al. 39 spin-coated a MoS2 layeron the Cu2O， which enhanced the separation of electron-holepairs and then improved the photocurrent density to 6.5mA·cm?2.

In a word， surface modification is the most commonly usedmethod for catalyst modification. Loading cocatalysts orintegrating ETL not only improves the efficiency ofphotogenerated carriers’ separation， but also reduces thecorrosion of electrons and holes on Cu2O. With extensive research， more and more noble-metals free catalysts have beenused for PEC water splitting.

4.3 Substrate modification

As a substrate of photocathodes， it is highly desired that it is able to accept photogenerated holes from semiconductors andtransfer holes fast， to avoid the accumulation of holes insemiconductors. As mentioned above， although Au substrate hasa matching work function with Cu2O valence band energy levelto form an ideal Ohmic junction， Au is not a hole-selectivecontact and leading to recombination of electrons and holes.Therefore， it is necessary to explore new materials as the backcontact. Würfel et al. 40 put forward that the quasi- Fermienergies drive charge carriers’ transfer， which provide a guidanceto choose the right substrate materials. Contrary to the electrontransport layer on surface of Cu2O， the substrate should be alayer with high mobility of holes， the so-called hole transferlayer.

Yang et al. 41 evaluated the effects of different substrates onthe performance of Cu2O electrodes. With a work function of4.84 eV， Cu2O cannot form Ohmic junction with manyconductive substrates. Combining with the work function andvalence band potential， FeOOH 42， NiO 43 and CuSCN 44，45 havebeen used as hole transport layers for Cu2O. Among thesematerials， CuSCN/Cu2O cathode delivered a higherphotocurrent density and better stability， due to a matching bandstructure of CuSCN valence band and Cu2O， as shown in Fig. 7.In this structure， the holes could be transported smoothly throughband-tail states while electrons were injected into Cu2O easilybecause of the large conduction band offset.

4.4 Morphology/structure engineering

All the strategies mentioned above aim at improving theperformance of Cu2O by introducing other elements orcompounds. Surface morphology and structure of Cu2O are alsocrucial for the photoelectrochemical performance， for differentexposing facets are with different atom arrays and electron distribution.

By modifying preparation conditions， it could easily gainCu2O with different morphologies， such as regular polyhedralstructures （cubes， cuboctahedra， octahedra）， nanowires， hollowstructures and so on. To synthesis Cu2O with regular polyhedralstructures， it is necessary to add surfactants. Murphy et al. 46fabricated Cu2O nanocubes from a solution containing CuSO4，sodium ascorbate， NaOH and cetyltrimethylammonium bromide（CTAB） at 55 °C for 15 min. Even if the same surfactant is used，the change of solution composition， temperature and otherconditions could also affect the morphology of Cu2O. Cao etal. 47 synthesized Cu2O nanotubes and nanorods， with CTAB asthe structure-directing surfactant， and Cu（OH）42? as an inorganicprecursor. Furthermore， with the help of poly（vinyl pyrrolidone）（PVP） 48， sodium dodecyl sulfate （SDS） 49 and other surfactants，Cu2O with different morphologies could be prepared.

Nanostructured photoelectrodes， especially nanowire arrays，are considered effective to solve the contradiction betweenoptical absorption and carrier diffusion 29. Cu2O NW could beprepared by chemical bath deposition 50 or electrochemicalanodization （showed in Fig. 8） 51 on the Cu substrates. Luo etal. 29 sputtered a Cu film on FTO glass and then Cu waselectrochemically anodized to Cu2O NW. Benefiting from therigid FTO glass substrate， Cu2O cracks are avoided， so the photocorrosionis reduced， leading to improved stability.

Consulted from strategies of photocatalysis， core-shell andyolk-shell nanostructures could ensure long-term stability andthe shell could provide active sites and then improve theperformance. Lai et al. 52 designed Au@Cu2O core-shell andAu@Cu2Se yolk-shell nanocrystals， which was then drop-coatedon the FTO glass. Due to the protection of Cu2O and Cu2Seshells， Au core/yolk avoided the aggregation or detachment. Aucould promote charge separation and provide catalytic sites forPEC reaction. The fast transfer of the photogenerated chargesaway from Cu2O protect it to be reduced， as a result， thephotocurrent density exceeded 40 μA?cm?2， which is higher thaneach separate component and decayed only 6%–8% upon 11 hof continuous irradiation.

