Metal–organic framework-derived heterojunctions as nanocatalysts for photocatalytic hydrogen production

Abstract

Based on various well-defined components and structures, high specific surface areas and adjustable pore sizes, metal–organic frameworks (MOFs) have not only been widely used for gas sorption and separation as well as catalytic and sensing materials but have also been applied as the precursors for nanomaterials with promising applications such as in energy storage and conversion. In this review, we reported a summary of the recent progress of MOF-derived heterojunctions as nanocatalysts for photocatalytic hydrogen production from water splitting, which is organized in two main categories: (i) using the pore chemistry of MOFs to accommodate small species for MOF-derived heterojunctions; (ii) using the surface chemistry of MOFs to combine with other functional species for MOF-derived heterojunctions. The fundamentals of the rational design and synthesis of MOF-derived heterojunctions for photocatalytic hydrogen production were discussed in detail.

---

1 Introduction

With the intensification of the global energy dilemma, the massive burning and overuse of fossil fuels, it has been imminent to find new, green, renewable and efficient energy sources, among which hydrogen energy is considered to be an ideal alternative source in the 21st century owing to its characteristics of extensive sources, high calorific value of combustion, pollution-free nature and so on. To produce hydrogen, photocatalytic water splitting is a very eco-friendly technology with abundant water used as the raw material and the endless solar energy as the driving force.^1,2

In principle, the overall water splitting contains two half-reactions: the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER). As an uphill reaction with a high Gibbs free energy change of 237.13 kJ mol⁻¹, it requires photons with energy higher than 1.23 eV.³ There are two types of water splitting catalysts, i.e., homogeneous catalysts and heterogeneous catalysts, in which the semiconductor catalysts as heterogeneous catalysts have received significant attention because of their adjustable catalytic properties, high thermal stability and strong anti-toxicity.⁴ The photocatalytic water splitting process by semiconductor-based photocatalysts mainly consists of three steps: (i) when the photoenergy absorbed by the photocatalyst exceeds its band-gap, the electrons in the valence band (VB) can be excited to the conduction band (CB) and simultaneously leave behind positive holes. (ii) The separated charge carriers migrate to the surface of the photocatalyst. (iii) The protons of water molecules accept photo-generated electrons and are reduced to hydrogen; meanwhile, the holes and oxygen atoms of the water molecules combine, causing the production of oxygen (Fig. 1).⁵ For these two half-reactions, the bottom potential of CB is required to be more negative than the redox potential of the H⁺/H₂ energy level (0 V vs. NHE, pH = 0), while the top potential of VB must be higher than the oxidation potential of the O₂/H₂O energy level (1.23 V vs. NHE, pH = 0). As a result, the minimum band-gap required for photocatalytic water splitting is 1.23 eV.⁶ Furthermore, it should be noted that the potential positions of CB and VB of a photocatalyst determine the ability of reduction and oxidation; thus, an appropriate band structure in well-designed semiconductors is the key for the better performance of the photocatalytic system.

Fig. 1 The basic principle of semiconductor-based photocatalyst for water splitting.

However, because of the inefficiency of light utilization, the lack of a suitable energy-band structure and the rapid recombination of photo-generated electron–hole pairs, it is still vital to obtain highly effective nanocatalysts for water splitting under sunlight, which demands a deeper understanding of the fundamentals for the rational design and synthesis of high-performance water splitting nanocatalysts.

In a practical system, semiconductor photocatalysts,⁷ such as TiO₂,⁸ ZnO,⁹ CdS,¹⁰ and g-C₃N₄,^11,12 have greatly attracted researchers’ attention for their catalytic properties. Nevertheless, the efficiency of semiconductor photocatalysts remains low because of weak light harvesting properties, which are mostly in the ultraviolet region, and also the fast electron–hole recombination rate.¹³ To solve these problems, the construction of a heterojunction by the combination of two semiconductors is a promising way. Heterojunctions can be divided into conventional heterojunctions (Type-I,¹⁴ Type-II,¹⁵ and Type-III¹⁶), p–n heterojunctions¹⁷ and Z-scheme heterojunctions,¹⁸ as shown in Fig. 2.^19,20 However, the synthesis of these heterojunctions usually needs complicated processes as well as fine-tuning of the compositions and morphologies of these nanomaterials to optimize the catalytic properties, which is highly challenging.²¹ Under these circumstances, the development of new synthetic methods for high-performance heterojunction-based photocatalysts for water splitting is desired.

Fig. 2 Diagrams of Type-I (a); Type-II (b); Type-III (c), p–n heterojunction (d), Z-scheme (e) heterojunction. 1, 2, p, and n refer to semiconductor 1, semiconductor 2, p-semiconductor, and n-semiconductor, respectively.

