Recent advances in interface engineering strategy for highly-efficient electrocatalytic water splitting

Abstract

The hydrogen energy generated by the electrocatalytic water splitting reaction has been established as a renewable and clean energy carrier with ultra-high energy density, which can well make up for shortcomings of conventional renewable energy sources, such as geographical limitations, climatic dependence, and energy wastage. Notably, the introduction of electrocatalysts can enhance the efficiency of the water splitting process to generate hydrogen. Particularly, the heterostructure electrocatalysts constructed by coupling multiple components (or phases) have emerged as the most promising option for water splitting due to the well-known electronic and synergistic effects. The existing reviews on interface engineering for electrocatalyst design mostly focus on the relationship between the heterostructures and specific electrocatalytic reactions. However, a comprehensive overview of the integration of model building, directional synthesis, and electrocatalytic mechanism has been rarely reported. To this end, in this review, the development of heterostructure catalysts is systematically introduced from the perspective of interface classification, interface growth and synthesis, and regulation of electrocatalytic performance based on the interfacial microenvironment (bonding, electronic configuration, lattice strain, etc.), thereby offering useful insights on the design and construction of interfacial models. Besides, combined with the current development and applications of interface engineering strategies, the challenges of future heterostructure catalysts are discussed and relevant solutions are proposed. Overall, this review can serve as a useful theoretical reference for the integration of interfacial model building, directional synthesis, and electrocatalytic mechanism, which can further promote the development of hydrogen production technologies with low energy consumption and high yield.

INTRODUCTION

The “Renewables 2021 Global Status Report” released by the renewable energy policy network for the 21st century (REN 21) states that the installed capacity of renewable power generation will continue to increase and may set new records, indicating that the competition between renewable energy sources and fossil fuels is continuously increasing.¹ However, the drawbacks of renewable energy sources, such as high energy consumption during transmission, serious power abandonment, and low utilization efficiency, limit their further applications.² Surprisingly, the energy carriers driven by low-voltage electricity effectively address this issue. This energy carriers are primarily divided into electrochemical energy storage (batteries, capacitors, etc.) and chemical energy storage (hydrogen evolution reaction) devices.³ In particular, the hydrogen energy with the merits of zero carbon emissions and ultra-high energy density (645 000 000 MJ kg⁻¹), which can be obtained from renewable energy sources through overall water splitting driven by low-voltage electricity (see Figure 1), can achieve multi-energy complementation with renewable energy and promote the global energy transformation toward green, clean, and low-carbon energy systems.^4,5 The common industrial technologies for producing hydrogen by overall water splitting include alkaline water electrolysis, proton exchange membrane (PEM) water electrolysis, and solid oxide electrolytic cells (SOECs). Among them, the alkaline water electrolysis technology has become the primary means of industrial hydrogen production due to its mature development and broad application range.

FIGURE 1. The actual situation of renewable energy power generation and its drawbacks in 2020, and put forward the way of energy storage and the significance of realizing multi-energy complementation in response to this problem.

Specifically, the overall water splitting includes hydrogen evolution reaction (HER) and oxygen evolution reaction (OER).^6–8 The electron-rich cathode easily adsorbs H⁺ (or H₃O⁺) to generate hydrogen, while the electron-deficient anode easily adsorbs electron-carrying OH⁻ to release oxygen.^9,10 Notably, the Gibbs free energy change of the water splitting reaction is 237.2 kJ mol⁻¹ in an ideal experimental environment, and the required applied theoretical voltage is 1.23 V versus RHE. However, the concentration polarization, electrode polarization, and slow mass transfer in practical operation greatly hinder the reaction kinetics of water splitting and raise the reaction energy barrier, leading to a higher applied voltage requirement for water splitting. Moreover, the hydrogen production cost in the water electrolysis industry primarily stems from the consumption of excess electricity. To overcome the above issues, electrocatalysts have been developed to reduce the energy barrier and overpotential of electrocatalytic reactions by optimizing the hydrophilic capacity of electrodes and the adsorption or desorption energies of reaction intermediates, thereby enhancing the hydrogen production efficiency and reducing its cost.¹¹

Based on the latest achievements in advanced catalysts, an ideal electrocatalyst must possess the following characteristics to achieve exceptional performance¹²: (1) ultrahigh intrinsic activity that relies on the chemical composition and electronic configuration; (2) abundant active sites that depend on the exposed electrochemically active area and the varieties of active species; (3) high-efficiency electron transport that relies on the intrinsic conductivity of the material; (4) high-efficiency mass transport ability, which is closely linked to the morphology and hydrophilicity; (5) excellent electrochemical and structural durability; (6) low cost. To achieve the above goals, morphology regulation, elemental doping, and interface engineering are applied as the fundamental optimization strategies for electrocatalysts (see Figure 2).¹³ Compared with the other two optimization strategies, the interface engineering strategy has rich realization means, wide action range, complex action mechanism, and outstanding electrocatalytic effect, making it a promising technique to achieve high-performance electrocatalysts.

FIGURE 2. The characteristics of the ideal catalyst and its optimization strategy.

It is worth noting that the interface design is different for various electrocatalytic reactions.^14,15 For HER, the heterointerface constructed by combining active components with conductive materials can simultaneously improve the electrocatalytic performance and conductivity.^16,17 Additionally, a heterogeneous interface of “highly active component/highly stable component” can be constructed through phase transition to drive the electron migration at the interface, thereby promoting the stability of the active component. For OER, the outermost layer of the catalyst inevitably undergoes in situ oxidation during oxygen generation to form an oxide or (oxy) hydroxide shell with excellent OER performance. Such heterointerface derived from spontaneous surface reconstruction has become the main research focus to achieve enhanced OER activity.^17,18 To realize high-efficiency overall water splitting, electrocatalysts with bifunctional activity must be used, and the heterostructure catalysts assembled from HER-active species and OER-active species precisely meet this requirement.¹⁹ Accordingly, several reviews have systematically elaborated the development of heterostructure catalysts.^20–23 However, a comprehensive review on the integration of model building, directional design, and electrocatalytic mechanism for the construction of electrocatalysts based on the interface engineering strategy has not been reported yet.

To this end, this review focuses on the directional design of electrocatalyst interface for overall water splitting, and the interface engineering strategies for preparing highly active electrocatalysts are examined from the perspectives of interface composition (classification), construction strategies, and electrocatalytic action mechanisms. Finally, combined with the current development stage and applications of interface engineering strategies, the challenges of future heterostructure catalysts and prospective directions are discussed. Overall, this work can serve as a theoretical guide for the directional design of electrocatalyst interface and further promote the development of hydrogen production technologies with low energy consumption and high yield.

INTERFACE CLASSIFICATION

Over the recent years, electrocatalysts have undergone a radical transformation from a single-component homogeneous structure to a multi-component hybrid structure, which has significantly enhanced the electrocatalytic performance. Remarkably, the interface engineering strategy enables the elaborate design and optimization of multi-component hybrid structures by constructing unique interfaces. Therefore, a clear understanding of the interfacial structure is necessary for the directional design and development of high-performance electrocatalyst materials in the future. Notably, the interfaces of hybrid electrocatalysts can be classified according to the morphological configuration and crystal structure (Figure 3). Based on the morphological structures of hybrid materials, the interfaces can be divided into the following representative configuration models: supported structure, core–shell structure, and heterogeneous structure. Besides, considering the diversity of crystal structures and the complex lattice binding condition of the as-formed interfaces, the interfaces can be divided into phase interfaces or grain boundaries of the same crystal, heterointerfaces of different crystals, and crystalline/amorphous heterointerfaces.^24–26

FIGURE 3. Schematic depiction of classification of the interfaces in electrocatalysts in terms of morphology and the crystal structure. Reproduced with permission.24 Copyright 2022, Elsevier. Reproduced with permission.25 Copyright 2021, Wiley. Reproduced with permission.26 Copyright 2021, American Chemical Society.

Classification based on the morphological structure Supported structure

Supported structure refers to the loading of an active species with a smaller size on the surface of a larger-sized species, such as zero-dimensional (0D) nanodots on one-dimensional (1D) nanowires,²⁷ 0D nanoparticles on two-dimensional (2D) nanosheets,²⁸ 0D nanoparticles on three-dimensional (3D) blocks,^29,30 2D nanosheets on 3D current collectors,³¹ 2D nanosheets on 1D nanowires,³² and so on. Moreover, the heterogeneous interface of the support structure is primarily determined by the number of components with smaller dimensions. Qin's group developed a hollow PtCu alloy nanosphere distributed on a WO₃ nanosheet array on a Cu foam substrate by electrochemical deposition (Figure 4A).³³ The hollow PtCu alloy spheres effectively improved the utilization and reactivity of Pt atoms. It is worth noting that the HER mainly occurred at the interface of PtCu and WO₃, which was due to the transfer of hydrogen ions from the Pt surface with a strong H* adsorption capability to the surface of the conductive WO₃, triggering the Volmer–Heyrovsky reaction mechanism.

