p-type ZnO for photocatalytic water splitting

In recent years, energy demand and the requirement for innovation in clean and environmentally friendly technologies have become increasingly urgent. Renewable energy sources such as tidal energy, solar energy, wind energy, ocean energy, and geothermal energy have gradually come into people’s vision with the view of replacing traditional non-renewable fossil energy.^1–6 Among these new energies, solar energy is extremely abundant. Photocatalytic water splitting to produce H₂ and O₂ using sunlight energy has the advantages of being clean, pollution-free, and widely distributed, occupying a unique position in the field of renewable energy technologies.⁷ Since 1972, the first experiment on the light-driven photocatalytic hydrogen production by using TiO₂ has promoted scientific exploration, bringing convenience and sustainable development to human society.⁸ However, it still faces fundamental and technological challenges to efficiently convert solar energy to hydrogen (STH) energy and finalize it in cost-effective methods on a large scale.

Over the past five decades, various photocatalytic materials have been developed to split water into H₂ and O₂ under ultraviolet (UV) and visible-light radiation, among which ZnO has drawn much attention from researchers since it has a similar bandgap to TiO₂, faster electron mobility, and fewer defects.⁹ Moreover, ZnO has both great semiconductor properties and piezotronic effects. The piezotronic effect will generate a built-in electric field under the mechanical energy, thus promoting charge separation and transport, which provides a new way to solve the low efficiency of photocatalysis. Although ZnO has various superior properties, it still lacks high-efficiency ZnO-based optoelectronic systems for industrial production mainly because of the difficulty of formation of stable p–n junctions, which is essential for achieving high performance of ZnO.^10,11 The current situation boils down to the challenge of obtaining p-type ZnO with good stability, thus hindering the rapid development of ZnO-based systems. In general, the prepared ZnO often exhibits a natural n-type characteristic due to its intrinsic donor defects and hydrogen participation. Therefore, it is highly desired to focus more attention on constructing p-type ZnO, which can provide more possibilities for boosting the photocatalytic water splitting system. As an appendant advantage, p-type ZnO is also promising to fabricate functional heterojunctions and homojunctions.^12–15 For example, benefiting from successfully getting p-type Ag-doped ZnO (Ag:ZnO) nanorods (NRs) through a low-temperature solution procedure, the homojunction of n-type ZnO NRs/p-type Ag:ZnO NRs was prepared and worked effectively for photocatalytic water splitting.¹⁶ 

Intrinsic donor defects are the primary inducements that cause the as-prepared ZnO to reflect n-type conductivity; among them, Zn interstitials (Zn_i) and oxygen vacancies (V_O) are the main types of ionic defects.^17–19 It is more difficult to increase the acceptor concentration due to the limitation of the solid solubility of acceptor impurities and self-compensation in ZnO. In addition, most of acceptor levels are relatively deep, which also leads to the fact that the majority of the doped acceptors are unable to ionize and provide holes.^20–22 Therefore, effective p-type doping should meet the requirements of shallow acceptor level and large acceptor doping concentration. According to recent experimental and theoretical reports, diverse elements doping can be used to prepare the p-type ZnO nanostructure, including doping group-I elements into Zn sites and doping group-V elements and co-doping group III and V elements into ZnO.^23–28 The engineering and operation of the doping method are increasingly playing a vital role in achieving p-type ZnO since the doping process could promote carrier transfer and change the band structure such as narrowing down the bandgap energy, which helps to expand the light adsorption of ZnO to visible light and increase the conversion efficiency of photocatalytic water splitting.

With the aim of stimulating new thoughts to address crucial challenges, in this Perspective, our attention predominantly concentrates on the preparation and application of p-type ZnO for water splitting. Starting with a brief introduction to achieve and improve p-type conductivity in ZnO by various synthetic techniques in Sec. II, different elements as dopant sources for forming p-type ZnO will be briefly discussed. Then, we move on to the recent developments of getting high-efficiency and stable p-type ZnO for photocatalytic water splitting systems in Sec. III; meanwhile, some methods and related mechanisms for improving photocatalytic performance will be also involved. Finally, we give our comments and outlook on p-type ZnO for photocatalytic water splitting by disclosing the remaining challenges and future research directions in Sec. IV.