Metal–organic frameworks (MOFs), as a set of crystalline porous materials fabricated by metal nodes and organic linkers, have shown great potentials in numerous fields such as gas sorption and separation, catalysis, and sensors due to their varied and customizable structures, size/shape adjustable nanopores or nanochannels and high surface areas.^22,23 As water splitting catalysts, MOFs are used as catalysts directly²⁴ or as precursors for nanocatalysts.²⁵ For example, in 2009, Mori et al. reported the first example of MOF, namely, [Ru₂(1,4-BDC)₂]_n (1,4-BDC = 1,4-benzene-dicarboxylic acid) as a photocatalyst for HER under the irradiation of visible light.²⁶ By the calcination of the nanoscale MIL-101 octahedral particles coated with a titanium dioxide shell, a core–shell nanostructure of Fe₂O₃@TiO₂ was obtained for HER by Lin et al.²⁷ A timeline depicting the milestones of MOF-derived heterojunctions as the nanocatalysts for photocatalytic hydrogen production is illustrated in Fig. 3.^27–36 Because of the unique structures and morphologies of the nanocatalysts derived from MOFs, the catalytic performances of these nanomaterials have been significantly improved.³⁷ In the literature, reviews of the nanomaterials derived from MOFs have been summarized.^38–40 In 2018, Xu et al. summarized MOF-based nanocatalysts for hydrogen production from water and for batteries.^41,42 Tang et al. summarized MOF-based heterogeneous catalysts for CO₂ conversion.⁴³ Wu et al. summarized MOFs and their derived materials as electro- or photo-catalysts for CO₂ reduction.⁴⁴ In this review, we summarized the developing strategy of using the pore chemistry and surface chemistry of MOFs to make MOF-based composites and then via the pyrolysis of the composites to obtain heterojunctions as nanocatalysts for photocatalytic hydrogen production from water splitting. The advantages of this strategy and the applications of these photocatalysts were discussed.

Fig. 3 Timeline of the milestones of MOF-derived heterojunctions as nanocatalysts for photocatalytic hydrogen production.

2. Mechanism of heterojunctions for photocatalytic hydrogen production

A heterojunction generally refers to the interfacial region formed by two different semiconductors, satisfying lattice matching and thermal matching; it is considered to be one of the most applicable photocatalysts for water splitting.¹⁹ Because of the differences in band structures and material properties, when a heterojunction is formed, an interfacial potential difference will also be created, leading to the generation of a built-in electric field. Under the influence of this electric field, the photo-generated electrons and holes will migrate rapidly, promoting the separation of electron–hole pairs and effectively suppressing recombination.

The energy-band structure of a Type-I heterojunction is defined by the VB and CB of semiconductor 2 located in the forbidden band range of semiconductor 1. The CB potential of semiconductor 1 is more negative than that of semiconductor 2; thus, the photoelectrons on semiconductor 1 can easily pass on to the CB of semiconductor 2. Meanwhile, the holes will also migrate from the VB of semiconductor 1 to the VB of semiconductor 2 (Fig. 2a).

In a Type-II heterojunction, both the CB and VB potentials of semi-conductor 1 are more negative than that of semiconductor 2. This band structure facilitates the transport of electrons from the CB of semiconductor 1 to the CB of semiconductor 2, and the holes from the VB of semiconductor 2 migrate to the VB of semi-conductor 1 simultaneously (Fig. 2b).

The CB and VB of a type-III heterojunction have similarities to those of a type-II heterojunction except the extremely staggered gap, leading to the mismatch of the band-gaps. Hence, the electron–hole separation and migration between the two semiconductors cannot take place, making it unfavourable for the separation of the electron–hole pairs for photocatalytic applications (Fig. 2c).

For a p–n heterojunction, due to the different concentrations of the electrons and holes in the two semiconductors, the reversible diffusion of the electrons and holes occurs at the interface of the p–n heterojunction; this results in the formation of a spatial charge region that creates a built-in electric field from the n-region to the p-region, guiding the electrons to quickly move from the p-type semiconductor to the n-type semiconductor. The holes also rapidly migrate backwards, effectively implementing the spatial separation of the electron–hole pairs (Fig. 2d).

The Z-scheme heterojunction was first proposed in 1979.⁴⁵ Its unique energy structure and charge transport mode allow electrons and holes to be spatially segregated, and the recombination rate of the electrons and holes is low in the photocatalytic process. As shown for the band structure in Fig. 2e, the electrons on the CB of semiconductor 2 can rapidly migrate to the VB of semiconductor 1 and recombine with holes, thereby ensuring that the holes on the VB of semiconductor 2 and the electrons on the CB of semiconductor 1 do not recombine easily, hence increasing the catalytic efficiency significantly.

The construction of a heterojunction by the combination of two semiconductors can notably broaden the light-harvesting range, enhance the spatial separation and moreover suppress the recombination of electron–hole pairs; thus, attention should be paid to each type of a heterojunction. However, the specific band structure and obviously different charge migration modes have made the Z-scheme photocatalyst the most promising system.⁴⁶ In addition, the p–n heterojunction requires p- and n-type semiconductors, which has distinct limitations in semiconductor selection.

3. MOFs for MOF-derived nanocatalysts

In the past few decades, MOFs have been developed very rapidly in most of the fields in chemistry and materials science due to their powerful pore chemistry. Nevertheless, MOFs are less documented in the study of photocatalysis on account of the relatively poor stability and low electronic conductivity.⁴⁷ However, recent studies have shown that some MOFs can be excellent precursors for the synthesis of heterojunction-based nanocatalysts, exhibiting a high photocatalytic HER performance due to the following features: (i) the compositions and the ordered framework structures of MOFs can not only be the sources of metal compounds of the heterojunctions but also the in situ generated allotropes of carbon to enhance the electrical conductivity; (ii) the porosity of MOFs permits small species to be quantificationally accommodated in the pores or channels of MOF-based composites with well-defined compositions as precursors; (iii) the surfaces of the nanoparticles of MOFs can be fine-tuned by structural modifications to be linked with specific species to form MOF-based composites with well-defined compositions as precursors (Fig. 4).²²

Fig. 4 Selected examples of MOFs for MOF-derived nanocatalysts using the pore chemistry and surface chemistry of MOFs.

