A review on photocatalytic methane conversion systems: from fundamental mechanisms to the emerging role of ferroelectric materials

Abstract

Methane (CH₄) conversion is a promising strategy for reducing greenhouse gases, synthesizing high-value-added chemicals, and thus achieving carbon neutralization. However, due to its inertness, CH₄ requires high energy input to initiate the conversion process, which always leads to excessive energy consumption and catalyst deactivation. Along these lines, the use of sunlight as energy input has demonstrated enormous potential to help overcome the uphill thermodynamics of methane conversion under mild reaction conditions. Nevertheless, the high recombination rate of photoinduced charge carriers still hampers photocatalytic efficiency. In the past years, ferroelectric photocatalysts have attracted much attention for dealing with sluggish charge separation/transfer dynamics, presenting themselves as excellent materials for enhancing photocatalytic CH₄ conversion rates under mild conditions. Consequently, this work reviews and discusses four pivotal photocatalytic CH₄ conversion routes, including (i) dry reforming of methane (DRM), (ii) partial oxidation of methane (POM), (iii) non-oxidative coupling of methane (NOCM), and (iv) oxidative coupling of methane (OCM). In particular, special attention has been paid to the recent advances in ferroelectric-assisted photocatalytic CH₄ conversion to elucidate the influence of polarized ferroelectrics in the charge transfer mechanisms. Finally, the main challenges in the field are highlighted, also presenting possible strategies to overcome them in order to encourage more in-depth research on ferroelectric-assisted methane conversion in the future.

---

Yiming Lei

Yiming Lei is a PhD student in the Department of Chemistry at Universitat Autònoma de Barcelona, Spain (2023–current). He is awarded by the China Scholarship Council (CSC). His current research focuses on the functionalization of 2D Xene materials represented by 2D germanene derivatives for real implementations including (bio)sensor and catalysis.

Jose Muñoz

Dr Jose Muñoz received his PhD in chemistry from Universitat Autònoma de Barcelona (UAB). Afterward, he carried out two postdoctoral stays: the first one at the Materials Science Institute of Barcelona (ICMAB, 2016–2019) with a “Juan de la Cierva” fellow, and the second one at the Central European Institute of Technology (CEITEC, 2020–2022) with an MSCA-IP grant. In 2023 he returned to the UAB as an independent researcher with a “Ramón y Cajal” senior grant. His research interests combine 3D printing technology, 2D materials, and molecular engineering for the development of biosensors and multiresponsive devices.

---

1. Introduction

The rapid global population growth and rising living standards cause the high consumption of fossil fuels, leading to a non-negligible energy crisis and environmental issues.^1,2 It is predicted that compared to 2018, the total energy demand will increase by about 40% in 2024.³ Fossil fuels as the dominant energy source will be depleted after 70 years, meaning that the current energy-industry system cannot be maintained without developing an alternative energy source.⁴ Besides, a series of environmental issues like climate change and air pollution owing to the over-exploitation of natural resources are the most complex challenges faced today.⁵ Therefore, the development of alternative energy resources is extremely urgent due to the increasing growth in energy demand and the necessity to reduce environmental problems.^6,7 With a sufficient reserve of methane (CH₄) hydrate and shale gas, the utilization of methane for industrial production becomes a reliable strategy for replacing petroleum.^8,9 Besides the abundant natural resources, the diversity of synthesis pathways to obtaining CH₄ is also a great advantage to utilizing this high-quality fuel gas. For instance, natural gas is expected to increase from 35.57 Tcf (trillion cubic feet) in 2024 to 42.07 Tcf by 2050, meaning that the chemical synthesis of high-value-added products will be possible on a large scale if suitable methane conversion technologies are realized.^10,11 In addition, biogas—composed of ∼60% CH₄—can be generated from animal excrement, wastewater, and food debris via anaerobic digestion, which is a clean and green chemistry route.¹² In this regard, CH₄ from biogas can be purified to obtain biomethane as a promising renewable energy for further chemical transformation. Another environmentally friendly production route for preparing CH₄ is CO₂ reduction (or hydrogenation) reactions. CH₄ can be made from CO₂ captured from industrial sources, combined with hydrogen obtained from water splitting using surplus renewable energy. This offers additional opportunities for large-scale CH₄ conversion while reducing the use of fossil fuels and the global greenhouse effect. Thus, the diverse sustainable sources of CH₄ are beneficial for its utilization towards the production of high value-added chemicals, oxygenated products including alcohols and aldehydes chemicals, and even longer chain compounds like long chain alkane.

Methane conversion to added-value chemical products can be classified into two different strategies: indirect and direct routes. While indirect routes include (i) dry reforming of methane (DRM) and (ii) steam reforming of methane (SRM), direct routes are based on (iii) partial oxidation of methane (POM), (iv) oxidative coupling of methane (OCM), and (v) nonoxidative coupling of methane (NOCM). The main difference between these two strategies is the fact that indirect routes need to firstly produce syngas (a mix of CO and H₂) for the production of high-value-added products from methanol (CH₃OH) conversion, as summarized in Table 1.¹³ Importantly, these strategies have been shown to be useful in addressing excess greenhouse gas emissions without disrupting current infrastructure.^14–19

Table 1 Main products from diverse methane conversion processes

Conversion process	Product	Reaction pathway	ΔG^0 (298 K) [kJ mol^−1]
DRM	CO + H2 (syngas)	CH4 + CO2 → 2CO + 2H2	170.7
	CH3COCH3	2CH4 + CO2 → CH3COCH3 + H2O	114.1
SRM	CO + H2 (syngas)	CH4 + H2O → CO + 3H2	142.1
	CH3OH	CH4 + H2O → CH3OH + H2	117.3
	C2H5OH	2CH4 + H2O → C2H5OH + 2H2	161.7
POM	CO + H2 (syngas)	2CH4 + O2 → 2CO + 4H2	−173
	CH3OH	2CH4 + O2 → 2CH3OH	−222.6
	HCHO	2CH4 + O2 → 2HCHO + 2H2	−103.6
OCM	C2H4	2CH4 + O2 → C2H4 + 2H2O	−287.6
	C2H6	4CH4 + O2 → 2C2H6 + 2H2O	−320
NOCM	C2H6	2CH4 → C2H6 + H2	68.6
	C6H6	6CH4 → C6H6 + 9H2	433.9

---

However, because of the thermodynamic stability of methane (440 kJ mol⁻¹ for C–H bond cracking) due to its symmetrical tetrahedral geometry, it is difficult to convert methane into high-density energy fuels or chemicals taking into account other small hydrocarbon molecules with low chemical bond energy like 413 and 412 kJ mol⁻¹ in propane and ethane, respectively.^14,20,21 To date, the thermocatalytic reaction, carried out at high temperature or pressure, is still the main industrial route of methane conversion.²² It is undeniable that thermocatalytic processes often cause high energy consumption and catalyst poisoning (generally due to the rapid carbon deposition and catalyst sintering).²³ To overcome these hurdles, increasing attention is paid to photocatalytic methane conversion technologies.

Compared to traditional thermocatalysis and its reaction conditions, photocatalysis is a promising technology that can be carried out under ambient conditions, using photons to drive chemical processes.^24–27 Light illumination causes energetic charge carriers to move across the photocatalyst, significantly reducing the activation energy and overcoming the thermodynamic barrier.^28–31 Thus, photocatalytic methane conversion with solar energy as the only energy input during the reaction process can reduce energy consumption to the greatest extent possible. Typically, over the conventional semiconductor surfaces, the light-driven catalytic reactions can be viewed as three fundamental steps as follows (Fig. 1a):^32,33 (i) when the light irradiates the semiconductor-based photocatalysts, if the energy of the photons is equal to or greater than the semiconductor band gap (E_g), photons can be captured to excite photoelectrons. (ii) Then, the photoelectrons transfer and occupy the conduction band (CB). And the holes stay at the valence band (VB). Afterward, electrons migrate to the reduction centres, while holes transfer to oxidation sites. (iii) Finally, the adsorbed reactants will interact with electrons and holes to compete with the photocatalytic redox reactions. And several possible products should be desorbed from the photocatalyst surface to ensure the subsequent reaction cycles. Notably, several factors can affect the whole photocatalytic efficiency. For example, many semiconductor-based photocatalysts have a narrow light response range, which means that they only absorb ultraviolet light and little visible light.^34–37 Therefore, visible-infrared light energy (∼90% of the solar energy) is wasted, resulting in a low quantum efficiency.^38–40 In addition, poor adsorption/activation of reactants at the few active sites or high charge transfer resistance in photocatalysts can lead to the recombination of photoexcited carriers before they can participate in the redox process, which is also a major challenge.^41–44 In addition, difficulties in the absorption–desorption also limit the photocatalytic efficiency.^45,46 Among these issues, the sluggish charge separation inhibits the solar conversion and hinders the catalytic process, so improving the charge transfer dynamics is a must to improve the photocatalytic activity. Although the design of co-catalysts^47–50 and heterojunction structures^51–54 got a lot of attention to accelerate the charge transfer process, the efficiency is still not satisfactory for commercial use. Consequently, rationally designed photocatalysts with intrinsic controllable photoinduced charge transfer behaviour are crucial. Under these circumstances, ferroelectrics have been identified as potential materials that might offer new opportunities to further promote charge separation and thus improve photocatalytic activity.

Fig. 1 (a) Three basic steps of photo-driven catalytic reactions. Step 1: inducing photoexcited electrons and holes; Step 2: charges migration; Step 3: photo-driven redox reactions. (b) An illustration showing the impact of the polarization effect in the separation of charge carriers and subsequent redox reactions. (c) The band bending phenomenon in ferroelectric photocatalysts owing to polarization effect. (d) The blue bar chart is publications of ferroelectric, photocatalysts according to the search result in “Web of Science”. Key words: “photocatalysis” and “ferroelectric”. The red curve is “publications of photocatalytic methane conversion via ferroelectric photocatalysts”.

In general, ferroelectric materials are polar non-centrosymmetric (NCS) materials with domain walls.⁵⁵ These domain walls can generate an electric field induced by non-coincide positive–negative charge centres. And the NCS crystal structure endows ferroelectrics with spontaneous polarization ability even without external stimulation.^56–61 Thus, an external electric field can simply adjust the polarization directions. However, high-temperature conditions might lead to the symmetric structure of the ferroelectric crystal, causing the materials to lose their ferroelectricity. The built-in electric field associated with the asymmetric domain wall (DW) polarity in ferroelectric materials facilitates the vectorial migration of photoexcited charge carriers (Fig. 1b).^56–61 In addition, by controlling the polar facets, ferroelectric materials are able to further improve the ability of reactant adsorption and product desorption.^41,62,63 Also, the selectivity of products as one of the most important factors for highly efficient photocatalytic methane conversion is adjusted by polarity switching at DWs.^63,64 More importantly, the adsorbed reactants at specific active sites can generate several intermediates and charges with different polarity properties. In this case, their accumulation will lead to the band bending in different directions (Fig. 1c).⁶⁵ Specifically, the positive polarity always caused an electron accumulation and thus a downward band bending. This can promote the flow of electrons and restrict hole movement.^66,67 Conversely, the surface with negative polarity should lead to an upward band bending, since there is an electron depletion layer. Obviously, in this case, the electron transfer will be difficult and hole transfer is accelerated owing to the different concentrations of charges on both sides of the interface.^68,69 Accordingly, the controllable properties of charge transfer in ferroelectric materials are beneficial for photocatalytic redox reactions with methane conversion no exception. Among the extensive library of ferroelectric materials—which can be mainly divided into three families, viz. inorganic, organic, and organic–inorganic hybrid ferroelectrics—inorganic and organic–inorganic hybrid ferroelectrics made of perovskite and its halide derivatives show, simultaneously, strong spontaneous polarization and optoelectrical properties.^70–72 Thus, special interest has been focused on the implementation of perovskite-based ferroelectrics for photocatalytic approaches.

Although many publications on ferroelectric-assisted photocatalysis have been recently reported, there are fewer works on their implementation for photocatalytic methane conversion applications to advance toward the efficient production of value-added chemicals and energy sources (Fig. 1d). Therefore, this review aims to provide a comprehensive tutorial about the recent progress in ferroelectric-assisted photocatalytic methane conversion, pointing out the remarkable materials and merits of the most common CH₄ conversion technologies. Finally, the main challenges in this emerging field and the corresponding solutions are listed. All in all, this review is hoped to provide in-depth insights to encourage the scientific community to further explore ferroelectric materials as potential photocatalysts for their implementation in the field of energy conversion in general, and methane conversion in particular.

2. Routes for photocatalytic methane conversion

2.1. Prominent semiconductors for photocatalytic methane conversion

Before reviewing the fundamental reaction mechanisms and key studies on different photocatalytic methane conversion strategies, it is essential to first introduce prominent examples of semiconductor photocatalysts capable of facilitating this conversion. Among the broad spectrum of semiconductor materials, two main families stand out due to their stability under redox catalytic reactions:^73–75 metal oxides and nitrides. Indeed, the intrinsic properties of these materials are crucial for understanding why the C–H bond in CH₄ can be cleaved photocatalytically (see Fig. 2).

Fig. 2 Energy band structures of semiconductors and the redox potentials of reactants and active species.

2.1.1. Metal oxides. As some of the most conventional semiconductors, several metal oxides exhibit large band gaps, which suggest strong redox capabilities for complete photocatalytic methane conversion reactions (Fig. 2).⁷⁶ For instance, the suitable CB and VB positions exhibited by TiO₂ and ZnO have been widely studied for both indirect and direct photocatalytic methane conversion process.^16–19,77 From a selectivity perspective, WO₃ (ref. 78) and BiVO₄ (ref. 79 and 80) are potential semiconductors to preferentially forming CH₃OH owing to their CB positions are positioned just between the redox potential of CH₄/CO and CH₄/CH₃OH.^81,82 Moreover, SrTiO₃ and CuO are potential candidates to obtain high-value-added chemicals; however, their energy band structures are defective.^83,84 To address this issue, surface/energy band engineering is widely utilized.⁸⁵ Alternatively, introducing high temperatures to promote CH₄ activation has also shown to be a feasible strategy.⁸⁶ Another key aspect to be considered is the ability of the employed photocatalyst to adsorb/desorb the substrate. In particular, porous semiconductor materials, such as SiO₂ (ref. 87) and zeolites⁸⁸ have been shown to boost the photocatalytic methane conversion reaction by not only taking advantage of their large surface areas and abundant active sites but also accommodating alternative metal oxide semiconductors, such as Ga₂O₃.⁸⁹

2.1.2. Nitrides. Research on nitride-based photocatalytic methane conversion systems has primarily focused on carbon nitride (C₃N₄) derivatives. Due to their polymeric properties, a diverse range of C₃N₄ materials with tailored properties have been prepared by selecting different monomer precursors for their synthesis.⁹⁰ Additionally, their photocatalytic ability has been further modified through the incorporation of different metal nanoparticles.⁹¹ Relevant examples include Ru/Zn-modified C₃N₄ for CH₄ coupling with CO₂,⁹² La-modified C₃N₄ with tube shape for DRM⁹³ and POCM,⁹⁴ and Cu-modified C₃N₄ for CH₄ oxidation to CH₃OH.⁹⁵ Interestingly, the interaction between Cu and C₃N₄ on Cu-modified C₃N₄ also allowed the conversion of CH₄ into ethanol (CH₃CH₂OH) via the combination of –CH₂–OH on Cu atoms and –O–CH₃ on C atoms. Another potential nitride for methane conversion is GaN, which has demonstrated benzene production under UV light irradiation.⁹⁶ Specifically, Si-doped GaN nanowires, characterized by n-type semiconductor properties, have exhibited good selectivity toward benzene (96.8%), with an evolution rate of 0.019 μmol h⁻¹. However, the limited number of studies on GaN limits the understanding of the reaction mechanism on its surface. To date, a key challenge in nitride-based materials is that replacing O atoms with N atoms broadens the light response range, but the energy band structures of the resulting metal nitrides may not be optimal for methane conversion. Therefore, achieving a balance between enhanced optical properties and strong redox capabilities remains a crucial issue to be addressed.

2.2. Indirect methane conversion

Indirect methane conversion pathways encompass the oxidation of CH₄ to syngas through either SRM or DRM technologies, in which the resulting mixture of CO and H₂ serves as the precursor for their subsequent conversion into upgraded products. Generally, SRM utilizes steam (H₂O) to produce syngas according to the formula CH₄ + H₂O → CO + 3H₂,⁹⁷ DRM involves the reaction of CH₄ with CO₂ by following the formula CH₄ + CO₂ → 2CO + 2H₂.⁵⁵

2.2.1. Steam reforming of methane (SRM): mechanism and kinetics. Indirect methane conversion technology still attracts much attention, because of the reforming ability that convert the CH₄ into syngas including CO and H₂, which can be further used as a fuel to produce downstream chemicals of significant economic value (e.g., methanol, butanal, and kerosene).⁹⁸ These products from the downstream process could be important raw materials for high-value-added industrial products, such as plastics from methanol via formaldehyde as intermediate,⁹⁹ biodiesel precursor from butanal via catalytic hydroxy alkylation/alkylation reaction,¹⁰⁰ and aviation kerosene from long-chain alkanes via Fischer–Tropsch.¹⁰¹ Therefore, with abundant CH₄ and so many CO₂ sources, such as fossil fuel combustion or agricultural waste, their consumption and reuse are promising to sufficiently supply the chemical industry.^102,103

Since the high H/C ratio of CH₄ among hydrocarbons, SRM is a reliable technology to provide hydrogen energy at high operating temperatures.¹⁰⁴ In brief, in the presence of CH₄ and water, hydrogen production can be produced as shown in eqn (2-1). And some oxygenated products like CH₃OH can also be directly formed by SRM through another reaction pathway shown in eqn (2-2). SRM is an endothermic reaction, which is always operated at temperatures of >700 °C, indicating that the catalysts for SRM should be stable under high-temperature conditions.¹⁰⁵

CH₄ + H₂O → CO + 3H₂ (2-1)

CH₄ + H₂O → CH₃OH + H₂ (2-2)

Several excellent works have tried to construct kinetic models for specific reaction conditions. Because of the endothermic nature, the fast transfer of both heat and mass is important to a highly efficient SRM process. In traditional industrial SRM reactors, the heat energy always is transferred from the furnace to the tube reactor, resulting in rapid reaction rates and thus very low effectiveness factor (η), compared to the diffusion rate of reactants and products over pellet-type catalysts.¹⁰⁵ Abbas's group simulated and studied the SRM process based on a one-dimensional (1D) theoretical model.¹⁰⁶ Theoretically, the activation energy for SRM was 257.01 kJ mol⁻¹. High temperature, low pressure, and high H₂O/CO₂ ratio contributed to rapid CH₄ conversion and high purity of H₂. This work is consistent with the previous report from Xu et al., also showing a low η (<0.03) when using a small portion of catalyst in industrial-scale conditions.¹⁰⁷ An early theoretical work studying the influence of SRM reactors on kinetics has also been investigated.¹⁰⁸ Two computational fluid dynamics (CFD) models were developed for simulating the steam reforming and combustion of methane in microchannel reactors. The models included 20 mm in length and 0.7 mm side with 4 or 20 square microchannels. The results illustrated that the gaseous reactants were heated isothermally so that the complete CH₄ combustion occurred at the channel entry within a short distance. The cross-flow disposition led to 80 °C of maximum temperature differences and an easier feed stream distribution, which was beneficial for the rapid reaction kinetics. In addition to the unique reaction vessels, the reaction temperature was also investigated from a kinetic perspective. Kechagiopoulos et al. clarified the reaction kinetics of SRM at low temperatures in conjunction with isotopic investigations.¹⁰⁹ Temperature-programmed experiments of CH₄ reforming and decomposition modes with isotopic investigations showed the cleavage of the C–H bond was the rate-determining step, but the steam-derived intermediates had a low influence on the whole reaction rate. The authors also proposed a thermodynamically consistent microkinetic model with the consideration of several surface pathways, indicating that –CH₃ dehydrogenation on the catalyst surface could also limit the reaction rate.

Although SRM has become a mature industrial technology for obtaining hydrogen fuel, it still has several disadvantages such as mass/heat transfer and carbon deposition (coking). Coking often blocks reactors or breaks catalyst particles so as to cause deactivation. Experimentally, when the ratio of steam to carbon (S/C) is lower than 1, rapid and severe coke will inevitably occur.¹¹⁰ Meanwhile, the high temperature and pressure are also related to the carbon deposition problem.¹¹¹ In the process of SRM industrialization, researchers try to use a high S/C ratio (2.5–3.0) to avoid the coke deposition and catalyst deactivation problem. However, the excessive partial pressure of steam reduces energy efficiency. In order to develop catalysts with high stability, the Boudouard reaction (including CH₄ and CO cracking) should be considered to suppress coke formation. Additionally, the ratio and selectivity of products from SRM also need to be further controlled. In this case, the catalytic transformation of CH₄ and CO₂ into syngas via DRM might be an ideal alternative technology to SRM. DRM has the potential ability to achieve real carbon neutrality via the conversion of solar energy to chemical energy and mitigate the greenhouse effect. Considering the fact that the study about ferroelectric-assisted SRM systems is still a blank field, the SRM will not be further discussed in this review article. Some excellent works have comprehensively summarized and discussed the traditional thermal and advanced photo(thermal) catalytic SRM implementations. Interested readers are recommended to read ref. 97 and 112.