FIGURE 4. (A) Schematic depiction of PtCu/WO3@CF. Reproduced with permission.33 Copyright 2022, Wiley. (B,C) TEM image and schematic illustration of the RuO2-WC NPs. Reproduced with permission.34 Copyright 2022, Wiley. (D,E) Schematic illustration and the high-angle annular dark-field (HAADF)-STEM image of Ni3Fe-CO32− LDH-Pt nanosheet. Reproduced with permission.35 Copyright 2021, Royal Society of Chemistry.

Owing to their good thermal stability, corrosion resistance, and Pt-like electronic configuration, metal carbides are often used as the supporting material to facilitate a strong catalyst-substrate interaction in the HER process. Xu et al. constructed a 0/2D composite structure with WC as the substrate combined with RuO₂ nanocomposite particles (RuO₂-WC NPs).³⁴ It was found that the WC substrate had a good lattice matching with RuO₂, stimulating the catalyst-substrate interaction that was beneficial to the catalytic reaction (Figure 4B,C).

Apart from materials such as particles and nanospheres, single atoms with high activity and ultra-small size are also popular components for constructing hybrid materials with supporting structures. In particular, single-atom catalysts (SACs) exhibit immense potential in heterogeneous catalysis with the merits of ultra-high atomic utilization and unique electronic structures. However, the general synthetic methods of such SACs are quite challenging. Wang's group proposed a general strategy to nest Pt atoms on Ni₃Fe layered double hydroxide support (Ni₃Fe-CO₃²⁻ LDH-Pt) using a space-confined electroreduction technique (Figure 4D,E).³⁵ For HER, the Pt single atoms and Ni₃Fe-LDH acted as the catalytic site and auxiliary catalytic support, respectively. Among them, Ni₃Fe-LDH accelerated the dissociation of water molecules in an alkaline electrolyte, and the electron transfer ability was greatly improved by the interlayer Pt single atoms, resulting in a remarkable catalyst-driven HER performance by a mixed (Heyrovsky–Volmer and Tafel–Volmer) mechanism. On the other hand, the OER active site was not Ni₃Fe-LDH but the electro-oxidation-induced Ni_2+δFe_3+ζO_xH_y species. Notably, the interlayer Pt single atoms induced phase transition from Ni₃Fe-LDH to Ni_2+δFe_3+ζO_xH_y with superior OER activity.

Core–shell structure

Similar to the supported structure, core–shell structures are composite materials in which one component is grown on the surface of another component. However, the difference is that the outermost component completely wraps the inner component. Zhang's team synthesized the Au@Pd nanorods electrocatalyst with a unique core–shell structure via a wet chemical method, which exhibited excellent electrocatalytic activity.³⁶ Simultaneously, his team prepared the heterophase Pd₄₅@Ir₅₅ nano-dendrites with core–shell structure by a similar method, showing remarkable HER activity under acidic conditions.³⁷ Additionally, heteroatom-doped carbon-coated materials have been extensively studied.^38–40 Liu et al. synthesized N, S co-doped carbon-coated Fe-Co₉S₈ (Fe-Co₉S₈@SNC) catalysts via a rapid annealing method.⁴¹ Notably, the N, S co-doped carbon layer not only facilitated additional OER active sites, but also acted as an “armor” to protect the corrosion of Fe-Co₉S₈ structure from the electrolytes. Furthermore, Chen's team employed density functional theory (DFT) calculations to deeply explore the effect of distinct kinds of N dopants on the HER kinetics of MoP@NCHSs.⁴² It was found that the strong interaction between pyridine nitrogen-doped carbon and MoP increased the electron density on the carbon shell and weakened the Mo*H bond strength, both of which contributed to the acceleration of HER.

Apart from the above-mentioned powdered catalyst with core–shell structure, binder-free core–shell catalysts have also been extensively examined. Kanatzidis et al. designed interacting MoS₂ and Co₉S₈ nanosheets attached Ni₃S₂ nanorod arrays on nickel foam (NF) to form a hierarchical co-assembly electrocatalyst, which exhibited excellent conductivity and superior OER activity. The excellent OER activity was attributed to the electron transfer from Co₉S₈ to MoS₂ at the interface. Importantly, this was the first report on the electrocatalytic overall water splitting of metal sulfides in the full pH range.⁴³ Furthermore, our group designed and synthesized Ni₂P and Ni₃S₂ hybrid nanorod arrays grown on NF substrates and covered with defect-rich 1T-MoS₂ nanosheets by P and W co-doping activation strategy and interface engineering strategy.⁴⁴ Simultaneously, they were coupled to form a two-electrode system for the overall water splitting, and only an ultra-low voltage of 1.53 V was required to achieve a current density of 100 mA cm⁻².

Heterogeneous structure

The heterogeneous structure is a hybrid structure formed by multi-components with comparable size that are tightly connected on a specific face, and this heterostructure configuration is widely used in hybrid materials. Mai's group used a simple solvothermal method to synthesize a FePc/CoPc nanorod electrocatalyst with a heterogeneous phase. This unique heterostructure induced the extension of FeN bonds and increased the electron density of the surrounding Fe site, leading to an excellent electrocatalytic activity and stability.⁴⁵ Interestingly, Yu's group assembled different metal–organic framework (MOF) building blocks to obtain rare ternary MOF-on-MOF hybrid heterostructures by adopting a multiple selective assembly strategy and using cake-like MIL-125 as the MOF matrix. This strategy laid the foundation for the construction of unique heterostructure materials.⁴⁶

Classification based on the crystalline structure Phase or crystalline interfaces

For single-component catalysts with a specific phase, the chemical composition and number of interfaces between different crystal planes have a significant effect on the electrocatalytic activity.⁴⁷ Qin et al. synthesized a GB-(Fe_xCo_1−x)₂B electrocatalyst with controllable grain boundary density by a facile ball milling method.²⁴ The DFT calculation results illustrated that the introduction of Fe atoms into the Co₂B catalyst resulted in a higher grain boundary density and tuned the electronic structure (Figure 5A). Benefitting from the tuning of grain boundary density and electronic structure, the optimal GB-(Fe_xCo_1−x)₂B required an overpotential of only 221 mV to deliver an OER current density of 10 mA cm⁻² while exhibiting electrochemical stability for over 100 h during the alkaline OER process. Therefore, the regulation of the number of grain boundaries is a vital step in the interface engineering strategy.

FIGURE 5. (A) Schematic depiction of the experimental procedure for the GB-(FexCo1−x)2B and their related TEM images. Reproduced with permission.24 Copyright 2022, Elsevier. (B) Diagrammatic figure, growth process, and growth mechanism illustration of the heterophase MoTe2 with a phase boundary. Reproduced with permission.48 Copyright 2022, Wiley. (C) HAADF-STEM images and their relevant FFT patterns of fcc-2H-fcc heterophase Pd66@Ir34 and Pd68@Ir22Co10 nanoparticles. Reproduced with permission.37 Copyright 2022, Wiley.

In addition, the phase-engineered construction of transition metal dichalcogenides (TMDs) is considered to be a promising strategy for enhancing the catalyst performance.^49,50 Cho et al. successfully adjusted the area density of the 2H/1T′ heterophase region of MoTe₂ thin films by chemical vapor deposition (CVD; Figure 5B).⁴⁸ The experimental results proved that the 2H/1T′-MoTe₂ boundary exhibited band bending and local charge accumulation. Consequently, the Te atomic sites at the heterophase boundary exhibited an extremely large turnover frequency (TOF; 317 s⁻¹), which was 103 times that of pure 1T′-MoTe₂. However, the controllable fabrication of noble metal heterophase interfaces through phase engineering remains challenging.^36,51 Zhang's group has extensively explored the directional synthesis of fcc-2H-fcc heterophase noble metal-based materials (Figure 5C).³⁷ They found that the Pd₄₅@Ir₅₅ nanodendrites with fcc-2H-fcc heterophase exhibited excellent catalytic performance for acidic HER, which highlights the vital role of crystalline phase in the electrocatalysis.