ZnO-based nanomaterials have been intensively studied in many emerging fields especially in photovoltaics, electronics, and photocatalysis due to their excellent stability, good mechanical strength, and high electron mobility. However, the difficulty of preparing durable and reproducible p-type ZnO has limited its widespread development. Investigators have tried many methods to successfully obtain p-type ZnO mainly through doping various elements into ZnO. In this section, we will analyze and discuss the selectivity of p-type dopants and their fabrication technology in the recent representative articles.

In theory, dopants at cation sites in semiconductors typically generate shallower acceptor levels than dopants at anion sites.^18,29,30 For example, the low acceptor levels can be obtained when introducing group-I elements to replace Zn sites in ZnO. Group-I elements include group-IA elements (Li, Na, and K) and group-IB elements (Ag, Cu, and Au) among which Li is the most studied element. Group-I elements doping into ZnO not only occupy the lattice position of Zn as an acceptor but also tend to become interstitial atoms as an electron donor.^29,31

Lee et al. developed an effective hydrothermal method to fabricate vertically oriented p-type Li:ZnO nanowires (NWs) by using the lithium nitrate aqueous solution as a dopant source followed by post-annealing.²⁴ Systematic experiment results showed that Li could occupy the vacant sites of Zn, and the replacement of Li was stemming from the thermally induced migration by post-annealing. Furthermore, after annealing the Li:ZnO NWs (annealed Li:ZnO NWs), the photoluminescence (PL) spectra presented two free excitonic peaks (FX_A and FX_B) and two acceptor-bound exciton peaks (A₁X and A₂X), indicating the behavior of p-type conductivity [Fig. 1(a)]. More specifically, Huang et al. tried to analyze p-type Li:ZnO from the aspects of energy band theory and ionic crystal localized state electronic theory.³² According to the energy band theory [Fig. 1(b)], the valence band of Li:ZnO would not be filled in since Li only has one valence electron. Consequently, the defects of Li-substitutional Zn (Li_Zn) could get electrons and introduce a shallow acceptor level to achieve p-type conductivity. Through the ionic crystal localized state electronic theory, the abstract concept could be concluded that the hole was O⁻ and the unexcited and excited acceptors are LiO₄⁶⁻ and LiO₄⁷⁻, respectively [Figs. 1(c) and 1(d)]. Therefore, directly building up a lattice with O⁻ was the goal in order to fabricate p-type ZnO. Based on the above theory, they came up with a multi-step thermodynamic strategy for the preparation of p-type ZnO [Fig. 1(e)]. The main step was introducing Li atoms into the ZnO lattice by the electrochemistry process for converting vacancies of Zn (V_Zn) to Li_Zn in CH₃Li atmosphere and annealing in a weak oxidizing atmosphere at low temperature to get rid of the relatively unstable interstitial Li (Li_i) for obtaining stable Li_Zn. From this understanding, we can choose other group-I elements using the multi-step thermodynamic process for fabricating p-type ZnO.