4.5 Protecting layer

To improve the stability of Cu2O， one of the most direct and effective way is to introducing a protecting layer. The protectinglayer is usually formed by chemical or physical approaches toslow down and passivate the corrosion reaction by separatingCu2O with H2O （resource of protons） existing in aqueouselectrolytes.

The pioneer research could be traced back to 1978， Pande etal. 53 utilized selective-anodization techniques to passivate GaAswith ammonium borate. With the formation of a thick anodicoxidation layer， the leakage current was reduced by a factor of105–106. However， the overall photocurrent is simultaneouslydecreased significantly due to the light-shielding of theoverlayer. Thus， the thickness of protecting layers should becontrolled at a nanometer level. As a nanofabrication technology，atomic layer deposition （ALD） is one of the most ideal ways toprepare protective layer. According to the composition of thetarget film， the required compound is deposited on the substrateby two or more surface reactions step by step 54. However，limited by the expensive cost of equipment and also the complexcontrolling system， ALD has not been used widely. Thus， otherprotective layers have been applied to Cu2O， such as carbonlayer 51 and polymer layers 55. Li et al. 56 applied dense andhydrophobic polyimide （PI） layer on the surface of Cu2Othrough a facile spin-coating method， which showed the highestcurrent density value of 1.8 mA·cm?2 after 20 min withoutattenuation. Compared with ALD， this method provides a lowcoststrategy and could be extended to other systems with photocorrosionproblems. However， with further enhancing the PIamount， the cathode would lead to a very stable but lower currentdensity because of the larger charge transfer resistance andhigher capacitance in the space charge layer.

After extensive research with the above strategies， theperformance of Cu2O photocathodes has been improved greatly.Some novel and excellent Cu2O photocathodes and theirperformance are compared in Table 1. Among of them，F(xiàn)TO/Au/Cu2O/ZnO：Al/TiO2/Pt delivers the highestphotocurrent at 0 V vs. RHE， ?7.6 mA cm?2， lasting at least 1 h，benefiting from the elaborately designed multi-layer structure，and this structure is the most stable and has been widely adoptedcurrently. What’s more， substituting the Au substrate with a holetransport layer （such as FeOOH， NiOx， CuSCN et al.） couldenhance cathodes stability by enabling better separation ofcharge carriers. Nevertheless， the solar-to-hydrogen （STH） ofthe Cu2O-based photocathodes， which is rarely reported in theseworks， is still pretty low （lt; 5%）， which requires more efficientdesign of photoelectrodes.

5 Advanced characterization techniques

With the development of modern research， especially the fastprogress on nanoscience， traditional characterization tools， suchas electron microscopy， X-ray photoelectron spectroscopy （XPS）and X-ray diffraction （XRD）， are far more than enough toanalyze the fine structures of electrodes and understand theunderlying mechanism. In this section， some newly emerging advanced characterization techniques in the field ofphotoelectrochemistry， such as X-ray absorption spectroscopy（XAS）， surface photovoltage microscopy （SPVM）， timeresolvedtwo-photon photoemission spectroscopy （TR-2PPE），etc.， will be briefly introduced.

As an advanced technology to probe the electronic andgeometric structures， X-ray absorption spectroscopy providesmore detailed information on samples. Yilmaz et al. 57 studiedthe effect of deposition parameters on surface and bulk and usedthe high resolution X-ray absorption near edge structure（XANES） to analyze precisely the chemical state of Cu in Cu2Ofilms. By changing deposition time and potential， Cu2O filmswith different morphologies were obtained. The results showedthat the Cu2+ fraction decreased with negatively shifting thedeposition potential or increasing the electrodeposition time. Itwas concluded that Cu2+ （specifically CuO） could not onlyenhance the charge separation and light absorption， but alsoinfluence the long-term stability of Cu2O depending on thecontent of Cu2+. So， a suitable amount of Cu2+ is much importantfor applying Cu2O to PEC system.