2.2.2. Dry reforming of methane (DRM): mechanism and kinetics. Comparing SRM (eqn (2-1)) and DRM (eqn (2-3)), DRM is particularly attractive because it uses two abundant greenhouse gases as feedstocks. However, this technology is also because of its inherent endothermic nature and the deactivation during the process triggered by carbon deposition at high-temperature conditions.¹¹³ One of the advantages of the DRM process is the high energy efficiency compared to SRM, since SRM uses water as feed gas, meaning lots of energy has to be used to evaporate the water.⁹⁸ In DRM reactions, the feed reactants are gaseous, indicating that the energy could be entirely involved in redox chemical reactions rather than physical evaporation as in SRM. And during the DRM process, the flow rate and ratio of reactants are easily controlled. Importantly, DRM has a higher CO/H₂ ratio (1 : 1) than SRM (1 : 3). Thus, further downstream processes are contributed by the CO-rich syngas from DRM.¹¹⁴

CH₄ + CO₂ → 2CO + 2H₂ (2-3)

Generally, both traditional thermal catalysts and photocatalysts tend to activate reactants (CH₄ and CO₂) via a bifunctional route: (i) after CH₄ is absorbed on reduction sites, dehydrogenation will occur (Fig. 3a). However, during the CH₄ activation step, the cracking of the strong CH₃–H_(g) bond is very difficult because of the high binding energy (439.3 kJ mol⁻¹), meaning the first step of CH₄ dissociation is one of the rate-determining step (DRS).¹¹⁵ Regarding the CO₂ molecules, as shown in Fig. 3b, they can be specifically adsorbed on three different reactive sites involving the metal (cocatalyst) surface, metal–support interface, and support. Although the conversion of CO₂ to CO or carbonate precursors is simple and direct,^116–118 the intrinsic properties of CO₂ like non-polarity and thermodynamic stability lead to robust C=O bond with ∼750 kJ mol⁻¹ of dissociation energy barrier. From the thermodynamic view, the activation of CO₂ with high energy input is another RDS factor.^103,119 Further, the strong chemical bonds in CH₄ and CO₂ make a high-temperature condition (>800 °C) necessary during the conventional thermal catalytic DRM process, resulting in the unavoidable sintering and carbon deposition on transition metal-based catalysts.^120–122 It has been confirmed that the photocatalytic DRM can be carried out at low temperatures using semiconductor photocatalysts, which may be an effective strategy to avoid sintering and coking of catalysts.¹²³

Fig. 3 Illustration of reactants activation during DRM process: (a) CH₄ adsorption and dissociation (cracking), generating H₂ product. (b) CO₂ adsorption on three different active sites including metal surface, metal–support interface, and support, and then direct generation and desorption of CO within one step.

For the kinetic studies, several kinetic models have been proposed based on different adsorption states of the reactants, so that the rate-determining step (RDS) could be studied in different DRM systems. Here, three mainstream kinetic models for predicting the reaction rates are presented.

(1) Power-law model: the power-law model was constructed to provide reaction parameters without including reaction mechanisms. This kinetic model only assumed one situation, in which catalyst surfaces tend to adsorb more CH₄ instead of CO₂ so that the CH₄ adsorption dominated the whole reaction. Thus, the parameters from this simple model are rough and sometimes are not consistent with the actual experimental results.

(2) Eley–Rideal (ER) model: the ER model investigated the DRM kinetics in the temperature range of 700–850 °C with the pressure ratio of CH₄ to CO₂ was 1. Besides, other competitive or reverse reactions were also taken into consideration.¹²⁴ In this model, only one reactant (CH₄ or CO₂) is adsorbed on the catalyst surface to interact with another unadsorbed gaseous reactant. The ER model showed that both conversion processes, CH₄ to H₂ and active carbon (C) to CO (C + CO₂ → CO), determine the reaction rate.

(3) Langmuir–Hinshelwood–Hougen–Watson (LHHW) model: although the ER model has considered a lot of factors including crucial parameters and common side reactions, the LHHW model is more accurate because the mechanistic steps in the LHHW model are consistent with practical experimental conditions and results.¹²⁵ The greatest difference between LHHW and other models is that in the LHHW model, two reactants should be adsorbed on actives at the same time. In this case, the LHHW model regards both activation and cracking processes as the RDS. And another step is thermodynamic equilibrium. The LHHW model is proposed based on the activation process of two reactants and the interaction between CH_x and O, which finally generates the desired syngas product. It has been confirmed that CH_xO as an intermediate could be rapidly dissociated to CO and H₂.¹²⁶ However, water was always detected even though no water or steam was present in the initial reactants owing to the unavoidable RWGS (reverse water–gas shift) side reaction.⁹⁸ Fortunately, although a small amount of water had a slight effect on the ratio of syngas (H₂ : CO), the RWGS with water generation leads to more CH_xO active species and thus the yield rate of syngas.

2.2.2.1. Thermal catalytic DRM. Considering the fact that many photo(thermal)catalysts have been developed based on previous investigations of mature thermal DRM technology, some common thermal catalysts for conventional DRM reactions should be introduced before showing the advantages of ferroelectric-assisted photocatalytic DRM technology. During the thermal catalytic process, the selection of catalysts is important to achieve long-term and high conversion efficiency of methane. Although the cost of catalysts cannot be neglected, noble metals (e.g. Rh, Ru, Pt, Ir, and Pd) could be considered as suitable candidates due to their high activity and stability under harsh conditions for the DRM process.^120–122 An early work reported a Ru/Al₂O₃ catalyst. Ru/Al₂O₃ exhibited 46% CH₄ conversion and 77% H₂ selectivity. Kinetic and temporal studies indicated that the Ru metals induced both CH₄ and CO₂ conversion processes. In addition, H and O species as intermediates were continuously provided by the Al₂O₃ substrate, so that catalyst aging (or active Ru oxidation) could be inhibited. Another example is the Ru-loaded lanthanum oxide (Ru/La₂O₃) for DRM application published by Cornaglia and co-workers.¹²⁷ Ru/La₂O₃, which showed ∼15% conversion efficiency of CH₄, had excellent stability for more than 80 h at high temperatures (823–903 K).

Furthermore, when considering industrial production, other requirements have been proposed, such as maintaining high activity while reducing the cost of catalysts, or achieving high stability and efficiency at the same time. In this case, transition metals have 3d orbitals with unpaired d-electrons. Especially, Ni-based catalysts have become the best choice owing to their low cost, high adsorption energy towards O and C atoms, and excellent activity for DMR.^128–130 However, the strong cracking process of CH₄ over Ni surface always leads to carbon deposition (CH₄ → C* + 4H*), which can cover the active sites on the catalyst surface and finally result in inactivation. Because long-term stability is also pivotal for catalyst industrialization, several advantages have been proposed to improve the stability of Ni-based catalysts. For instance, the fabrication of bimetallic thermal catalysts has been used in Ni-based catalysts.^131–133 The Ni-based alloy can resist carbon formation than single Ni.¹³⁴ Zhang et al. reported Ni_xAl₁O_2−δ mesoporous catalyst.¹³⁵ The oxygen vacancy formed by the partially inverted structure of NiAl₂O₄ spinel can be used to activate CO₂. Then CO₂ reacted with deposited carbon atoms to avoid carbon deposition so that the stability was improved up to 50 h at 800 °C. Besides, in recent years, it has been reported that the position of carbon atoms could be tailored in interstitial or surface sites in/over Ni crystalline.¹³⁶ And the interstitial atoms affect the electronic properties of surface atoms, further adjusting the catalytic behaviour. This provides a new idea to eliminate carbon deposition. In addition to reactants activation and long-term stability, other factors also determine the activity and selectivity of DRM implementation. Several works have discussed the impact of important factors in the DRM efficiency and product selectivity.^137–139 However, the photocatalytic DRM systems did not get enough attention from the past works.

2.2.2.2. Photocatalytic DRM. Photocatalysis can reduce the harsh conditions of the traditional thermal DRM process by taking advantage of light illumination, as the energetic photoexcited charges effectively dissociate the C–H and C=O bonds and prevent carbon deposition under mild conditions instead of high temperature and pressure conditions to maintain the activity of the catalysts after several cycles.^140–143 However, low-energy photons can hardly be used to directly excite photocatalysts to trigger reactions, so the infrared light energy has been overlooked for a long time.^39,40,144 The photothermal DRM approach has led to an increase in DRM activity by integrating solar energy and thermal energy in the same reactor,^145–147 and more studies have attempted to utilize both light and thermal energy from light irradiation to further improve the conversion rate of CH₄. A representative work is the one reported by Shoji and co-workers employing SrTiO₃-supported rhodium catalyst (Rh/STO).⁸⁶ During DRM experiments, the limit of thermal catalytic DRM over Rh/STO was determined, suggesting the maximum H₂ generation under dark and high-temperature conditions (Fig. 4a). However, under light irradiation without external thermal energy input, Rh/STO could overcome the limitation in the thermal system even at a mild temperature. On the other hand, no H₂ could be detected from bare STO, meaning Rh was an important active centre for photothermal catalytic DRM. Rh remained at 1–2 nm of particle size before and after the DRM, indicating that there was no aggregation of Rh particles (Fig. 4b and c). And no carbon deposition was detected by STEM images. Thus, catalytic DRM with light energy as the only input energy source effectively enhanced the longevity of catalyst via inhibiting aggregation and coking. The charge transfer pathway during the DRM reaction was discussed by in situ EPR measurements. Trapped hole peaks in the VB were clearly observed under light irradiation (Fig. 4d). However, the trapped electrons in the CB were not obtained in Rh/STO, suggesting that the photoexcites electrons were injected into Rh. In N₂, CO₂/CH₄ or CH₄ atmosphere, the signal intensity of the hole decreased, indicating that the holes reacted with CH₄ instead of CO₂ (Fig. 4e). The subtract potential profile from Kelvin probe force microscope (KPFM) analysis implied that the electron could transfer from STO to Rh due to the gap in their Fermi levels (Fig. 4f–h). According to isotopic trace analysis, in the Rh/STO catalyst, the lattice oxygens (O₂⁻) acted as mediators to drive DRM instead of protons. Thus, CH₄ was selectively oxidized by surface lattice oxygen ions of STO, while CO₂ was reduced by photoelectrons from STO over Rh particles, avoiding the recombination of charge carriers in the interior of STO (Fig. 4i).

Fig. 4 (a) H₂ evolution efficiency and the limitation of the thermal DRM system. Reaction conditions: 150 W Hg–Xe lamp, 1% CH₄ and 1% CO₂ mixture in argon balance gas (4.5 μmol min⁻¹). HAADF-STEM images of Rh/STO (b) before and (c) after DRM. ESR spectra of Rh/STO (d) in an N₂ atmosphere; (e) in N₂, CO_2, or CH₄/CO₂ atmosphere. KPFM analysis of Rh/STO (f) before and (g) during UV light irradiation, and (h) potential difference between light-on and light-off conditions. (i) Charge carriers and oxygen ions transfer mechanisms.⁸⁶ Copyright © 2020, Springer Nature.

Photo(thermal) catalytic DRM technology realizes a highly efficient CH₄ conversion under mild conditions because it can overcome the thermodynamic energy barrier with external light energy input and assistance.^148,149 From this perspective, the charge-carrier recombination instead of CH₄/CO₂ activation will become the main challenge for the whole DRM process, since the random collision of electrons and holes will be aggravated by high-temperature conditions from external heating or light illumination.³⁸ The collision of electrons and holes leads to their recombination and the waste of the captured photon energy before the charges take part in the photo(thermal) catalytic redox process.^150,151 In this case, ferroelectric materials are able to inhibit the photoinduced charge recombination and promote charge transfer by constructing a built-in electric field during the photo(thermal)catalytic DRM process.

2.3. Direct methane conversion

Direct methane conversion technologies, including POM, OCM, and NOCM, rely on converting methane into added-value products without the need of producing syngas as intermediate fuel.¹⁵² Thus, these technologies are especially appealing when compared to traditional indirect methane conversion approaches since reduce energy consumption, improve selectivity, and minimize greenhouse gas emissions.¹⁵³

2.3.1. Partial oxidation of methane (POM): mechanism and kinetics. Another major and well-established reaction route is the partial oxidation of methane (POM), which consists of the combustion of methane to produce syngas or methanol, which can be further utilized for the production of olefins, middle distillates, acetic acid, etc.¹⁵⁴ Thus, POM holds the ability to supply to the downstream fine chemical industry.¹⁵⁵ Different reaction mechanisms of POM via homogenous and heterogenous catalysts have been proposed. In detail, for homogeneous catalysts, the metal centres are always considered as active sites for C–H activation by several methods (e.g. oxidative addition, metal-radical activation, 1,2-addition over the metal centres, etc.).^156,157 On the other hand, in heterogeneous systems, there are two main transition statuses involving ˙CH₃ radicals or M–CH₃.¹⁵⁸ The CH₄ activation has two processes including dehydrogenation and deprotonation of C–H. In the dehydrogenation process, a hydrogen atom is separated from CH₄ by M–O or other active electrophilic oxygen atoms and then forms ˙CH₃ and ˙OH radicals as a transition state (Fig. 5).^156,159 In the deprotonation process, this mechanism involves the generation of CH₃⁻ and H⁺ by the dissociative adsorption of CH₄.^160,161 The CH₃⁻ coordinates to the catalyst surface and forms an M–C σ-bond via oxidative addition, electrophilic activation, and σ-bond metathesis. H⁺ interacts with the metal sites/O species in heterogeneous catalysts and molecules/ligands in homogeneous catalysts.^161–163 Under these conditions, the carbon centre will always maintain a formal oxidation state (FOS) showing a tetrahedral geometry with sp³ hybridization during the formation of reaction intermediates.¹⁶¹ The FOS of the carbon centre is not changed, confirming that the methane activation step is an acid-based reaction according to the deprotonation mechanism, suggesting that the deprotonation tends to occur in some homogeneous systems with highly polar solvents (e.g. H₂O or concentrated acids).¹⁵⁶ In addition to two aforementioned mechanisms, free radicals (e.g. ˙OH from the Fenton reaction) can initiate the POM reactions,¹⁶⁴ and plasma has also been used to oxidize CH₄ to various C₁ species.¹⁶⁵ The high stability and thermodynamic inertness of CH₄ also make the oxidation reaction an extremely difficult task. The conventional industrial thermocatalytic POM processes require high temperature (700–1000 °C) and pressure (2.5–3.5 MPa), which can cause the oxidation of catalysts and non-negligible risk of explosion due to the presence of O₂.¹⁵⁵ To overcome the possible hazards during POM, photocatalytic technology once again becomes a promising strategy once again to complete POM under mild conditions, reducing energy consumption, and largely avoiding coke formation and sintering.

Fig. 5 Dehydrogenation and deprotonation mechanism for CH₄ activation during POM reactions.

Several studies have discussed the kinetic model of POM over different catalysts. An early study elucidated the kinetics of the catalytic reactions over Pt catalysts,¹⁶⁶ where the apparent activation energy of CH₄ and O₂ conversion is only 10 kJ mol⁻¹, suggesting that the POM reaction is not temperature dependent, but completely determined by the transport behavior. Due to the excess of CH₄ in this system, the efficiency of oxygen consumption, which is strongly influenced by mass transfer to the surface, appears to have become the RDS. However, the selectivity of the products (e.g. CO and H₂) is strongly influenced by the temperature at the surface of the Pt catalysts.¹⁶⁷ In addition to the noble metal catalysts, a microkinetic model of the CH₄ oxidation reaction over PdO catalysts was proposed based on DFT results under atmospheric to moderate pressures, rare fuel, and low-temperature conditions.¹⁶⁸ The adsorption of activated CH₄ via dehydrogenation step over unsaturated surface Pd–O site pairs (denoted as Pd_cus–O_cus) is a RDS during the initial stage of the POM reaction. Another focus of this work was the strong water-inhibiting effect, which persists up to high temperatures.¹⁶⁹ The water inhibition could be considered as hydroxide formation leading to reduced availability of the Pd_cus–O_cus site pairs. Thus, the presence of water vapor interacting with the catalyst particles was one of the main factors causing sluggish kinetics and even catalyst deactivation. Another kinetic model of Ni-based catalysts was also developed by comparing the numerical simulations with the experimental data in a flow reactor.¹⁷⁰ The results showed that the initial temperature for CH₄ conversion to CO₂ and H₂O was 723 K. However, there was no significant syngas evolution up to 880 K. At temperatures above 880 K, the formation of H₂ and CO was increased, resulting in an equilibrium composition with CO₂ and H₂O. This behaviour indicated the indirect reaction pathway of syngas formation via the initial generation of H₂O and CO₂ during the CH₄ oxidation process.^171–173 Thus, the competition between the first and second steps determined the efficiency of the overall POM reaction, meaning that the CO₂ dissociation and CO desorption were the RDSs for CO evolution at high temperatures. This review only shows some representative studies, including noble metal-/metal oxide-/transition metal-based catalysts and their kinetic models, which determine the reaction rate. More kinetic models of POM and discussion of reaction pathways over different catalysts have been well demonstrated by other comprehensive works and are therefore not shown in this review.^174–176

2.3.1.1. Thermal catalytic POM. Various thermal catalysts using different oxidants for POM have been developed and reported over the last few decades. Early works used VO_x and MoO_x to convert CH₄ to formaldehyde (desired product) and methanol (by-product) at 450 °C.¹⁷⁷ The C–H bond activation over Pt and Pd sites has also been investigated via Periana's system using sulfuric acid (H₂SO₄) as the oxidant.¹⁵⁷ The H₂SO₄ oxidant stabilized the methyl bisulfate product so that Pt²⁺ complex-based catalysts achieved ∼90% conversion and 81% selectivity. In addition to the noble metal-based catalysts, under mild conditions, zeolites with transition metal centres (e.g. Cu and Fe) could easily dissociate C–H bonds and were therefore developed for the oxidization of CH₄ with N₂O or O₂ as oxidant (Fig. 6a).^178,179 Metal zeolite was activated in O₂ at 450 °C but reacted with CH₄ without O₂ at 200 °C. This stepwise process avoided the over-oxidation of CH₄ at high temperatures (the first step) and thus achieved over 90% CH₃OH selectivity in the second step (Fig. 6b). In addition, metal oxide-based supports have also been extensively studied for POM reactions. Some reports showed the real implementations of oxide-based thermal catalysts such as SiO₂, Al₂O₃, MgO, ZrO₂, and CeO₂.^180–184 For example, the Ni-loaded CeO₂ catalysts were suitable for long-term POM at high temperatures, excellent redox capacity, and coke resistance behaviour. Bal et al. reported Ni/CeO₂ catalyst, which could activate and totally convert methane at 400 and 800 °C, respectively.¹⁸⁵ Due to the cooperation between Ni and CeO₂, Ni/CeO₂ could resist coke and produce syngas (H₂ : CO = 2, Fig. 6c). Besides CeO₂, iron oxides were also investigated as oxygen carrier candidates,¹⁸⁶ but the oxygen species in iron oxide-based supports were too active to avoid the overoxidation of CH₄ to CO₂ and H₂O.¹⁸⁷ In addition, tungsten oxides (WO₃) showed their superior stability and selectivity for specific products (e.g. syngas).¹⁸⁸ However, one of the main challenges for WO₃ thermal catalysts has been poor reducibility, suggesting that a high temperature is required to utilize the lattice oxygen in WO₃ and initiate reactions.¹⁸⁹ Despite the above advances in POM applications, the high stability and thermodynamic inertness of CH₄ also make the oxidation reaction an extremely difficult task. Conventional industrial catalytic POM processes require high temperature (700–1000 °C) and pressure (2.5–3.5 MPa), which can cause oxidation of catalysts and non-negligible explosion risks due to the presence of O₂.¹⁵⁵

Fig. 6 (a) Active site in enzymes and zeolites with Fe or Cu atoms. (b) Conceptual activation of CH₄ at a high-spin high-valent metal-oxo species.¹⁷⁸ Copyright © 2018, American Chemical Society. (c) Selectivity of Ni–CeO₂ during POM reaction.¹⁸⁵ Copyright © 2016, Elsevier B.V.