As for Na-doped p-type ZnO, several methods such as pulsed laser deposition, chemical vapor deposition, or sol-gel processes have been published for the material fabrication.^33–36 As shown in Fig. 1(f), Jaisutti et al. doped Na atoms to synthesize p-type Na:ZnO nanoflowers (NFs) by a moderate solution process at low temperature.³⁸ Through optimization of the composition, the Na:ZnO NFs presented high sensitivity to acetone and various alcohol gases. Theoretically, the efficiency of p-type doping via the group-IA elements is affected by the formation of compensatory interstitials.³⁹ In this term, group-IB elements should be better dopant candidates than group-IA elements because they are not interested in occupying the interstitial sites of ZnO and possess small self-compensation.⁴⁰ Choosing Cu as dopants should be an effective way to form p-type Cu:ZnO because the radii of Cu⁺ (77 pm) and Cu²⁺ (73 pm) ions were suitable for Zn²⁺ (74 pm) ion substitution. Xu et al. illustrated that the introduced acceptor energy level induced by Cu doping was 0.45 eV above the original valence band of ZnO from which it can be deduced that the valence level of Cu may be +1 or +2 depending on its chemical environment.⁴¹ Experimentally, a p–n ZnO NW homojunction was synthesized by vertical doping with Cu under the furnace system with hotwire assistance [Fig. 1(g)]. The Hall effect proved that ZnO NWs showed p-type conductivity after Cu doping.⁴² Furthermore, other p-type Cu:ZnO films were successfully grown on c-sapphire substrates by plasma-assisted molecular beam epitaxy.⁴³ The results of PL measurement indicated that a shallow acceptor level was introduced in the forbidden zone, which was 0.15 eV above the edge of the valence band, and the strong p-type behavior was attributed to the replacement of the Zn site by Cu⁺ state [Fig. 1(h)]. However, almost all investigated p-type ZnO samples were unstable and tend to transform to n-type ZnO finally; the intrinsic reason may be mainly originated from the charge compensation from exterior defects in ZnO. In conclusion, the stability of p-type ZnO remains a great challenge.

According to theoretical calculations of the electronic band structure, group-V elements have the potential to be better doping sources than group-I elements for producing p-type ZnO, especially the N element, which is considered to be the most hopeful dopant to obtain p-type ZnO via replacing O sites with N. The advantages of N doping include the similar atomic size and the close energy of the N 2p-orbitals to the O states as well as high solubility of N and its low formation energy.^44–46 However, based on theoretical contributions, it is impossible to obtain stable shallow acceptor levels in a Zn-rich environment by doping N alone.⁴⁸ Therefore, Chavillon et al. prompted a strategy to address the unstable problem;^49–50 they supposed that p-type ZnO should be obtained if the process of N insertion under the conditions of O-rich and Zn-poor environment is controlled.⁵² As a result, N:ZnO nanostructures were successfully fabricated using ZnO₂ as the precursor and then ammonolysis under NH₃ atmosphere, as shown in Fig. 2(a). The photoelectrochemical (PEC) measurements indicated that the N:ZnO sample prepared at 250 °C had obvious p-type characteristics, while the sample prepared with the Zn-rich precursor (ZnO) still exhibited n-type conductivity [Fig. 2(b)]. This work demonstrated that only combining V_Zn with N doping can induce a stable p-type ZnO nanostructure. More interestingly, the obtained p-type ZnO nanoparticles remained stable for more than 2 years under ambient conditions. In another similar work, N was also used as a doping source to grow p-type ZnO on a silicon substrate by chemical vapor deposition without any metal catalyst.⁵³ The x-ray photoelectron spectroscopy (XPS) result showed two N 1s related peaks, the one at low energy was assigned to N substituting for O sites, which was the ideal approach for p-type doping [inset of Fig. 2(c)]. Another peak at high energy indicated that these NWs were grown under the O-rich condition, which was consistent with the above research results that the O-rich condition can restrain the intrinsic donors (Zn_i and V_O) and contribute to p-type conductivity.^53,54

The other group-V elements denoted as X elements (P, As, and Sb) were inclined to substitute O sites to form deep acceptors X_O (P_O, As_O, and Sb_O) with high acceptor ionization energies according to theoretical results due to their larger ionic radii than that of O, which made X elements impossible to be incorporated into the lattice of O effectively.^30,55 With the help of first-principle calculations, Limpijumnong et al. proposed a model of the X_Zn–2V_Zn acceptor complex to reveal the doping of large-size elements into ZnO, owing to the lower formation energy required at that case.⁵⁷ Specifically, X elements are energetically preferred to replace Zn and subsequently induce two V_Zn to form the X_Zn–2V_Zn complexes. A p–n ZnO NW homojunction was successfully constructed by a simple two-step hydrothermal method using P₂O₅ as a doping source [Fig. 2(d)].⁵⁸ In addition, Nagar et al. has synthesized p-type P-doped ZnO via the plasma immersion ion implantation technique instead of the well-studied conventional method.⁵⁹ The experimental results certified that there was a gradual conversion that happened on ZnO conductive behavior, which was converted from n-type to p-type when the Zn sites were substituted by P elements, and the results further supported the universality of the P_Zn–2V_Zn acceptor model in p-type P:ZnO.