To understand the photo charges transfer process， Li group 58constructed surface photovoltage microscopy to analyze theprocess of charge transfer. By measuring the surfacephotovoltage at different facets of Cu2O particles， they foundthat the aggregation of photogenerated electrons is different ateach facet， as is showed in Fig. 9. More specifically， the electronsaccumulated more on the {001} facet than {111} facet， whichis explained by that the {001} facet has a more significant ptypecharacter and more copper vacancy. In a word， they provedthat the photogenerated charge separation is more efficient at thesurface with anisotropic defects.

Recently， time-resolved two-photon photoemissionspectroscopy （TR-2PPE） has been developed to obtainfemtosecond response， which allows detecting photo chargestransfer process as short as femtosecond possible. TR-2PPE usedtwo femtosecond laser pulses to gain carriers in two differentstates， which could provide information on the energy of theintermediate， normally unoccupied， states at the surface （Fig.10）. Borgwardt et al. 59 used TR-2PPE to investigate the electrondynamics and energetics at surface and interface of Cu2O. Andthe results demonstrated that the electrons lost a lot of energy forbeing trapped into bulk defect states. Furthermore， a Pt layercould accelerate the extraction of electrons and increase thephotocurrent.

In addition to the techniques mentioned above， manycharacterization techniques could also be used to obtaininformation of light adsorption， morphologies， element valence，and even the conversion process of intermediates， such asspectroscopy （ultraviolet visible spectrum （UV）， Infraredspectroscopy （IR） and Raman spectroscopy）， X-ray Diffraction（XRD）， X-ray Photon-electron Spectroscopy （XPS）， ScanningElectron Microscope （SEM）， Transmission Electron Microscope（TEM） and so on. Furthermore， by virtue of in situ electrochemical spectroscopy techniques， the reactionintermediates could be detected， which is indispensable for understanding reaction mechanism. In summary， advancedcharacterization technologies help to better understand theunderlying mechanism of the PEC process， which gives rationalguidance for photocathode design.

6 Conclusions and perspective

Cu2O as a representative copper-based compounds， hasdisplayed promising performance in PEC water splitting， yetsuffers seriously from photo-corrosion， leading to unsatisfactorystability. So， there is still a long way to go before overcoming thechallenge in both activity and stability. In the future， the researchdirections as follows are promising and needs further research.

（1） It is difficult to achieve both high PEC performance andlong-term stability for single Cu2O. Construction of multi-layerstructure is necessary and efficient， such as introducing a holetransfer layer as substrate， n-type semiconductor or electrontransfer layer as a buffer layer， introducing co-catalyst on the toplayer， etc. Currently， the most effective way to improve stabilityis to form heterojunctions with suitable n-type semiconductors，which could be expected to not only improve the separationefficiency of charge carriers but also provide a protecting layerto isolate Cu2O from water.

（2） Although some techniques like ALD are efficient toprepare thin-layer electrodes， the high cost of devices hamper itsextensive application in labs. Developing low cost andconvenient techniques， such as bottom-up chemical routes， torealize controllable synthesis of photoelectrodes are highlydesired.

（3） Developing advanced characterization techniques iscrucial for understanding the underlying photoelectrochemicalmechanism， yet full of challenges， limited by one side theultrafast process of photogenerated charge creation andseparation/transfer， and the other side the highly complicatedsolid/liquid interfaces of photoelectrochemical reactions.

（4） To obtain high performance Cu2O-based photocathodes，the onset potential should be emphasized， which could beadjusted by introducing a buffer layer with appropriate bandedges. Meanwhile， it is also necessary to guarantee rapid transferof photogenerated electrons to surface to participate in waterreduction rather than to reduce Cu2O.

（5） Although Cu2O is with significant advantages asphotocathodes for water splitting， there is still a long way forpractical application. More efforts are required for Cu2O-basedphotocathodes to improve the stability， photocurrent， STH andso on.

Author Contribution： Conceptualization， L.H. Jiang andG.B. Liu; Writing-Original Draft Preparation， H.S. Lu， S.X.Song and Q.S. Jia; Writing-Review amp; Editing， L.H. Jiang;Supervision， L.H. Jiang; Project Administration， L.H. Jiang.