Different crystalline heterointerfaces Precious metal catalysts

Owing to the incompletely filled d-electron orbital and a modest ability to adsorb reactants and desorb products, noble metal electrocatalysts exhibit ultra-high intrinsic electrocatalytic activity. Due to the satisfactory HER and OER activities, Pt/C and IrO₂ (or RuO₂) have been established as the benchmark electrocatalysts for HER and OER, respectively, providing a reference for boosting the performance of electrocatalysts.^36,37 For example, well-designed Pt/PtTe_x heterojunctions were obtained via the electrochemical activation of PtTe₂ nanorods to modulate the reaction intermediate adsorption for HER. As anticipated, the Pt/PtTe_x catalyst presented 19 and 5 folds higher mass activity than that of Pt/C benchmark material in acidic and alkaline electrolytes at a voltage of 50 mV, respectively. Further, the DFT results confirmed that the Gibbs free energy of H* adsorption (ΔG_H*) value of the Pt/PtTe_x heterointerface could be adjusted to approximately 0 eV. Meanwhile, the adsorption capacity of OH* at the well-designed heterointerface was significantly enhanced.⁵² Furthermore, a novel antioxidative electrocatalyst with the Pd₄S/Pd₃P_0.95 heterostructure was rapidly fabricated by interface engineering modulation, which greatly exposed the active sites while retaining the remarkable conductivity of each metal-rich component. The DFT calculations revealed that the optimal Pd₄S/Pd₃P_0.95 heterostructure significantly reduced the dissociation energy barrier of water molecules and accelerated the HER reaction kinetics.⁵³

Compared with the other noble metals, Ru-based catalyst is widely regarded as a promising electrocatalyst for total water splitting due to its excellent performance, superior thermal stability, and proper binding energy for adsorbing H* and oxygen-containing intermediates.^54,55 Mu's group firstly prepared Ru/RuS₂ heterostructures with layered structures by simultaneous reduction and sulfurization in a eutectic-salt system. The relevant DFT calculations revealed that the excellent conductivity of Ru/RuS₂ heterostructure originated from the larger state density at the Fermi level, and the rearranged charges around the Ru/RuS₂ heterointerface regulated the adsorption behavior of the intermediate, eventually leading to an ultrahigh theoretical activity toward acidic water electrolysis.⁵⁶ Based on the above work, the S source was replaced with the P source to obtain the Ru–Ru₂P heterostructure catalyst using the same design idea. The DFT calculation results revealed that the Ru–Ru₂P heterointerface induced electron transfer, which was consistent with the above Ru/RuS₂ sample.⁵⁷

Considering the adverse characteristics of noble metal materials, such as high cost, easy agglomeration, and easy dissolution, it is of great significance to combine them with transition metals to prepare noble metal/non-precious metal composite catalysts.³⁵ It is worth noting that the introduction of non-precious metal materials not only reduces the production cost of catalyst materials, but also effectively enhances the utilization efficiency of noble metal atoms and the stability of materials.³⁰ It is well known that the Volmer step is the rate-determining step of the alkaline HER process. Therefore, expediting the Volmer step by optimizing the adsorption of OH⁻ is crucial for the improvement of electrocatalytic efficiency. Markovic found that the controllable arrangement of nanoscale Ni(OH)₂ clusters on the surface of the Pt electrode could increase the HER activity of the original Pt electrode by eight times. The edges of the Ni(OH)₂ clusters promoted water dissociation to form H intermediates, which were then adsorbed on the Pt surface at the Ni(OH)₂/Pt interface and recombined to form hydrogen molecules. This work had a profound impact on the use of interface engineering strategies to promote the dissociation of small molecules.⁵⁸ Furthermore, the Ru/HfO₂ heterostructure was constructed by anchoring Ru nanoparticles on the oxygen-vacancy-rich HfO₂ substrate, where the Ru nanoparticles and HfO₂ substrate interacted through the interfacial RuOHf bonds and further promoted the water breakdown.⁵⁹ Besides, the formation of composite containing noble metal-based compounds and carbon materials is also a remarkable interface engineering strategy. This structure not only increases the conductivity of the material, but also protects the active phase from rapid oxidative deactivation and prevents metal ions from escaping during the electrocatalytic process, thereby improving the chemical and mechanical stability.

Non-precious metal catalyst

Over the recent years, the non-precious metal materials (e.g., phosphides, phosphates, sulfides, selenides, etc.) have received considerable research interest due to their low cost, simple preparation, and easy control.^60–62 However, the limited intrinsic activity restricts their wide application. The construction of non-precious metal heterostructures can make up for this deficiency to a certain extent.⁶³ Luo et al. designed a specific Co₄N/Co₂P heterostructure catalyst through a one-pot method and proposed that the dynamic flow of electrons from Co₂P to Co₄N at the interface formed a unique electronic communicating vessel (ECV) to drive superior alkaline HER (η₁₀ = 40 mV).⁶⁴ The experimental results and DFT calculations confirmed that the particular EVC configuration optimized the electronic structure of Co₄N/Co₂P and the adsorption strength of *H species. Ren's group constructed in-plane NiP₂/Ni₅P₄ heterostructure by controlling the P content, whose HER activity (30 mV at 10 mA cm⁻²) was comparable to that of the Pt/C.⁶⁵ Notably, it required an overpotential of only 247 mV to reach a high current density of 2 A cm⁻². The DFT calculation results demonstrated that the modest overlap of density states between the P 2p and H 1s orbitals in the NiP₂/Ni₅P₄ heterostructure optimized the adsorption of P sites for H*, which offered a new avenue for the design of non-noble metal heterostructure catalysts.

Notably, the Mott–Schottky (M–S) heterostructures composed of non-precious metals and semiconductors, such as Co/CoMoN, Co/Co₂C, and Ni–MoN, and so on, have emerged as a representative heterostructure in the field of electrocatalysis.^66–70 Ren et al. prepared Ni–MoN HER catalysts with ultralow overpotential (136 mV) at 1 A cm⁻².⁶⁸ Further, the related DFT results revealed that the integration of MoN and Ni promoted the charge redistribution at the Ni/MoN interface and caused Ni–MoN to be in a metallic state, thereby enhancing the water adsorption and dissociation ability of Mo sites. Wang et al. constructed the CoMn/CoMn₂O₄ heterostructure with the unique built-in electric field, which induced spontaneous electron transfer and caused local charge polarization at the CoMn/CoMn₂O₄ interface.⁶⁹ The DFT results further confirmed that H₂O was more inclined to be adsorbed on the CoMn₂O₄ side of the M–S heterostructure rather than on the pure CoMn₂O₄, and the adsorption capacity of H₂O was greatly optimized in this Schottky barrier region.

Besides, the construction of composites containing non-noble metal-based compounds and carbon materials (e.g., graphene, graphdiyne, carbon nanotubes [CNTs], the carbon layer or substrate originating from the MOFs) is also a remarkable interface engineering strategy. This structure not only increases the conductivity of the material, but also protects the active phase from rapid oxidative deactivation and prevents metal ions from escaping during the electrocatalytic process, thereby improving the chemical and mechanical stability.^71,72 Lee et al. verified that the specific CNT architecture and the coupling of inner CNTs and outer NiP₂/NiP nanosheets jointly facilitated the exceptional water splitting performance of CNTs@NiP₂/NbP.⁷³

Crystalline/amorphous heterointerfaces

Amorphous materials with disordered arrangements and unsaturated coordination structures have received extensive attention in the electrocatalysis field due to their abundant active sites and superior intrinsic activity than the crystalline materials.^74,75 Although amorphized catalysts can enrich highly active sites, the low electrical conductivity and structural stability originating from their disordered atomic arrangement greatly limit the electrocatalytic activity.^71,76

Therefore, the fabrication of amorphous/crystalline (a/c) heterostructures is crucial to simultaneously elevate the electrocatalytic activity, conductivity, and durability. However, there are limited reports on the amorphous/crystalline heterostructure catalysts, and their directional design and construction are still challenging.^77,78 Fei and co-workers prepared a core–shell electrocatalyst consisting of an amorphous/crystalline heterophase NiFe alloy encapsulated by ultrathin graphene layers (a/c-NiFe-G) with abundant amorphous/crystalline interfaces and enriched active sites. X-ray photoelectron spectroscopy (XPS) and extended x-ray absorption fine structure (EXAFS) were used to reveal higher metal oxidation state and unsaturated coordination configuration in a/c-NiFe-G. The above features enhanced its catalytic kinetics and tuned the electronic structure, resulting in ultrahigh OER active catalytic sites.³ Yang et al. synthesized ternary metal oxides (CoFeCeO_x) with amorphous/crystalline heterostructures through the lattice matching effect (Figure 6A–C).⁷⁹ The CeO₂ acted as a “buffer” to reduce the lattice mismatch and dislocation between Fe₂O₃ and CoO. Consequently, the CeO₂–CoO heterostructure with small lattice mismatch was conducive to reduce the energy barrier for charge transfer during the catalytic process (Figure 6D). Liu et al. reported an amorphous-crystalline CrOx–Ni₃N heterostructure electrocatalyst with abundant oxygen defects and heterointerfaces (Figure 6E).⁸⁰ The theoretical and experimental results suggested that the electrons at the interface were significantly coupled due to the synergistic effect of the crystalline and amorphous structures, reducing the HER kinetic barrier, thereby enhancing the HER activity (Figure 6F,G). Additionally, abundant heterojunction interfaces and oxygen-deficient sites exposed numerous active centers, which optimized its electrocatalytic performance significantly.