Similar to P, Sb as a considerable large-size mismatched dopant in ZnO is now widely used for the preparation of high-quality p-type ZnO. Sb:ZnO NWs with variable aspect-ratios were recently fabricated by Huo et al. via the low-temperature hydrothermal method.⁶⁰ They found that the defect peak of PL spectra caused by oxygen defects (V_O) reduced with the increasing doping concentration of Sb, which was the direct evidence for Sb substituting Zn sites (Sb_Zn) and generating V_Zn. Then, V_Zn could attract adjacent valence electrons to form a strong negative charge center [Figs. 3(a) and 3(b)]. Interestingly, they found that the prepared p-type Sb:ZnO films had a piezotronic effect. Pradel et al. also designed a similar p-type Sb:ZnO with an ultralong NW morphology via a modified solution growth method.⁶¹ By controlling the process of self-nucleation of ZnO, the NW length could reach up to 60 µm [Fig. 3(c)]. Besides, a chemical vapor deposition method was also employed for the fabrication of p-type Sb:ZnO NWs.⁶² The obtained Sb:ZnO NWs presented great uniformity and consistency in the oblique-view SEM picture, and the length of the NWs could reach an average of 5 µm [Fig. 3(d)]. The near band edge (NBE) PL emission spectrum confirmed the p-type behavior of the Sb:ZnO NWs [Fig. 3(e)]. From the schematic of energy level emission [the inset of Fig. 3(e)], it seemed that the emission-related acceptor level played the dominant role, which indicated the formation of durable and reliable Sb_Zn–2V_Zn complexes.

Generally, one-dimensional nanostructures are attractive for basic research, but their applications in products remain limited because the prepared materials do not exhibit morphological uniformity and severely lack reproducibility. To produce sufficient dimensional nanostructures, Laurenti et al. prepared high surface area p-type Sb:ZnO with a nanobranched structure by a multi-step synthetic process including deposition, wet-impregnation, and thermal oxidation approaches.⁶³ At first, the introduction of Sb dopants could generate additional shallow donor energy states (Sb_Zn–2V_Zn) in the ZnO band structure, resulting in the reduction of the energy bandgap (E_g) of ZnO from 3.30 to 3.25 eV after Sb doping [Fig. 3(f)]. Then, Hall effect measurements further verified that p-type behavior of ZnO was obtained after inserting Sb dopants, and the hole mobility and concentration were calculated to be 1.44 cm² V⁻¹ s⁻¹ and 2.8 × 10²⁰ cm⁻³, respectively, which were believed to be quite overestimated for other p-type Sb:ZnO due to the unique nanobranched morphology. Moreover, the two-dimensional p-type Sb:ZnO/n-type ZnO homojunction was constructed based on Sb:ZnO films grown on undoped non-polar ZnO substrates.⁶⁴ The experimental results confirmed the existence of the Sb_Zn–2V_Zn doping pattern, which induced the p-type characteristic of Sb:ZnO film. Surprisingly, this simple fabricated non-polar p–n ZnO homojunction exhibited excellent durability (more than 12 months); the material shows great promise in the application of photocatalytic water splitting for hydrogen generation. Using a similar approach, Pradel et al. achieved ultralong p-type Sb-doped ZnO NWs, which presented stability of over 18 months by the modified hydrothermal method.⁶⁵ Based on the acquirement of these stable p-type ZnO NWs, bulk homojunction films of special core–shell ZnO structures were grown with either n-type or p-type ZnO as cores and the opposite type as shells. As presented in Figs. 3(g) and 3(h), it could be seen that the diameter and density of NWs increased remarkably with the envelopment of the shell layer, making them grow until they collided with each other to form homojunction films. In this construction, p-type Sb:ZnO core–shell structures have high interfacial area, leading to the easy formation of electron–hole and high stability, which gives the guarantee to get high photocatalytic performance.