To use these features of MOFs for the heterojunction-based nanocatalysts, the synthesis methods of the selected MOFs should be facile and performed on a large scale. Organic bridging ligands are usually the costliest part of the total cost of MOFs; well-used and relatively cheap ligands such as 1,4-benzenedicarboxylate (H₂BDC), 1,3,5-benzenetricarboxylic acid and 2-methylimidazole (2-mIM) were selected for discussing in detail in the following section.^32,35,48

4. MOF-derived nanocatalysts for photocatalytic hydrogen production

Pyrolysis of the well-designed composites is a useful way for the synthesis of nanocatalysts.⁴⁹ For the heterojunction-based nanocatalysts, due to the relatively complex compositions, it is highly challenging to combine different compositions uniformly and control the morphology of the products by calcination. Based on the well-defined framework components, long-range ordered structures, and varied pore and surface chemistry of MOFs, the chemical compositions of the MOF-based precursors can be well adjusted at the molecular level, which is beneficial for the chemical fine-tuning of the compositions and morphology of the nanocatalysts.

4.1 Using pore chemistry of MOFs for MOF-derived nanocatalysts

The pores or channels of MOFs can accommodate specific small species and hence have been well studied for gas storage and separation, catalysis, and so on.⁵⁰ For chemistry, the nano-sized pores or channels of MOFs can also be used to combine with specific species, such as thioacetamide,^32,33 phosphorus⁵¹ and H₃PMo₁₂O₄₀,⁵² to produce well-defined composites followed by pyrolysis to generate nanomaterials. Very recently, this strategy was successfully applied to synthesize nanocatalysts for photocatalytic HER (Table 1).

Table 1 A summary of using pore chemistry of MOFs for selected MOF-derived nanocatalysts

Photocatalyst	Template	Light	Sacrificial agent	Cocatalyst/sensitizer	Catalyst dose (mg)	Activity (μmol g^−1 h^−1)	Stability/AQE	Ref.
CdS@Cd(II)-MOF@TiO2 (7)	Cd-MOF	Visible light (λ > 400 nm)	0.025 M Na2SO3/0.1 M Na2S	Pt	20	1416	2.05% at 420 nm	54
CdS@Cd(II)-MOF@TiO2 (8)						1604	2.31% at 420 nm
Pt-ZnO-Co3O4	ZnCo-ZIF	UV-Vis light	10% methanol	Pt	100	7800	5 times recycled	30
Pt-ZnS-CoS						8210	30 h
Pt-Zn3P2-CoP						9150
Co-Doped Zn0.5Cd0.5S	ZnCo-ZIF	Visible light (λ > 420 nm)	0.35 M Na2S/0.25 M Na2SO3	—	5	17 360	6 times recycled	55
							30 h
							39.2% at 420 nm
30 wt% CdS/ZnxCo3−xO4	ZnCo-ZIF	Visible light (λ > 420 nm)	5 mL lactic acid	—	10	3978.6	4 times recycled	56
							16 h
NiS/Zn0.5Cd0.5S	Ni-ZnxCd1−x-MOF	Visible light (λ > 420 nm)	35 mM Na2S/25 mM Na2SO3	—	50	16 780	5 times recycled	32
							20 h
Zn0.5Cd0.5S/ZnO/Zn0.5Cd0.5S cages	ZIF-8	300 W Xe lamp	0.75 M Na2S/1.05 M Na2SO3	—	20	28 600	6 times recycled	34
							30 h
ZnO/ZnS-30	MOF-5	Visible light (λ > 420 nm)	10 mM Na2S/Na2SO3	—	50	435	8 times recycled	33
							40 h
NiS/CdS/h-TiO2	NH2-MIL-125	Visible light (λ > 420 nm)	0.35 M Na2S/0.25 M Na2SO3	—	50	2149.15	4 times recycled	53
							16 h

---

Using newly developed bimetallic ZnCo-ZIF as template, Pt-ZnO-Co₃O₄, Pt-ZnS-CoS, and Pt-Zn₃P₂-CoP photocatalysts were prepared through the oxidation, sulfurization, and phosphorization of ZnCo-ZIF with Pt doping.³⁰ More importantly, suitable band matching and strong electron coupling were achieved in these heterojunctions, greatly facilitating effective electron–hole separation and transmission. Moreover, the Pt nanoparticles dispersed on porous heterojunctions as electron traps can provide abundant redox active sites for HER. Correspondingly, the rates of HER for three photocatalysts were found to be 7800, 8210, and 9150 μmol g⁻¹ h⁻¹, which were higher than those of the respective monometallic ZIF-8 or ZIF-67-based derivatives.

In 2018, a non-noble metal-based cocatalyst/solid solution heterojunction of NiS/Zn_xCd_1−xS with superior photocatalytic HER activity was reported by us (Fig. 5a).³² The chemical composition and band-gap of Zn_xCd_1−xS were well-tuned, and the light capturing capacity and photocatalytic activity were also optimized by regulating the doped Zn²⁺ in the precursor of Cd-MOF (Fig. 5b–e). As a cocatalyst, NiS played a crucial role in further improving the catalytic activity of the heterojunction. NiS/Zn_0.5Cd_0.5S exhibited an optimum HER rate (16 780 μmol g⁻¹ h⁻¹) and retained excellent stability and recyclability under visible-light irradiation (Fig. 5f–i) because of its high light absorption, suitable band gap as well as the position of CB and structural stability.