FIGURE 6. (A) TEM, (B) HRTEM, and (C) FFT images of CoFeCe-2. (D) Diagrammatic model of CeO2/CoO heterointerface. Reproduced with permission.79 Copyright 2021, Elsevier. (E) Growth process illustration of the CrOx-Ni3N electrocatalyst. (F,G) Overall water splitting performances and their Faradaic efficiency. Reproduced with permission.80 Copyright 2022, Wiley.

Overall, the alkaline HER and OER performances of the aforementioned heterogeneous electrocatalysts are highlighted in Table 1. Compared with the single-component (or single-phase) catalysts, the heterostructure catalysts constructed by interface engineering exhibit significantly improved activities for both HER and OER, confirming the significant role of interface engineering in surmounting the inherent activity limitations. Simultaneously, the comparison between the properties of different types of heterointerfaces reveals that the types and numbers of phase (or grain) boundaries are the vital controlling factors for interface engineering. Notably, crystalline/amorphous heterostructure catalysts generally exhibit superior activity than the pure crystalline heterostructure catalysts.

TABLE 1 Comparison table of the electrocatalytic performance of the heterostructural catalysts for HER and OER under an alkaline environment (1.0 M KOH)

[Table omitted. See PDF]

ADVANCED INTERFACE ENGINEERING STRATEGIES Construction strategy

The growth of heterostructures can mainly occur through the following ways: (1) extension growth of a specific component; (2) ion exchange of specific templates; and (3) in situ separation of the active phases.

Extension growth of a specific component

Epitaxially grown heterostructures refer to the epitaxial growth of a second component on the surface of the original material. In general, fine control of the epitaxially grown second component along a particular crystal plane can be achieved by tuning the size, morphology, structure, and chemical state of the existing material (Figure 7A).^81,84 Zhang's group obtained a series of heterogeneous Pd@Ir and Pd@IrCo electrocatalysts with diverse shapes by the phase-selective epitaxial growth of heterophase Ir-based material on the outer layer of Pd seeds.³⁷ Among them, the Pd₄₅@Ir₅₅ nanodendrites displayed excellent acidic HER performance, reaching a current density of 10 mA cm⁻² with an overpotential of only 11.0 mV, outperforming the ordinary Pd₄₇@Ir₅₃ counterpart, commercial Pt/C, and Ir/C samples.

FIGURE 7. (A) The phase-selective epitaxial growth route of Pd-based catalysts. Reproduced with permission.81 Copyright 2020, American Chemical Society. (B–D) Optical graphics of the pristine MoS2-OH, the lateral and vertical heterostructures of WS2/MoS2. Reproduced with permission.82 Copyright 2020, American Chemical Society. (E–G) Synthesis illustration, TEM, and HRTEM images of the MoSe2–NiSe hybrids. Reproduced with permission.83 Copyright 2016, American Chemical Society.

Notably, Zhang and coworkers firstly proposed to introduce a hydroxide-assisted process to control the selective growth of lateral and vertical MoS₂/WS₂ heterostructures, confirming the guiding role of hydroxides in the controllable growth of MoS₂/WS₂ heterostructures (Figure 7B–D).⁸² This work provided useful insight on the selective growth of MoS₂/WS₂ with lateral and vertical heterostructures and the precise regulation of the sulfide heterostructures. Moreover, Zhou and colleagues utilized hot injection to achieve the epitaxial growth of NiSe crystallites on the surface of MoSe₂ nanosheets and obtained a MoSe₂–NiSe heterostructure (Figure 7E–G), which laid the experimental foundation for the subsequent growth of other materials on MoSe₂ nanosheets.⁸³

Ion exchange of specific templates

Ion exchange method can be primarily divided into anion exchange method and cation exchange method. Cation exchange refers to retaining the original anion sublattice and replacing the cations in the existing materials to form new components (Figure 8A), which can effectively control the material composition to prepare a variety of heterostructures.⁸⁵ Yan's team obtained cobalt-vanadium-iron (oxy) hydroxide (CoV-Fe_0.28) nanosheets with amorphous/crystalline heterostructures by replacing some Co atoms in the CoV nanosheets with Fe atoms by the cation exchange method.⁸⁹ The CoV-Fe_0.28 catalyst with amorphous/crystalline heterostructure effectively increased the exposure of active sites. Furthermore, the interaction between Co, V, and Fe cations in CoV-Fe_0.28 subtly modulated the local coordination environment, accelerating the reaction kinetics of OER. Besides, Zhang's group sequentially performed cation exchange treatment and alloying treatment on the Cu_3−XP nanoplates precursor, replaced all Cu²⁺ in Cu_3−XP with Pd²⁺ and further alloyed the nanoplates with Ni, thereby forming a crystalline-Pd-Ni-P@amorphous-Pd-Ni-P nanoplate catalyst with a core–shell structure (Figure 8B,C).⁸⁶

FIGURE 8. (A) Schematic formation process of the Ag/Ag2S catalyst and the SEM pictures of Ag/Ag2S gained via adding various Ag+ concentrations solution. Reproduced with permission.85 Copyright 2021, Wiley. (B) Synthesis illustration of the c-Pd-P@a-Pd-P nanoplate. (C) TEM and related mapping images of the c-Pd7P3@a-Pd-P nanoplate. Reproduced with permission.86 Copyright 2020, Wiley. (D) Ion exchange route from ZnO to ZnS. (E,F) FFT and TEM images of ZnO/ZnS nanoparticles. Reproduced with permission.87 Copyright 2009, American Chemical Society. (G,H) Synthesis illustration and the FFT image of the Ni(CN)2/NiSe2 catalyst. Reproduced with permission.88 Copyright 2021, Wiley.

Unlike the cation exchange, which can well maintain the original shape of the template, anion exchange is often accompanied by the Kirkendall effect that affects the template morphology, resulting in obvious changes in the template morphology with the formation of polycrystals (Figure 8D–F).⁸⁷ For example, Lou et al. demonstrated that unique Ni(CN)₂/NiSe₂ heterostructures could be created via partial anion exchange using Ni(CN)₂ as the original template (Figure 8G,H).⁸⁸ Importantly, the Ni(CN)₂/NiSe₂ heterostructure was assembled by Ni(CN)₂ single-crystal nanoplates and NiSe₂ nanoparticles aligned along with the crystal. Moreover, the Ni(CN)₂/NiSe₂ heterostructure with the optimal Se content exhibited excellent OER performance.

In situ separation of the active phases

Phase separation is also an effective method to construct heterostructures. The phase separation of precursors can be induced by doping heteroatoms, controlling the crystal phase of precursors, and increasing the difference in the decomposition temperature of different phases, leading to the directional synthesis of heterostructures. Qin et al. confirmed via DFT calculations that Fe doping can induce grain boundaries in the Co₂B catalyst, optimize its electronic structure, and promote the formation of the OER intermediate species (MOO⁻). Notably, this high grain boundary density enriched the active sites for the OER reaction, which in turn enhanced the OER activity of GB-(Fe_0.66Co_0.34)₂B.²⁴ Li and co-workers realized the structural transformation of Mo₂C to MoC by doping Zn atoms and finally obtained a hierarchical Zn-doped MoC/Mo₂C electrocatalyst (Figure 9A–C) with enhanced HER kinetics.⁹⁰ The doping of Zn effectively weakened the adsorption energy of Mo for H* and optimized the electronic structure of Mo-C species. Further, all the sites in the Zn-MoC/Mo₂C catalyst could be activated, thereby significantly enhancing the number of active sites. Zhang's group proposed a new in situ hybrid heterojunction strategy. Specifically, a porous Co-doped Ni/Ni₃N heterostructure was fabricated by the pyrolysis of partially Co-substituted nickel-zeolitic imidazolate framework (Ni-ZIF) nanosheets (Figure 9D–H).⁹¹ The Co dopant not only promoted the formation of the Ni/Ni₃N heterostructure, but also optimized its electronic structure, thereby lowering the thermodynamic energy barrier and accelerating the catalytic kinetics. Wu's team proposed an effectual continuous phase inversion strategy and successfully realized the epitaxial growth of 1D Ni₂P-Ni₁₂P₅ hetero-nanorods (Ni₂P-Ni₁₂P₅@Ni₃S₂/NF) on the surface of conductive Ni₃S₂ films.⁹² Importantly, the introduction of the Ni₃S₂ intermediate phase effectively tuned the phase composition and morphology of Ni₂P-Ni₁₂P₅. Moreover, the as-prepared Ni₂P-Ni₁₂P₅ with abundant Ni-P species optimized the atomic ratio of Ni to P and enhanced the electrical conductivity (Figure 9I–K). Besides, Han et al. utilized the huge difference in the thermal decomposition temperature between iron oxalate and copper oxalate to synthesize mesoporous Fe₂O₃/CuO heterostructures anchored on NF by a facile one-step method (Figure 9L).⁹³ The Fe₂O₃/CuO heterostructure with abundant Fe₂O₃/CuO interfaces and Fe–O–Cu bridges effectively enhanced the oxygen binding energy and facilitated the adsorption/desorption of oxygen intermediates, which resulted in its outstanding performance and long-term durability for HER and OER. Strikingly, this work provided a facile fabrication route for materials with mesoporous heterostructures. Furthermore, Zhai's team employed Cu₂SnS₃ nanosheets as precursors to construct SnO₂@CuS/Cu₂O nucleus-shell nanosheets with a strong coupling interface through in situ rapid phase separation during the electro-reduction process.⁹⁴ Subsequently, the electronic structure of the SnO₂@CuS/Cu₂O nanosheets with heterostructure was thoroughly examined, and the strong electronic interaction between SnO₂ and CuS/Cu₂O at the interface of the heterojunction was confirmed.