References

（1） Grossmann， W. D.; Grossmann， I.; Steininger， K. W. Renew. Sust.Energ. Rev. 2014， 32， 983. doi： 10.1016/j.rser.2014.01.003

（2） Fujishima， A.; Honda， K. Nature 1972， 238 （5358）， 37.doi： 10.1038/238037a0

（3） Khaselev， O.; Turner， J. A. Science 1998， 280 （5362）， 425.doi： 10.1126/science.280.5362.425

（4） Wang， G.; Wang， H.; Ling， Y.; Tang， Y.; Yang， X.; Fitzmorris， R. C.;Wang， C.; Zhang， J. Z.; Li， Y. Nano Lett. 2011， 11 （7）， 3026.doi： 10.1021/nl201766h

（5） Hisatomi， T.; Domen， K. Faraday Discuss. 2017， 198 （0）， 11.doi： 10.1039/C6FD00221H

（6） Pinaud， B. A.; Benck， J. D.; Seitz， L. C.; Forman， A. J.; Chen， Z.;Deutsch， T. G.; James， B. D.; Baum， K. N.; Baum， G. N.; Ardo， S.; etal. Energy Environ. Sci. 2013， 6 （7）， 1983. doi： 10.1039/C3EE40831K

（7） Huang， Q.; Ye， Z.; Xiao， X. J. Mater. Chem. A 2015， 3 （31）， 15824.doi： 10.1039/C5TA03594E

（8） Bagal， I. V.; Chodankar， N. R.; Hassan， M. A.; Waseem， A.; Johar， M.A.; Kim， D.-H.; Ryu， S.-W. Int. J. Hydrog. Energy 2019， 44 （39），21351. doi： 10.1016/j.ijhydene.2019.06.184

（9） de Jongh， P. E.; Vanmaekelbergh， D.; Kelly， J. J. J. Electrochem. Soc.2000， 147 （2）， 486. doi： 10.1149/1.1393221

（10） Chen， Z.; Jaramillo， T. F.; Deutsch， T. G.; Kleiman-Shwarsctein， A.;Forman， A. J.; Gaillard， N.; Garland， R.; Takanabe， K.; Heske， C.;Sunkara， M.; et al. J. Mater. Res. 2010， 25 （1）， 3.doi： 10.1557/JMR.2010.0020

（11） Paracchino， A.; Laporte， V.; Sivula， K.; Gr?tzel， M.; Thimsen， E. Nat.Mater. 2011， 10 （6）， 456. doi： 10.1038/nmat3017

（12） Toe， C. Y.; Scott， J.; Amal， R.; Ng， Y. H. J. Photochem. Photobiol. C2019， 40， 191. doi： 10.1016/j.jphotochemrev.2018.10.001

（13） Carp， O.; Huisman， C. L.; Reller， A. Prog. Solid State Chem. 2004，32 （1）， 33. doi： 10.1016/j.progsolidstchem.2004.08.001

（14） Tench， D.; Warren， L. F. J. Electrochem. Soc. 1983， 130 （4）， 869.doi： 10.1149/1.2119838

（15） Zhang， W.; Wen， X.; Yang， S.; Berta， Y.; Wang， Z. L. Adv. Mater.2003， 15 （10）， 822. doi： 10.1002/adma.200304840

（16） Soon， A.; Todorova， M.; Delley， B.; Stampfl， C. Phys. Rev. B 2007，75 （12）， 125420. doi： 10.1103/PhysRevB.75.125420

（17） Aveline， A.; Bonilla， I. R. Sol. Energy Mater. 1981， 5 （2）， 211.doi： 10.1016/0165-1633（81）90033-2

（18） Ishibashi， S.; Higuchi， Y.; Ota， Y.; Nakamura， K. J. Vac. Sci. Technol.A 1990， 8 （3）， 1403. doi： 10.1116/1.576890

（19） Qin， C.; Chen， X.; Liang， R.; Jiang， N.; Zheng， Z.; Ye， Z.; Zhu， L. ACSAppl. Energy Mater. 2022， 5 （11）， 14410. doi： 10.1021/acsaem.2c02974

（20） Eisermann， S.; Kronenberger， A.; Laufer， A.; Bieber， J.; Haas， G.;Lautenschl?ger， S.; Homm， G.; Klar， P. J.; Meyer， B. K. Phys. StatusSolidi A-Appl. Mat. 2012， 209 （3）， 531. doi： 10.1002/pssa.201127493

（21） Jeong， S.; Aydil， E. S. J. Vac. Sci. Technol. A 2010， 28 （6）， 1338.doi： 10.1116/1.3491036

