Introduction
Carbon dioxide (CO2) released in the combustion of fossil fuels, such as oil, coal, and natural gas, makes up the majority of greenhouse gas emissions, which is the major contributor to global warming [1,2]. Therefore, it is imperative to develop technologies to reduce CO2 emissions. To date, the main approaches developed to reduce CO2 emissions include ① CO2 capture and sequestration and ② CO2 conversion and utilization. CO2 capture and sequestration is a set of technologies developed for reducing CO2 emissions from fossil-fueled power plants. CO2 capture is to separate CO2 from gas mixtures via chemical absorption using an agent such as monoethanol amine or physical adsorption using solid adsorbents such as activated carbons and metal organic frameworks (MOFs), membrane separation and cryogenic separation at a low temperature [3]. CO2 sequestration refers to long-term storage of CO2 in ocean, soils, plants, and geologic formations [3,4]. However, these technologies are relatively energy-intensive and thus are not cost-effective. An alternative sustainable approach to mitigating CO2 emissions is CO2 utilization and conversion. CO2 utilization describes the uses of CO2 in both physical processes such as welding medium and chemical processes such as chemical synthesis. The CO2 utilization technology has found applications in several industries such as carbonated drinks, dry ice, solvent, food preservation, refrigerant, and fire extinguisher. These direct CO2 utilization applications, however, have a small effect on the overall CO2 emission reduction due to the limited usage [5]. CO2 utilization can also be employed indirectly in industries to promote a process such as in enhanced gas recovery, enhanced oil recovery, and enhanced geothermal systems. CO2 conversion refers to the transformation of CO2 into valuable products such as fuels and chemicals. CO2 can be utilized directly as a feedstock to react with other components to form chemical products such as urea and formic acid under heat and/or pressure. CO2 can also be utilized indirectly as a building block of a chemical product [6]. It is worth noting that CO2 is a highly stable molecule. The C-O bond strength in CO2 molecule is 364 kJ/mol and the carbon atom has the highest oxidation state, therefore, CO2 conversion into valuable chemicals generally requires a significant input of energy and the use of a catalyst [7]. The main approaches to converting CO2 into valuable products include photocatalysis [8,9], chemical fixation [10], hydrogenation [11], and electrocatalysis [12]. Among these methods, photocatalytic reduction of CO2 attracts much attention because it converts CO2 into useful products by utilizing solar energy via a clean and sustainable route. Under sunlight irradiation, photocatalysts can induce CO2 reduction and convert it into fuels and chemicals, mainly including CO, HCHO, HCOOH, CH3OH, C2H5OH, and CH4, etc., which is determined by the number of electrons and protons (e-/H+) transferred in the reactions. The selectivity of product and efficiency of CO2 reduction may have been affected by the thermodynamic reduction potentials and reaction conditions. H2O is commonly used as a solvent for CO2 photocatalytic reduction because it is of low cost and is natural abundance of hydrogen. The photoreduction reactions of CO2 in aqueous solution at pH= 7 and their reduction potentials with reference to the normal hydrogen electrode (NHE) at 25°C and 1 atm are given in Table 1 [13,14].
Tab.1 Photoreduction reactions of CO2 in aqueous solution at pH= 7 and their reduction potentials with reference to normal hydrogen electrode (NHE) at 25°C and 1 atm |
| Reaction | Thermodynamics potential (V) vs. NHE |
|---|---|
| CO2 + 2H+ + 2e-->HCOOH | -0.61 |
| CO2 + 4H+ + 4e-->HCHO+ H2O | -0.52 |
| CO2 + 2H+ + 2e-->CO+ H2O | -0.48 |
| CO2 + 6H+ + 6e-->CH3OH+ H2O | -0.38 |
| CO2 + 12H+ + 12e-->C2H5OH+ 3H2O | -0.33 |
| CO2 + 8H+ + 8e-->CH4 + 2H2O | -0.24 |
| H2O ->1/2O2 + 2H+ + 2e- | +0.82 |
| 2H+ + 2e-->H2 | -0.41 |
Photocatalysis is a process that converts solar energy to chemical energy, where solar energy is the driving force for the excitation and transfer of holes and electrons to induce oxidation and reduction reactions. The photocatalytic CO2 reduction process can be explained as follows: first, under light illumination with energy greater than the band gap of the photocatalysts, electrons are excited from the valence band (VB) to the conduction band (CB), generating an equal number of holes in the VB. Secondly, electron-hole pairs separate from each other and move to the surface of photocatalyst. Finally, the electrons reduce CO2 into chemical products, while the holes oxidize a sacrificial agent or H2O. To induce CO2 photoreduction, it requires that reduction potential and oxidation potential of the reaction is less negative and less positive than the CB edge and the VB edge of the photocatalyst, respectively.