FIGURE 9. (A) Formation mechanism diagram of Mo2C and Zn-MoC/Mo2C. (B,C) The related HRTEM images of Zn-MoC/Mo2C. Reproduced with permission.90 Copyright 2022, Elsevier. (D) Synthetic illustration and (E–H) corresponding electron microscope photographs of the Co-Ni/Ni3N architecture. Reproduced with permission.91 Copyright 2021, Wiley. (I) Synthetic procedures illustration of the phase conversion-induced phosphides. (J,K) TEM and HRTEM images of Ni2P-Ni12P5@Ni3S2. Reproduced with permission.92 Copyright 2022, Wiley. (L) Synthetic procedures illustration of the heterostructured Fe2O3/CuO. Reproduced with permission.93 Copyright 2022, Wiley.

Synthetic method

To date, various synthetic approaches have been proposed for heterostructure catalysts, such as hydrothermal reaction, steam transfer method, electrochemical reaction, and so on.⁹⁵ Actually, the specific synthetic methods can affect the size, chemical composition, and morphology of the as-prepared materials and further interfere with their interface distribution, electronic configuration, and energy band structure.⁹⁶ Therefore, it is necessary to select an appropriate synthetic method according to the anticipated structure of the catalyst interface.

Steam transfer method

The steam transfer method has received extensive attention due to its high crystalline quality, high-quality interfaces, and simple operation. This method is mainly characterized by gas–solid reaction, chemical vapor deposition (CVD), and magnetron sputtering reaction.

CVD is a technique in which a solid is grown by the chemical reaction of gaseous precursors, which is directly deposited on the surface of the substrate material to form the final product.^97–99 Remarkably, the growth sequence of heterostructures can be optimized by adjusting parameters such as pressure, time, temperature, and the ratio of mixed gases during the CVD process, eventually realizing the controllable fabrication of high-quality 2D or 3D heterostructure materials.¹⁰⁰ It is noteworthy that the substrate material not only acts as a host for the newly formed products during the growth of heterostructures, but also as a structural and morphological guiding material to influence the morphology and electrocatalytic properties of the final product. Presently, the substrate materials commonly used in the CVD route include metal-based substrates,¹⁰¹ SiO₂/Si substrate materials,¹⁰² and carbon-based substrates.¹⁰³

Unlike the CVD method, the gas–solid reaction mainly includes the deposition of new components on the solid precursor by adjusting the gas atmosphere of the tube furnace.^104,105 In this method, the solid precursor is generally a material rich in metal elements, and the atmosphere of the tube furnace is usually divided into a redox atmosphere and an inert gas.^106–108 When the atmospheric gas has certain oxidizing or reducing properties (such as oxygen, hydrogen, PH₃, H₂S, NH₃), it can undergo redox reactions, such as reduction,¹⁰⁹ sulfurization,¹¹⁰ phosphorization, nitridation,^70,111 and selenylation, with solid materials to form new components. When the atmospheric gas is an inert gas (such as Ar, N₂), in situ thermal decomposition of solid materials, such as carbonization process, can be achieved.^67,112,113 Since the reducing atmospheres are often corrosive and toxic, their storage and utilization have certain leakage risks, so the thermal decomposition of specific materials to obtain their substitute atmospheres is also a popular method. Typically, some powders (such as urea, melamine, red phosphorus, sodium hypophosphite, Se powder, S powder, etc.) can be used as self-sacrificial reactants for inducing the redox reaction.¹¹⁴ This reaction is fast and safe, which can be tuned by adjusting the number of self-sacrificing reactants and calcination temperature. Further, the degree of gas–solid reaction and the density of the heterostructure can also be controlled.

Additionally, the magnetron sputtering method is an attractive physical vapor deposition technique.¹¹⁵ Specifically, the magnetic field drives the ionized Ar atoms to collide with the target material, promoting the escape of reactive molecules/atoms from the target and their deposition on the surrounding substrate material. Qiao et al. sputtered Pt nanoparticles on the outer surface of CoO nanowires by an industrialized magnetron sputtering method to fabricate a Pt/CoO composite electrocatalyst with the atomically matched interfaces. Obviously, the particle size of Pt could be precisely controlled by magnetron sputtering time and current. Notably, the atomic coupling between Pt and CoO could be used to precisely optimize the electronic structure of Pt, making it three times more active than the commercial Pt/C benchmarks.¹¹⁶

Hydrothermal reaction

The hydrothermal method (or solvothermal method) is a bottom-up synthesis method in which the precursors are reacted in a harsh environment of high temperature and high pressure.¹¹⁷ The chemical composition and morphology of the final product can be controlled by adjusting the input ratio of reactants, reaction temperature, time, and other parameters.¹¹⁸ Compared with the steam transfer method, the hydrothermal (solvothermal) method is more suitable for the industrial production of heterostructured catalysts owing to its characteristics of simple operation, mild reaction conditions, and high product yield.¹¹⁹

However, this method has limited control over the nucleation and growth process of crystalline material, and the final product often suffers from the disadvantages of stacking, poor dispersion, and unclear heterogeneous interfaces. Moreover, this method relies on a high-pressure production environment and production equipment, which limits its large-scale application.^120,121 Therefore, the hydrothermal (solvothermal) method still needs to be improved.

Electrochemical reaction

Electrochemical reaction primarily includes electrodeposition method and electrochemical reconstruction. This method has the advantages of low cost, mild reaction conditions, and short reaction time, which makes it the most effective route for preparing high-performance catalysts.^122,123

The electrodeposition method mainly relies on the potential difference between the redox reactions of different substances to drive the deposition of new components. Specifically, the composition, morphology, and thickness of the deposited components can be adjusted by optimizing the parameters such as the type and concentration of the electrolyte, deposition temperature and time, and the applied current or voltage.¹²⁴ Song et al. reported that the properties of many catalyst materials can be optimized during the electrocatalytic process, especially the OER process, and the surface of the material inevitably undergoes phase transition after the electrocatalytic process.¹²⁵ This phase transition on the surface of the material is widely known as the surface reconstruction phenomenon, which mainly originates from the applied potential during the electrocatalysis process that drives the oxidation or reduction reaction on the surface of the material.¹²⁶ At the same time, the parameters of the electrocatalytic reaction as well as the structural composition of the catalyst affect the surface reconstruction phenomenon. Furthermore, Huang's team constructed IrOx and Ir shells on the surface of IrTe₂ by regulating the oxidation voltage value, revealing the important effect of the applied oxidation potential on the chemical composition of the surface reconstruction products.¹²⁷ Interestingly, Xu et al. employed electrochemical hydrogenation/oxidation to induce the surface reconstruction of c-Ni₂P₄O₁₂/a-NiMoOx/NF, traced the interface evolution by in situ Raman spectroscopy, and found that the real active phase for the HER was c-Ni₂P₄O₁₂/a-NiMoO₄, while that for the OER was c-Ni₂P₄O₁₂/a-NiOOH.¹²⁸

Photodeposition reaction

The traditional methods for the synthesis of noble metal single-atom or cluster material generally involve harsh conditions, techniques, or expensive equipment. The photodeposition method with the merit of the simple operation, mild reaction conditions, and fast processing is a popular method to prepare heterostructured catalysts containing highly dispersed noble metal clusters (or nanoparticles).¹²⁹ Driven by the xenon lamp, the light-absorbing substrate material generates photoelectrons to promote the reduction of nearby free noble metal ions to metal elements and deposit them on the surface of the substrate. Based on the anchoring of oxygen vacancies in the MoO₂ nanosheets to Pt clusters, Zhou et al. developed a room-temperature photoreduction method to synthesize a highly efficient catalyst for water electrolysis and hydrogen production with Pt clusters supported on MoO₂ nanosheets. This work confirmed the formation mechanism of Pt clusters, and the synergistic catalytic effect of Pt and MoO₂ was revealed using DFT calculations.¹³⁰ Further, this team successfully deposited Ru nanoparticles on MoO₂ nanosheets using the same photodeposition method.¹³¹