Fig. 5 (a) The synthetic route of NiS/Zn_xCd_1−xS. (b) PXRD of NiS/Zn_xCd_1−xS. (c) TEM and particle size distribution (inset), (d) elemental mapping and (e) HRTEM of NiS/Zn_0.5Cd_0.5S. (f) EIS and (g) time-dependent photocurrent of NiS/Zn_xCd_1−xS. (h) Photocatalytic HER rates of NiS/Zn_xCd_1−xS under visible light. (i) HER cycling test for NiS/Zn_0.5Cd_0.5S under visible light. Reprinted with permission from ref. 32.

A set of ZnO/ZnS nanostructured heterojunctions was fabricated for photocatalytic HER under visible light by coupling solvothermal synthesis with MOF-5 as the host and thioacetamide (TAA) as the sulfur source via two-step calcination in different atmospheres.³³ In the heterojunctions, ZnO and ZnS were evenly integrated due to the homogeneous sulfidation of the MOF-5 precursor at the molecular level and then the oxidization of ZnS@C (Fig. 6a–g). This facile synthesis method increased the interfacial catalytic active sites and the light absorption capability as well as charge separation efficiency by optimizing the calcination time of ZnS@C in air, thus obtaining optimized ZnOS-30, which possessed excellent photocatalytic HER activity (435 μmol g⁻¹ h⁻¹) and recyclability without any co-catalyst (Fig. 6h–k).

Fig. 6 (a) The synthetic route of ZnO/ZnS heterostructures. (b) TEM and (c) HRTEM images of ZnS@C. TEM images of (d) ZnOS-15; (e) ZnOS-30; (f) ZnOS-45 and (g) ZnOS-60. (h) Time-dependent photocatalytic HER under visible-light irradiation of ZnOS-n. (i) Comparison of photocatalytic HER of ZnS/ZnO mixture, ZnS and ZnOS-n under visible-light irradiation. (j) Cycling test of HER for ZnOS-30 under visible-light irradiation. (k) Cycling test of HER for recycled and reprocessed ZnOS-30 after 40 h cycling test. Reprinted with permission from ref. 33.

Sandwich-shelled ZnCdS/ZnO/ZnCdS cages were synthesized based on ZIF-8.³⁴ ZnS and ZnO were produced concurrently during sulfidation by TAA, in which ZnS acted as an obstruction to centralize the ZnO particles filling up the spaces between the adjacent ZnS layers. For photocatalysis, ZnCdS resulting from the cation-exchanged ZnS and ZnO can be more advantageous for the formation of staggered band structures by fine-tuning the composition. Benefiting from the enhanced light utilization and high surface area, the Zn_0.5Cd_0.5S/ZnO/Zn_0.5Cd_0.5S cage showed an HER rate of 28 600 μmol g⁻¹ h⁻¹ and long-time steadiness, resulting in supreme activities among the ZnCdS and ZnO families.

In 2019, Zhou et al. reported non-noble metal-containing hollow CdS/TiO₂ (CdS/h-TiO₂) decorated with an NiS cocatalyst nanohybrid photocatalyst obtained by the use of NH₂-MIL-125 as the MOF template.⁵³ It presented no prominent impact on the morphology of the composites after loading few NiS species. The final material not only maintained the original framework of pristine MOF, but also formed hollow morphology. Owing to the porous structure, the as-prepared composite exhibited enhanced photocatalytic HER efficiency under visible-light irradiation in contrast to TiO₂, CdS/TiO₂, or even CdS/TiO₂ with Pt as the cocatalyast. By regulating the weight ratio of CdS and NiS, the obtained NiS/CdS/h-TiO₂ photocatalyst with optimized amounts of 30% CdS and 0.3% NiS showed the highest photocatalytic HER activity of 2149.15 μmol g⁻¹ h⁻¹ due to the synergistic effect between the heterojunctions for quick charge separation and accelerated surface redox reaction by the NiS cocatalyst.

Using the pores or channels of MOFs to accommodate specific small species for photocatalytic water splitting is developing rapidly in the last few years because of the size and structure of the pores or channels and compositional diversities of MOFs. Further efforts should be put for developing high-efficiency and well-defined heterojunction-based photocatalysts that cannot be easily constructed from a direct reaction of the two semiconductor components from the well-matching of the pores or channels of the MOFs and guest species.

4.2. Using surface chemistry of MOFs for MOF-derived nanocatalysts

Besides, for the pore chemistry of MOFs, the surface of the nanoparticles of MOFs can also have affinity for specific species to produce MOF-based nanocomposites. The pyrolysis of the well-defined nanocomposites can generate heterojunction-based nanocatalysts for photocatalytic HER (Table 2).

Table 2 A summary of using surface chemistry of MOFs for selected MOF-derived nanocatalysts