Since Fujishima et al. reported on the feasibility of using TiO2 for photoelectron chemical water splitting under ultraviolet (UV) irradiation [15], many different photocatalysts have been developed, most of which are inorganic semiconductors such as TiO2 [16], CdS [17], ZnO [18], ZnS [19], Fe2O3 [20], g-C3N4 [21], Ag3PO4 [22] and their composites. Some cocatalysts such as Pt have also been explored [23,24]. However, the wide band gap, high recombination rate of electron‐hole pairs, low adsorption capacities for CO2, and less tunable structure of the conventional semiconductors limit their practical applications. For example, TiO2 is mainly used for UV light photocatalysis because of its wide band gap (3–3.2 eV) [25]. ZnO and CdS are not stable under irradiation in an aqueous solution, making them inactive over time [26,27]. Therefore, it is imperative to develop new photocatalytic materials with finely tunable energy band structures and high stability.
Metal-organic frameworks (MOFs), a new class of inorganic-organic hybrid material composed of inorganic metal clusters and organic linkers which are connected through coordination bonds, have attracted much interest due to their large surface areas, tunable structures and high porosity [28–34]. These excellent properties enable their wide applications in many fields, such as gas storage and separation [35], catalysis [36], drug delivery [37], water treatments [38], and sensing [39]. Recently, MOFs have found applications in CO2 photoreduction, degradation of organics, and chemical synthesis as photocatalytic materials [40–42]. In photocatalytic reaction, MOFs undergo a similar process to traditional semiconductors, but differently, the VB and CB are described as the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) in MOFs, respectively [36,43]. In general, the HOMO and LUMO energy levels are associated with the redox potential energy levels of the organic linker and the metal-oxo cluster, respectively [42]. The tunable organic linkers and/or metal clusters in MOFs can act as antennas to harvest light to generate electron-hole pairs for photocatalysis. Most of the photocatalytic reactions in MOFs reported to date are ascribed to a localized metal-to-ligand charge transfer (MLCT), a ligand-to-metalcharge transfer (LMCT), or a p–p* transition of the aromatic ligand [36]. MOFs are utilized for photocatalytic reduction of CO2 because of their attributes as follows: ① both the organic linkers and metal clusters can serve as the light harvesting sites and they can be tailored to tune the optical absorption range of MOFs [44], ② some MOFs have a catalytic activity resulting from the catalytically active organic linkers and/or unsaturated metal sites [45,46], ③ MOFs have a large surface area and a high CO2 adsorption capacity, and the high CO2 concentration in the pores can facilitate the photocatalytic reactions, and ④ the three-dimensional porous structures and high surface areas enable MOFs to incorporate foreign photoactive species into their frame works, through which photocatalytic reactions can be enhanced by the synergistic cooperation of the metal clusters, organic linkers and the incorporated active sites [47]. Based on these unique properties, MOFs are promising photocatalysts for CO2 reduction.
MOF photocatalysts can be classified into two categories: ① Pristine MOF photocatalysts, which areal so called “opportunistic” photocatalysts. The photocatalytic properties of some pristine MOFs are a consequence of the catalytically active organic linkers and unsaturated metal sites [48], and some MOFs are semiconductors [49–51] whose photocatalytic reactions are completed through LMCT. Most pristine MOFs as photocatalysts, however, have large band gaps and barely absorb visible light, which limits their practical applications. These MOFs generally have a low photon energy utilization efficiency, and are typically employed for photocatalytic degradation of organics which have high thermodynamic driving forces and low kinetic barriers [36]. ② Modified MOF photocatalysts, which refer to MOFs functionalized to increase light harvesting and decrease the recombination rate of the photo-generated charge carriers for photocatalytic activity enhancement. To date, many strategies have been developed for enhancing the photocatalytic properties of MOFs under solar illumination, including the decoration of the metal clusters and organic linkers [52,53], incorporation of foreign photocatalytic species such as metal particles, semiconductors [54–56], and photosensitizers [57–59].