Other reactions

Recently, microwave-assisted synthesis has garnered considerable attention due to its extremely short reaction time, high yield, simple operation, and low energy consumption, which make it suitable for the preparation of functional nanomaterials, such as high-quality graphene, CNTs, and metal carbides. Our team prepared Ru-Mo₂C hetero-particles with a size of 3.5 nm on CNTs by microwave reaction for only 100 s. Such nano-scale Ru-Mo₂C particles effectively shortened the reaction path of the catalytic process, while the strong metal-support interaction improved its stability.¹³² Fei et al. employed the characteristics of ultrahigh heating and cooling rate of microwave thermal shock strategy for the first time to prepare amorphous/crystalline heterophase NiFe alloy (a/c-NiFe-G) catalysts encapsulated in graphene layers.²⁶

Notably, the raw materials used for microwave reactions must be able to absorb microwave energy, which limits the application of microwave-assisted synthesis methods to a certain extent. Compared to the microwave-assisted methods, the ball milling method has fewer limitations. Oh et al. prepared La_0.5Sr_0.5CoO_3−δ (LSC) and MoSe₂ heterostructure catalyst by a high-energy ball milling method with an optimal LSC:MoSe₂:Ketjen black (KB) mass ratio of 6:3:1, which exhibited excellent synergistic effects and electrocatalytic activity for HER.¹³³

Figure 10 summarizes the advantages and disadvantages of the aforementioned synthetic methods. For industrial hydrogen production, the simplicity, low energy consumption, and high yield of the synthesis route of electrocatalyst are equally important as the electrocatalytic activity, which must be considered for the selection of appropriate synthesis methods. Remarkably, the hydrothermal method with the merits of low cost, strong versatility, and high yield has emerged as the most suitable method for the industrial production of non-precious metal-based catalysts. Simultaneously, to maximize the utilization efficiency of noble metal reactants, the ball milling method is the most promising method for the industrial preparation of noble metal-based catalysts owing to its simplicity, efficiency, and economy.¹³⁴ We believe that the industrial synthesis and application of heterostructure catalysts with exceptional activity will be further boosted in the future.

FIGURE 10. Schematic comparison of advanced synthetic methods of interface engineering to realize splendid electrocatalytic activities.

IMPACT OF INTERFACE ENGINEERING ON ELECTROCATALYSIS

The functional indicators to measure the electrocatalytic performance mainly include intrinsic activity, number of active sites, conductivity, and so on. Generally, interfacial engineering regulates these functional indicators by tuning the interfacial microenvironment of electrocatalysts (bonding, electronic configuration, lattice strain, synergistic effects, etc.). Notably, due to the structural complexity and specificity of the well-designed catalysts, their electrocatalytic activity is often affected by more than a single interfacial microenvironment.^22,96

Intrinsic activity

Intrinsic activity is the inherent ability of the active components to adsorb and desorb reactants and reaction intermediates, and it is the fundamental factor determining the activity of the catalyst.

The intrinsic activity of the interface-engineered catalysts is mainly governed by electronic effect, interface bonding, and lattice strain. The local changes in the geometry and electronic structure can tune the band structure and density of states (DOS) of the catalyst, affect the relative filling of bonding and antibonding orbitals, and ultimately optimize the intrinsic activity.

Electronic effect

Among the heterostructure electrocatalysts, the most common mechanism is electron interaction, which is also known as electron (or charge) redistribution, charge transfer, and so on.¹³⁵ This is mainly due to the disparate energy band arrangements of different phases in the heterostructure, which is beneficial for the surface electron modulation of the heterostructure. Studies have shown that the electronic effect is a ubiquitous mechanism in electrocatalysts, and the crux to optimizing the electrocatalytic performance is to tune the electronic arrangement of the active sites.^136,137 Generally, a proper electronic arrangement can be used to optimize the adsorption/desorption energies of the heterogeneous materials for the adsorbate/product, essentially enhancing the catalyst activity. Liu et al. constructed the atomic-scale Co₃O₄/CeO₂ interface with enhanced OER activity.¹³⁸ Through electron energy loss spectroscopy (EELS), it was confirmed that the heterointerface was rich in free electrons and the carrier density was as high as 3.8 × 10¹⁴ cm⁻², leading to a high electron mobility. The DOS results revealed a partial transfer of electrons from Co to Ce atoms at the Co₃O₄/CeO₂ interface (Figure 11A), reflecting the accumulation of excess electrons at the heterointerface.

FIGURE 11. (A) DOS of the Co3O4/CeO2 heterostructure and its charge density distribution around the Fermi level. Reproduced with permission.138 Copyright 2019, Wiley. (B) Schematic representations of Ni3N@for 2M-MoS2water splitting (H—green; O—red; Mo—black; S—yellow; N—silver; Ni—blue). (C) The charge control effect on Ni3N@2M-MoS2. (D–F) The projected DOS image of the pristine 2M-MoS2, Ni3N, as well as the well-designed Ni3N@2M-MoS2. Reproduced with permission.139 Copyright 2022, Wiley.

Recently, Huang et al. designed a novel Ni₃N@2H-MoS₂ electrocatalyst, which exhibited excellent overall water splitting performance at a high current density (1.644 V@1000 mA cm⁻²).¹³⁹ Moreover, the NiS interfacial bond with metallic properties accelerated the transfer of charges from Ni₃N to 2H-MoS₂, regulating the electronic state of Ni/N atoms at the heterointerface near the Fermi level and reducing the overpotentials of HER and OER (Figure 11B–F). Simultaneously, Yin and co-workers designed a PtCo@PtSn catalyst based on the characteristics of Pt-based bimetallic alloys and heterojunctions, effectively adjusting the electronic structure of the Pt surface and accelerating the electron transfer ability.²⁵ The differential charge density and Bader charge results verified that the electron migration at the PtCo@PtSn heterointerface was from PtSn to PtCo, forming an electron-poor/rich region. Further, this charge redistribution at this heterointerface expedited the Volmer step of HER and water dissociation. The DOS results signified that the strong interaction between PtCo and PtSn caused the d-band center of Pt to move down, reducing the strength of the PtH bond to obtain the optimal ΔG_H*, effectually raising the HER intrinsic activity.

Interface bonding

When two components with similar energy levels are in close contact to form a heterostructure, their atomic orbitals may overlap or collide, resulting in the formation of chemical bonds across the interface, that is, interface bonding.¹⁴⁰ Fundamentally, the interfacial bonds primarily originate from the mutual filling of bonding and antibonding states between different components at the interface, which inevitably interferes with the local electronic arrangement and accelerates the generation and shift of the reaction intermediates at the heterointerface. Thus, it can also be regarded as a unique electronic modulation mechanism.¹⁴¹

Wei et al. constructed Ir-MoO₂ heterostructure catalysts with interfacial IrMo bonds, IrO bonds, and mixed IrMo/IrO bonds by manipulating the interfacial bonds.¹⁴² It was found that the diverse interfacial bonds exerted different effects on the surface (or interface) structure, species adsorption strength, and the electrocatalytic activity of the electrocatalyst. The DFT test verified that the IrO bond at the Ir/MoO₂ interface was localized (Figure 12A,B). Ir atoms were more receptive to the residual charge on the surface, which caused downward movement of the d-band center of Ir, thereby weakening the adsorption of H* and OH*. This work revealed that the modulation of interfacial bonding is an effective strategy to enhance the catalytic activity. Zhou's team prepared Au-loaded MoO₂ nanosheets by a facile photoreduction technique, demonstrating that a strong coupling could be realized at the Au/MoO₂ heterointerface through the AuO bond, thereby regulating the local electromagnetic field and greatly boosting the HER activity (Figure 12C).¹⁴³ Qiao's group hybridized Ni-MOF nanosheets and Pt nanocrystals to construct a Ni-MOF/Pt catalyst with splendid overall water splitting activity.¹⁴⁴ It is worth noting that the newly-formed NiOPt bond at the heterointerface induced charge redistribution, increased the OH* adsorption energy, decreased the H* adsorption energy, and accordingly promoted the HER and OER (Figure 12D). This work provided valuable insights on the relationship between heterointerfaces and catalytic activity. Additionally, Peng and co-workers anchored noble metal nanoparticles on MOF materials (M@Ni-MOF, M = Ru, Ir, Pd) via a spontaneous redox strategy, which exhibited enhanced HER performance over the full pH range.¹⁴⁶ Importantly, this delicate interface modulated the charge distribution in the HER catalyst mainly through the NiOM bond bridge at the interface, optimizing the adsorption capacity of water and H* intermediates. Meanwhile, Li and coworkers found that the OER catalytic performance of Fe-doped Ni-MOFs/FeOOH catalyst was 9.3 fold higher than that of the pristine Ni-MOFs owing to the appearance of abundant FeONiOFe bonds.¹⁴⁵ Further, the DFT results elucidated that the FeONiOFe bonds located at the interface shortened the NiO bond in Ni-MOFs and optimized the coordination structure of the Ni active site (Figure 12E–G), thereby accelerating the OER kinetics.