（22） Das， C.; Ananthoju， B.; Dhara， A. K.; Aslam， M.; Sarkar， S. K.;Balasubramaniam， K. R. Adv. Mater. Interfaces 2017， 4 （17），1700271. doi： 10.1002/admi.201700271

（23） Liu， H.; Nguyen， V. H.; Roussel， H.; Gélard， I.; Rapenne， L.;Deschanvres， J.-L.; Jiménez， C.; Mu?oz-Rojas， D. Adv. Mater.Interfaces 2019， 6 （3）， 1801364. doi： 10.1002/admi.201801364

（24） Güneri， E.; Aker， D.; Henry， J.; Billur， C. A.; Saat?i， B. Phase Transitions 2022， 95 （10）， 679. doi： 10.1080/01411594.2022.2104161

（25） Aref， A. A.; Xiong， L.; Yan， N.; Abdulkarem， A. M.; Yu， Y. Mater. Chem.Phys. 2011， 127 （3）， 433. doi： 10.1016/j.matchemphys.2011.02.029

（26） Liu， M.; Xue， D. J. Phys. Chem. C 2008， 112 （16）， 6346.doi： 10.1021/jp800803s

（27） Luo， C.; Xue， D. Langmuir 2006， 22 （24）， 9914. doi： 10.1021/la062193v

（28） Xiong， L.; Yu， H.; Yang， G.; Qiu， M.; Chen， J.; Yu， Y. Thin Solid Films 2010， 518 （23）， 6738. doi： 10.1016/j.tsf.2010.05.117

（29） Luo， J.; Steier， L.; Son， M.-K.; Schreier， M.; Mayer， M. T.; Gr?tzel，M. Nano Lett. 2016， 16 （3）， 1848. doi： 10.1021/acs.nanolett.5b04929

（30） Jiang， D.; Zhang， Y.; Li， X. Chin. J. Catal. 2019， 40 （1）， 105.doi： 10.1016/S1872-2067（18）63164-X

（31） Zhang， Z.; Song， R.; Cao， T.; Huang， W. J. Energy Chem. 2016，25 （6）， 1086. doi： 10.1016/j.jechem.2016.09.012

（32） Wang， Y.; Miska， P.; Pilloud， D.; Horwat， D.; Mücklich， F.; Pierson， J.F. J. Appl. Phys. 2014， 115 （7）， 073505. doi： 10.1063/1.4865957

（33） Musa， A. O.; Akomolafe， T.; Carter， M. J. Sol. Energy Mater. Sol.Cells 1998， 51 （3）， 305. doi： 10.1016/S0927-0248（97）00233-X

（34） Visibile， A.; Wang， R. B.; Vertova， A.; Rondinini， S.; Minguzzi， A.;Ahlberg， E.; Busch， M. Chem. Mater. 2019， 31 （13）， 4787.doi： 10.1021/acs.chemmater.9b01122

（35） Nolan， M.; Elliott， S. D. Chem. Mater. 2008， 20 （17）， 5522.doi： 10.1021/cm703395k

（36） Tseng， C. C.; Hsieh， J. H.; Liu， S. J.; Wu， W. Thin Solid Films 2009，518 （5）， 1407. doi： 10.1016/j.tsf.2009.09.116

（37） Chen， D.; Liu， Z.; Guo， Z.; Yan， W.; Xin， Y. J. Mater. Chem. A 2018，6 （41）， 20393. doi： 10.1039/C8TA07503D

（38） Li， X.; Liu， B.; Chen， Y.; Fan， X.; Li， Y.; Zhang， F.; Zhang， G.; Peng，W. Nanotechnology 2018， 29 （50）， 505603.doi： 10.1088/1361-6528/aae569

（39） Shinde， P. S.; Fontenot， P. R.; Donahue， J. P.; Waters， J. L.; Kung， P.;McNamara， L. E.; Hammer， N. I.; Gupta， A.; Pan， S. J. Mater. Chem.A 2018， 6 （20）， 9569. doi： 10.1039/C8TA01771A

（40） Würfel， U.; Cuevas， A.; Würfel， P. IEEE J. Photovoltaics 2015， 5 （1），461. doi： 10.1109/JPHOTOV.2014.2363550