To obtain high-quality p-type ZnO, Yamamoto et al. introduced the co-doping method based on ab initio electronic band-structure calculations.^65,66 They found that combining acceptors and donors as co-dopants with acceptor/donor ratio of 2:1 could boost the integration of acceptors, reduce the acceptor levels, and lift the donor levels. Interestingly, donors could decrease the ionization energy of acceptors and induce to form acceptor–donor–acceptor complexes because of the strong interaction between acceptors and donors. Afterward, varieties of co-doping combinations were extensively studied in experiment and theory, including the acceptor–donor co-doping,^67,68 dual-acceptor co-doping,^69,70 and acceptor-hydrogen co-doping.⁷² Mannam et al. prepared p-type ZnO thin films with P and N dual-acceptor co-doping through pulsed laser deposition.²⁶ A high hole concentration of 1.14 × 10²¹ cm⁻³ with a mobility of 1.2 cm² V⁻¹ s⁻¹ was accomplished. The XPS and PL experiments indicated that P occupied Zn sites and N replaced O sites, resulting in the presence of V_Zn in co-doped ZnO [Fig. 4(a)]. The p-type conductivity and high electrical performance were attributed to the introduced effective shallow acceptor complex (P_Zn–V_Zn–4N_O) in P and N co-doped ZnO films [Fig. 4(b)]. Sharma et al. adopted Li and P as dual-acceptor dopants to produce stable p-type ZnO by dual ion beam sputtering and the following annealing treatment.²¹ Similar to the aforementioned results on mono-acceptor doping, they proposed that Li and P as the dual-acceptor acted simultaneously on ZnO for different roles [Fig. 4(c)], namely, Li could substitute the Zn site for the production of the shallow acceptor level (Li_Zn) while P could incorporate into ZnO for forming the P_Zn–2V_Zn complex. Moreover, Saravanakumar et al. employed the acceptor–donor co-doping combination in which Al and N were applied as the dopants for preparing p-type ZnO films by the sol-gel assisted spin coating method.⁷³ It could be observed that the bandgap was decreased to 3.1 eV after co-doping N and Al into ZnO films, which provided favorable conditions for enhancing the photocatalytic efficiency [Fig. 4(d)]. Besides, the p-type conductivity of Al–N:ZnO was attributed to the O-rich environment created by the co-doping of Al and N. For better reviewing, we summarized the relevant literature and presented some reported dopants and basic parameters of prepared p-type ZnO in Table I.

The demand for developing renewable energy has attracted rising attention among the massive study; converting and storing solar energy into chemical energy (such as H₂) via photocatalytic technique shows highly promising results. At present, the efficiency of photocatalytic water splitting is limited by the light absorption property of the photocatalyst, the separation and transfer of photogenerated carriers, and interfacial charge-participated chemical reactions.⁸⁷ As a kind of photocatalysts, ZnO is easy to be nanostructured with various morphologies, which are expected to be a promising candidate in water splitting applications. Generally, ZnO presents poor photocatalytic performance due to its wide bandgap, resulting in the absorption of only ultraviolet (UV) light while neglecting the main visible-light irradiation in sunlight. Different strategies of narrowing the bandgap of ZnO have been investigated to boost photocatalytic performance. Construction of the homojunction and heterojunction of ZnO is one of the important approaches to engineer its bandgap, and the prerequisite is achieving stable p-type ZnO. Doped p-type ZnO could give rise to absorption of light with broader wavelengths because p-type doping will introduce a new acceptor energy level between the conduction band (CB) and the valence band (VB) to modify the electronic band structure, which leads to light absorption with broader wavelengths. A successful example of the ZnO homojunction was prepared by p-type Cu:ZnO/n-type ZnO,⁴² as illustrated in the energy band diagram of Fig. 5(a); the interface band bending was fundamentally determined by the natural characteristics of n-type and p-type ZnO. The photoinduced electrons at the conduction band and the holes at the valence band can be separated and transferred effectively, thus improving the photocatalytic efficiency of ZnO.