2.3.1.2. Photocatalytic POM. Many conventional semiconductor photocatalysts, such as TiO₂, WO₃, BiVO₄, etc., have shown remarkable activity in CH₄ oxidation.^80,190,191 For example, Morante's group investigated the selectivity of BiVO₄ photocatalyst in the presence of nitrite ions.¹⁹² The low concentration of nitrite was beneficial for the selective oxidation of CH₄ to CH₃OH. By reducing the free ˙CH₃ radicals, nitrite as a ˙OH scavenger limited the subsequent undesirable oxidation process of CH₃OH, thus greatly reducing the formation of CO₂. Meanwhile, the nitrite ions could become a UV-filtering species, inhibiting water photolysis, which ultimately forms ˙OH. In this way, more than 90% selectivity towards CH₃OH could be achieved over BiVO₄ (Fig. 7a). As an ultrawide-bandgap semiconductor, strontium tantalate (Sr₂Ta₂O₇) with Pd nanoparticles has also been used to perform POM reactions under low-temperature conditions.¹⁹³ In this composite, O₂ could accept the electrons from Sr₂Ta₂O₇ to form radical species. In addition, the layered perovskite structure made Sr₂Ta₂O₇ has high stability for immobilizing Pd nanoparticles (Fig. 7b). The hot electrons from Pd nanoparticles facilitated the activation of O₂ and CH₄. The photothermal effect was triggered by hot carrier relaxation. Thus, light illumination lowered the initial temperature of POM reactions to below 423 K and improved the catalytic performance to below 873 K to produce syngas. Other previous research has also demonstrated the excellent ability of photocatalysis to reduce the large heat energy during the POM process over various semiconductor-based photocatalysts such as g-C₃N₄, TiO₂, and CuMoO₄/SiO₂.^94,194 In addition, the recombination between electron–hole pairs is an unavoidable defect in the photocatalytic process. Compared to some photocatalytic reactions with simple products (e.g. water splitting, organic degradation), the selectivity of POM products is also a major challenge due to the diversity of products. In this case, ferroelectric materials show their ability to overcome these problems.

Fig. 7 (a) Selectivity for POM towards CH₃OH product. Reaction conditions: 450 W Hg lamp, 20% CH₄ in He (∼22.4 mL min⁻¹), 55 °C.¹⁹² Copyright © 2015, Royal Society of Chemistry. (b) Band alignment mechanism in a metal NP/semiconductor system.¹⁹³ Copyright © 2019, Wiley.

2.3.2. Oxidative coupling of methane (OCM): mechanism and kinetics. The OCM technology is a new technology that directly converts methane to olefins without an intermediate process.^8,195 The OCM has relatively low-temperature reaction conditions and cleaner product distribution, so it has recently attracted wide attention for industrial production.¹⁹⁶ The OCM is generally considered to be a mixed catalytic and gas-phase process.^197–201 In principle, the reaction pathways of OCM involve three steps on the catalyst surface (Fig. 8).¹⁹⁶ First, CH₄ reacts with O₂ present in the feed to form radicals on the catalyst surface, which is a heterogeneous step. This is followed by a homogeneous step, during which is coupled to produce ethane (C₂H₆). Finally, the production of ethylene (C₂H₄) is achieved by subsequent dehydrogenation of C₂H₆. During these steps, one of the difficulties is the energy required to dissociate the strong C–H bond (∼438 kJ mol⁻¹), which is also present in other CH₄ conversion technologies, suggesting that a sufficiently low activation energy barrier is required for the generation of radicals and the subsequent coupling or dehydrogenation process.²⁰² Overcoming the thermodynamic barriers often requires the introduction of O₂ and high temperatures, which leads to the production of unwanted products, such as CO₂ and graphitic carbon.⁸⁵ In addition, C₂H₄ is more unstable than CH₄, so consecutive oxidation of C₂H₄ is also a major challenge.^203,204 Moreover, it is difficult to achieve efficient CH₄ conversion while maintaining high selectivity for C₂ products, because the complete and partial oxidation of methane to CO_x and the combustion of C₂H₆ and C₂H₄ can occur simultaneously. Thus, the low selectivity and evolution rates of C₂ products are major challenges for large-scale utilization.

Fig. 8 OCM reaction steps in the gas phase and over the catalyst surface.

The actual reaction steps can be more complex than the theoretical three basic steps because both gas phase and catalyst surface reactions should be considered simultaneously to discuss the kinetic models and the RDSs.^205–207 However, the currently proposed kinetic models are still incomplete and not able to fully explain the obtained experimental data.^208,209 However, two typical assumptions based on the absence and presence of catalysts have been studied, leading to the development of many kinetic models in different catalytic systems. A representative kinetic scheme reflecting the OCM reaction under gas-phase conditions without heterogeneous catalysts has been demonstrated.²¹⁰ According to this kinetic model, the time dependence of the compound concentrations could be calculated, thus determining the product selectivity as a function of the degree of conversion. Specifically, when there was an excess of CH₄ in the gas phase, the concentration of active oxygen species interacting with to form carbon oxides was low, resulting in a high selectivity for C₂ hydrocarbon products. In contrast, under high O₂ concentration conditions (e.g. CH₄ : O₂ = 2), the formation of oxygen-containing radicals such as and ˙OH was promoted, leading to the low evolution rate of C₂ hydrocarbon products. The authors summarized the optimum reaction conditions for achieving the maximum selectivity of C₂ products. The reaction temperature should be 1000–1100 K, and the pressure ca. 10 bar while the ratio of CH₄ to O₂ was not less than 5–7. A total pressure higher than 10 bar was advantageous to produce the primary product C₂H₆ instead of the desired C₂H₄. A lower ratio between CH₄ and O₂ would make the non-selective reaction dominant. Based on the above study, further work has been carried out through dedicated experimental measurements, regression, sensitivity, and parameter contribution analyses were carried out to explore the optimal OCM reaction conditions.^{198,211–213} An interesting paper by Simoní Da Ros and co-workers summarizes some details of these works.²⁰⁸ In short, the updated kinetic model of OCM gas phase reaction systems has contributed to the optimization of the reaction parameters, and thus the operation of the product distribution.

For the kinetic model in the presence of catalysts, Stansch et al. described a 1D homogeneous plug flow model comprising 10 reaction steps according to the experimental data over a La₂O₃/CaO catalyst with high activity and selectivity towards C₂ hydrocarbons in a packed bed reactor.²⁰⁷ An important assumption in this kinetic model was the small effect in gas phase reactions, meaning that only a few cases where the gas phase reaction can be neglected, such as short contact time and low gas volume, were suitable for this model. Despite this drawback, the dynamics of the oxidative reaction were successfully represented by Hougen–Watson-type rate equations (eqn (3-1)). The power law rate equations determined the rates of dehydrogenation, C₂H₄/H₂O reforming, and the WGSR (water–gas shift reaction, eqn (3-2) to (3-5)). The effects of O₂ and CO₂ on enhancing C₂H₄ selectivity via inhibiting C₂H₆ were also described (eqn (3-6)).

r₁ = k_0,1 e^−E_a,1/RT p_C2H6 (3-2)

r₂ = k_0,2 e^−E_a,2/RT p_C2H4^m₂p_H2O^n₂ (3-3)

r₃ = k_0,3 e^−E_a,3/RT p_CO^m₃p_H2O^n₃ (3-4)

r₄ = k_0,4 e^−E_a,4/RT p_C2H4^m₄p_H2O^n₄ (3-5)

where E_a,L is the activation energy (J mol⁻¹), ΔH_ad,CO2 is the adsorption enthalpy for CO₂ (J mol⁻¹), ΔH_ad,O2 is the adsorption enthalpy for O₂ (J mol⁻¹), k_0,L is the pre-exponential factor, K_0,L is the adsorption constant (Pa⁻¹), n is the number of experiments, n_j is the reaction order, p is the partial pressure (Pa) and R is the gas constant (J mol⁻¹ K⁻¹). The author emphasized that the proposed kinetic model could be used for all reaction steps over other OCM catalysts with similar properties involving basic carrier and non-reducible catalytic components. According to the above kinetic model, other updated models considering the gas-phase reactions were proposed later.^214–217 These works included at most 39 gas-phase reactions, which played important roles in explaining the selectivity of products, the inhibiting effect of excess O₂, and the prediction of OCM reactors.

2.3.2.1. Thermal catalytic OCM. Considerable efforts have been made to develop catalysts for OCM to C₂ hydrocarbons. For thermal catalysts, an early result was reported by Ito et al., who developed a lithium-doped magnesium oxide (Li/MgO) catalyst that had a high activity for the formation of ˙CH₃ radicals in the presence of O₂, resulting in high selectivity for C₂ compounds (C₂H₄ and C₂H₆) and 28% conversion of CH₄.²¹⁸ Cu- and Fe-contained zeolite thermal catalysts have been widely studied with their dehydrogenation mechanism. For example, the Cu-/Fe-based catalysts have been used for OCM using N₂O or H₂O₂ as strong oxidants.^219,220 The Cu–O–Cu sites have been confirmed as the main active platform to activate CH₄ via hydrogen abstraction, convert ˙CH₃ and ˙OH radicals to CH₃OH, and avoid catalyst inactivation by rapid desorption of products (Fig. 9a).²²¹ Recently, IrO₂ has been shown to be an effective catalyst for OCM application. Nørskov's group claimed that Ir coordinatively saturates with surface oxygen,¹⁵⁹ which could be considered as an oxygen-promoted surface contributing to the high activity via the dehydrogenation process.²²² However, another research with the opposite idea suggested that the (110) facet over IrO₂ is able to dissociate C–H at low-temperature via the deprotonation mechanism. In addition to the controversial reaction mechanism, single IrO₂ was not suitable for long-term OCM because overoxidation was unavoidable, resulting in undesired CO and CO₂. Besides IrO₂, other common thermal catalytic systems with the deprotonation process are noble metal-contained catalysts, which always exist in a low oxidation state so that proton acceptors can be accessible.^156,223,224 Taking Pt as an example, an organic ligand-stabilized Pt^II complex was designed for the conversion of CH₄ into methyl bisulfate (CH₃OSO₃H).¹⁵⁷ Using H₂SO₄ as a strong oxidant, this complex achieved 90% conversion of CH₄ and 81% selectivity of CH₃OSO₃H at 220 °C. Mechanistic studies showed that platinum-methyl, the intermediate of methyl ester, was produced via C–H activation of CH₄ over Pt^II. And Pt^II was the most active oxidation of Pt for OCM (Fig. 9b). Other works on homogeneous Pt-based catalysts reported a similar mechanism, in which Pt oxidized to Pt^II active sites and Pt^IV–CH₃ bond in an orderly manner. Finally, the reductive elimination produced CH₃OH and regenerated Pt^II active sites.^225–227 Although recent advances have been made in traditional thermocatalytic OCM, high temperatures (typically >600 °C) are always required for reactant activation, resulting in undesired products such as CO_x due to the high energy output property (exothermicity) of hydrocarbon combustion.²²⁸ Meanwhile, the stability of OCM catalysts is also an inevitable issue due to the excessive heat energy released during the reaction.

Fig. 9 (a) Transformation mechanism of CH₄ into CH₃OH on the Cu/Na-ZSM-5 zeolite.²²¹ Copyright © 2019, American Chemical Society (b) OCM mechanism.¹⁵⁷ Copyright © 1998, The American Association for the Advancement of Science.

2.3.2.2. Photocatalytic OCM. Photocatalysis is a promising approach to achieving OCM reactions in mild conditions.^21,28,229 However, studies of photocatalysis technology for realizing OCM are rare. Thus, in this section, a few noticeable and representative works will be summarized to display the great potential of photocatalytic OCM application. For instance, Tang's group designed a Pt/CuO_x-loaded TiO₂ photocatalyst with a 60% selectivity of C₂H₆ product in a gas flow reactor (Fig. 10a).⁸⁵ In this nanohybrid, Pt as an electron acceptor facilitated the charge separation, while CuO_x was an active site for photoexcited holes. Thus, deep dehydrogenation and overoxidation could be effectively avoided because of the low oxidation potential of photoinduced holes over the VB of CuO_x clusters (Fig. 10b). A similar work to Pt/CuO_x-loaded TiO₂ photocatalyst was an Au–ZnO/TiO₂ photocatalyst with an ethane production rate of ∼5000 μmol g⁻¹ h⁻¹ and 90% selectivity (Fig. 10c).²²⁸ These values are the highest efficiency in the previously published works. The modification of Au and TiO₂ not only accelerated the charge transfer but benefited the O₂ activation owing to the low adsorption energy and activation barrier at oxygen vacancies (Fig. 10d and e).²³⁰ The comprehensive investigation demonstrated that Au as a cocatalyst promotes the formation of ˙CH₃ radicals and inhibits the overoxidation of CH₄ to CO₂ through the OCH₃ intermediate (Fig. 10f–g). Another representative work was reported in 2019 by Liu et al.²³¹ A plasma-assisted photocatalyst was developed for the oxidation of CH₄ to C₂H₄ and CO via a new reaction route with CH₄ and CO₂ as reactants. Experimental analyses indicated that CO₂ and CH₄ tend to adsorb on TiO₂ and Ag surfaces, separately. Because of the localized surface plasma resonance (LSPR) effect on Ag and photoelectric effect in TiO₂, under a mild condition (simulated solar irradiation, 84.2 mW cm⁻²), the evolution rates of CO and C₂H₄ were 1149 and 686 μmol g⁻¹ h⁻¹, respectively. Apparently, the above studies successfully elucidate the great capacity of photocatalysis for OCM to obtain C₂ products with high selectivity.

Fig. 10 Copyright © 2020 Wiley. (a) C₂ production and selectivity of photocatalytic OCM. Reaction conditions: O₂/CH₄ = 1 : 400, GHSV = 2400 h⁻¹, 40 W 365 nm LED 40 W, 40 °C.⁸⁵ (b) Photocatalytic OCM process.⁸⁵ (c) Photocatalytic OCM activity. Reaction conditions: 69 mL min⁻¹ of CH₄ and 1 mL min⁻¹ of 20% O₂/N₂, 300 W Xe lamp. Theoretical energy diagrams for (d) oxidizing CH₄ to C₂H₆ or CO₂ and (e) two key competing pathways. Proposed reaction process on (f) Au–ZnO and (g) Pt–ZnO.²²⁸ Copyright © 2021 Springer Nature.

2.3.3. Non-oxidative coupling of methane (NOCM): mechanism and kinetics. Since the C–C coupling reaction is able to produce more useful products than simple CO products, NOCM has become one of the main CH₄ conversion routes. Compared to the OCM, NOCM is more attractive since no oxidizing agent is necessary during this process. Meanwhile, NOCM can avoid too active oxidative products, which are often generated by the OCM process, and get high selectivity to desired products like H₂. The NOCM reaction mechanism is simpler compared to the above POM and DRM reactions. In brief, the activation process of CH₄ is direct during NOCM, during which the CH bond is dissociated. The absorbed H species was reduced to H₂, and two methyl radicals coupled with each other to form C₂H₆.

Several microkinetic models during NOCM have been proposed based on different kinds of catalysts. Dumesic's group studied the kinetic model of NOCM on PtSn surfaces with SiO₂ or H-ZSM-5 supports via DFT calculation.²³² Their model predicted that the full Pt surfaces adsorbed CH and C fragments and C–C coupling had the highest rate during the whole NOCM process. The RDS was the CH₄ activation and product (H₂) desorption over Pt species. Besides, the desorption of C₂H₆ as another main product also influenced the whole reaction rate to a large extent.

Kopyscinski et al. built another kinetic model based on a metal-free GaN/SBA-15 catalyst.²³³ Their kinetic model was proposed via a simple 1D model with the same observed rate as the forward rate, reflecting the reaction process away from the chemical equilibrium and without mass transfer limitation. The kinetic studies and also the isotope labelling experiments confirmed that the possible reaction pathway should be through a fast transformation from CH₄ to . Subsequently, the 2nd step was slow bond cleavage of C–H, forming and finally coupling to C₂H₆. However, the author believed that the RDS should be the conversion step from to instead of the H–CH₃ bond cracking since the apparent order of NOCM was ∼0 in CH₄, indicating a fast adsorption process. More importantly, the authors claimed that active sites always fully adsorbed species. Their idea might be a basis for understanding and developing other metal-free catalysts.

With the appearance of single-atom catalysts (SACs), lots of SACs have been used for NOCM so some related kinetic studies have been conducted. For instance, Pohar's group explored the micro-kinetics of NOCM to C₂H₆ over Pt/CeO₂ catalyst.²³⁴ In the SAC system, the RDS step was similar to the above metallic catalysts, in which the abstraction of the first H˙ from a CH₄ molecule and the sorption/desorption of the H₂ product. However, one of the deficiencies in this model was the authors assumed the reactions only performed on Pt sites. Some works have proved that the active sites could also be the interface between Pt with CeO₂.²³⁵ In the future, the synergistic effects between CeO₂ support and Pt sites should be accounted for. Kim et al. also studied the kinetic model on SAC containing Fe single-atom.²³⁶ Similarly, they also claimed the formation of limited the whole NOCM efficiency. More importantly, they emphasized in the case of production, the activation barriers of CH₄ needed abundant conditions. However, the C–C coupling required higher hydrogen transfer rates on carbon-attached structures under deficient conditions. This conflict suggested the optimal conditions for these two steps (CH₄ conversion and C–C coupling) were difficult over a single catalyst. Therefore, the above works showed that there is still no perfect kinetic theory explaining all the catalytic NOCM systems. It is urgent to develop more theories and models to supply this area.

2.3.3.1. Thermal NOCM. The high thermodynamic barrier for NOCM means the high energy is essential to C–C coupling. Therefore, thermal catalytic technology is one of the common strategies to realize the NOCM reaction.^237,238 Since the NOCM was first reported by Wang's group,²³⁹ different kinds of thermal catalysts have been designed and prepared for highly efficient and selective NOCM implementation. Metal- and alloy-based catalysts are one of the mainstream materials for NOCM. For instance, Soulivong's group reported a SiO₂-supported HTa catalyst with an ethane selectivity of 98%. However, in this work, the CH₄ conversion was less than 0.5%, which was not enough for industrialization.²⁴⁰ Similarly, Guo et al. realized CH₄ conversion of 48% with C₂H₄, C₆H₆, and C₁₀H₈ with carbon selectivity of 53%, 22%, and 25% on Fe/SiO₂ catalyst at 950 °C.²⁴¹ The high CH₄ conversion rate and good selectivity are attractive, but how to separate such diverse products seemly will be another problem. Varma prepared Pt–Bi bimetallic catalyst supported on ZSM-5 zeolite for NOCM, achieving more than 90% selectivity at 873 K, suggesting the potential of polymetallic or alloy systems for NOCM.²⁴² Recently, due to the rapid development of molybdenum titanium carbide (MXene) family materials, noble metals presented by Pt have been anchored on the MXene substrate.²⁴³ It has been confirmed that the metal/MXene system could activate the C–H bond in CH₄ molecular, then forming methyl radicals, which favoured desorption over subsequent dehydrogenation, so that coke deposition could be suppressed. In this case, the stability and CH₄ conversion rate could be enhanced.

Although so many works have been reported in the thermal catalytic NOCM field. As mentioned before, the inevitable disadvantage of NOCM, the high-temperature conditions, still cannot be avoided, which causes large energy consumption, low selectivity, and short longevity of thermal catalysts. Hence, the photocatalytic NOCM at mild conditions attracts more and more attention for exploring a new NOCM route and thus achieving carbon neutrality.

2.3.3.2. Photocatalytic NOCM. Photocatalytic NOCM can convert CH₄ to C₂H₆ and H₂ using light energy instead of thermal energy, thereby reducing the reaction temperature and thus avoiding catalyst coking or deactivation. In addition, the energy barrier for CH₄ activation can be lowered by photocatalysts. In 1998, Yoshida's group first reported the photocatalytic NOCM reaction over SiO₂/Al₂O₃ composite.²⁴⁴ The SiO₂/Al₂O₃ photocatalyst exhibited the highest CH₄ conversion rate of 0.328 μmol h⁻¹ under UV light irradiation. This study proved that the energy transfer from active sites (highly surface-dispersed Al species) to absorbed CH₄ might be the crucial step for CH₄ activation. Subsequently, a series of silica-based, metal oxide-based, and zeolite-based catalysts have been widely investigated in the photocatalytic NOCM field.^89,245–247 Xiong et al. reported that the incorporation of Pd single atoms (SAs) into the lattice could overcome the limitations since the effect of O sites in VB was reduced.²⁴⁸ In this work, TiO₂ was used as a proof-of-concept model. DFT calculations suggested that the TiO₂ (001) facet was the best surface for the SAs modification because of the loose arrangement of O atoms that can form the M–O bond (Fig. 11a). In order to demonstrate the activity and selectivity toward CH₄ to C₂H₆ conversion, a series of metal SA-modified TiO₂ were prepared for photocatalytic NOCM (Fig. 11b). The highest activity and selectivity were observed for Pd SA-loaded TiO₂. Therefore, the Pd₁/TiO₂ photocatalyst was further investigated. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) revealed the atomically dispersed Pd elements on the TiO₂ support (Fig. 11c). The coordination of Pd was investigated by synchrotron radiation-based X-ray absorption near-edge structure (XANES) spectroscopy (Fig. 11d). Only a single peak at 1.53 Å could be detected on Pd/TiO₂ corresponding to the coordination between Pd and O (Pd–O), indicating the presence of isolated Pd atoms in the photocatalyst. Thus, the coordination environment of Pd–O₄ was formed through Pd SA binding with O atoms in TiO₂ crystalline. Under ambient conditions with Xe lamp illumination, the photocatalytic NOCM results showed pristine TiO₂ nanosheets exhibited a very low production rate of C₂H₆ (Fig. 11e). This could be attributed to the involvement of the lattice oxygen in CH₄ conversion, resulting in a susceptible overoxidation to produce CO₂ (Fig. 11f). In contrast, Pd/TiO₂ exhibited a very high C₂H₆ production rate during the NOCM reactions. The improvement of the photocatalytic activity was due to Pd active sites instead of the lattice oxygen in TiO₂, which suppressed the overoxidation of CH₄ and accelerated the desorption of C₂H₆. To investigate the role of Pd SAs in CH₄ dissociation and intermediate coupling via photogenerated holes, the energy barrier of the CH₄–CH₄ coupling on the active sites of Pd/TiO₂ was studied by DFT calculation (Fig. 11g–h). Under dark conditions (Uh = 0.00), CH₄ required higher energy for activation. Meanwhile, 1.74 eV of activation energy from CH₄ to TS state was so high that hindered the whole NOCM. However, photogenerated holes would provide additional potential, so that the energy barrier of the step turns down (IM3 to IM4). In addition, the energy barriers of C–C coupling and C₂H₆ desorption via Pd SAs were also lowered. Thus, the photoexcited charge and the atomic-level reaction platform were advantageous for triggering the photocatalytic NOCM reaction with lower energy input and mild conditions. This work suggested the great advantages of photocatalysis in the activity and selectivity of NOCM.