So far, a number of good literature reviews have summarized the photocatalytic applications of MOFs [60–69], most of which have discussed the role of MOFs and their performances in water splitting [70–73], degradation of organic pollutants [74–77], hydrogen evolution [36,78–80], and CO2 conversion [36,43,66,78,80–85]. Therefore, this review is mainly focused on reviewing the recent progress of MOF-based photocatalysts in CO2 photoreduction. Besides, it discusses modification strategies of MOFs and their photocatalytic activities in CO2 reduction. Moreover, it explores the challenges and future perspectives of MOFs-based photocatalysts in CO2 photoreduction.
Modification of pristine MOFs as photocatalysts
Compared to traditional inorganic semiconductors, it is more convenient to tune the optical properties and consequently the photocatalytic properties of MOFs by modification of the metal clusters or organic linkers. It is reported that the linker decoration can change the energy band gap of MOFs by shifting the photo absorption edge from the UV to visible light region [86]. Many strategies have been developed to functionalize organic linkers and metal clusters [44,87–92], such as amine and photosensitizer functionalization of organic linkers and incorporation of foreign metal cations into organic linkers and metal sites (e.g., Ti–O, Fe–O, and Zr–O clusters). The performances of recent photocatalytic MOFs for CO2 photoreduction are summarized in Table 2.
Tab.2 Performances of recent photocatalytic MOFs for CO2 photoreduction |
| Functionalization | MOF | Irradiation | Solvent/sacrificial agent | Main product | Photocatalytic reactivity | Reaction time/h | Ref. |
|---|---|---|---|---|---|---|---|
| – | MIL-125(Ti) | UV | MeCN/TEOA | HCOO- | 2.41 mmol | 10 | [53] |
| – | MIL-125(Ti) | Visible | 0 | ||||
| NH2-functionalized linker | MIL-125(Ti)-NH2 | 8.14 mmol | |||||
| NH2-functionalized linker | NH2‐UiO‐66(Zr) | Visible | MeCN/TEOA | HCOO- | 13.2 mmol | 10 | [93] |
| (NH2)2‐UiO‐66(Zr) | 20.7 mmol | ||||||
| NH2-modified with partial Ti cation substitution | NH2-UiO-66(Zr/Ti) | Visible | MeCN/TEOA BNAH | HCOO- | 22.23 mmol | 6 | [94] |
| (NH2)2‐UiO‐66(Zr/Ti) | 31.57±1.64 mmol | ||||||
| – | MIL-101(Fe) | Visible | MeCN/TEOA | HCOO- | 59.0 mmol | 8 | [95] |
| NH2-functionalized linker | MIL-101(Fe)-NH2 | 178 mmol | |||||
| – | MIL-53(Fe) | 29.7 mmol | |||||
| NH2-functionalized linker | MIL-53(Fe)-NH2 | 46.5 mmol | |||||
| – | MIL-88B(Fe) | 9.0 mmol | |||||
| NH2-modified with partial metal ion substitution | MIL-88B(Fe)-NH2 | 30 mmol | |||||
| NH2-modified with partial Ti cation substitution | NH2-UiO-66(Zr/Ti) | Visible | MeCN/TEOA | HCOO- | 3.4 mmol/mol | 10 | [96] |
| NH2-UiO-66(Zr/Ti)-100-4 | 4.2 mmol/mol | ||||||
| NH2-UiO-66(Zr/Ti)-120-16 | 5.8 mmol/mol | ||||||
| Porphyrin- functionalized linker | Rh-PMOF-1(Zr) | Visible | MeCN/TEOA | HCOO- | 6.1 mmol/mmol | 18 | [97] |