Photocatalyst	Template	Light	Sacrificial agent	Cocatalyst/sensitizer	Catalyst dose (mg)	Activity (μmol g^−1 h^−1)	Stability/AQE	Ref.
Fe2O3@TiO2/Pt	MIL-101(Fe)	Visible light (λ > 420 nm)	20/1 v/v H2O/TEOA	Pt	0.5	625	3 times recycled	27
							48 h
Hollow α-Fe2O3-TiO2-PtOx	MIL-88B	Visible light (λ > 420 nm)	Lactic acid	—	20	1100	5 times recycled	58
							25 h
NC-PA 2 wt%Co3O4/TiO2	Co-MOF	400 W xenon arc lamp	15 vol% CH3OH	—	20	6400	2.4% at 420 nm	28
CdS/2 wt% MCTMPs	ZIF-67	Simulated solar light	3 mL lactic acid	—	1	136 770	5 times recycled	57
							60 h
							40.6% at 425 nm
14.6 wt% MoS2@TiO2	NH2-MIL-125(Ti)	Visible light (λ ≥ 420 nm)	TEOA	Fluorescein	5	10 046	3 times recycled	29
							30 h
1 wt% Cu/TiO2-400	MOF-199	UV-light	5 vol% glycerol	—	2.5	15 130	4 times recycled	31
							16 h
20 wt% CdS/Co-C@Co9S8	ZIF-67	Simulated solar light	3 mL lactic acid	—	1	26 690	4 times recycled	59
							20 h
							7.82% at 425 nm
CdS/MOF-5 (25)	MOF-5	Visible light (λ ≥ 420 nm)	0.35 M Na2S/0.25 M Na2SO3	—	50	11 620	4 times recycled	18
							16 h
							11.09% at 470 nm
g-C3N4/MOF/MoS2 (20%)	ZIF-67	5 W LED white light	15% TEOA	10 mg Eosin Y	20	4012.5	3 times recycled	60
							12 h
							5.04% at 425 nm
Ni-MOF-74/CdS/Co3O4 (10 wt% Co)	Ni-MOF-74	M300 xenon lamp	10% lactic acid	—	20	5810	4 times recycled	61
							20 h
							14.7% at 475 nm
Core–shell CdS/MoS2-0.6	Cd-MOF (Cd-ZIF-8)	300 W Xe lamp	35 mM Na2S/18 mM Na2SO3	—	30	5587	2 times recycled	62
Mixed MoS2-CdS						91	—
CdS/MoS2-DP hybrid						2102
g-C3N4/α-Fe2O3-2	MIL-100(Fe)	Visible light (λ > 420 nm)	10 vol% TEOA	3 wt% Pt	10	2066.2	5 times recycled	35
							20 h
							7.43% at 420 nm
Ni2P/Ni@C/g-C3N4-550	Ni-MOF	Visible light (λ ≥ 420 nm)	10% TEOA	1.0 mM Eosin Y	10	18 040	5 times recycled	36
							25 h
							58.1% at 420 nm
				—		210	—
TiO2-Ti3C2-CoSx	TiO2-ZIF-67	UV-Vis light	10 mL methanol	—	20	950	5 times recycled	63
							15 h

---

For photocatalytic HER, an ideal system of Z-scheme heterojunctions can separate the photo-induced charges for exceedingly stronger catalytic ability. Wang et al. adopted a technique to synthesize Z-scheme CdS/ZnS heterojunctions with Zn vacancies derived from CdS/MOF-5, with a diverse loading weight of CdS by a one-step hydrothermal method, in which MOF-5 could be directly converted into ZnS with a large number of zinc vacancy defects in the aqueous solution with Na₂S and Na₂SO₃ as the sacrificial agents at an ambient temperature.¹⁸ Its two-photon absorption behavior could be realized by the zinc vacancies of ZnS under visible-light irradiation. The interfacial charge transfer of CdS/ZnS transformed from a traditional type-I heterojunction to Z-scheme because of the ohmic contact caused by zinc vacancy defects. The maximum photocatalytic HER rate of 11 620 μmol g⁻¹ h⁻¹ was obtained with AQE of 11.09% at 470 nm of the CdS/MOF-5(25) heterojunction under visible-light irradiation.

A facile, tunable and inexpensive MOF-templated strategy for synthesizing Fe₂O₃@TiO₂ was reported by the use of MIL-101(Fe) as a template (Fig. 7a), which showed a higher H₂ production yield than single Fe₂O₃ and TiO₂ materials.²⁷ For synthesis, the amorphous shell of titania was coated on the outside of the MIL-101 nanoparticles by the acid-catalyzed hydrolysis and condensation of a titanium(IV) bis-(ammonium lactato)-dihydroxide water solution, in which the concentration of the acid and the reaction time determined the thickness of the TiO₂ coating. HRTEM showed that each octahedral shell consisted of roughly spherical single crystalline domains of 10 nm TiO₂ and Fe₂O₃ blended together. Lattice fringes were visible with the lattice d-spacings of 0.354 nm of the {101} plane of anatase and 0.251 nm of the {110} plane of hematite (Fig. 7b–e). The optimized HER rate was 625 μmol g⁻¹ h⁻¹ with Pt as the cocatalyst (Fig. 7f).

Fig. 7 (a) The synthetic route of Fe₂O₃@TiO₂ composite with Pt particles. HRTEM images of Fe₂O₃@TiO₂: as-synthesized (b–d) and after Pt particles deposited (e). (f) H₂ produced by Fe₂O₃@TiO₂ with a 420 nm filter. Reprinted with permission from ref. 27.

A Co₃O₄/TiO₂ p–n heterojunction was prepared using Co-MOFs as sacrificial template by Bala et al.²⁸ They demonstrated that their strategy of using MOFs can construct high-performance p–n Co₃O₄/TiO₂ heterojunctions for photocatalytic HER, in which the nanocomposite containing 2 wt% Co loading showed supreme activity (6400 μmol g⁻¹ h⁻¹) under UV-vis light irradiation. The formation of the pony-size heterojunction and the cocatalytic character of Co₃O₄ should be ascribed to fast interfacial charge transfer and effective electron–hole separation in the photocatalytic reaction.