Fig. 11 (a) Pristine TiO₂, Pd₁/TiO₂, and Pd_n/TiO₂ models. (b) The metal contribution in VBM and photocatalytic NOCM performance. Reaction conditions: 0.1 MPa CH₄, 300 W Xe lamp, room temperature. (c) STEM image and (d) Pd K-edge EXAFS spectra in R space of Pd₁/TiO₂. (e) The C₂H₆ production rates (f) and the corresponding selectivity of C₂H₆, C₂H₄, and CO₂ in 0.1 MPa CH₄. (g) Free-energy diagrams for C₂H₆ production on Pd₁/TiO₂. (h) Coordination environment during the DFT simulation.²⁴⁸ Copyright © 2022, Springer Nature.

3. Key role of ferroelectric materials in photocatalytic reactions

3.1. Ferroelectric materials

Ferroelectricity was first discovered in rock salt in 1921.²⁴⁹ The ferroelectric material is one of the piezoelectric materials, which belongs to pyroelectrics and then dielectric materials in sequence. Unlike pyroelectric and piezoelectric materials, ferroelectrics have spontaneous dipole moments without an external heating source or applied stress activation. The dipole moments generate spontaneous polarization electric fields, which offer a sustainable driving force for the separation and transfer of charge carriers from bulk to the surface of catalysts. In general, ferroelectric materials can be classified into three main families, viz. inorganic, organic, and organic–inorganic hybrids. What follows summarizes the most representative ferroelectric-based photocatalysts utilized for CH₄ conversion approaches.

3.1.1. Inorganic ferroelectric materials.
3.1.1.1. Ferroelectric oxides. According to different structures, ferroelectric oxides have four basic types: perovskite, pyrochlore, tungsten-bronzes, and layer-type bismuth compounds (also called Aurivillius).²⁵⁰ Among them, perovskite possesses higher polarization strength. Therefore, they are widely exploited in photocatalysis and photovoltaics. In ferroelectric oxides with perovskite structure, the ferroelectricity is from the electron clouds hybridization mechanism between adjacent ions, called displacive ferroelectricity.²⁵¹ The photovoltaic effect in ferroelectric perovskite oxides was discovered in the 1950s and still gaining increasing attention.²⁵² The homogeneous bulk region in ferroelectric semiconductors with inherent non-centrosymmetry could generate photocurrent and even photovoltage beyond the bandgap.^253,254 Therefore, the photovoltage from ferroelectrics depended on the energy band structure instead of the built-in potential induced by p–n junctions or Schottky junctions at the interface.⁵⁵ This property could be not only used in photovoltaics but also extended to photocatalysis. In fact, several studies have applied photoinduced charges and polarization from ferroelectric perovskite oxides to improve photocatalytic efficiency.^255–257 These works confirmed that perovskite oxides were effective in both photocatalytic reduction and oxidation reactions, providing experimental support for fully utilizing ferroelectric oxides in various photocatalytic fields. More details about published work in photocatalytic methane conversion fields will be further introduced in the following sections. Currently, the main deficiency in conventional ferroelectric oxide perovskites is the narrow light responding range and unsatisfactory conductivity, limiting the photovoltaic and photocatalytic efficiency. In this case, the bandgap engineering promotes the development of ferroelectric perovskite into organic–inorganic halide perovskites as shown below.

3.1.2. Organic ferroelectric materials.
3.1.2.1. Liquid crystals (LCs). LCs are a crucial part of the family called soft condensed matter materials. Thermodynamically, the state of LCs is located between an isotropic liquid and a crystalline phase.²⁵⁸ They are partially ordered but have anisotropic physical properties, such as electric conductivity or magnetic susceptibility while maintaining flowability.²⁵⁹ The mainstream methods for producing LCs are based on thermotropic phases and lyotropic phases, which are achieved by changing the intensive variable of state and the dopant concentration in the isotropic solvent, respectively. The ferroelectric properties in the LC family should be attributed to the chiral smectic C (SmC) phase, which is one of the molecular arrangements in LCs (the other two structures are nematic and smectic A phase).²⁶⁰ The chiral SmC is similar to SmC structure, but owing to the chiral properties, Meyer et al. presented the idea that a chiral smectic C mesophase should be ferroelectric, which has been experimentally confirmed in 1975 via reporting practical ferroelectric LCs.²⁶¹ Owing to the particular chiral SmC phase with helical structure, ferroelectric LC shows spontaneous polarization and molecular switching in microsecond order. These attractive properties lead to wide applications in modern photonic fields including grating,²⁶² lensing,²⁶³ beam steering,²⁶⁴ etc. Unfortunately, the aforementioned promising advantages of ferroelectric LCs are still not utilized in the whole catalytic field. Considering the fact that more and more studies tried to regulate the ferroelectricity/ferromagnetism in LCs,^265,266 the ferroelectric LC family could be a potential candidate for achieving efficient photocatalytic implementations.
3.1.2.2. Ferroelectric polymers. In addition to ferroelectric LCs, the organic ferroelectric solid is another great family belonging to the class of organic ferroelectric materials. In this family, ferroelectric polymer is one of the most widely developed organic ferroelectrics and has been used in organic solar cells,²⁶⁷ ferroelectric memory,²⁶⁸ piezoelectric devices,²⁶⁹ and electrocaloric refrigeration devices.²⁷⁰ Among them, poly(vinylidene fluoride) (PVDF) and corresponding copolymers/terpolymers are the most representative ferroelectric polymers. PVDF and the co-/ter-polymers had high dielectric constant so that they are regarded as promising thin film-type materials for high-energy-density capacitors.²⁷¹ In the photocatalytic field, PVDF is also developed as the source of ferroelectric fields, which are used to improve charge-carrier transfer behaviours. In brief, through the ferroelectric effect, PVDF-based photocatalysts have promoted different photocatalytic redox reactions, such as overall water splitting,^38,272 CO₂ photoreduction,²⁷³ and advanced oxidation.²⁷⁴ Although there is no work on PVDF-based photocatalytic systems for methane conversion implementation, in consideration of the great ability of PVDF and other ferroelectric polymers, organic ferroelectric polymers should be applied as substrates or cocatalysts for controlling charge migration pathways in future studies.

Although the low-molecular-mass compound is also a kind of organic ferroelectrics, since it is difficult for the steric to reorient dipoles, causing an extremely larger coercive field than that of other ferroelectrics, the number of the low-molecular-mass compound with ferroelectricity is rare so that the real-implementation of the low-molecular-mass compound is limited.²⁷⁵ Therefore, the low-molecular-mass compound represented by thiourea,²⁷⁶ the 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) free radical,²⁷⁷ and 1,6-bis(2,4-dinitrophenoxy)-2,4-hexadiyne²⁷⁸ are not listed as an independent sub-section.

3.1.3. Organic–inorganic hybrid ferroelectric materials.
3.1.3.1. Ferroelectric metal–organic complex (MOC). Since Valasek reported the ferroelectrics in Rochelle salt,²⁷⁹ tremendous efforts have been made to design and fabricate novel molecular-based ferroelectric materials. Although the ferroelectric oxides family is well-known due to their strong intrinsic ferroelectric effect, they are heavy, weak mechanically flexible, and not environmentally friendly.²⁸⁰ Owing to the ferroelectricity of organic molecules, organic ferroelectric materials gradually enter researchers' version as a new generation of ferroelectric material. If a metal–organic complex (MOC) can show ferroelectricity via modulating dipole moments of its components, this MOC belongs to the family of organic ferroelectric materials. Further, according to their crystal structures, ferroelectric MOC can be divided into (i) discrete (0D), (ii) 1D chain, (iii) 2D layer structure, and (iv) 3D framework. It is true that several ferroelectric MOCs have been reported by molecular design based on the addition of chiral/polar molecules/ions or structurally flexible ligands or ions, but the ferroelectric MOC is still at the early step since the number of developed MOCs is not abundant. Some works have reviewed the advances of ferroelectric MOC and also predicted their potential real-applicat ion. Detailed descriptions can be seen in these excellent review articles.^280,281 However, in the catalytic field, studies about ferroelectric MOC are scarce, meaning that it is urgent to perform some theoretical studies based on thermodynamic theories and crystallographic requirements to offer fundamental support for exploring and developing ferroelectric MOC in photocatalytic areas.
3.1.3.2. Organic–inorganic perovskites. To a large extent, the prosperity of basic theory and practical applications in the field of ferroelectric materials should be attributed to the conventional inorganic ferroelectric perovskites with high stability and tunable ferroelectric properties.²⁸² Subsequently, molecular ferroelectrics are flexible, cheap, and environmental-friendly, and also have biocompatibility and homochirality so the appearance of molecular ferroelectrics attracted our version, causing a novel family of hybrid organic–inorganic perovskite ferroelectric (HOIPF) materials.^283,284 Compared to traditional inorganic ferroelectrics, HOIPFs have larger piezoelectricity coefficients, lower toxicity, and controllable chemical structures although their ferroelectricity is weaker. Besides, the ferroelectric quantum theory and Berry phase calculation also clarify their potential in semiconductor areas.²⁸⁵ Semiconductor properties transform HOIPFs into a new generation of photonic device platforms.²⁸⁶ For instance, some Pb-based or lead-free HOIPFs had tunable bandgap and high mobility similar to common inorganic ferroelectric semiconductors.^287,288 A series of functional HOIPFs including (cyclohexylaminium)₂PbBr₄,²⁸⁹ (R)- and (S)-3-(fluoropyrrolidinium)MnBr₃ (ref. 290) possess outstanding luminescent properties, such as high-purity luminescence, wider luminescence range, and high photoluminescence efficiency.²⁹¹ Besides, the HOIPFs-based optoelectronic applications (e.g. bulk photovoltaic and self-powered photodetection) are also gradually explored with tremendous efforts on the photoresponse effects from HOIPFs and multifunctional hybrid ferroelectric systems.^292,293 Although HOIPFs still face some challenges like mechanical weakness, low stability, and short resistivity, the disadvantages do not stop the expansion of the applications of HOIPFs towards photocatalytic fields. Actually, some lead-free halide perovskites (e.g. Cs₂AgInCl₆ and Cs₃BiBr₆) have been used for photocatalytic advanced oxidation reactions,^294,295 photoreduction reactions,²⁹⁶ and even photocatalytic organic synthesis.²⁹⁷

Considering the number of practical examples of ferroelectric photocatalysts developed for methane conversion is very limited, it is necessary to emphasize the great effect of ferroelectric-assisted photocatalytic systems in improving methane conversion efficiency. In subsequent sections, the representative works about aforementioned ferroelectrics, especially ferroelectric perovskite and ferroelectric oxides, for methane conversion implementations will be introduced for further advancing the ferroelectrics-based catalytic community.

3.1.4. 2D ferroelectric materials. Although the traditional metal oxide-based ferroelectric materials have displayed stable spontaneous polarization, the high demands of micro/nano devices might result in depolarization and dangling bond problems during the reducing thickness process and finally eliminate the ferroelectricity. In order to deal with this issue, the exploration of 2D materials with stable ferroelectric properties attracts wide interest in the Material Science community. Especially, monolayered 2D materials open a new direction for developing 2D ferroelectrics. Recently, single-layered Sn–Te²⁹⁸ and α-In₂–Se₃ (ref. 299) have been confirmed as ferroelectric materials owing to their out-of-plane polarization under 270 K and ambient temperature, respectively. Through theoretical prediction, oxidized MXene family showed switchable polarization via the intermediate antiferroelectric phase.³⁰⁰ Some research also indicated the potential ferroelectricity of monoelemental 2D materials akin to graphene (known as 2D Xenes).³⁰¹ For instance, it has been forecasted that the chemical-functionalized phosphorene has ferromagnetism and ferroelectricity simultaneously,³⁰² however, the poor stability and high surface energy of 2D Xenes seemly become the challenge to realize this prediction. The quantum orders and pristine crystal structures of 2D Xenes should be well reserved when functionalizing them towards ferroelectrics. Fortunately, some works experimentally achieved the implantation of ferroelectricity into 2D Xenes, including CrI₃ (ref. 303 and 304) and CrGeTe₃,³⁰⁵ encouraging further investigation in this area. Beyond them, Kou et al. proposed a possible multiferroic or concurrent ferroic and topological order based on the CH₂OCH₃-functionalized germanene (2D-Ge).³⁰⁶ The p_z orbits of Ge were the resource of ferroelectricity. And the switchable polarization resulted from the electric field-induced ligand molecules (i.e. CH₂OCH₃) rotation. Considering the potential use of 2D-Ge derivatives in several fields, including electrocatalysis,³⁰⁷ biosensing,³⁰⁸ and (opto)electronics,³⁰⁹ 2D-Ge might be a suitable candidate to act as a ferroelectric-assisted system for photocatalytic energy conversion reactions.

3.2. Key role of ferroelectric materials in photocatalytic reactions

Ferroelectric materials play a crucial role in photocatalytic reactions owing to their inherent domain walls (DWs), which arise from breaking-symmetry. These DWs influence charge transfer,³¹⁰ local surface potential,³¹¹ and even surface adsorption/activation ability,³¹² all of which are essential for improving photocatalytic efficiency. Thus, this section summarizes the basic mechanism of DWs influencing photocatalytic reactions, supported by both experimental evidence and theoretical predictions.

3.2.1. Polarization/built-in electric field promoting charge separation. Ferroelectric materials exhibit spontaneous dipole moments that create an internal electric field that favours charge separation.³¹³ Additionally, they generate spontaneous polarization with potential differences. Then, perpendicular to DWs, these potential differences act as a driving force to establish a built-in electric field. The role of polarization, along with the corresponding built-in electric field, can be categorized into promoting bulk charge separation and surface charge separation, thereby reducing recombination and enhancing photocatalytic performance.⁶⁹

A prominent example for confirming the promotion of bulk charge separation is based on the SrBi₄Ti₄O₁₅ ferroelectric photocatalyst.²⁵⁵ In this work, Huang and co-workers visualized the intrinsic ferroelectric polarization of SrBi₄Ti₄O₁₅ nanosheet via KPFM, which was rationalized by DFT calculations. The KPFM images showed DWs with an amorphous shape (Fig. 12a). The bright and dark areas corresponded to different polarization directions. DFT calculations revealed that the effective masses of e⁻ and h⁺ along the [100] orientation were smaller than along other orientations, suggesting the higher migration rates of e⁻ and h⁺ along this axis, which was consistent with the polarization direction in SrBi₄Ti₄O₁₅. Therefore, ferroelectric polarization promoted charge transfer. Similarly, Park et al.³¹⁴ prepared K_0.5Na_0.5NbO₃ with a permanent polarization via a corona electric field polarization method. With external polarization, the time-resolved photoluminescence (PL) spectra of K_0.5Na_0.5NbO₃ presented a decrease in the decay time from 1.64 to 3.31 ns (Fig. 12b). This result suggests that the ferroelectric polarization field inhibited the recombination between electrons and holes.

Fig. 12 (a) Phase maps (top) and corresponding curve (bottom) of the piezoelectric response.²⁵⁵ Copyright © 2018 Elsevier. (b) Decay profile of PL from K_0.5Na_0.5NbO₃.³¹⁴ Copyright © 2014 Royal Society of Chemistry. (c) Transient photocurrent responses at different bias voltages.⁶⁷ Copyright © 2017 Elsevier.

Beyond the bulk charge separation, the ferroelectric effect can also promote charge transfer at the interface of heterojunctions. Considering the short time scale of interfacial charge migration, the electrochemical analysis becomes a reliable method to monitor the effect of ferroelectric materials. For example, Huang's group designed a heterojunction containing ferroelectric tetragonal-phase BaTiO₃ (T-BaTiO₃) and narrow-band-gap semiconductor BiOI (T-BaTiO₃/BiOI).⁶⁷ They applied photoelectrochemical (PEC) measurements to further investigate the polarization-induced field. Interestingly, with higher bias voltage, the photocurrent of T-BaTiO₃/BiOI samples showed clear changes. In contrast, control experiments displayed stable current intensity, indicating that polarization could adjust charge transfer and be adjusted by external bias voltage (Fig. 12c). Other similar works also prepared ferroelectrics-based heterojunctions for accelerating the surface charge mobility during different photocatalytic reactions.^58,315,316 In the field of methane conversion, although mechanistic studies are still at the early stage, state-of-the-art characterization methods like PL and PEC analysis have recently been used to clarify the role of ferroelectric materials in charge transfer kinetics. These results will be discussed in detail in Section 4.

3.2.2. Polar nano regions and surface active sites. The driving force from spontaneous polarization in ferroelectric materials leads to the opposite migration direction between electrons and holes, further resulting in the different polarity walls over the ferroelectric surface (viz. polar nano regions). Therefore, by controlling the switchable polarization via external fields or stimulations, ferroelectric materials can be explored as a bi-functional electrode for photocatalytic redox reactions.⁵⁶ For instance, Rohrer and co-workers used AFM characterization to elucidate the domain-specific reactivity of two samples of elemental-doped BiVO₄—named as (Bi_1−0.5xNa_0.5x)(V_1−xMo_x)O₄, where x = 0.05 and x = 0.175 was assigned to samples exhibiting monoclinic and tetragonal domains, respectively.³¹⁷ The AFM images of undoped monoclinic BiVO₄ showed the alternative properties between reduction and oxidation domains, corresponding to ferroelastic domains (Fig. 13a). This observation suggested the possibility of both reduction and oxidation to take place on the ferroelectric surface but with different polarization. Further, the authors employed piezoelectric force microscopy (PFM) to display the piezoresponsive nature of doped BiVO₄ (Fig. 13b). While the monoclinic BiVO₄ sample possessed a polar surface, the tetragonal BiVO₄ sample did not show polar domains according to the colour contrast in these images. After light illumination, AFM images presented the photo-deposition states over monoclinic and tetragonal BiVO₄ (Fig. 13c and d). The higher height of monoclinic BiVO₄ suggested that more particles were deposited on its surface, meaning stronger reactivity. Therefore, the ferroelectric effect was proved to improve the photochemical activity. Similarly, over a BaTiO₃ surface, AFM was also applied to monitor ferroelectric properties.³¹⁸ Interestingly, the product distribution matched the underlying domain structure of the BaTiO₃ matrix, once again confirming the role of the dipolar field in the improvement of reactivity (Fig. 13e and f). Besides the experimental evidence, some works used DFT calculations to show the polarization-dependent binding geometries and simulate the surface adsorption/desorption states over ferroelectric materials. Taking PbTiO₃ as a case study, the positively polarized electrons were beneficial for capturing reactants (i.e., O₂ and NO molecules) for subsequent redox reactions (Fig. 13g–i).³¹⁹ Once the surface was saturated, the polarization states were not stable, and the products were easily desorbed, thus returning the polar surface to the pristine status. This fact not only ensured the long-term reaction cycles but also avoided the negative effect of poisoning molecules or overreaction towards undesired products.

Fig. 13 (a) Surface topography after Ag⁺ reduction (left) and Mn²⁺ oxidation (right). (b) PFM phase images of monoclinic and tetragonal elemental-doped BiVO₄. Surface topographic images (left), inverse pole figure maps (mid), and topographic line profiles (right) of (c) monoclinic and (d) tetragonal elemental-doped BiVO₄.³¹⁷ Copyright © 2016 Royal Society of Chemistry. (e) AFM images of rutile TiO₂ supported by BaTiO₃ before (left) and after (right) a photochemical reaction, and (f) their heights.³¹⁸ Copyright © 2010 American Chemical Society. (g) O₂ dissociative binding and (h) NO binding on negatively (left) and positively (right) poled RuO₂–PbTiO₃. (i) Space electron redistribution, presenting covalent binding between NO and positively poled CrO₂–PbTiO₃.³¹⁹ Copyright © 2015 American Chemical Society.

Another potential role of ferroelectric materials is to provide high-density surface active sites as discussed in Section 1. The conversion of CO₂ into CO₂˙⁻ bidentate radical is promoted by the electrons on positively polarized surfaces/sites.¹⁴⁷ Further, based on the CO₂˙⁻ intermediate, the interaction between CO₂˙⁻ and H⁺ to form HCOO˙⁻ and finally HCOOH is achieved via a proton-coupled electron mechanism.¹⁵⁴ Accordingly, CH₄ dissociation/cracking should tend to occur on negatively polarized surfaces/sites. In the case of ferroelectrics, the reduction and oxidation reactions will take place at different polarized surfaces/sites, inhibiting the recombination between photoexcited electrons and holes.