As mentioned above, Na is a good dopant for p-type doping ZnO. Wu et al. successfully obtained p-type Na:ZnO ultrathin hollow spheres through economic ion adsorption and templating approach.⁸⁸ As an intermediate, the carbon spheres without selective absorption property were beneficial to the uniform distribution of Zn and Na ions on the surface; then, p-type Na:ZnO ultrathin hollow spheres were acquired after the removal of the carbon cores by calcination [Fig. 5(b)]. The results of ultraviolet photoelectron spectroscopy (UPS) measurement verified that the Na doping effectively shortened the related energy level position of the Fermi level in the bandgap of p-type ZnO [Fig. 5(c)], which enabled hydrogen production via photocatalytic water reduction. The yield of hydrogen reached 1380 µmol·g⁻¹·h⁻¹, which was 51 times compared with undoped ZnO hollow spheres; furthermore, the quantum efficiency of Na:ZnO could reach 13% [Figs. 5(d) and 5(e)]. Such high production of hydrogen was attributed to more catalytic active sites and shorter carrier diffusion lengths obtained from the special morphology of ZnO and effective Na doping. However, the stability of Na:ZnO decreased sharply within 5 h, while the performance of pure ZnO could remain unchanged for 8 h. From the detailed experiment, the dissolution of Na ions from Na:ZnO into the solution assumed the main responsibility of the rapid decay during the durable test. A p-type N:ZnO was investigated by Bhirud et al. for converting hazardous H₂S into hydrogen energy under solar irradiation.⁸⁹ After N doping, the light adsorption redshifted obviously and the bandgap narrowed to a minimum of 2.3 eV, which would greatly facilitate the photocatalytic reaction due to the possibility of absorbing more phonons [Fig. 5(f)]. As expected, the p-type N:ZnO showed a higher hydrogen evolution rate (3957 µmol·h⁻¹) and better stability in comparison with the undoped ZnO [Fig. 5(g)]. Their works further identify that control of p-type doping of ZnO could enhance photocatalytic activity through band engineering.

Aiming at the enhancement of photocatalytic performance of ZnO, choosing Sb as the dopant could be a promising way since the Sb doping can produce the shallow acceptor complex energy level in the ZnO bandgap. The idea was confirmed by Nasser et al. who fabricated p-type Sb:ZnO nanocrystals through an economical sol-gel method for the photocatalytic reaction.⁹⁰ The introduction of Sb created a new acceptor complex level above the VB of ZnO, which lowered the bandgap from 3.35 to 3.15 eV [Fig. 6(a)]. The p-type Sb:ZnO exhibited excellent degradation of RhB with the degrading efficiency of 98% in 100 min under solar irradiation, while the degradation rate was only 68% on undoped ZnO under the same condition [Fig. 6(b)]. This superior photocatalytic activity was contributed to the appearance of the defect complex (Sb_Zn–2V_Zn), which expanded light absorption and promoted charge separation. In our recent work, highly efficient and stable p-type ZnO NWs were successfully obtained by employing Sb as the dopant using the atomic layer deposition (ALD) and hydrothermal preparation method.⁹¹ It is worth noting that the energy bandgap of Sb-doped ZnO was reduced by 0.15 eV, relative to the undoped ZnO. The synchrotron-based x-ray absorption near-edge structure (XANES) spectra and theoretical simulations revealed that the Sb_Zn–2V_Zn complex appeared after doping Sb into ZnO NWs. The PL diagram showed that the photoinduced electrons preferred direct excitation from the Sb_Zn–2V_Zn complex energy level to the CB rather than from the VB to the CB, thus leading to the decreased bandgap in the Sb-doped ZnO NWs [Fig. 6(c)]. In terms of PEC performance of water splitting, the photocurrent density of Sb-doped p-type ZnO was up to five times higher than that of undoped ZnO, indicating effective charge separation/transfer and better photocatalytic activity after Sb doping [Fig. 6(d)]. Moreover, the Sb:ZnO exhibited remarkable stability with no significant photocurrent decrease in 10 h under continuous illumination, and the p-type characteristic had no change [Fig. 6(e)]. In addition, the unique piezotronic effect of p-type ZnO was found in our work, which could build an internal electric field under mechanical energy to properly regulate carrier transport and thus improve the efficiency of photocatalytic water splitting. Compared with p-type ZnO without strain, applying different strains could achieve effective modulation of photocatalytic hydrogen generation activity, where the superior H₂ production rate on p-type ZnO under 0.6% tensile strain could reach 559.5 µmol·cm⁻² after 1 h with the Faradaic efficiency (FE) of 86.3% [Fig. 6(f)]. Oppositely, a significant decrease in hydrogen production (25.4%) was observed by applying compressive strains on p-type ZnO. Therefore, the photocatalytic performance of p-type ZnO could be effectively regulated by the piezotronic effect, which provides a new strategy for future research. Compared with single n-type or p-type ZnO photocatalysts, Wang et al. constructed a p–n ZnO homogeneous junction for enhancing photocatalytic activity.⁹² In their work, the p–n ZnO homojunction was fabricated by decorating p-type ZnO on n-type ZnO nanostructures using the in situ secondary-growth method [Fig. 6(h)]. Through TEM images, homogeneous junctions could be directly observed in the overlapping parts of p-type and n-type ZnO, where the lattice fringes were distorted [Fig. 6(g)]. The special V-shaped Mott–Schottky plots further confirmed the existence of the p–n ZnO homojunction. The PEC results showed that the p–n ZnO homojunction possessed superior photocatalytic performance for water reduction with the photocurrent density 3.3 times higher than that of p-type ZnO [Fig. 6(i)]. The enhanced mechanism was elaborated in Fig. 6(j) because of the huge difference in the Fermi level between p- and n-type ZnO; a strong internal electric field existed on the interface that drove the photoexcited electrons/holes to separate efficiently, thus achieving better photocatalytic performance.