FIGURE 12. (A) Diagram of electrocatalytic activity of Ir/Mo-MoO2 system. (B) Three models (Ir/Mo-MoO2, Ir/O-MoO2, and Ir/MoO2) were established by DFT calculation. Reproduced with permission.142 Copyright 2021, American Chemical Society. (C) The specific binding locations on Au-MoO2 and their related DOS images of MoO2 and Au-MoO2 (Au-O2). Reproduced with permission.143 Copyright 2021, Elsevier. (D) The atomic Pt-NC/Ni-MOF heterostructure model was established by DFT calculation and its adsorption energy comparison for H* and OH*. Reproduced with permission.144 Copyright 2019, Elsevier. (E–G) Diagrammatic drawing of the electronic coupling among FeOFe, NiONi, and FeONiOFe bonds. Reproduced with permission.145 Copyright 2022, Wiley.

Lattice strain

Strain effects usually occur during the pseudo-growth of a well-structured crystal on another crystal with a different lattice constant. The long-range ordered arrangement of atoms is disrupted at the interface.¹⁴⁷ Further, this rearrangement of atoms directly disturbs the geometric structure and coordination environment of the active sites.¹⁴⁸ Some studies have found that a tensile strain of 1.5% on the crystal surface can elevate the dissociation constant of the adsorbate by more than three orders of magnitude. Therefore, the fabrication and monitoring of effective strain in the material are crucial for the regulation of its intrinsic activity.^149,150

For heterostructured catalysts, different chemical compositions and crystal structures induce lattice strain in the material, such as tension and compression. These strains modulate the d-bandwidth of the material by affecting the overlap degree of the wave function, optimizing the adsorption capacity of the catalytic site for reaction intermediates. Consequently, the electrocatalytic performance is significantly enhanced.^151,152 For tensile strain, the increase in the interatomic distances of the late transition metals and early transition metals at the heterointerface can cause the d-band center to shift up and down, respectively.¹⁵³ Importantly, the upward shift of the d-band center increases the antibonding state energy in the catalyst-adsorbate system (Figure 13A) and reduces the antibonding state filling, which in turn enhances the interaction between the catalyst and the reactants (or reaction intermediates).¹⁵⁴

FIGURE 13. (A) Schematic diagram of the impact of a tensile strain or reduced coordination on the width and location of the d-bands of early transition metals. Reproduced with permission.154 Copyright 2010, American Physical Society. (B) TEM image (inset is the FFT image) and (C) the corresponding TEM mapping of PdCu/Ir nanocrystals. Reproduced with permission.155 Copyright 2021, Wiley. (D) DFT calculations of Ru/MoS2 for the strain effects. Reproduced with permission.156 Copyright 2021, Nature Publishing Group.

Guo's group prepared a PdCu/Ir nanocrystalline with a core–shell structure (Figure 13B,C). Through physical characterization, it was revealed that the surface of PdCu nanocrystals was covered by Ir shells with a thickness of four atomic layers.¹⁵⁵ This PdCu/Ir catalyst had a compressive strain of 3.60% for Ir due to the lattice difference between the PdCu core and the Ir shell. Moreover, under the influence of compressive strain, electrons migrated from Ir to Pd atoms at the PdCu/Ir interface, which effectively weakened its adsorption energy for the reaction intermediates. The DFT results also confirmed that the strain energy on the Ir shell reduced the adsorption strength of the catalyst for oxygen-containing intermediates and promoted the OER. Furthermore, Tan et al. anchored Ru single atoms on nanoporous MoS₂ (np-MoS₂) with double triangular structure to construct Ru/np-MoS₂ heterostructure catalysts and precisely induced strain by tuning the curvature of nanoporous MoS₂ (Figure 13D).¹⁵⁶ Through a series of tests, it was found that the strain in Ru/np-MoS₂ effectively modulated the electronic structure of Ru. The concentration of reactants on the surface of MoS₂ was enhanced. Additionally, the dissociation of water molecules was accelerated during the electrocatalysis.

To recapitulate, the interfacial microenvironment (interface bonding, electronic configuration, lattice strain, etc.) has a huge impact on the intrinsic activity of heterostructured catalysts. Simultaneously, the typical characterization techniques are critical for the effective detection of the real interfacial microenvironment. Specifically, XPS, x-ray absorption spectroscopy (XAS), including x-ray absorption near-edge structure (XANES) and x-ray absorption fine structure (XAFS) and Raman spectroscopy are often employed to monitor the formation of specific interfacial bonds in the as-prepared samples.^{142–144,146} Notably, the shifts of the signal peaks in the XPS and XAS spectra also reveal the valence state changes and electron transfer in the materials, making them a common means of characterizing the electronic effects. Besides, the electronic effects are characterized through EELS, DOS, and differential charge density.^22,124 Furthermore, the detection techniques for lattice strain mainly include four-dimensional scanning transmission electron microscopy (4D-STEM) nano-diffraction, high-angle annular dark field (HAADF) STEM imaging, and geometric phase analysis (GPA).^157–159 These methods can be used for quantifying the strain inside the catalyst. Specifically, the 4D-STEM obtains the lattice strain parameters with sub-picometer precision by combining the 2D real space data obtained by the actual STEM scanning and the 2D momentum space data captured by the probe. Further, the distribution and intensity of bright spots in the HAADF-STEM images can truly reflect the atoms or atom pairs, which can then be used to obtain atomic-level strain information. Additionally, GPA can be used to qualitatively evaluate the microscopic strain by performing certain geometric operations on the obtained HRTEM images.

Active sites

Unlike the single-component compounds, the heterostructures constructed by the interfacial engineering strategy can often exhibit a remarkable synergistic effect in the catalytic process, fundamentally increasing the number of active sites, thereby optimizing the electrocatalytic performance.¹⁶⁰ This effect mainly originates from the cooperation of multiple components with different functions in the heterostructure. Specifically, multiple active components interact with each other through the heterogeneous interface. Motivated by the differences in physical and chemical properties between the interconnected components, the interface electrons are rearranged, effectively inducing more active sites and accelerating the decomposition of reactants and the formation of products in the catalytic process.¹⁶¹

Generally, the synergistic effects caused by interface engineering are mainly divided into two categories. The first one is for a single catalytic reaction, combining the components with excellent ability to adsorb reactants (or reaction intermediates) and desorb products. Markovic's team prepared Ni(OH)₂/Pt heterostructure catalyst and observed that the Ni(OH)₂ at the heterointerface contributed to the dissociation of water molecules and the generation of H* intermediates. The Pt atoms adsorbed newly-generated H* and released hydrogen.⁵⁸ Huang et al. fabricated Pt nanoparticles grown on the edge of MOF nanosheets (Pt/MOF-O) by precisely tuning the reduction kinetics of metal ions.¹⁶² Importantly, the Pt/MOF-O heterointerface caused the migration of electrons from Pt atoms to O atoms, resulting in the formation of electron-rich O atoms and electron-poor Pt atoms. Among them, the electron-rich O atoms facilitated the adsorption of intermediates (H* or OH*), while electron-poor Pt atoms promoted the electron transfer and H desorption during HER, both of which synergistically accelerated the HER process.

The other category is bifunctional catalysts for total water splitting reaction, which couple components (noble metal Pt, sulfides, phosphides, selenides, nitrides, etc.) with excellent HER activity and components (RuO₂, IrO₂, metal oxides, metal hydroxide, perovskites, etc.) with high OER activity. Wang and collaborators reported a Ni₃Fe NiFe layered double hydroxide (LDH) catalyst embedded with Pt single atoms and revealed the important contribution of the synergy between Pt and Ni₃Fe LDH to the total water splitting process. The Pt single atoms promoted the adsorption of H*, and the Ni₃Fe LDH support accelerated the dissociation of water molecules and promoted the formation of high OER active components (Ni_2+δFe_3+ζO_xH_y).³⁵ Moreover, Fu et al. combined Ni with high HER activity and NiO with high OER activity to prepare a Ni/NiO nanosheet catalyst with various Ni species for high-efficiency water splitting.¹⁶³ Under the effect of the two valence states of Ni, Ni/NiO exhibited ultra-high total water splitting activity (1.64 V at 10 mA cm⁻²) and stability.