4. Ferroelectric-assisted photocatalytic methane conversion

4.1. Ferroelectric-assisted photocatalytic DRM

Recently, a Fe₂O₃-loaded BaTiO₃ photocatalyst (denoted as BFxy, where xy is the BaTiO₃/Fe₂O₃ molar ratio) is reported for the DRM reaction.³²⁰ The photoluminescence signals, a feature of the recombination rate of charge carriers, emitted by all the samples were measured (Fig. 14a). BaTiO₃ showed the lowest charge recombination rate due to its intrinsic ferroelectric property. On the contrary, Fe₂O₃ showed the maximum charge recombination rate due to the narrow band gap of Fe₂O₃. The heterostructure and charge transfer across to interface of BF were confirmed by PL spectra. In this case, the ferroelectric effect from BaTiO₃ endowed BF composites with a higher charge separation/transfer efficiency than Fe₂O₃. Thus, the DRM activity of BF31 showed the highest reactant conversion efficiency (Fig. 14b). Besides BaTiO₃, TaN as another ferroelectric material was also developed for photocatalytic DRM reaction.¹⁴⁶ In this work, Liu's group explored the effect of TaN as the photocatalyst substrate with a Pt NPs loading (Pt/TaN). Under visible light illumination, Pt/TaN showed improved photocatalytic DRM activity. The convergent beam electron diffraction (CBED) characterization exhibited the polarization phenomenon in the TaN substrate (Fig. 14a). Obviously, the polarity would lead to a built-in electric field, facilitating the photoexcited charge separation. On the contrary, TiN and ZrN as the reference samples did not show any polarized property and thus had lower activity than Pt/TaN (Fig. 14b and c). Accordingly, this study also proved the ineligible enhancement role of ferroelectrics with electrostatic field in photocatalytic efficiency.

Fig. 14 (a) PL spectra and (b) CO₂ conversion efficiency of BaTiO₃, Fe₂O₃, and BFxy. Reaction conditions: 40 psi containing 45% CO₂, 45% CH₄, and 10% He, 125 W Hg lamp, 35 °C.³²⁰ Copyright © 2019, Springer Nature. (a) CBED pattern of TaN. Catalytic DRM performance of (b) Pt/ZrN and (c) Pt/TiN. Reaction conditions: 300 W Xe lamp, CH₄/CO₂ = 1 (20.0 mL min⁻¹), 500 °C.¹⁴⁶ Copyright © 2018, Wiley.

Until now, a large number of ferroelectric materials have been developed in photocatalytic fields.^86,321 Unfortunately, in photocatalytic DRM implementations, the previous reports prefer to pay more attention to the impact of cocatalysts or the heterojunction with controllable energy band structure in the DRM process instead of investigating the ferroelectric effect and enhanced charge transfer dynamics. With the development of ferroelectric materials, the ferroelectric effect is expected to be maintained even under harsh reaction conditions (high temperatures or pressures),¹⁵⁴ meaning that ferroelectrics can be widely used in different photo(thermal) catalytic applications. Therefore, it is hoped to report more in-depth research on the improvement of DRM reactions from the view of the ferroelectric effects soon. Our previous literature also discussed and emphasized the role of the ferroelectric field especially in DRM applications, the potential interested audience is recommended to browse ref. 55.

4.2. Ferroelectric-assisted photocatalytic POM

As mentioned above, the permanent polarization generated from ferroelectrics creates a built-in electric field. The strong and reliable driving force in an electric field adjusts the transfer path of electrons and holes to different active sites, separately. As a result, more charge carriers tend to complete the redox reaction instead of recombination. An early work was reported by Yi's group in 2013,³²² where NaNbO₃ polar semiconductors were used to enhance the POM activity. Scanning electron microscopy (SEM) images showed the preparation of NaNbO₃ nanocuboids with ordered b-axis orientation (NNB-1, Fig. 15a and b). To investigate the role of ferroelectric property, the disordered random distribution of NaNbO₃ (NNB-8) was also obtained (Fig. 15c). In general, the NaNbO₃ crystal showed antiferroelectric and ferroelectric properties in the direction of the c-axis and b-axis, respectively, implying an internal polarization in NaNbO₃ nanocuboids along the b-axis.³²³ Instead, NaNbO₃ nanocuboids with a disordered nanostructure would induce a randomly distributed local bias electric field. In photocatalytic CH₄ oxidation experiments, with Pt loading as cocatalyst via impregnation and photodeposition methods (denoted as Pt-NNB-1, Pt-NNB-1-LD, Pt-NNB-8, and Pt-NNB-8-LD), the CO₂ evolution rate of NNB-1 was much higher than that of NNB-8 (Fig. 15d). Interestingly, pure NNB-1 had higher activity than that Pt-NNB-1 and Pt-NNB-1-LD. The reason was investigated by SEM, which showed that the Pt loading process caused the destruction of the ordered structure of NNB-1 (Fig. 15e). Therefore, the structure orderliness was a crucial factor for CH₄ oxidation because it was a prerequisite for maintaining the strong ferroelectric effect (Fig. 15f). And the above results also confirmed the great ability of polar semiconductors with built-in electric fields to enhance CH₄ oxidation compared to conventional cocatalysts (e.g. Pt). To further investigate the ferroelectric effect, photoelectrochemical (PEC) measurements were used to detect the photocurrent responses of different NaNbO₃ samples. As shown in Fig. 15g and h, the authors used photoelectrochemical measurements, which displayed higher photovoltage and photocurrent response on ordered NNB-1 compared to disordered NNB-8, to confirm the built-in electric field of each crystal unit with consistent direction was beneficial for the directional charge transfer. Thus, the advantage of ordered ferroelectrics in charge separation/transfer dynamics was proved in this work.

Fig. 15 SEM images of NaNbO₃ samples with (a) and (b) ordered, and (c) disordered nano cuboid assemblage. (d) CO₂ evolution from POM reaction. Reaction conditions: 1 mL mixture of 10% CH₄ and 90% O₂, 300 W Xe lamp. Along the direction of the arrow, the samples are without catalyst, Pt-NNB-1, Pt-NNB-1-LD, NNB-8, Pt-NNB-8, Pt-NNB-8-LD, and NNB-1, respectively. (e) SEM image of Pt–NNB-1 sample prepared by the impregnation method. (f) Illustration of charge distribution in single crystal with ordered and disordered arrangement structures. (g) Open circuit photovoltage curves. (h) Photocurrent responses of ordered NNB-1 and disordered NNB-8.³²² Copyright © 2014, Royal Society of Chemistry.

Furthermore, to achieve the high selectivity of POM over photocatalysts, precise control of the surface polarity on ferroelectric materials is a credible strategy after the separation of electron–hole pairs is completed. In the CH₄ to CH₃OH conversion process, CH₄ can interact with ˙OH on the oxidizing surface with negative polarity to produce CH₃OH.²² The state of the surface is then switched to positive polarization with reducibility, allowing the O₂ molecules to complete the dissociation. Due to the addition of O₂, the surface is ready to oxide CH₄ again. The whole cycle is therefore advantageous for the partial oxidization of CH₄ to CH₃OH instead of other products. However, Arvin Kakekhani et al. suggested that this theory also has some potential challenges according to DFT results. Firstly, the surface oxygens could couple with the neighbouring oxygen to form O₂. In fact, the barrier to O₂ leaving the surface is much lower than that for direct oxidation of CH₄, suggesting that the dissociated oxygen may not participate in the oxidation reaction. Secondly, before CH₃OH leaves the ferroelectric materials, the products undergo a reverse reaction to dissociate back into CH₄ and surface oxygens when the surface state changes to positive polarization. Some groups have attempted to overcome these shortcomings in various photocatalytic applications by using monolayers of transition metal oxides.³²⁴ As an example, ZnO monolayer on PbTiO₃ ferroelectric photocatalysts. Through the dynamically tuned catalysis mechanism, the energy barrier for CO₂ conversion was obviously lowered.³²⁵ In addition, the CrO₂-based monolayer also promotes NO_x dissociation and CO oxidation over the PbTiO₃ substrate. And the switchable polarity between strong and weak binding achieved rapid dissociation and easy desorption simultaneously, avoiding oxygen and sulfur poisoning.³¹⁹ Recently, Yi et al. investigated the influence of polar and non-polar facets in methane oxidation based on ZnO ferroelectric photocatalysts.³²⁶ The ZnO nanosheets and nanorods were prepared to compare the influence of facets with different polar states (Fig. 16a and b). Compared to nanosheets with polar facets, ZnO nanorods mostly contained non-polar facets (Fig. 16c). The photocatalytic oxidation of CH₄ showed distinct morphology dependence characteristics (Fig. 16d). Due to the polarization in ZnO nanosheets, the methane could be rapidly oxidated. However, ZnO nanorods showed a lower rate of photocatalytic CH₄ oxidation under the same conditions. The CH₄ oxidation followed the quasi-first-order reaction kinetics (Fig. 16e). The rate constants were 0.01 and 8.01 × 10⁻⁴ min⁻¹ for ZnO nanosheets and nanorods, respectively. These results indicated the significant enhancement effect of polar facets over ZnO nanosheets in CH₄ oxidation. High-energy polar facets are thermodynamically more reactive than non-polar facets, inducing polarity in stable and symmetric CH₄ molecules. During the PEC measurements, ZnO nanosheets had higher photocurrent (Fig. 16f), which could be attributed to more polar facets in ZnO nanosheets. Furthermore, the authors proposed the relationship between CH₄ activation and polar facets in ZnO nanostructures. As a ferroelectric material, the polar semiconductor property of ZnO could be viewed as alternately stacking multilayers of Zn²⁺ and O²⁻ ions with tetrahedral coordination structure along the c-axis. This unique structure endowed ZnO with a built-in electric field. When CH₄ was adsorbed on the m-plane of ZnO, the H₃C^δ−–H^δ+ was polarized due to the interaction between ˙CH₃ and ˙H and active sites over the m-facets. Meanwhile, the light illumination activated the C–H bond in CH₄, the formation of free ˙CH₃ and ˙H. On the other hand, the generation of oxygen species (activation of O₂ molecules) would take place on the polar (0001) plane, forming ˙O₂⁻ radicals. Finally, ˙O₂⁻ would oxidize the intermediate to produce the products (CO₂ and H₂O). Although the products in this work were not high value-added chemicals, the above work indeed demonstrates the potential of an epitaxial monolayer of a transition metal oxide in the activation of reactants over the specific facet with polarization. Thus, the development of specific surface layers is beneficial to fully exploit the ferroelectric photocatalysts in POM application.

Fig. 16 FESEM of the ZnO (a) nanosheets and (b) nanorods. (c) Growth models of ZnO. (d) Time course and (e) reaction kinetic plots of POM under simulated solar light over the ZnO. Reaction conditions: 200 ppm CH₄, Xe lamp (200 mW cm⁻²). (f) Photocurrent responses of ZnO.³²⁶ Copyright © 2019, American Chemical Society.

4.3. Ferroelectric-assisted photocatalytic OCM

Compared to the conventional thermal catalytic OCM process, photocatalytic OCM focuses on the inhibition of competitive reaction and overoxidation under mild conditions. From this perspective, ferroelectric materials are a suitable candidate for enhancing photocatalytic OCM efficiency. However, few studies reported that ferroelectric materials are used in the thermal catalytic or photocatalytic OCM field. Thus, in this section, the feasibility and possible advantages of ferroelectric-assisted OCM systems will be discussed to encourage more ferroelectrics to be developed for OCM applications. Firstly, both ferroelectric-based photocatalysts and ferroelectric cocatalysts could form an electric field due to intrinsic spontaneous polarization. The electric field would function as the above noble metal cocatalysts that become the charge acceptor via adjusting the electrons and holes to transfer onto the surface with different polarized states. Secondly, ferroelectric polarization could avoid some unfavorable effects from O₂ by actively toggling the surface–absorbate interaction. Using PbTiO₃ as an example, after the dissociation of O₂ in the reactants, each oxygen adsorbate could bind to the surface oxygen atom, forming a surface O₂ molecule. In negative polarization, the surface O₂ molecules have very low dissociative binding energy. On the contrary, the interaction between positively polarized surface and surface O₂ molecules is strong. Table 2 exhibits the binding energies for the surface O₂ molecule.²² A positive number means the binding is favourable. Conversely, a negative number indicates a metastable state that is more unfavourable than O₂. Indeed, this is meaningful for photocatalytic OCM applications because the overoxidation and inactivation from adsorbed O₂ could be eliminated by controlling the polar state on ferroelectric catalysts (or cocatalysts). Finally, similar to the above-mentioned plasma-assisted photocatalysts, some novel reaction pathways of the OCM process might be found in the ferroelectric-assisted photocatalytic systems.²³¹ Because the local surface state was changed by polarization, the localized electric field in ferroelectric materials probably led to some different intermediates or initiated OCM reactions by other oxidants except O₂ and H₂O₂. Thereby, it is feasible to efficiently convert CH₄ into high-value-added C₂ products over photocatalysts containing ferroelectric materials.

Table 2 Intact and dissociative binding energies toward one O₂ molecule on PTO as a function of polarization relative to O₂ in a vacuum²²

Polarization status	Intact binding energy [eV]	Dissociative binding energy [eV]
Positive	2.4	2.6
Paraelectric	0.2	−0.8
Negative	0	0.4

---

4.4. Ferroelectric-assisted photocatalytic NOCM

Although photocatalytic NOCM has many advantages, unfortunately, the evolution rate of C₂H₆ from the photocatalytic NOCM reaction is almost an unexplored field. It is very urgent to design more efficient photocatalysts for NOCM. Several aspects should be considered for the reasonable development of NOCM photocatalysts. First, a strong local electric field may be needed to easily polarize and nucleophilically or electrophilically attack CH₄. From this perspective, ferroelectric material is a suitable candidate to improve the efficiency of photocatalytic NOCM. However, few studies used ferroelectric materials in the NOCM field. Second, metallic co-catalysts are dynamically necessary for the occurrence of the half-reaction of H₂ generation because the large overpotential inhibits the release of H₂ across the bare oxide surface. In order to simultaneously address the above challenges, an Au/ZnO porous nanosheet with a local electric field and polar facet was designed for photocatalytic NOCM by Long's group.³²⁷ A ZnO crystal was tailored to expose [001] facet with polarized property, beneficial for the CH₄ activation (Fig. 17a). Au/m-ZnO showed a bandgap absorption at ca. 395 nm and a broad absorption peak from plasmonic Au NPs (Fig. 17b). The as-prepared Au/m-ZnO samples achieved the highest evolution rates of C₂H₆ and H₂. Further, the reaction mechanism of NOCM was investigated under different reactive atmospheres. Compared to the pure CH₄ atmosphere, the C₂H₆ production was increased by O₂ and H₂O and decreased by CO₂. However, the H₂ evolution rate was decreased in the order of pure CH₄ > CH₄ + H₂O > CH₄ + CO₂ > CH₄ + O₂. Thus, it could be assumed that CO₂ did not favor the NOCM because CO₂ is competitive for photoelectrons and inhibits the dehydrogenation process, which is an important process for CH₄ activation. The decrease in H₂ production could be attributed to the side reaction CH₄ + CO₂ → CH₃COO⁻ + H⁺, which would cause the stoichiometric mismatch between H₂ and C₂H₆. Under O₂-free conditions, C₂H₆ and H₂ yields could remain stoichiometrically matched (C₂H₆ : H₂ = 1 : 1), indicating that the photocatalytic NOCM proceeds by a dehydrogenative process. In situ electron spin resonance (ESR) results showed the coupling between electron-occupied oxygen vacancy (V_O⁺) and diamagnetic oxygen ion (O⁻) according to the characteristic peaks at 1.960 and 1.957 (Fig. 17c and d). However, light illumination did not change the V_O⁺ signal, indicating the lattice O atoms have no effect on the NOCM process. Thus, the dehydrogenative coupling occurred at the Zn²⁺ atoms. CH₄ was polarized by the Zn plane, resulting in H^δ+–CH₃^δ− polarization to generate ˙CH₃ radical and H⁺ via the photoelectron transfer between ZnO and CH₄. The plasmonic hot electrons on Au NPs promoted the H₂ formation. DFT studies calculated the energy barrier of NOCM on Au/ZnO photocatalyst (Fig. 17e). The results showed the advantage of the polar Zn(001) surface in the generation of C₂H₆. The charge analysis showed the interaction in charge transfer between ZnO and Au, which means that Au activated the ZnO surface and enhanced the CH₄ chemisorption. Meanwhile, the CH₄-TPD results proved the enhancement of the CH₄ dissociation on the (001) facet by the electric field coupling (Fig. 17f). Therefore, this work provided convincing evidence for the important role of the synergistic effect between ZnO ferroelectric material and Au plasma in accelerating the CH₄ activation and charge-carrier interaction.

Fig. 17 (a) SEM image of Au/m-ZnO-4.8. (b) UV-vis spectra. In situ ESR spectra of (c) Au/m-ZnO-4.8 under CH₄ atmosphere and (d) O deficiency under CH₄/H₂O atmospheres. (e) Energy barrier of NOCM on Zn₃O₃H₃(001)/Au(111). (f) CH₄-TPD patterns.³²⁷ Copyright © 2018, Royal Society of Chemistry.

5. Potential role of ferroelectric effect during photocatalytic methane conversion

Based on the previous studies with convincing evidence of the influence of ferroelectric materials on photocatalytic processes, the potential effects of ferroelectrics on the photocatalytic methane conversion field can be summarized as follows:

5.1. Promoting charge transfer

As a general effect, the polarization and corresponding built-in electric field are always mentioned and widely utilized when researchers purposely introduce ferroelectric materials into different photocatalytic systems. Considering that photocatalytic methane conversion systems must follow the general steps of photocatalysis, the sustainable internal electric field derived from ferroelectric properties should enhance charge-carrier transfer kinetics in this specific process.³²⁸ Meanwhile, as mentioned before (Section 3.3.2), different polar surfaces will lead to the separation of reduction and oxidation processes. This also adjusts the transfer routes of electrons and holes towards different surface sites, decreasing the recombination possibility. In addition, the spatial selectivity of charges over the ferroelectrics surface leads to energy band bending at the interface when the polarized surface contacts with solution.³²⁹ This characteristic has been widely used in both photocatalytic and electrocatalytic systems, because the negative surface/electrode causes upward band bending, promoting the charge migration from electrolyte/reactants to catalysts. On the contrary, the positive surface/electrode leads to downward band bending, enhancing charge transfer from ferroelectric catalysts to electrolyte/reactants.³³⁰

5.2. Promoting reactant adsorption/activation

Although light illumination can directly promote chemical bond cracking, the high C–H bond energy of this small hydrocarbon molecule (CH₄: ∼440 kJ mol⁻¹) might arouse a suspicion: is the light energy enough to complete the CH₄ association? In this case, the photoexcited radical species (˙OH and h⁺) should be more reliable in order to activate CH₄ molecules. From this view, ferroelectric materials are able to create a negatively polarized surface, which is beneficial to oxygen-containing active species, and further oxidizes CH₄ to intermediates or CH₃OH.¹⁴⁷ Depending on the atmosphere in the reaction systems, the intermediates (i.e. ˙CH₃) will take part in the NOCM or OCM to obtain CO, CH₃OH, or C₂ products like C₂H₆ as reviewed in the previous sections. Another potential effect for accelerating reactant activation over ferroelectric surfaces might be the adsorption–desorption enhancement. It has been reported that the adsorption–desorption properties of ferroelectric materials can be tuned according to the molecules polarity.^331,332 Even if CH₄ is a non-polar molecule, products like CO and CH₃OH are polar molecules. This suggests the surface states can be automatically switched during CH₄ conversion processes, promoting reactant adsorption and subsequent activation, ensuring long-term reaction cycles. Further, DFT calculations have confirmed that 2D ferroelectric materials are able to control the adsorption energies, electron transfer, and magnetic moment of adsorbed reactant molecules so that there was a polarization-dependent gas adsorption behaviour over the monolayer surface.³³³

5.3. Improving selectivity

Ferroelectric materials are able to switch reaction pathways to obtain desired products without stopping the reaction process and changing catalysts, demonstrating the great potential for achieving high selectivity during methane conversion systems.³³⁴ One approach to control the selectivity is the surface polarization. When the absorbance is saturated, the negative surface state will be switched to a positive state reducing oxygen species so that the overoxidation can be inhibited. Therefore, complex methane conversions can achieve high selectivity over a polarized surface.³³⁵ Besides, the selective adsorption of reactants on the polarized surface can effectively lift the limitations imposed by Sabatier's principle.³³⁶ For instance, the ferroelectric BiFeO₃ could reduce the charge recombination rate from 17 to 0.6 s⁻¹ so that increased 4.4 times the photoelectric efficiency.³³⁷ The improvement should be attributed to the compensation effect through electronic reconstruction when a built-in electric field promotes electron–hole separation.^319,338 The changes in electron distribution lead to different abilities of polarized surfaces towards oxidation or reduction reactions. Thus, the switchable ferroelectric polarization allows us to control the surface affinity to desired redox reactions. Because ferroelectric materials always have large band gaps and dielectric constants, using ferroelectrics as triggers to control the dissociation of specific small hydrocarbon molecules has been reported as a strategy to tune the chemical selectivity.³²⁵ The ferroelectric matrix could dynamically reduce the reaction barrier through polarization manipulation to affect the local reaction pathways.³³⁹ In addition to changing the surface properties and reaction routes, recently, a novel idea to control the selectivity based on ferroelectric materials is using a ferroelectric field to immobilize active sites (i.e. cocatalyst on photocatalyst).^340,341 To be specific, during methane conversion reactions, depositing at cocatalyst with a strong ability towards CH₄ cracking over a holes-accumulated surface under the assistance of a ferroelectric field promoted the directional oxidation of CH₄ molecules and the reversible reactions could be inhibited. However, the real implementation of ferroelectric photocatalysts with ideal selectivity is still challenging due to the large gap between theoretical predictions and practical paradigms. Up to now, these predictions have been mainly realized in CO₂RR and NRR systems or ferroelectric electrocatalysts. These works make us hold strong confidence that ferroelectric systems should be widely and practically utilized in photocatalytic methane conversion reactions in the near future.