| Porphyrin- functionalized linker | Zn/PMOF | UV-visible | H2O vapor | CH4 | 10.43 mmol | 4 | [98] |
| Porphyrin- functionalized linker | Al/PMOF | Visible | H2O/TEOA | CH3OH | 37.5 ppm /(g·h) | – | [99] |
| Porphyrin- functionalized linker with partial Cu cation substitution | Cu-Al/PMOF | 262.6 ppm /(g·h) | |||||
| Porphyrin- functionalized linker | MOF-525 | Visible | MeCN/TEOA | CO | 64.02 mmol /(g·h) | 6 | [100] |
| CH4 | 6.2 mmol /(g·h) | ||||||
| Porphyrin- functionalized linker with embedded Zn cations | MOF-525-Zn | CO | 111.7 mmol /(g·h) | ||||
| CH4 | 11.64 mmol /(g·h) | ||||||
| Porphyrin- functionalized linker with embedded Co cations | MOF-525-Co | CO | 200.6 mmol /(g·h) | ||||
| CH4 | 36.76 mmol /(g·h) | ||||||
| Porphyrin- functionalized linker | PCN-222 | Visible | MeCN/TEOA | HCOO- | 30 mmol | 10 | [101] |
| Photosensitizer functionalization | Eu-Ru(phen)3-MOF | Visible | MeCN/TEA | HCOO- | 47 mmol | 10 | [102] |
| Photosensitizer functionalization | UiO-67-ReI(CO)3(5,5′-dcbpy)Cl | Visible | MeCN/TEA | CO | TON= 5.0 | 6 | [103] |
| H2 | TON= 0.5 | ||||||
| CO | TON= 10.9 | 20 | |||||
| H2 | TON= 2.5 | ||||||
| – | ReI(CO)3(5,5′-dcbpy)Cl | CO | TON= 5.6 | 6 | |||
| H2 | TON= 0.3 | ||||||
| CO | TON= 7.0 | 20 | |||||
| H2 | TON= 1.0 | ||||||
| Photosensitizer functionalization | Zr6(O)4(OH)4[Re(CO)3Cl(bpydb)]6 | Visible | MeCN/TEA | CO | TON= 6.44 | 6 | [104] |
| H2 | TON= 0.40 | 6 | |||||
| Photosensitizer functionalization | UiO-67-ReI(CO)3(5,5′-dcbpy)Cl | Visible | TEA | CO | 0.5 mmol/(g·h) | 6 | [52] |
| UiO-67-ReI(CO)3(5,5′-dcbpy)Cl-NH2 (33% (mol)) | CO | 1.5 mmol/(g·h) | 6 | ||||
| Photosensitizer functionalization | UiO-67-Cp*Rh(5,5′- dcbpy)Cl2 (10%) | Visible | ACN/TEOA | HCOO- | TON= 47 | 10 | [105] |
| H2 | TON= 36 | ||||||
| – | Cp*Rh(5,5′- dcbpy)Cl2 | HCOO- | TON= 42 | ||||
| H2 | TON= 38 | ||||||
| – | [Ru(bpy)3]Cl2 | HCOO- | TON= 125 | ||||
| H2 | TON= 55 | ||||||
| Photosensitizer functionalization | MOF-253-Ru(CO)2Cl2 | Visible | MeCN/TEOA | HCOO– | 0.67 mmol | 8 | [106] |
| CO | 1.86 mmol | ||||||
| H2 | 0.09 mmol | ||||||
| Ru(bpy)2Cl2- sensitized MOF-253-Ru(CO)2Cl2 | HCOO– | 4.84 mmol | |||||
| CO | 1.85 mmol | ||||||
| H2 | 0.72 mmol | ||||||
| MOF-253-Ru(bpy) 2Cl2 | HCOO– | 0.46 mmol | |||||
| CO | 0.21 mmol | ||||||
| H2 | 0.07 mmol | ||||||
| – | Ru(bpy)2Cl2 | HCOO– | 0.27 mmol | ||||
| CO | 0.18 mmol | ||||||
| H2 | 0 mmol | ||||||
| Photosensitizer functionalization | Y[Ir(ppy)2(4,4′‐dcbpy)]2[OH] | Visible | MeCN/TEOA | HCOO– | 118.8 mmol/(g·h) | 6 | [107] |
| Photosensitizer functionalization | [Cd2[Ru(4,4’-dcbpy)3]·12H2O]n nanoflower | Visible | MeCN/TEOA | HCOO– | 77.2 mmol/(g·h) | 8 | [108] |
| [Cd2[Ru(4,4’-dcbpy)3]·12H2O]n microflake | 52.7 mmol/(g·h) | ||||||
| [Cd2[Ru(4,4’-dcbpy)3]·12H2O]n bulk crystals | 30.6 mmol/(g·h) | ||||||