Conductivity

The conductivity (or charge transfer rate) is another significant indicator to measure the intrinsic resistance of the catalyst material and the kinetics of the catalytic reaction. A high conductivity can greatly improve the electron utilization efficiency of the catalytic system and reduce the energy barrier of the reaction.¹⁶⁴ Theoretically, the ability of charge transfer is mainly governed by the Fermi level. Specifically, the higher the accumulation of electrons near the Fermi level, the stronger the electron transport ability of the material.¹⁶⁵ Actually, the rate of charge transfer inside the catalyst can be enhanced by coupling the active material with a highly conductive substrate material (carbon material or metal current collector, etc.). Due to the difference in the chemical composition and electronegativity on both sides of the interface, there is a clear flow of electrons at the interface, which inevitably optimizes the electronic configuration and Fermi level of the material, ultimately leading to ultrahigh conductivity.¹⁶⁶

Stability

The stability of the catalyst is of great significance to the actual industrial hydrogen production as it affects both the efficiency and cost of hydrogen production. Specifically, the stability is primarily reflected in two aspects: mechanical stability (morphology, size) and chemical persistence (composition, performance). The two parameters complement each other and are of great value in the electrocatalytic process.¹⁶⁷

In the actual catalysis process, the active components can easily react with the highly acidic (or highly basic) electrolyte. Therefore, the active components are dissolved and agglomerated, which can damage their morphology. In addition, the gas generated by the catalytic reaction may promote the exfoliation and passivation of the catalyst material on the electrode.¹⁶⁸ Therefore, the use of interfacial engineering strategies to construct core–shell structures or carbon material coatings can effectively prevent the corrosion and dissolution of active materials during the catalytic process, thereby enhancing the stability of the catalysts.^169,170

Wettability

Often, the electrocatalysts are required to be hydrophilic and air repellent for ensuring better water (or reactant) absorption and targeted hydrogen (or oxygen) release.¹⁷¹ Piao's group synthesized Sn₄P₃/Co₂P electrocatalyst with a “stem-cap” nanorod structure.¹⁷² Especially, the “cap”-like Co₂P material on top of the nanorod greatly enhanced the hydrophilicity and gas repellency of the Sn₄P₃/Co₂P catalyst, preventing HER- and OER-generated air bubbles from accumulating and adhering to the surface of the electrode (Figure 14A,B). Simultaneously, the amorphous a-Rh(OH)₃/crystalline NiTe catalyst displayed excellent hydrophilicity (Figure 14C,D).¹⁷³ This facilitated the mass transfer and rapid release of H₂, endowing the a-Rh(OH)₃/NiTe material with efficient HER catalytic activity in acidic, neutral, and alkaline electrolytes. Considering that the Mg²⁺ ions with hydration effect can promote the adsorption of water on the catalyst, Wang et al. constructed the MgO/NiCo₂S₄ heterostructure in situ on carbon cloth.¹⁷⁴ Compared with the pure NiCo₂S₄ catalyst, the contact angle of this heterostructured catalyst with water was reduced to 0°, indicating that the composite of MgO and NiCo₂S₄ promoted the super hydrophilic state of the catalyst, which in turn improved the neutral OER activity. Wang's team exploited the strong hydrophilicity and reducibility of MXene to spontaneously prepare MXene@Pt catalysts with excellent water-binding energy.¹⁷⁵ This material displayed high stability during HER for 800 h, and a high current density of 230 mA cm⁻² was achieved with an applied voltage of only 50 mV.

FIGURE 14. (A) Contact angle measurements of Sn4P3/Co2P and its contrast samples. (B) Schematic diagram of gas discharge from disparate electrodes. Reproduced with permission.172 Copyright 2022, Elsevier. (C) The contact angles of a-Rh(OH)3/NiTe and its contrast samples. (D) Schematic diagram of the adhesion behavior. Reproduced with permission.173 Copyright 2022, Elsevier.

CONCLUSION AND PERSPECTIVES

The development of catalysts with ultrahigh current density, low toxicity, high stability, and low cost is a crucial prerequisite for industrial hydrogen production, and most of the reported well-designed catalysts cannot meet all the above requirements simultaneously. Therefore, further research and improvements are still needed to realize large-scale industrial production and application.^57,176 Over the recent years, the interface engineering strategy has emerged as an extremely efficient approach to construct electrocatalysts with abundant and highly-active interfaces, which has broad application prospects. This review provided a systematic overview of the promising interface engineering strategies. First, the specificity and electrocatalytic performance of various interfaces were discussed from the perspective of the structure and composition, and it was established that the type and number of phase (or grain) boundaries have a significant effect on the performance. The activity of crystalline heterostructure catalysts can be improved by the enrichment of interfacial density and proper introduction of amorphous species. Then, the specific implementation details of the interface engineering were explained from the perspective of the growth method and experimental realization of heterointerfaces. Combined with the actual operation and industrial hydrogen production requirements, the characteristics and application prospects of synthesis methods were examined, providing directions for the design and synthesis of heterostructure catalysts. Finally, the functional advantages of interfacial engineering on the electrocatalytic activity were comprehensively analyzed, and the intimate connection between the interface microenvironment (bonding, electronic configuration, lattice strain, etc.) and the intrinsic activity was proposed, laying the foundation for the directional design of ideal catalysts and clarifying the connection between the interface structure and catalytic activity.

Overall, interfacial engineering can be used to construct heterostructure catalysts with exceptional performance or complementary functions by combining multiple active components. Remarkably, the highly active sites of well-designed heterostructure catalysts are generally distributed at the heterointerface, and their intrinsic activity is mainly affected by the interface microenvironment (bonding, electronic configuration, lattice strain, etc.), implying that the monitoring and regulation of the interface microenvironment should be the research focus of interface engineering. During the construction of heterointerface, materials with structural or functional differences should be selected for compounding (especially, appropriate amorphous substances should be introduced in the crystal). Simultaneously, the changes in the interface microenvironment should be closely examined, and the electrocatalytic performance can be further optimized by building interfacial bonds, creating lattice strain, and accelerating electron migration. To optimize the microenvironment of the interface, other design strategies (such as heteroatom doping and defect manufacturing) have also been adopted in interface engineering.^126,177,178 However, the complex interface can complicate the catalytic reaction mechanism. Notably, the research in this area is still in the exploratory stage, which will be useful for boosting the practical applications of interface engineering in the future.

Considering the development and application of interfacial engineering strategies in the field of electrocatalysis, the outlooks and challenges for future heterostructure catalysts are summarized as follows:

1. For industrial hydrogen production, the electrocatalysts must have exceptional activity, low cost, corrosion resistance, hydrophilicity, and gas repellency. Notably, the interfacial engineering strategy can be used to fabricate electrocatalysts with ultra-high overall water splitting activity by wrapping materials with superior OER activity on the surface of catalysts with ultra-high HER catalytic activity. Especially, the outer layer material can not only protect the inner active components, but also enhance the hydrophilicity and gas repellency. Simultaneously, the regulation of the interfacial structure and microenvironment can be used to further optimize the intrinsic activity of the catalyst. Therefore, the interface engineering strategy exhibits immense potential in the industrial application of catalysts. However, obtaining a clear understanding of the actual structure of the heterointerface and its microenvironment is difficult, which warrants further research.
2. Theoretical calculations (modeling, fitting, and computation) can effectively guide the structural design and optimization of catalysts with specific interfaces.^179–184 However, there is still a certain gap between the fitted interface model and the actual situation of the synthesized catalyst due to the simplification and idealization of the constructed model. Therefore, bridging the gap between theoretical models and actual interface structures is crucial, which can facilitate a clear understanding of the influence mechanism and the optimization of electrocatalytic performance. This should also be emphasized in the interface engineering strategy and even in the field of electrocatalysis.
3. Particularly, in situ monitoring techniques can be used to examine the dynamic evolution of the active sites and interfaces of electrocatalysts during the electrochemical reactions, which is conducive to obtain the intrinsic link between the structure and performance of catalysts. For example, in situ TEM techniques can be used to examine the changes in the morphology during electrochemical reactions.¹⁸⁵ Simultaneously, in situ XRD and Raman spectroscopy can be employed to track the structural reconstruction of materials under electrochemical conditions.^186–188 To summarize, the development and combined use of a variety of in situ techniques can provide a comprehensive understanding of the properties and working mechanism of the heterostructure electrocatalysts from multiple perspectives, which is of great significance for exploring complex heterostructured materials.

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (52072197, 21971132), Outstanding Youth Foundation of Shandong Province, China (ZR2019JQ14), Major Scientific and Technological Innovation Project (2019JZZY020405), and Major Basic Research Program of Natural Science Foundation of Shandong Province under Grant (ZR2020ZD09).

CONFLICT OF INTEREST

The authors declare no conflict of interest.