6. Conclusions and outlook

The utilization of CH₄ is one of the crucial steps in the transition from a fossil fuel energy system to a sustainable energy system. The mainstream sources for renewably obtaining CH₄ are currently based on anaerobic digesters, which convert cow manure into natural gas, and CO₂-based green chemistry transfer routes, such as biochemical, (photo)electrochemical, and photo(thermal) catalytic methods.³⁴² Globally, more than 50% of H₂ is produced through the reforming of CH₄.³⁴³ Additionally, by producing higher hydrocarbons, various conversion routes of methane are crucial to reducing the greenhouse gas and achieving carbon neutralization. Therefore, clean production and sustainable utilization of CH₄ seem like a complete and ideal strategy to deal with energy shortages and environmental problems.

However, the C–H bond strength in CH₄ is the biggest obstacle to converting CH₄ into added-value products. As mentioned above, many thermal catalysts have been developed to overcome this thermodynamic challenge. Kinetic models and reaction pathways in these thermal catalytic systems are continuously being elucidated. Nevertheless, the high temperature needed for CH₄ conversion processes leads to the inactivation of thermal catalysts and even causes excessive CO₂ emissions. Fortunately, with the extensive studies of environmentally friendly photocatalysts, the harsh reaction conditions needed in thermal catalytic methane conversion are being gradually reduced, indicating that carbon neutralization becomes possible. Currently, an important problem to be solved in photocatalytic methane conversion is the insufficient efficiency mainly due to the slow charge transfer. In this regard, the ferroelectric-based photocatalysts are envisioned to play a key role in manipulating the evolution efficiency and selectivity of products from CH₄ conversion by regulating the transfer kinetics. Many efforts have been made in this interdisciplinary field so far, but some obstacles and deficiencies still exist as follows.

6.1. Development of novel photocatalysts using ferroelectric materials

An increasing number of ferroelectric-assisted photocatalytic systems have been designed and fabricated to participate in reactions that produce the desired chemicals and clean energy fuels (e.g., H₂ from overall water splitting and CO from CO₂ reduction). However, in the catalytic methane conversion field, research based on ferroelectric materials is still rare. Thus, the excellent ability of ferroics should be further developed for improving CH₄ conversion efficiency. One of the possible reasons is that some ferroelectric-based semiconductors do not have suitable energy band structures to generate photoinduced charge carriers with high redox capability. In addition, the work function of certain ferroelectric materials is not sufficient to promote the reduction process when electrons are transferred to the ferroelectric co-catalyst surface. Thus, energy band engineering is important for the design of ferroelectrics-supported photocatalytic systems. However, it is not practical to synthesize and experimentally test all candidate ferroelectrics to obtain the details of the energy band and work function. DFT calculations and hybrid functional calculations are useful to determine the electronic structure and predict ideal ferroelectric materials with sufficient redox ability and polarized state. However, the high computational cost of these methods should be considered for high throughput calculations. In addition, machine learning (ML) has recently become a highly efficient technique in materials data engineering to estimate various parameters (e.g., E_g, CB, and VB) in materials. Thus, with the improvement and supplementation of organic/inorganic databases, the ML method could provide guidance for the discovery and design of novel energy materials.

6.2. Emphasize the role of the ferroelectric effect

Ferroelectric materials have been widely used in various photo(thermal)catalytic applications, but some works paid too much attention to other properties (e.g., optical properties, absorption/desorption abilities) in as-designed ferroic-based photocatalysts instead of their intrinsic ferroelectric effect. In particular, some outstanding works shown in this review also used ferroelectric materials such as BiVO₄ and SrTiO₃,^86,192 but the authors only systematically investigated the influence of loaded co-catalyst or elemental doping on the photocatalytic activity or selectivity. To some extent, the reason for this problem is that the ferroelectric effect was not considered in the key points during the initial design process. However, it is known that the photocatalytic activity could be affected by various factors, and even a small factor could have a non-negligible effect on the overall photocatalytic efficiency. Therefore, the function of the natural ferroelectric effect and the accompanying built-in electric field should gain more interest when using ferroelectric materials not only for methane conversion reactions but also for photocatalytic applications in general.

6.3. Exploring new effects from ferroelectric materials

During the photocatalytic reactions, the electron spin phenomenon plays an important role in electron migration and thus photocatalytic efficiency. The electron spin could be regulated by external fields like magnetic and electric fields.³⁴⁴ In detail, the spin–orbit coupling is able to change the spin orientation of both electrons and holes from semiconductors. Then the recombination between photoinduced charges will be inhibited. The electrons scatter can gradually decrease the impact from spin polarization, meaning that the spin polarization with a longer lifetime can effectively promote charge separation and transfer. From this perspective, a series of ferroelectrics with a magnetic response called the magnetoelectric effect, could be a new path toward ferroelectrics-based control of electron spin. It has been confirmed that some single-phase ferroelectric materials like BiFeO₃ could adjust the spin behaviour via the interaction between intrinsic coupling between electric and magnetic order parameters.^345,346 Some works also discussed the influence of spin polarization in the photocatalytic process.³⁴⁴ However, in the photocatalytic field, the investigation on the control of photoexcited electrons spin via ferroelectric materials is still scarce, and this blank is no exception in methane conversion areas. With the development of theories about ferroelectric materials, the spin-related mechanisms in ferroelectrics should obtain more attention for full utilization in practical photocatalytic reactions.

6.4. Design of advanced reaction vectors

To further exploit the advantage of ferroelectric materials for boosting photocatalytic methane conversion, the design or update of reactors must be as important as the development of ferroelectric materials. It has been confirmed that different polarization states lead to the variation of adsorption/desorption behaviour towards reactants on ferroelectric materials. Thus, if the updated reactors are able to periodically control the surface state of ferroelectric photocatalysts, the activity, and selectivity for specific hydrocarbon products, especially C₂ products, will be significantly improved. Similarly, there are some successful examples of piezoelectric/pyroelectric-assisted photocatalytic systems, proving that this proposal is reasonable and feasible. For example, the introduction of an external ultrasonic device to excite the piezoelectric effect, which forms a strong electric field in photocatalysts and thus improves the photocatalytic efficiency. In addition to the external auxiliary device, advanced flow reactors are more effective than conventional batch reactors. Although the operation, cost, and safety are major challenges for the design of flow reactors, some mature flow systems in thermal catalysis are ideal references for photocatalytic gas–solid two-phase reactions. Additionally, some membrane reactors also have unique advantages in separating catalysts from products after reaction, lowering the operating temperature, breaking the restriction of thermodynamic balance, and improving the selectivity and activity. Membrane reactors are expected to achieve the coupling of different photocatalytic applications. Therefore, if more effort is devoted to the development of reactors and assisted devices, ferroelectric-assisted photocatalytic CH₄ conversion will realize industrial application at an early date.

6.5. Coupling of other photocatalytic reactions

Nowadays, (i) most products of methane conversion are raw materials for downstream reactions, and (ii) ferroelectric-based materials have been widely used for photocatalytic applications beyond methane conversion. Taking these two points into account, if some advanced reactors could separate specific products of methane conversion, at least two photocatalytic reactions including methane conversion could be simultaneously and continuously conducted on ferroelectric photocatalysts. In this case, the cost of industrial production could be significantly reduced by eliminating the transportation and storage of intermediate raw materials. Meanwhile, other labour costs and space occupied by chemical production equipment, which are difficult to calculate, could also be saved by reaction coupling. However, to realize this goal, it is necessary to fabricate excellent (ferroelectric-based)photocatalysts and the aforementioned novel reaction vectors to overcome the complex requirements of various catalytic reactions.

6.6. Investigation of enhancement and reaction mechanism

Compared with much scientific research focused on the development of novel photocatalysts, few mechanisms are known about each methane conversion technology. The transient intermediates and local reaction pathways need to be studied in detail because the surface state of ferroelectric materials may be changed during CH₄ conversion. Beyond the reaction mechanism, little evidence has been provided to directly show the effect of polarization on charge transfer dynamics at the nano or micro-scale. Although some works claimed that electrons and holes would separate to different locations under the influence of the ferroelectric effect, more characterization data should be provided to demonstrate the migration or bias of photoexcited charge carriers. For this problem, a series of advanced characterization methods, such as in situ Fourier transform infrared spectroscopy (FT-IR), in situ X-ray photoelectron spectroscopy, piezoresponse force microscopy (PMF), and molecular probe technology, were integrated with DFT-based calculations to clarify both the enhancement and the reaction mechanism in the photocatalytic field. It is expected these advanced characterization methods will provide a deeper insight into the ferroelectric-assisted methane conversion field.

It is undeniable that some high-level methane conversion routes are still at a very early stage and that ferroelectric-assisted catalysis, especially in the photocatalytic field, is far from industrial applications. However, this multidisciplinary field could develop rapidly with the presence of a new generation of advanced characterization techniques and chemical synthesis strategies. It is expected that more studies on ferroelectric-assisted photocatalytic methane conversion will be carried out in the coming years.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Y. L. acknowledges the program of China Scholarships Council (No. 202206250016). X. S. thanks ICREA for the ICREA Academia Prize 2020. X. S. and J. G.-A. thank MINECO/FEDER for financial support (PID2019-104171RB-I00). J. M. is indebted to the Ramón y Cajal Program (RYC2021-033820-I fellowship) funded by MCIN/AEI/10.13039/501100011033 and by the European Union “NextGenerationEU/PRTR”.