Another Z-scheme heterojunction of g-C₃N₄/α-Fe₂O₃ was prepared by using MIL-100 (Fe) as a self-sacrificing template. By changing the loading amount of MIL-100 (Fe), nanocomposites with diverse α-Fe₂O₃ and g-C₃N₄ ratios were obtained (Fig. 8a).³⁵ The proportion of α-Fe₂O₃ was also regulated by changing the calcination temperature of g-C₃N₄/MIL-100-2 in air atmosphere. As shown in Fig. 8b, granular α-Fe₂O₃ nanoparticles tightly adhere to the flakes of g-C₃N₄. In addition, the obvious lattice spacing of 0.367 nm can be attributed to the {012} plane of α-Fe₂O₃ (Fig. 8c and d). The elemental mapping images of g-C₃N₄/α-Fe₂O₃-2 show that all the elements are homogeneously dispersed throughout the composite, and the α-Fe₂O₃ nanomaterials evenly adhere to g-C₃N₄ (Fig. 8e and f). Owing to the well separated photo-excited charge carriers, the HER performance of the optimized g-C₃N₄/α-Fe₂O₃-2 photocatalyst was 2066.2 μmol g⁻¹ h⁻¹ (Fig. 8g–j).

Fig. 8 (a) The synthesis route of the g-C₃N₄/α-Fe₂O₃ hybrid. (b) SEM, (c) TEM, (d) HRTEM, and (e) elemental mapping images before reaction and (f) after 5 cycling tests of g-C₃N₄/α-Fe₂O₃-2. (g) Photocatalytic HER of g-C₃N₄ and g-C₃N₄/α-Fe₂O₃ hybrids. (h) Photocatalytic HER of pure g-C₃N₄ and g-C₃N₄/α-Fe₂O₃ hybrids. (i) Photocatalytic performance of g-C₃N₄/α-Fe₂O₃-2-T hybrids. (j) HER cycling test for g-C₃N₄/α-Fe₂O₃-2. Reprinted with permission from ref. 35.

Ni₂P/Ni nanoparticles (NPs) were wrapped in carbon/g-C₃N₄ hybrids, originated from the in situ pyrolysis and phosphatization of the Ni-MOF/g-C₃N₄ precursor, which served as the photocatalyst for HER under visible-light irradiation.³⁶ By adjusting the pyrolysis temperature under an N₂ atmosphere of the Ni-MOF/g-C₃N₄ powder, a set of Ni@C/g-C₃N₄ composites was obtained. The highest HER rate of Ni₂P/Ni@C/g-C₃N₄-550 was 18 040 μmol g⁻¹ h⁻¹ with 1.0 mM of EY-sensitization, while 210 μmol g⁻¹ h⁻¹ was obtained without EY-sensitization, and AQE at 420 nm was 58.1%. This improved photocatalytic activity can be assigned to the effective and quick transfer of the photo-generated charges from excited EY and g-C₃N₄ to Ni₂P/Ni, where carbon as an electron transfer bridge, each component in close contact, interlaced band alignment among g-C₃N₄, Ni and Ni₂P, and the acceleration of the proton reduction reaction by Ni₂P/Ni NPs all play synergistic roles.

Kim's group fabricated polyporous multicomponent transition metal phosphides (MCTMPs) from layered double hydroxide double-shelled nanocages through MOF templates.⁵⁷ CdS semiconductor nanorods combined with these unique MCTMP nanostructures could exhibit exceptionally high stability and prominent HER rate up to 136 770 μmol g⁻¹ h⁻¹, with quantum efficiency of 40.6% for the sample with 2 wt% MCTMP loading, which can be ascribed to the large surface area, superior light absorption, and outstanding separation ability of the electron–hole pairs.

In light of the structural multiformity and tailorability of MOFs, the controllable surface modification of MOFs may inaugurate a new pathway to the synthesis of advanced MOF-based nanocatalysts for enhancing photocatalytic H₂ production. This strategy will give more opportunities for the construction of efficient photocatalysts to inspire the research on new photocatalysts and to obtain a more efficient visible-light response for sunlight-driven hydrogen evolution by water splitting.

5. Summary and outlook

The recent progress of MOF-derived heterojunction-based nanocatalysts using their pore chemistry or surface chemistry for photocatalytic water splitting was summarized. These MOF-derived nanocatalysts have the following advantages: (i) well-defined composition from the parent MOFs or MOF-based composites; (ii) high porosity with ordered crystallized pores and (iii) effective integration of multifarious functional materials.

However, there are still problems that need particular attention, such as how to realize the controllable and facile synthesis of MOF-derived nanocomposites with desired pore structures and active sites and how to achieve the efficient use of visible even near-infrared light by the combination of two or more components to improve the utilization of light energy.

In addition, in photocatalytic water splitting, HER is a half-reaction and takes advantage of sacrificial reagents to accomplish sacrificing holes as the oxidation half-reaction; thus, the following two valuable applications are suggested to be studied by taking the synthetic advantages of the above-mentioned MOF-derived nanocatalysts: (i) Splitting water by sunlight without sacrificial reagents. For example, Wang's team reported that the RGO nanosheets can serve as a solid-state electron mediator to link Fe₂O₃ nanoparticles with PCN nanosheets by chemical bonding and π–π stacking, forming an Fe₂O₃/RGO/PCN ternary heterojunction and thus enhancing the photocatalytic efficiency.⁶⁴ (ii) Replacing the sacrificial agent consumption with organic compound oxidation to realize the coupling of H₂ production with useful organic reactions. As the first example of MOF-based photocatalysts to simultaneously perform HER and organic conversion, Jiang's group used 2.3 wt% Pt/PCN-777 as the photocatalyst for HER and selected benzylamine instead of TEOA as the sacrificial agent to produce N-benzylbenzaldimine.⁶⁵

Overall, the MOF-templated strategy with the intrinsic features of MOFs has become a powerful method for the synthesis of nanocomposites and their related nanocatalysts, which can convert solar energy into hydrogen energy effectively and become the most potential alternatives to fossil fuels.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (21622105, 21861130354 and 21421001), the Natural Science Foundation of Tianjin (18JCJQJC47200) and the Ministry of Education of China (B12015). W. S. acknowledges the receipt of a Newton Advanced Fellowship from Royal Society.