（41） Yang， W.-Y.; Rhee， S.-W. Appl. Phys. Lett. 2007， 91 （23）， 232907.doi： 10.1063/1.2822403

（42） Zhou， M.; Guo， Z.; Liu， Z. Appl. Catal. B 2020， 260， 118213.doi： 10.1016/j.apcatb.2019.118213

（43） Wei， Y.; Chang， X.; Wang， T.; Li， C.; Gong， J. Small 2017， 13 （39），1702007. doi： 10.1002/smll.201702007

（44） Pan， L.; Liu， Y.; Yao， L.; Dan， R.; Sivula， K.; Gr?tzel， M.; Hagfeldt， A.Nat. Commun. 2020， 11 （1）， 318. doi： 10.1038/s41467-019-13987-5

（45） Liu， G.; Lu， H.; Xu， Y.; Quan， Q.; Lv， H.; Cui， X.; Chen， J.; Jiang， L.;Behm， R. J. Chem. Eng. J. 2023， 455， 140875.doi：10.1016/j.cej.2022.140875

（46） Gou， L.; Murphy， C. J. Nano Lett. 2003， 3 （2）， 231.doi： 10.1021/nl0258776

（47） Cao， M.; Hu， C.; Wang， Y.; Guo， Y.; Guo， C.; Wang， E. Chem.Commun. 2003， No. 15， 1884. doi： 10.1039/B304505F

（48） Kim， M. H.; Lim， B.; Lee， E. P.; Xia， Y. J. Mater. Chem. 2008，18 （34）， 4069. doi： 10.1039/B805913F

（49） Kuo， C. H.; Chen， C. H.; Huang， M. H. Adv. Funct. Mater. 2007，17 （18）， 3773. doi： 10.1002/adfm.200700356

（50） Lin， C.-Y.; Lai， Y.-H.; Mersch， D.; Reisner， E. Chem. Sci. 2012，3 （12）， 3482. doi： 10.1039/C2SC20874A

（51） Zhang， Z.; Dua， R.; Zhang， L.; Zhu， H.; Zhang， H.; Wang， P. ACSNano 2013， 7 （2）， 1709. doi： 10.1021/nn3057092

（52） Lai， T.-H.; Tsao， C.-W.; Fang， M.-J.; Wu， J.-Y.; Chang， Y.-P.; Chiu，Y.-H.; Hsieh， P.-Y.; Kuo， M.-Y.; Chang， K.-D.; Hsu， Y.-J. ACS Appl.Mater. Interfaces 2022， 14 （36）， 40771. doi： 10.1021/acsami.2c07145

（53） Pande， K. P.; Hsu， Y. S.; Borrego， J. M.; Ghandhi， S. K. Appl. Phys.Lett. 1978， 33 （8）， 717. doi： 10.1063/1.90513

（54） George， S. M.; Ott， A. W.; Klaus， J. W. J. Phys. Chem. 1996， 100 （31），13121. doi： 10.1021/jp9536763

（55） Li， Y.; Zhong， X.; Luo， K.; Shao， Z. J. Mater. Chem. A 2019， 7 （26），15593. doi： 10.1039/C9TA04822G

（56） Li， J.; Li， W.; Deng， G.; Qin， Y.; Wang， H.; Wang， Y.; Xue， S. Ionics2023， 29 （2）， 685. doi： 10.1007/s11581-022-04827-6

（57） Yilmaz， M.; Handoko， A. D.; Parkin， I. P.; Sankar， G. J. Catal. 2020，389， 483. doi： 10.1016/j.jcat.2020.06.021

（58） Chen， R.; Ren， Z.; Liang， Y.; Zhang， G.; Dittrich， T.; Liu， R.; Liu， Y.;Zhao， Y.; Pang， S.; An， H.; et al. Nature 2022， 610 （7931）， 296.doi： 10.1038/s41586-022-05183-1

（59） Borgwardt， M.; Omelchenko， S. T.; Favaro， M.; Plate， P.; H?hn， C.;Abou-Ras， D.; Schwarzburg， K.; van de Krol， R.; Atwater， H. A.;Lewis， N. S.; et al. Nat. Commun. 2019， 10 （1）， 2106.doi： 10.1038/s41467-019-10143-x

國家自然科學基金（22179067， 22279069）和山東省自然科學基金（ZR2022ZD10）資助項目