References

1. X. Jiang, X. Nie, X. Guo, C. Song and J. G. Chen, Chem. Rev., 2020, 120, 7984.
2. S. N. Habisreutinger, L. Schmidt-Mende and J. K. Stolarczyk, Angew. Chem., Int. Ed., 2013, 52, 7372.
3. H. Weldekidan, V. Strezov and G. Town, Renewable Sustainable Energy Rev., 2018, 88, 184.
4. J. O. Metzger and A. Hüttermann, Naturwissenschaften, 2009, 96, 279.
5. S. Morales, R. Miranda, D. Bustos, T. Cazares and H. Tran, J. Anal. Appl. Pyrolysis, 2014, 109, 65.
6. J. Ran, H. Zhang, J. Qu, J. Shan, K. Davey, J. M. Cairney, L. Jing and S. Qiao, Small, 2021, 17, 2100296.
7. B. Xia, Y. Zhang, J. Ran, M. Jaroniec and S. Qiao, ACS Cent. Sci., 2021, 7, 39.
8. B. L. Farrell, V. O. Igenegbai and S. Linic, ACS Catal., 2016, 6, 4340.
9. E. McFarland, Science, 2012, 338, 340.
10. U.S., Energy Information Administration (EIA), Annual Energy Outlook 2019, 2019.
11. X. Meng, X. Cui, N. P. Rajan, L. Yu, D. Deng and X. Bao, Chem, 2019, 5, 2296.
12. D. E. Holmes and J. A. Smith, Adv. Appl. Microbiol., 2016, 1–61.
13. X. Li, C. Wang and J. Tang, Nat. Rev. Mater., 2022, 7, 617.
14. P. Schwach, X. Pan and X. Bao, Chem. Rev., 2017, 117, 8497.
15. H. Song, X. Meng, Z. J. Wang, Z. Wang, H. Chen, Y. Weng, F. Ichihara, M. Oshikiri, T. Kako and J. Ye, ACS Catal., 2018, 8, 7556.
16. X. Yu, V. De Waele, A. Löfberg, V. Ordomsky and A. Y. Khodakov, Nat. Commun., 2019, 10, 700.
17. M. A. Gondal, A. Hameed, Z. H. Yamani and A. Arfaj, Chem. Phys. Lett., 2004, 392, 372.
18. H. Yoshida, K. Hirao, J. I. Nishimoto, K. Shimura, S. Kato, H. Itoh and T. Hattori, J. Phys. Chem. C, 2008, 112, 5542.
19. W. Zhang, Y. Yu and Z. Yi, J. Mater. Sci., 2017, 52, 5106.
20. N. J. Gunsalus, A. Koppaka, S. H. Park, S. M. Bischof, B. G. Hashiguchi and R. A. Periana, Chem. Rev., 2017, 117, 8521.
21. L. Yuliati and H. Yoshida, Chem. Soc. Rev., 2008, 37, 1592.
22. A. Kakekhani and S. Ismail-Beigi, Phys. Chem. Chem. Phys., 2016, 18, 19676.
23. Y. Song, E. Ozdemir, S. Ramesh, A. Adishev, S. Subramanian, A. Harale, M. Albuali, B. A. Fadhel, A. Jamal, D. Moon, S. H. Choi and C. T. Yavuz, Science, 2020, 367, 777.
24. D. Kong, Y. Zheng, M. Kobielusz, Y. Wang, Z. Bai, W. Macyk, X. Wang and J. Tang, Mater. Today, 2018, 21, 897.
25. C. Jiang, S. J. A. Moniz, A. Wang, T. Zhang and J. Tang, Chem. Soc. Rev., 2017, 46, 4645.
26. S. J. A. Moniz, S. A. Shevlin, D. J. Martin, Z.-X. Guo and J. Tang, Energy Environ. Sci., 2015, 8, 731.
27. K. Shimura and H. Yoshida, Catal. Surv. Asia, 2014, 18, 24.
28. J. Xie, R. Jin, A. Li, Y. Bi, Q. Ruan, Y. Deng, Y. Zhang, S. Yao, G. Sankar, D. Ma and J. Tang, Nat. Catal., 2018, 1, 889.
29. Y. Wang, H. Suzuki, J. Xie, O. Tomita, D. J. Martin, M. Higashi, D. Kong, R. Abe and J. Tang, Chem. Rev., 2018, 118, 5201.
30. H. Li, J. Shang, Z. Ai and L. Zhang, J. Am. Chem. Soc., 2018, 140, 526.
31. Y. Lei, Z. Jia, H. Hu, L. Liu, J. Ye and D. Wang, Catalysts, 2022, 12, 1323.
32. J. Wen, J. Xie, X. Chen and X. Li, Appl. Surf. Sci., 2017, 391, 72.
33. H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri and J. Ye, Adv. Mater., 2012, 24, 229.
34. T. M. Khedr, K. Wang, D. Kowalski, S. M. El-Sheikh, H. M. Abdeldayem, B. Ohtani and E. Kowalska, J. Environ. Chem. Eng., 2022, 10, 107838.
35. P. Jiménez-Calvo, V. Caps and V. Keller, Plasmonic Au-based junctions onto TiO2, gC3N4, and TiO2-gC3N4 systems for photocatalytic hydrogen production: Fundamentals and challenges, Renewable Sustainable Energy Rev., 2021, 149, 111095.
36. Z. Zhao, H. Tan, H. Zhao, Y. Lv, L. J. Zhou, Y. Song and Z. Sun, Chem. Commun., 2014, 50, 2755.
37. R. Leary and A. Westwood, Carbon, 2011, 49, 741.
38. B. Dai, J. Fang, Y. Yu, M. Sun, H. Huang, C. Lu, J. Kou, Y. Zhao and Z. Xu, Adv. Mater., 2020, 32, 1906361.
39. L. Liu, S. Ouyang and J. Ye, Angew. Chem., Int. Ed., 2013, 52, 6689.
40. X. Xu, F. Luo, W. Tang, J. Hu, H. Zeng and Y. Zhou, Adv. Funct. Mater., 2018, 28, 1804055.
41. L. Yin, M. Zhao, H. Hu, J. Ye and D. Wang, Chin. J. Catal., 2017, 38, 1307.
42. J. Tan, X. Wang, W. Hou, X. Zhang, L. Liu, J. Ye and D. Wang, J. Alloys Compd., 2019, 792, 918.
43. M. Ye, X. Wang, E. Liu, J. Ye and D. Wang, ChemSusChem, 2018, 11, 1606.
44. J. Li, Z. Wang, H. Chen, Q. Zhang, H. Hu, L. Liu, J. Ye and D. Wang, Catal. Sci. Technol., 2021, 11, 4953.
45. P. Li, L. Liu, W. An, H. Wang, H. Guo, Y. Liang and W. Cui, Appl. Catal., B, 2020, 266, 118618.
46. A. Meng, L. Zhang, B. Cheng and J. Yu, Adv. Mater., 2019, 31, 1807660.
47. C. C. Hu and H. Teng, J. Catal., 2010, 272, 1.
48. C. Liu, J. Qian, Y. Ye, H. Zhou, C.-J. Sun, C. Sheehan, Z. Zhang, G. Wan, Y.-S. Liu, J. Guo, S. Li, H. Shin, S. Hwang, T. B. Gunnoe, W. A. Goddard and S. Zhang, Nat. Catal., 2020, 4, 36.
49. Z. Jia, S. Ning, Y. Tong, X. Chen, H. Hu, L. Liu, J. Ye and D. Wang, ACS Appl. Nano Mater., 2021, 4, 10485.
50. H. Zhao, P. Jiang and W. Cai, Chem.–Asian J., 2017, 12, 361.
51. C. Han, L. Ge, C. Chen, Y. Li, X. Xiao, Y. Zhang and L. Guo, Appl. Catal., B, 2014, 147, 546.
52. Q. Yu, Y. Dai, Y. Ling, Q. Wu, Z. Zhang and B. Feng, J. Environ. Chem. Eng., 2022, 10, 108693.
53. J. Yang, T. Liu, H. Zhou, W. Cao, C. Chen, X. He, C. Jiang, Y. Li and Y. Wang, J. Alloys Compd., 2022, 900, 163440.
54. H. Gao, P. Zhang, J. Hu, J. Pan, J. Fan and G. Shao, Appl. Surf. Sci., 2017, 391, 211.
55. Y. Lei, J. Ye, J. García-Antón and H. Liu, Chin. J. Catal., 2023, 53, 72.
56. P. S. Bednyakov, T. Sluka, A. K. Tagantsev, D. Damjanovic and N. Setter, Sci. Rep., 2015, 5, 15819.
57. X. Liu, F. Zhang, P. Long, T. Lu, H. Zeng, Y. Liu, R. L. Withers, Y. Li and Z. Yi, Adv. Mater., 2018, 30, 1801619.
58. H. Li, X. Quan, S. Chen and H. Yu, Appl. Catal., B, 2017, 209, 591.
59. H. Simons, A. B. Haugen, A. C. Jakobsen, S. Schmidt, F. Stöhr, M. Majkut, C. Detlefs, J. E. Daniels, D. Damjanovic and H. F. Poulsen, Nat. Mater., 2018, 17, 814.
60. B. Dai, Y. Yu, Y. Chen, H. Huang, C. Lu, J. Kou, Y. Zhao and Z. Xu, Adv. Funct. Mater., 2019, 29, 1970105.
61. X. Chen, Y. Li, X. Pan, D. Cortie, X. Huang and Z. Yi, Nat. Commun., 2016, 7, 12273.
62. H. Kato, K. Asakura and A. Kudo, J. Am. Chem. Soc., 2003, 125, 3082.
63. J. C. Yang, Y. L. Huang, Q. He and Y. H. Chu, J. Appl. Phys., 2014, 116, 066801.
64. A. S. Pisat, G. S. Rohrer and P. A. Salvador, Semicond. Sci. Technol., 2017, 32, 103001.
65. Y. Cui, J. Briscoe and S. Dunn, Chem. Mater., 2013, 25, 4215.
66. H. Kanda, N. Shibayama, A. J. Huckaba, Y. Lee, S. Paek, N. Klipfel, C. Roldán-Carmona, V. I. E. Queloz, G. Grancini, Y. Zhang, M. Abuhelaiqa, K. T. Cho, M. Li, M. D. Mensi, S. Kinge and M. K. Nazeeruddin, Energy Environ. Sci., 2020, 13, 1222.
67. H. Huang, S. Tu, X. Du and Y. Zhang, J. Colloid Interface Sci., 2018, 509, 113.
68. C. Paillard, X. Bai, I. C. Infante, M. Guennou, G. Geneste, M. Alexe, J. Kreisel and B. Dkhil, Adv. Mater., 2016, 28, 5153.
69. F. Chen, H. Huang, L. Guo, Y. Zhang and T. Ma, Angew. Chem., 2019, 131, 10164.
70. W. Song, P. A. Salvador and G. S. Rohrer, ACS Appl. Mater. Interfaces, 2018, 10, 41450.
71. O. Kwon, D. Seol, D. Lee, H. Han, I. Lindfors-Vrejoiu, W. Lee, S. Jesse, H. N. Lee, S. V. Kalinin, M. Alexe and Y. Kim, Adv. Mater., 2018, 30, 1703675.
72. J. He, Y. Zhao, S. Jiang and S. Song, Sol. RRL, 2021, 5, 2000446.
73. P. Wang, R. Shi, J. Zhao and T. Zhang, Adv. Sci., 2024, 11, 2305471.
74. J. Wang and S. Wang, Coord. Chem. Rev., 2022, 453, 214338.
75. Q. Luo, C. Lu, L. Liu and M. Zhu, Green Energy Environ., 2023, 8, 406.
76. X. Yu, V. L. Zholobenko, S. Moldovan, D. Hu, D. Wu, V. V. Ordomsky and A. Y. Khodakov, Nat. Energy, 2020, 5, 511.
77. H. Song, X. Meng, S. Wang, W. Zhou, X. Wang, T. Kako and J. Ye, J. Am. Chem. Soc., 2019, 141, 20507.
78. K. Wang, L. Luo, C. Wang and J. Tang, Chin. J. Catal., 2023, 46, 103.
79. G.-X. Dong, M.-R. Zhang, K. Su, Z.-L. Liu, M. Zhang and T.-B. Lu, J. Mater. Chem. A, 2023, 11, 9989.
80. S. Murcia-López, K. Villa, T. Andreu and J. R. Morante, ACS Catal., 2014, 4, 3013.
81. J. Yang, J. Hao, J. Wei, J. Dai and Y. Li, Fuel, 2020, 266, 117104.
82. S. Murcia-López, K. Villa, T. Andreu and J. R. Morante, ACS Catal., 2014, 4, 3013.
83. H. Wei, J. Cai, Y. Zhang, X. Zhang, E. A. Baranova, J. Cui, Y. Wang, X. Shu, Y. Qin, J. Liu and Y. Wu, RSC Adv., 2020, 10, 42619.
84. F. P. Koffyberg and F. A. Benko, J. Appl. Phys., 1982, 53, 1173.
85. X. Li, J. Xie, H. Rao, C. Wang and J. Tang, Angew. Chem., Int. Ed., 2020, 59, 19702.
86. S. Shoji, X. Peng, A. Yamaguchi, R. Watanabe, C. Fukuhara, Y. Cho, T. Yamamoto, S. Matsumura, M.-W. Yu, S. Ishii, T. Fujita, H. Abe and M. Miyauchi, Nat. Catal., 2020, 3, 148.
87. G. A. Ozin and F. Hugues, J. Phys. Chem., 1982, 86, 5174.
88. W. Hill, B. N. Shelimov and V. B. Kazansky, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 2381.
89. L. Yuliati, T. Hattori, H. Itoh and H. Yoshida, J. Catal., 2008, 257, 396.
90. J. Wen, J. Xie, X. Chen and X. Li, Appl. Surf. Sci., 2017, 391, 72.
91. D. Masih, Y. Ma and S. Rohani, Appl. Catal., B, 2017, 206, 556.
92. N. Li, Y. Li, R. Jiang, J. Zhou and M. Liu, Appl. Surf. Sci., 2019, 498, 143861.
93. A. Muhammad, M. Tahir, S. S. Al-Shahrani, A. Mahmood Ali and S. U. Rather, Appl. Surf. Sci., 2020, 504, 144177.
94. S. Shi, Z. Sun, C. Bao, T. Gao and Y. H. Hu, Int. J. Energy Res., 2020, 44, 2740.
95. Y. Zhou, L. Zhang and W. Wang, Nat. Commun., 2019, 10, 506.
96. L. Li, S. Fan, X. Mu, Z. Mi and C.-J. Li, J. Am. Chem. Soc., 2014, 136, 7793.
97. H. Song, X. Meng, Z. Jun Wang, H. Liu and J. Ye, Joule, 2019, 3, 1606.
98. K. Wittich, M. Krämer, N. Bottke and S. A. Schunk, ChemCatChem, 2020, 12, 2130.
99. G. C. Behera, K. Parida, N. F. Dummer, G. Whiting, N. Sahu, A. F. Carley, M. Conte, G. J. Hutchings and J. K. Bartley, Catal. Sci. Technol., 2013, 3, 1558.
100. E. Ramírez, R. Soto, R. Bringué, M. Iborra and J. Tejero, Ind. Eng. Chem. Res., 2020, 59, 20676.
101. J. Li, Y. He, L. Tan, P. Zhang, X. Peng, A. Oruganti, G. Yang, H. Abe, Y. Wang and N. Tsubaki, Nat. Catal., 2018, 1, 787.
102. C. D. Thomas, A. Cameron, R. E. Green, M. Bakkenes, L. J. Beaumont, Y. C. Collingham, B. F. N. Erasmus, M. F. de Siqueira, A. Grainger, L. Hannah, L. Hughes, B. Huntley, A. S. van Jaarsveld, G. F. Midgley, L. Miles, M. A. Ortega-Huerta, A. Townsend Peterson, O. L. Phillips and S. E. Williams, Nature, 2004, 427, 145.
103. C. Hepburn, E. Adlen, J. Beddington, E. A. Carter, S. Fuss, N. Mac Dowell, J. C. Minx, P. Smith and C. K. Williams, Nature, 2019, 575, 87.
104. S. D. Angeli, G. Monteleone, A. Giaconia and A. A. Lemonidou, Int. J. Hydrogen Energy, 2014, 39, 1979.
105. Y.-L. Lee, K. Lee, C. Hyun Ko and H.-S. Roh, Chem. Eng. J., 2022, 431, 134299.
106. S. Z. Abbas, V. Dupont and T. Mahmud, Int. J. Hydrogen Energy, 2017, 42, 2889.
107. J. Xu and G. F. Froment, AIChE J., 1989, 35, 97.
108. G. Arzamendi, P. M. Diéguez, M. Montes, J. A. Odriozola, E. F. Sousa-Aguiar and L. M. Gandía, Chem. Eng. J., 2009, 154, 168.
109. P. N. Kechagiopoulos, S. D. Angeli and A. A. Lemonidou, Appl. Catal., B, 2017, 205, 238.
110. D. Pashchenko and I. Makarov, Energy, 2021, 222, 119993.
111. D. Pashchenko, Int. J. Energy Res., 2020, 44, 438.
112. Z. Zhu, W. Guo, Y. Zhang, C. Pan, J. Xu, Y. Zhu and Y. Lou, Carbon Energy, 2021, 3, 519.
113. W. Li, J. Zhang and W. Wang, Coord. Chem. Rev., 2024, 503, 215638.
114. P. Gronchi, P. Centola and R. D. Rosso, Appl. Catal., A, 1997, 152, 83.
115. B. Christian Enger, R. Lødeng and A. Holmen, Appl. Catal., A, 2008, 346, 1.
116. C. Papadopoulou, H. Matralis and X. Verykios, in Catalysis for Alternative Energy Generation, 2012, pp. 57–127.
117. A. Álvarez, M. Borges, J. J. Corral-Pérez, J. G. Olcina, L. Hu, D. Cornu, R. Huang, D. Stoian and A. Urakawa, ChemPhysChem, 2017, 18, 3135.
118. L. F. Bobadilla, V. Garcilaso, M. A. Centeno and J. A. Odriozola, ChemSusChem, 2017, 10, 1193.
119. S. Xie, Q. Zhang, G. Liu and Y. Wang, Chem. Commun., 2016, 52, 35.
120. R. O. da Fonseca, R. C. Rabelo-Neto, R. C. C. Simões, L. V. Mattos and F. B. Noronha, Int. J. Hydrogen Energy, 2020, 45, 5182.
121. B. Faroldi, J. Múnera, J. M. Falivene, I. R. Ramos, Á. G. García, L. T. Fernández, S. G. Carrazán and L. Cornaglia, Int. J. Hydrogen Energy, 2017, 42, 16127.
122. J. Yang, J. Wang, J. Zhao, Y. Bai, H. Du, Q. Wang, B. Jiang and H. Li, J. CO₂ Util., 2022, 57, 101893.
123. C. Agrafiotis, H. von Storch, M. Roeb and C. Sattler, Renewable Sustainable Energy Rev., 2014, 29, 656.
124. M. F. Mark, W. F. Maier and F. Mark, Chem. Eng. Technol., 1997, 20, 361.
125. Y. Kathiraser, U. Oemar, E. T. Saw, Z. Li and S. Kawi, Chem. Eng. J., 2015, 278, 62.
126. E. le Saché and T. R. Reina, Prog. Energy Combust. Sci., 2022, 89, 100970.
127. C. Carrara, J. Múnera, E. A. Lombardo and L. M. Cornaglia, Top. Catal., 2008, 51, 98.
128. J.-H. Park and T.-S. Chang, Catal. Lett., 2019, 149, 3148.
129. I. de Dios García, A. Stankiewicz and H. Nigar, Catal. Today, 2021, 362, 72.
130. K. Nagaoka, Appl. Catal., A, 2004, 268, 151.
131. B. Pawelec, S. Damyanova, K. Arishtirova, J. L. G. Fierro and L. Petrov, Appl. Catal., A, 2007, 323, 188.
132. M. García-Diéguez, I. S. Pieta, M. C. Herrera, M. A. Larrubia and L. J. Alemany, Catal. Today, 2011, 172, 136.
133. B. Steinhauer, M. R. Kasireddy, J. Radnik and A. Martin, Appl. Catal., A, 2009, 366, 333.
134. F. Menegazzo, M. Signoretto, F. Pinna, P. Canton and N. Pernicone, Appl. Catal., A, 2012, 439–440, 80.
135. S. Zhang, M. Ying, J. Yu, W. Zhan, L. Wang, Y. Guo and Y. Guo, Appl. Catal., B, 2021, 291, 120074.
136. K. Y. Kim, J. H. Lee, H. Lee, W. Y. Noh, E. H. Kim, E. C. Ra, S. K. Kim, K. An and J. S. Lee, ACS Catal., 2021, 11, 11091.
137. A. Abdulrasheed, A. A. Jalil, Y. Gambo, M. Ibrahim, H. U. Hambali and M. Y. Shahul Hamid, Renewable Sustainable Energy Rev., 2019, 108, 175.
138. O. Muraza and A. Galadima, Int. J. Energy Res., 2015, 39, 1196.
139. Y. Gao, J. Jiang, Y. Meng, F. Yan and A. Aihemaiti, Energy Convers. Manage., 2018, 171, 133.
140. A. Tavasoli and G. Ozin, Joule, 2018, 2, 571.
141. A. V. Tavasoli, M. Preston and G. Ozin, Energy Environ. Sci., 2021, 14, 3098.
142. F. Pan, X. Xiang, W. Deng, H. Zhao, X. Feng and Y. Li, ChemCatChem, 2018, 10, 940.
143. M. Mao, Q. Zhang, Y. Yang, Y. Li, H. Huang, Z. Jiang, Q. Hu and X. Zhao, Green Chem., 2018, 20, 2857.
144. L. Zhu, T. Ding, M. Gao, C. K. N. Peh and G. W. Ho, Adv. Energy Mater., 2019, 9, 1900250.
145. B. Han, W. Wei, L. Chang, P. Cheng and Y. H. Hu, ACS Catal., 2016, 6, 494.
146. H. Liu, H. Song, W. Zhou, X. Meng and J. Ye, Angew. Chem., Int. Ed., 2018, 57, 16781.
147. M. A. Boda, R. L. Withers, Y. Liu, J. Ye and Z. Yi, J. Mater. Chem. A, 2022, 10, 22977.
148. Y. Li, D. Li, H. Liu, Y. Lei, R. Zhao, D. He, Z. Zheng, H. Luo and A. Liu, Catal. Sci. Technol., 2024, 14, 2722.
149. H. Liu, X. Meng, W. Yang, G. Zhao, D. He and J. Ye, Chin. J. Catal., 2021, 42, 1976.
150. N. Ma, N. Tanen, A. Verma, Z. Guo, T. Luo, H. (Grace) Xing and D. Jena, Appl. Phys. Lett., 2016, 109, 212101.
151. Y. Zhao, J. Qiao, Z. Yu, P. Yu, K. Xu, S. P. Lau, W. Zhou, Z. Liu, X. Wang, W. Ji and Y. Chai, Adv. Mater., 2017, 29, 1604230.
152. H. Song and J. Ye, Trends Chem., 2022, 4, 1094.
153. X. Meng, X. Cui, N. P. Rajan, L. Yu, D. Deng and X. Bao, Chem, 2019, 5, 2296.
154. C. Peng, G. Reid, H. Wang and P. Hu, J. Chem. Phys., 2017, 147, 030901.
155. D. Hu, V. V. Ordomsky and A. Y. Khodakov, Appl. Catal., B, 2021, 286, 119913.
156. S. Yuan, Y. Li, J. Peng, Y. M. Questell-Santiago, K. Akkiraju, L. Giordano, D. J. Zheng, S. Bagi, Y. Román-Leshkov and Y. Shao-Horn, Adv. Energy Mater., 2020, 10, 2002154.
157. R. A. Periana, D. J. Taube, S. Gamble, H. Taube, T. Satoh and H. Fujii, Science, 1998, 280, 560.
158. R. Horn and R. Schlögl, Catal. Lett., 2015, 145, 23.
159. A. A. Latimer, A. R. Kulkarni, H. Aljama, J. H. Montoya, J. S. Yoo, C. Tsai, F. Abild-Pedersen, F. Studt and J. K. Nørskov, Nat. Mater., 2017, 16, 225.
160. A. A. Latimer, H. Aljama, A. Kakekhani, J. S. Yoo, A. Kulkarni, C. Tsai, M. Garcia-Melchor, F. Abild-Pedersen and J. K. Nørskov, Phys. Chem. Chem. Phys., 2017, 19, 3575.
161. N. J. Gunsalus, A. Koppaka, S. H. Park, S. M. Bischof, B. G. Hashiguchi and R. A. Periana, Chem. Rev., 2017, 117, 8521.
162. J. A. Labinger and J. E. Bercaw, Nature, 2002, 417, 507.
163. A. I. Olivos-Suarez, À. Szécsényi, E. J. M. Hensen, J. Ruiz-Martinez, E. A. Pidko and J. Gascon, ACS Catal., 2016, 6, 2965.
164. Á. Szécsényi, G. Li, J. Gascon and E. A. Pidko, ACS Catal., 2018, 8, 7961.
165. J. Kim, M. S. Abbott, D. B. Go and J. C. Hicks, ACS Energy Lett., 2016, 1, 94.
166. C. R. H. de Smet, M. H. J. M. de Croon, R. J. Berger, G. B. Marin and J. C. Schouten, Appl. Catal., A, 1999, 187, 33.
167. D. Hickman, J. Catal., 1992, 138, 267.
168. J. Haber and E. M. Serwicka, React. Kinet. Catal. Lett., 1987, 35, 369.
169. G. Groppi, W. Ibashi, M. Valentini and P. Forzatti, Chem. Eng. Sci., 2001, 56, 831.
170. K. Delgado, L. Maier, S. Tischer, A. Zellner, H. Stotz and O. Deutschmann, Catalysts, 2015, 5, 871.
171. R. Schwiedernoch, S. Tischer, C. Correa and O. Deutschmann, Chem. Eng. Sci., 2003, 58, 633.
172. V. R. Choudhary, A. M. Rajput and B. Prabhakar, J. Catal., 1993, 139, 326.
173. D. Dissanayake, J. Catal., 1991, 132, 117.
174. G. A. Karim and G. Zhou, Int. J. Hydrogen Energy, 1993, 18, 125.
175. A. Mosayebi, J. Taiwan Inst. Chem. Eng., 2020, 114, 36.