References

1. M. R. Hoffmann, S. T. Martin, W. Y. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69–96.
2. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520–7535.
3. W. Wang, X. M. Xu, W. Zhou and Z. P. Shao, Adv. Sci., 2017, 4, 1600371.
4. H. Kisch, Angew. Chem., Int. Ed., 2013, 52, 812–847.
5. K. Maeda and K. Domen, J. Phys. Chem. C, 2007, 111, 7851–7861.
6. J. Schneider, M. Matsuoka, M. Takeuchi, J. L. Zhang, Y. Horiuchi, M. Anpo and D. W. Bahnemann, Chem. Rev., 2014, 114, 9919–9986.
7. H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri and J. Ye, Adv. Mater., 2012, 24, 229–251.
8. M. Z. Ge, Q. S. Li, C. Y. Cao, J. Y. Huang, S. H. Li, S. N. Zhang, Z. Chen, K. Q. Zhang, S. S. Al-Deyab and Y. K. Lai, Adv. Sci., 2017, 4, 1600152.
9. C. B. Ong, L. Y. Ng and A. W. Mohammad, Renewable Sustainable Energy Rev., 2018, 81, 536–551.
10. Y. Xu, W. W. Zhao, R. Xu, Y. M. Shi and B. Zhang, Chem. Commun., 2013, 49, 9803–9805.
11. D. J. Martin, K. P. Qiu, S. A. Shevlin, A. D. Handoko, X. W. Chen, Z. X. Guo and J. W. Tang, Angew. Chem., Int. Ed., 2014, 53, 9240–9245.
12. L. Jiang, X. Yuan, Y. Pan, J. Liang, G. Zeng, Z. Wu and H. Wang, Appl. Catal., B, 2017, 217, 388–406.
13. A. Khurshida, M. Md, B. Nurlan, K. Sarkyt and N. Nurxat, J. Mater. Chem. A, 2018, 6, 21696–21718.
14. Y. P. Xie, Z. B. Yu, G. Liu, X. L. Ma and H. M. Cheng, Energy Environ. Sci., 2014, 7, 1895–1901.
15. D. P. Kumar, J. H. Choi, S. Hong, D. A. Reddy, S. Lee and T. K. Kim, ACS Sustainable Chem. Eng., 2016, 4, 7158–7166.
16. J. K. Hyun, S. Zhang and L. J. Lauhon, Annu. Rev. Mater. Res., 2013, 43, 451–479.
17. Y. Z. Chen, C. Y. Zhang, X. J. Zhang, X. M. Ou and X. H. Zhang, Chem. Commun., 2013, 49, 9200–9202.
18. X. Q. Hao, Z. W. Cui, J. Zhou, Y. C. Wang, Y. Hu, Y. Wang and Z. G. Zou, Nano Energy, 2018, 52, 105–116.
19. J. X. Low, J. G. Yu, M. Jaroniec, S. Wageh and A. A. Al-Ghamdi, Adv. Mater., 2017, 29, 1601694.
20. W. W. Shi and N. Chopra, Nanomater. Energy, 2013, 2, 158–178.
21. H. L. Wang, L. S. Zhang, Z. G. Chen, J. Q. Hu, S. J. Li, Z. H. Wang, J. S. Liu and X. C. Wang, Chem. Soc. Rev., 2014, 43, 5234–5244.
22. L. Jiao, Y. Wang, H. L. Jiang and Q. Xu, Adv. Mater., 2018, 30, 1703663.
23. Z. B. Liang, R. Zhao, T. J. Qiu, R. Q. Zou and Q. Xu, Energy Chem., 2019, 1, 100001.
24. Y. J. Xiao, Y. Qi, X. L. Wang, X. Y. Wang, F. X. Zhang and C. Li, Adv. Mater., 2018, 30, 1803401.
25. J. K. Sun and Q. Xu, Energy Environ. Sci., 2014, 7, 2071–2100.
26. Y. Kataoka, K. Sato, Y. Miyazaki, K. Masuda, H. Tanaka, S. Natio and W. Mori, Energy Environ. Sci., 2009, 2, 397–400.
27. K. E. deKrafft, C. Wang and W. B. Lin, Adv. Mater., 2012, 24, 2014–2018.
28. S. Bala, I. Mondal, A. Goswami, U. Pal and R. Mondal, J. Mater. Chem. A, 2015, 3, 20288–20296.
29. B. Ma, P. Y. Guan, Q. Y. Li, M. Zhang and S. Q. Zang, ACS Appl. Mater. Interfaces, 2016, 8, 26794–26800.
30. M. Lan, R. M. Guo, Y. B. Dou, J. Zhou, A. W. Zhou and J. R. Li, Nano Energy, 2017, 33, 238–246.
31. I. Majeed, M. A. Nadeem, A. Badshah, F. K. Kanodarwala, H. Ali, M. A. Khan, J. A. Stride and M. A. Nadeem, Catal. Sci. Technol., 2017, 7, 677–686.
32. X. X. Zhao, J. R. Feng, J. Liu, W. Shi, G. M. Yang, G. C. Wang and P. Cheng, Angew. Chem., Int. Ed., 2018, 57, 9790–9794.