In summary, p-type ZnO materials as photocatalysts have been increasingly drawing concern in terms of applications in photocatalytic water splitting. However, until now, it is still a great challenge to get stable p-type ZnO, although many strategies have been proposed and implemented. Among them, element doping is considered one of the most commonly used and effective methods. In this Perspective, we present an overview of the various elements as dopants incorporated into ZnO. Systematic experiment and theoretical results showed that reasonable doping can successfully obtain high-quality p-type ZnO. Besides, doping can introduce new band energy levels in ZnO, which extended the light adsorption from the UV region to visible-light wavelengths, thus enhancing the performance of photocatalytic water splitting. Meanwhile, the development of p-type ZnO boosts the fabrication of ZnO-based homojunction and heterojunction structures; the built-in electric field could accelerate charge separation and improve photocatalytic water splitting efficiency.

However, there are still some issues that need to be addressed in the future research of p-type ZnO. At first, not all band engineering via doping can generate effective new energy levels; instead, they may become new charge recombination centers to reduce photocatalytic performance. Second, the stability of as-prepared p-type ZnO at the present time is very poor compared to n-type ZnO and the mechanism is unclear. In general, p-type ZnO only keeps stability within 1 year under ordinary storage conditions; the longest time is about 2 years. The reason for unstability mainly stems from p-type ZnO being easily converted to n-type conductivity. As a result, p-type ZnO is not suitable for long-term photocatalytic applications at present; it is also difficult to construct p-type ZnO-based heterojunctions, which are reported to perform better in photocatalytic reactions. Moreover, it is worth noting the introduction of the unique piezotronic effect of ZnO could improve photocatalytic performance, but its microscopic reaction mechanism still needs to be specifically elucidated and optimized. Last but not the least, most of the reported studies of p-type ZnO focus their attention on the photocatalytic degradation reactions; it is quite desired for the investigations devoted to the water splitting reactions.

In conclusion, research of p-type ZnO puts forward an effective way to develop the high-efficiency photocatalytic water splitting system and offers fundamental guidance for designation of corresponding photocatalysts, but obtaining p-type ZnO with high photocatalytic activity and stability is still a big challenge and critical direction in the future research.

This work was supported by the National Natural Science Foundation of China (Grant No. 52162025) and the Hainan Provincial Natural Science Foundation of China (Grant No. 521CXTD439).

The authors have no conflicts to disclose.

The data that support the findings of this study are available within the article.