176. K. J. Warren and J. R. Scheffe, Mater. Today Energy, 2018, 9, 39.
177. A. V. de Vekki and S. T. Marakaev, Russ. J. Appl. Chem., 2009, 82, 521.
178. K. T. Dinh, M. M. Sullivan, P. Serna, R. J. Meyer, M. Dincă and Y. Román-Leshkov, ACS Catal., 2018, 8, 8306.
179. M. H. Groothaert, P. J. Smeets, B. F. Sels, P. A. Jacobs and R. A. Schoonheydt, J. Am. Chem. Soc., 2005, 127, 1394.
180. C. Berger-Karin, J. Radnik and E. V. Kondratenko, J. Catal., 2011, 280, 116.
181. L. Li, S. He, Y. Song, J. Zhao, W. Ji and C.-T. Au, J. Catal., 2012, 288, 54.
182. Y. Hu and E. Ruckenstein, J. Catal., 1996, 158, 260.
183. T. V. Choudhary and V. R. Choudhary, Angew. Chem., Int. Ed., 2008, 47, 1828.
184. C. T. Au, H. Y. Wang and H. L. Wan, J. Catal., 1996, 158, 343.
185. R. K. Singha, A. Shukla, A. Yadav, L. N. Sivakumar Konathala and R. Bal, Appl. Catal., B, 2017, 202, 473.
186. J. Saupsor, C. Pei, H. Li, S. Wongsakulphasatch, P. Kim-Lohsoontorn, S. Ratchahat, W. Kiatkittipong, S. Assabumrungrat and J. Gong, Energy Fuels, 2021, 35, 11580.
187. Y. Feng, N. Wang and X. Guo, Fuel, 2021, 290, 120094.
188. L. Zeng, Z. Cheng, J. A. Fan, L.-S. Fan and J. Gong, Nat. Rev. Chem., 2018, 2, 349.
189. S. Kim, J. Lauterbach and E. Sasmaz, ACS Catal., 2021, 11, 8247.
190. I. Grčić, J. Marčec, L. Radetić, A.-M. Radovan, I. Melnjak, I. Jajčinović and I. Brnardić, Environ. Sci. Pollut. Res., 2021, 28, 18354.
191. S. Sugiyama, Y. Hayashi, I. Okitsu, N. Shimoda, M. Katoh, A. Furube, Y. Kato and W. Ninomiya, Catalysts, 2020, 10, 559.
192. S. Murcia-López, K. Villa, T. Andreu and J. R. Morante, Chem. Commun., 2015, 51, 7249.
193. H. Jiang, X. Peng, A. Yamaguchi, S. Ueda, T. Fujita, H. Abe and M. Miyauchi, Sol. RRL, 2019, 3, 1900076.
194. K. R. Thampi, J. Kiwi and M. Grätzel, Catal. Lett., 1988, 1, 109.
195. F. Schüth, Science, 2019, 363, 1282.
196. C. Mesters, Annu. Rev. Chem. Biomol. Eng., 2016, 7, 223.
197. H. Cruchade, I. C. Medeiros-Costa, N. Nesterenko, J.-P. Gilson, L. Pinard, A. Beuque and S. Mintova, ACS Catal., 2022, 12, 14533.
198. B. Beck, V. Fleischer, S. Arndt, M. G. Hevia, A. Urakawa, P. Hugo and R. Schomäcker, Catal. Today, 2014, 228, 212.
199. C. Karakaya, H. Zhu, C. Loebick, J. G. Weissman and R. J. Kee, Catal. Today, 2018, 312, 10.
200. J. Sun, J. Thybaut and G. Marin, Catal. Today, 2008, 137, 90.
201. J. H. Lunsford, Angew. Chem., Int. Ed. Engl., 1995, 34, 970.
202. E. E. Wolf, J. Phys. Chem. Lett., 2014, 5, 986.
203. P. Schwach, X. Pan and X. Bao, Chem. Rev., 2017, 117, 8497.
204. Y. Gao, L. Neal, D. Ding, W. Wu, C. Baroi, A. M. Gaffney and F. Li, ACS Catal., 2019, 9, 8592.
205. Y. Simon, F. Baronnet, G. M. Côme and P. M. Marquaire, Stud. Surf. Sci. Catal., 2004, 147, 571.
206. P. M. Couwenberg, Q. Chen and G. B. Marin, Ind. Eng. Chem. Res., 1996, 35, 3999.
207. Z. Stansch, L. Mleczko and M. Baerns, Ind. Eng. Chem. Res., 1997, 36, 2568.
208. S. Da Ros, T. Barbalho Fontoura, M. Schwaab, N. Castro de Jesus and J. Pinto, Processes, 2021, 9, 2196.
209. M. Y. Sinev, Z. T. Fattakhova, V. I. Lomonosov and Y. A. Gordienko, J. Nat. Gas Chem., 2009, 18, 273.
210. H. Zanthoff and M. Baerns, Ind. Eng. Chem. Res., 1990, 29, 2.
211. Q. Chen, P. M. Couwenberg and G. B. Marin, Catal. Today, 1994, 21, 309.
212. Q. Chen, J. H. B. J. Hoebink and G. B. Marin, Ind. Eng. Chem. Res., 1991, 30, 2088.
213. Q. Chen, P. M. Couwenberg and G. B. Marin, AIChE J., 1994, 40, 521.
214. O. V. Buyevskaya, M. Rothaemel, H. W. Zanthoff and M. Baerns, J. Catal., 1994, 146, 346.
215. P. N. Kechagiopoulos, J. W. Thybaut and G. B. Marin, Ind. Eng. Chem. Res., 2014, 53, 1825.
216. R. H. Nibbelke, J. Scheerova, M. H. J. M. Decroon and G. B. Marin, J. Catal., 1995, 156, 106.
217. J. Sun, J. Thybaut and G. Marin, Catal. Today, 2008, 137, 90.
218. T. Ito and J. H. Lunsford, Nature, 1985, 314, 721–722.
219. B. Wood, J. Catal., 2004, 225, 300.
220. C. Hammond, N. Dimitratos, J. A. Lopez-Sanchez, R. L. Jenkins, G. Whiting, S. A. Kondrat, M. H. ab Rahim, M. M. Forde, A. Thetford, H. Hagen, E. E. Stangland, J. M. Moulijn, S. H. Taylor, D. J. Willock and G. J. Hutchings, ACS Catal., 2013, 3, 1835.
221. J.-F. Wu, X.-D. Gao, L.-M. Wu, W. D. Wang, S.-M. Yu and S. Bai, ACS Catal., 2019, 9, 8677.
222. Z. Liang, T. Li, M. Kim, A. Asthagiri and J. F. Weaver, Science, 2017, 356, 299.
223. N. Basickes, T. E. Hogan and A. Sen, J. Am. Chem. Soc., 1996, 118, 13111.
224. T. Zimmermann, M. Soorholtz, M. Bilke and F. Schüth, J. Am. Chem. Soc., 2016, 138, 12395.
225. I. H. Hristov and T. Ziegler, Organometallics, 2003, 22, 1668.
226. T. M. Gilbert, I. Hristov and T. Ziegler, Organometallics, 2001, 20, 1183.
227. S. Chempath and A. T. Bell, J. Am. Chem. Soc., 2006, 128, 4650.
228. S. Song, H. Song, L. Li, S. Wang, W. Chu, K. Peng, X. Meng, Q. Wang, B. Deng, Q. Liu, Z. Wang, Y. Weng, H. Hu, H. Lin, T. Kako and J. Ye, Nat. Catal., 2021, 4, 1032.
229. H. Song, X. Meng, Z. Wang, H. Liu and J. Ye, Joule, 2019, 3, 1606.
230. H. Li, J. Li, Z. Ai, F. Jia and L. Zhang, Angew. Chem., Int. Ed., 2018, 57, 122.
231. N. Li, R. Jiang, Y. Li, J. Zhou, Q. Ma, S. Shen and M. Liu, ACS Sustain. Chem. Eng., 2019, 7, 11455.
232. D. Gerceker, A. H. Motagamwala, K. R. Rivera-Dones, J. B. Miller, G. W. Huber, M. Mavrikakis and J. A. Dumesic, ACS Catal., 2017, 7, 2088.
233. K. Dutta, M. Shahryari, C.-J. Li and J. Kopyscinski, Appl. Catal. A Gen., 2023, 663, 119319.
234. D. Bajec, A. Kostyniuk, A. Pohar and B. Likozar, Chem. Eng. J., 2020, 396, 125182.
235. M. Kottwitz, Y. Li, R. M. Palomino, Z. Liu, G. Wang, Q. Wu, J. Huang, J. Timoshenko, S. D. Senanayake, M. Balasubramanian, D. Lu, R. G. Nuzzo and A. I. Frenkel, ACS Catal., 2019, 9, 8738.
236. S. K. Kim, H. W. Kim, S. J. Han, S. W. Lee, J. Shin and Y. T. Kim, Commun. Chem., 2020, 3, 58.
237. S. F. Moya, R. L. Martins, A. Ota, E. L. Kunkes, M. Behrens and M. Schmal, Appl. Catal., A, 2012, 411–412, 105.
238. H. Sheng, E. P. Schreiner, W. Zheng and R. F. Lobo, ChemPhysChem, 2018, 19, 504.
239. L. Wang, L. Tao, M. Xie, G. Xu, J. Huang and Y. Xu, Catal. Lett., 1993, 21, 35.
240. D. Soulivong, S. Norsic, M. Taoufik, C. Coperet, J. Thivolle-Cazat, S. Chakka and J.-M. Basset, J. Am. Chem. Soc., 2008, 130, 5044.
241. X. Guo, G. Fang, G. Li, H. Ma, H. Fan, L. Yu, C. Ma, X. Wu, D. Deng, M. Wei, D. Tan, R. Si, S. Zhang, J. Li, L. Sun, Z. Tang, X. Pan and X. Bao, Science, 2014, 344, 616–619.
242. Y. Xiao and A. Varma, ACS Catal., 2018, 8, 2735.
243. Z. Li, Y. Xiao, P. R. Chowdhury, Z. Wu, T. Ma, J. Z. Chen, G. Wan, T.-H. Kim, D. Jing, P. He, P. J. Potdar, L. Zhou, Z. Zeng, X. Ruan, J. T. Miller, J. P. Greeley, Y. Wu and A. Varma, Nat. Catal., 2021, 4, 882.
244. Y. Kato, H. Yoshida and T. Hattori, Chem. Commun., 1998, 2389–2390.
245. S. Murcia-López, M. C. Bacariza, K. Villa, J. M. Lopes, C. Henriques, J. R. Morante and T. Andreu, ACS Catal., 2017, 7, 2878.
246. Y. Kato, N. Matsushita, H. Yoshida and T. Hattori, Catal. Commun., 2002, 3, 99.
247. S. Wu, L. Wang and J. Zhang, J. Photochem. Photobiol., C, 2021, 46, 100400.
248. W. Zhang, C. Fu, J. Low, D. Duan, J. Ma, W. Jiang, Y. Chen, H. Liu, Z. Qi, R. Long, Y. Yao, X. Li, H. Zhang, Z. Liu, J. Yang, Z. Zou and Y. Xiong, Nat. Commun., 2022, 13, 2806.
249. T. Sakthivel, G. Venugopal, A. Durairaj, S. Vasanthkumar and X. Huang, J. Ind. Eng. Chem., 2019, 72, 18.
250. M. M. Vijatovic Petrovic and J. D. Bobic, in Magnetic, Ferroelectric, and Multiferroic Metal Oxides, Elsevier, 2018, pp. 35–49.
251. X. Deng, Y. Wu, Z. Gao and G. Zhou, J. Mater. Chem. A, 2024, 12, 567.
252. A. G. Chynoweth, Phys. Rev., 1956, 102, 705.
253. A. M. Glass, D. von der Linde, D. H. Auston and T. J. Negran, J. Electron. Mater., 1975, 4, 915.
254. A. M. Glass, D. von der Linde and T. J. Negran, Appl. Phys. Lett., 1974, 25, 233.
255. S. Tu, Y. Zhang, A. H. Reshak, S. Auluck, L. Ye, X. Han, T. Ma and H. Huang, Nano Energy, 2019, 56, 840.
256. Y. Wang, M. Zhang, J. Liu, H. Zhang, F. Li, C. Tseng, B. Yang, G. Smith, J. Zhai, Z. Zhang, S. Dunn and H. Yan, Adv. Energy Mater., 2020, 10, 2001802.
257. Y. AlSalka, L. I. Granone, W. Ramadan, A. Hakki, R. Dillert and D. W. Bahnemann, Appl. Catal., B, 2019, 244, 1065.
258. I. Dierking and S. Al-Zangana, Nanomaterials, 2017, 7, 305.
259. I. Dierking, Nanomaterials, 2018, 8, 453.
260. S. K. Gupta, D. Budaszewski and D. P. Singh, Eur. Phys. J.: Spec. Top., 2022, 231, 673.
261. R. B. Meyer, L. Liebert, L. Strzelecki and P. Keller, J. Phys., Lett., 1975, 36, 69.
262. S. K. Gupta, Z. Sun, H.-S. Kwok and A. K. Srivastava, Liq. Cryst., 2020, 47, 1162.
263. S. Mukherjee, Z. Yuan, Z. Sun, A. Li, C. Kang, H.-S. Kwok and A. K. Srivastava, Opt. Express, 2021, 29, 8258.
264. P. Berthele, E. Gros, B. Fracasso and J.-L. de B. de la Tocnaye, Ferroelectrics, 1998, 214, 117.
265. S. Al-Zangana, M. Turner and I. Dierking, J. Appl. Phys., 2017, 121, 085105.
266. A. Mertelj, D. Lisjak, M. Drofenik and M. Čopič, Nature, 2013, 504, 237.
267. Y. Yuan, T. J. Reece, P. Sharma, S. Poddar, S. Ducharme, A. Gruverman, Y. Yang and J. Huang, Nat. Mater., 2011, 10, 296.
268. B. B. Tian, J. L. Wang, S. Fusil, Y. Liu, X. L. Zhao, S. Sun, H. Shen, T. Lin, J. L. Sun, C. G. Duan, M. Bibes, A. Barthélémy, B. Dkhil, V. Garcia, X. J. Meng and J. H. Chu, Nat. Commun., 2016, 7, 11502.
269. Y. Liu, H. Aziguli, B. Zhang, W. Xu, W. Lu, J. Bernholc and Q. Wang, Nature, 2018, 562, 96.
270. B. Neese, B. Chu, S.-G. Lu, Y. Wang, E. Furman and Q. M. Zhang, Science, 2008, 321, 821.
271. H. Li, F. Liu, B. Fan, D. Ai, Z. Peng and Q. Wang, Small Methods, 2018, 2, 1700399.
272. Y. Cui, Z. Wang, B. Li, Y. Yan, R. Xu, M. Meng and Y. Yan, Nano Energy, 2022, 99, 107429.
273. Z. Wei, T. Ji, X. Zhou, J. Guo, X. Yu, H. Liu and J. Wang, Small, 2023, 19, 2304202.
274. C. Yang, P. Wang, J. Li, Q. Wang, P. Xu, S. You, Q. Zheng and G. Zhang, Chem. Eng. J., 2021, 417, 129340.
275. S. Horiuchi and Y. Tokura, Nat. Mater., 2008, 7, 357.
276. A. L. Solomon, Phys. Rev., 1956, 104, 1191.
277. D. Bordeaux, J. Bornarel, A. Capiomont, J. Lajzerowicz-Bonneteau, J. Lajzerowicz and J. F. Legrand, Phys. Rev. Lett., 1973, 31, 314.
278. P. Gruner-Bauer and E. Dormann, J. Phys.: Condens. Matter, 1992, 4, 5599.
279. J. Valasek, Phys. Rev., 1921, 17, 475.
280. Y. Sekine, R. Akiyoshi and S. Hayami, Coord. Chem. Rev., 2022, 469, 214663.
281. T. Hang, W. Zhang, H.-Y. Ye and R.-G. Xiong, Chem. Soc. Rev., 2011, 40, 3577.
282. T. Zhang, K. Xu, J. Li, L. He, D. W. Fu, Q. Ye and R. G. Xiong, Natl. Sci. Rev., 2023, 10, nwac240.
283. R. Naaman, Y. Paltiel and D. H. Waldeck, Nat. Rev. Chem., 2019, 3, 250.
284. X.-B. Han, C.-Y. Chai, B.-D. Liang, C.-C. Fan and W. Zhang, CrystEngComm, 2022, 24, 1507.
285. N. A. Spaldin, J. Solid State Chem., 2012, 195, 2.
286. H. Huyan, L. Li, C. Addiego, W. Gao and X. Pan, Natl. Sci. Rev., 2019, 6, 669.
287. W. Zheng, X. Wang, X. Zhang, B. Chen, H. Suo, Z. Xing, Y. Wang, H. Wei, J. Chen, Y. Guo and F. Wang, Adv. Mater., 2023, 35, 2205410.
288. D. Li, J. Shi, Y. Xu, Y. Luo, H. Wu and Q. Meng, Natl. Sci. Rev., 2018, 5, 559.
289. Z. Sun, X. Liu, T. Khan, C. Ji, M. A. Asghar, S. Zhao, L. Li, M. Hong and J. Luo, Angew. Chem., Int. Ed., 2016, 55, 6545.
290. J.-X. Gao, W.-Y. Zhang, Z.-G. Wu, Y.-X. Zheng and D.-W. Fu, J. Am. Chem. Soc., 2020, 142, 4756.
291. C. Ji, S. Wang, L. Li, Z. Sun, M. Hong and J. Luo, Adv. Funct. Mater., 2019, 29, 1805038.
292. K. T. Butler, J. M. Frost and A. Walsh, Energy Environ. Sci., 2015, 8, 838.
293. Y. Sun, G. Niu, W. Ren, X. Meng, J. Zhao, W. Luo, Z.-G. Ye and Y.-H. Xie, ACS Nano, 2021, 15, 10982.
294. C. Wang, Y. Ding, B. Liu, B. Weng, J. Hofkens and M. B. J. Roeffaers, Chem. Commun., 2023, 59, 3122.
295. K. Li, S. Li, W. Zhang, Z. Shi, D. Wu, X. Chen, P. Lin, Y. Tian and X. Li, J. Colloid Interface Sci., 2021, 596, 376.
296. J. San Martin, N. Dang, E. Raulerson, M. C. Beard, J. Hartenberger and Y. Yan, Angew. Chem., Int. Ed., 2022, 61, e202205572.
297. X. Zhu, Y. Lin, J. San Martin, Y. Sun, D. Zhu and Y. Yan, Nat. Commun., 2019, 10, 2843.
298. K. Chang, J. Liu, H. Lin, N. Wang, K. Zhao, A. Zhang, F. Jin, Y. Zhong, X. Hu, W. Duan, Q. Zhang, L. Fu, Q.-K. Xue, X. Chen and S.-H. Ji, Science, 2016, 353, 274.
299. W. Ding, J. Zhu, Z. Wang, Y. Gao, D. Xiao, Y. Gu, Z. Zhang and W. Zhu, Nat. Commun., 2017, 8, 14956.
300. A. Chandrasekaran, A. Mishra and A. K. Singh, Nano Lett., 2017, 17, 3290.
301. J. Muñoz, Adv. Mater., 2024, 36, 2305546.
302. Q. Yang, W. Xiong, L. Zhu, G. Gao and M. Wu, J. Am. Chem. Soc., 2017, 139, 11506.
303. K. L. Seyler, D. Zhong, D. R. Klein, S. Gao, X. Zhang, B. Huang, E. Navarro-Moratalla, L. Yang, D. H. Cobden, M. A. McGuire, W. Yao, D. Xiao, P. Jarillo-Herrero and X. Xu, Nat. Phys., 2018, 14, 277.
304. B. Huang, G. Clark, E. Navarro-Moratalla, D. R. Klein, R. Cheng, K. L. Seyler, D. Zhong, E. Schmidgall, M. A. McGuire, D. H. Cobden, W. Yao, D. Xiao, P. Jarillo-Herrero and X. Xu, Nature, 2017, 546, 270.
305. C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao, W. Bao, C. Wang, Y. Wang, Z. Q. Qiu, R. J. Cava, S. G. Louie, J. Xia and X. Zhang, Nature, 2017, 546, 265.
306. L. Kou, Y. Ma, T. Liao, A. Du and C. Chen, Phys. Rev. Appl., 2018, 10, 024043.
307. Y. Lei, X. Sala, J. García-Antón and J. Muñoz, Small Methods, 2024.
308. Y. Lei, Á. Campos-Lendínez, X. Sala, J. García-Antón and J. Muñoz, Small Struct., 2025, 6, 2400240.
309. J. Muñoz, M. Palacios-Corella, I. J. Gómez, L. Zajíčková and M. Pumera, Adv. Mater., 2022, 34, 2206382.
310. J. L. Giocondi and G. S. Rohrer, Top. Catal., 2008, 49, 18.
311. G. F. Nataf, M. Guennou, J. Kreisel, P. Hicher, R. Haumont, O. Aktas, E. K. H. Salje, L. Tortech, C. Mathieu, D. Martinotti and N. Barrett, Phys. Rev. Mater., 2017, 1, 074410.
312. K. Garrity, A. M. Kolpak, S. Ismail-Beigi and E. I. Altman, Adv. Mater., 2010, 22, 2969.
313. M. Du, W. Liu, N. Liu, Y. Ling and S. Kang, Nano Energy, 2024, 124, 109495.
314. S. Park, C. W. Lee, M.-G. Kang, S. Kim, H. J. Kim, J. E. Kwon, S. Y. Park, C.-Y. Kang, K. S. Hong and K. T. Nam, Phys. Chem. Chem. Phys., 2014, 16, 10408.
315. F. Wu, Y. Yu, H. Yang, L. N. German, Z. Li, J. Chen, W. Yang, L. Huang, W. Shi, L. Wang and X. Wang, Adv. Mater., 2017, 29, 1701432.
316. W. Li, F. Wang, M. Li, X. Chen, Z. Ren, H. Tian, X. Li, Y. Lu and G. Han, Nano Energy, 2018, 45, 304.
317. R. Munprom, P. A. Salvador and G. S. Rohrer, J. Mater. Chem. A, 2016, 4, 2951.
318. N. V. Burbure, P. A. Salvador and G. S. Rohrer, Chem. Mater., 2010, 22, 5831.
319. A. Kakekhani and S. Ismail-Beigi, ACS Catal., 2015, 5, 4537.
320. N. P. Zangeneh, S. Sharifnia and E. Karamian, Environ. Sci. Pollut. Res., 2020, 27, 5912.
321. Z. Du, C. Petru, X. Yang, F. Chen, S. Fang, F. Pan, Y. Gang, H.-C. Zhou, Y. H. Hu and Y. Li, J. CO₂ Util., 2023, 67, 102317.
322. L. Jiang, Y. Zhang, Y. Qiu and Z. Yi, RSC Adv., 2014, 4, 3165.
323. R. Von der Mühll, A. Sadel and P. Hagenmuller, J. Solid State Chem., 1984, 51, 176.
324. K. Garrity, A. Kakekhani, A. Kolpak and S. Ismail-Beigi, Phys. Rev. B: Condens. Matter Mater. Phys., 2013, 88, 045401.
325. B. O. Alawode and A. M. Kolpak, J. Phys. Chem. Lett., 2016, 7, 1310.
326. Z. Li, M. A. Boda, X. Pan and Z. Yi, ACS Sustain. Chem. Eng., 2019, 7, 19042.
327. L. Meng, Z. Chen, Z. Ma, S. He, Y. Hou, H.-H. Li, R. Yuan, X.-H. Huang, X. Wang, X. Wang and J. Long, Energy Environ. Sci., 2018, 11, 294.
328. B. Dai, Y. Chen, S. Hao, H. Huang, J. Kou, C. Lu, Z. Lin and Z. Xu, J. Phys. Chem. Lett., 2020, 11, 7407.
329. A. M. Schultz, Y. Zhang, P. A. Salvador and G. S. Rohrer, ACS Appl. Mater. Interfaces, 2011, 3, 1562.
330. H. S. Kushwaha, A. Halder and R. Vaish, J. Mater. Sci., 2018, 53, 1414.
331. Y. Yun and E. I. Altman, J. Am. Chem. Soc., 2007, 129, 15684.
332. M. H. Zhao, D. A. Bonnell and J. M. Vohs, Surf. Sci., 2008, 602, 2849.
333. X. Tang, J. Shang, Y. Ma, Y. Gu, C. Chen and L. Kou, ACS Appl. Mater. Interfaces, 2020, 12, 39561.
334. T. L. Wan, L. Ge, Y. Pan, Q. Yuan, L. Liu, S. Sarina and L. Kou, Nanoscale, 2021, 13, 7096.
335. J. Y. Y. Loh, N. P. Kherani and G. A. Ozin, Nat. Sustainability, 2021, 4, 466.
336. W. Ding, J. Lu, X. Tang, L. Kou and L. Liu, ACS Omega, 2023, 8, 6164.
337. J. Xie, C. Guo, P. Yang, X. Wang, D. Liu and C. M. Li, Nano Energy, 2017, 31, 28.
338. C. Noguera, J. Phys.: Condens. Matter, 2000, 12, R367.
339. A. M. Kolpak, I. Grinberg and A. M. Rappe, Phys. Rev. Lett., 2007, 98, 166101.
340. C. Zhen, J. C. Yu, G. Liu and H.-M. Cheng, Chem. Commun., 2014, 50, 10416.
341. F. Chen, Z. Ren, S. Gong, X. Li, G. Shen and G. Han, Chem.– Eur. J., 2016, 22, 12160.
342. A. J. Welch, I. A. Digdaya, R. Kent, P. Ghougassian, H. A. Atwater and C. Xiang, ACS Energy Lett., 2021, 1540.
343. J. Baltrusaitis, I. Jansen and J. D. Schuttlefield Christus, Catal. Sci. Technol., 2014, 4, 2397.
344. R. Li, L.-P. Qiu, S.-Z. Cao, Z. Li, S.-L. Gao, J. Zhang, S. Ramakrishna, Y.-Z. Long, R. Li, L.-P. Qiu, S.-Z. Cao, Z. Li, S.-L. Gao, J. Zhang, Y.-Z. Long and S. Ramakrishna, Adv. Funct. Mater., 2024, 2316725.
345. P. Rovillain, R. de Sousa, Y. Gallais, A. Sacuto, M. A. Méasson, D. Colson, A. Forget, M. Bibes, A. Barthélémy and M. Cazayous, Nat. Mater., 2010, 9, 975.
346. A. Kumar, J. F. Scott and R. S. Katiyar, Appl. Phys. Lett., 2011, 99, 062504.

---