33. X. X. Zhao, J. R. Feng, J. W. Liu, J. Lu, W. Shi, G. M. Yang, G. C. Wang, P. Y. Feng and P. Cheng, Adv. Sci., 2018, 5, 1700590.
34. J. M. Chen, Z. R. Shen, S. M. Lv, K. Shen, R. F. Wu, X. F. Jiang, T. Fan, J. Y. Chen and Y. W. Li, J. Mater. Chem. A, 2018, 6, 19631–19642.
35. J. Liu, X. X. Zhao, P. Jing, W. Shi and P. Cheng, Chem. – Eur. J., 2019, 25, 2330–2336.
36. J. X. Xu, Y. H. Qi and L. Wang, Appl. Catal., B, 2019, 246, 72–81.
37. S. Dang, Q. L. Zhu and Q. Xu, Nat. Rev. Mater., 2018, 3, 17075.
38. J. N. Hao, X. Y. Xu, H. H. Fei, L. C. Li and B. Yan, Adv. Mater., 2018, 30, 1705634.
39. K. Meyer, M. Ranocchiari and J. A. van Bokhoven, Energy Environ. Sci., 2015, 8, 1923–1937.
40. Y. Z. Chen, R. Zhang, L. Jiao and H. L. Jiang, Coord. Chem. Rev., 2018, 362, 1–23.
41. B. J. Zhu, R. Q. Zou and Q. Xu, Adv. Energy Mater., 2018, 8, 1801193.
42. R. Zhao, Z. B. Liang, R. Q. Zou and Q. Xu, Joule, 2018, 2, 1–25.
43. Z. D. Lei, Y. C. Xue, W. Q. Chen, W. H. Qiu, Y. Zhang, S. Horike and L. Tang, Adv. Energy Mater., 2018, 8, 1801587.
44. H. G. Zhang, J. Z. Li, Q. Tan, L. L. Lu, Z. B. Wang and G. Wu, Chem. – Eur. J., 2018, 24, 18137–18157.
45. A. J. Bard, J. Photochem., 1979, 10, 59–75.
46. Q. L. Xu, L. Y. Zhang, J. G. Yu, S. Wageh, A. A. Al-Ghamdi and M. Jaroniec, Mater. Today, 2018, 21, 1042–1063.
47. W. W. Zhan, L. M. Sun and X. G. Han, Nano-Micro Lett., 2019, 11, 1.
48. Q. Mao, J. M. Chen, H. R. Chen, Z. J. Chen, J. Y. Chen and Y. W. Li, J. Mater. Chem. A, 2019, 7, 8472–8484.
49. J. D. Xiao and H. L. Jiang, Small, 2017, 13, 1700632.
50. M. Alhamami, H. Doan and C. H. Cheng, Materials, 2014, 7, 3198–3250.
51. G. H. Li, H. Yang, F. C. Li, J. Du, W. Shi and P. Cheng, J. Mater. Chem. A, 2016, 4, 9593–9599.
52. J. S. Li, Y. J. Tang, C. H. Liu, S. L. Li, R. H. Li, L. Z. Dong, Z. H. Dai, J. C. Bao and Y. Q. Lan, J. Mater. Chem. A, 2016, 4, 1202–1207.
53. N. X. Lia, H. L. Huang, R. Bibi, Q. H. Shen, R. Ngulube, J. C. Zhou and M. C. Liu, Appl. Surf. Sci., 2019, 476, 378–386.
54. C. W. Zhao, Y. A. Li, X. R. Wang, G. J. Chen, Q. K. Liu, J. P. Ma and Y. B. Dong, Chem. Commun., 2015, 51, 15906–15909.
55. X. Tang, J. H. Zhao, Y. H. Li, Z. J. Zhou, K. Li, F. T. Liu and Y. Q. Lan, Dalton Trans., 2017, 46, 10553–10557.
56. W. X. Chen, J. S. Fang, Y. W. Zhang, G. L. Chen, S. Zhao, C. Zhang, R. Xu, J. H. Bao, Y. M. Zhou and X. Xiang, Nanoscale, 2018, 10, 4463–4474.
57. D. A. Reddy, H. K. Kim, Y. Kim, S. Lee, J. H. Choi, M. J. Islam, D. P. Kumar and T. K. Kim, J. Mater. Chem. A, 2016, 4, 13890–13898.
58. M. H. Pham, C. T. Dinh, G. T. Vuong, N. D. Ta and T. O. Do, Phys. Chem. Chem. Phys., 2014, 16, 5937–5941.
59. D. A. Reddy, H. Park, M. Gopannagari, E. H. Kim, S. Lee, D. P. Kumar and T. K. Kim, ChemSusChem, 2018, 11, 245–253.
60. Z. J. Wang, Z. L. Jin, G. R. Wang and B. Z. Ma, Int. J. Hydrogen Energy, 2018, 43, 13039–13050.
61. Y. K. Zhang, G. R. Wang, W. Ma, B. Z. Ma and Z. L. Jin, Dalton Trans., 2018, 47, 11176–11189.
62. L. Lin, S. Y. Huang, Y. X. Zhu, B. Du, Z. H. Zhang, C. Chen, X. W. Wang and N. Zhang, Dalton Trans., 2019, 48, 2715–2721.
63. J. H. Zhao, L. W. Liu, K. Li, T. Li and F. T. Liu, CrystEngComm, 2019, 21, 2416–2421.
64. Z. M. Pan, G. G. Zhang and X. C. Wang, Angew. Chem., Int. Ed., 2019, 58, 1–6.
65. H. Liu, C. Y. Xu, D. D. Li and H. L. Jiang, Angew. Chem., Int. Ed., 2018, 57, 5379–5383.

---