| Photosensitizer functionalization | [Cd3[Ru(5,5′-dcbpy)3]2·2(Me2NH2)]n | Visible | MeCN/TEOA | HCOO– | 67.5 mmol/(g·h) | 6 | [109] |
| [Cd[Ru(bpy)(4,4′-dcbpy)2]·3H2O]n | 71.7 mmol/(g·h) | ||||||
| Catechol- functionalized linker | UiO-66-CrIIIcatbdc | Visible | MeCN/TEOA/BNAH | HCOO– | TON= 11.22±0.37 | 6 | [110] |
| UiO-66-GaIIIcatbdc | TON= 6.14±0.22 | ||||||
| Anthracene- functionalized linker | NNU-28 | Visible | MeCN/TEOA | HCOO– | 183.3 mmol/(h·mmol) | 10 | [111] |
Notes: ACN–acetonitrile; BNAH–1-benzyl-1,4-dihydronicotinamide; bpy–2,2’-bipyridine; catbdc–2,3-dihydroxyterephthalic acid; Cp*–pentamethylcyclopentadiene; MeCN–acetonitrile; phen–phenanthroline; TEA–trimethylamine; TEOA–triethanolamine; TON–total turnover number; 5,5′-dcbpy–2,2’-bipyridine-5,5′-dicarboxylic acid; bpydb–4,4’-(2,2’-bipyridine-5,5′-diyl)dibenzoate; ppy–2-phenylpyridine; 4,4’-dcbpy–2,2’-bipyridine-4,4’-dicarboxylate |
Metal cluster nodes
MOFs have three-dimensional structures constructed from the interconnection of metal cluster nodes with organic linkers. Besides photoactive ligands, metal cluster nodes in MOFs can also drive photocatalytic CO2 reduction. To date, Ti-, Zr-, and Fe-based MOFs are among the most investigated photocatalytic MOFs whose photoactive metal nodes in metal-oxo clusters can initiate photocatalytic reaction by trapping photo-excited electrons and decreasing recombination rate of electron-hole pairs [83]. These MOFs share the similarity that they all contain metal ions with variable valence states (e.g., Ti4+/Ti3+ , Zr4+/Zr3+ , and Fe3+/Fe2+ ), which enables their effectiveness on photocatalytic reduction.
Of the Ti-based MOFs, MIL-125(Ti) (Ti8O8(OH)4(O2C-C6H4-CO2)6) is the most studied MOF, which is constructed from the Ti8O8(OH)4 secondary building units (SBUs) and 1,4-benzenediacarboxylate (BDC) ligands. Fu et al. investigated the CO2 photoreduction over MIL-125(Ti) [53]. 2.41 mmol HCOO- was detected in the acetonitrile (MeCN) solvent with TEOA under 365 nm UV light irradiation for 10 h. A number of strategies have been developed to functionalize MIL-125(Ti) to enhance the photocatalytic reduction of CO2, such as organic linker decoration and incorporation of foreign photocatalytic components into MOFs to form MOF composites. These strategies will be discussed in Section 2.2 and Section 3.
Zr-based MOFs, which have robust structures and excellent thermal and chemical stabilities [112,113], are another popular class of photocatalytic MOFs since their synthesis in 2008 [114]. The representative examples are UiO-66(Zr) and UiO-67(Zr), which are constructed by integrating Zr6O4(OH)4 SBUs with BDC ligands and 4,4′-biphenyldicarboxylic acid (BPDC) ligands, respectively [114,115]. Compared to Ti-based MOFs, Zr-based MOFs have more negative redox potential (Ti4+/Ti3+ (–0.1 V), Zr4+/Zr3+ (–1.06 V)) [116,117]. However, UiO-66 exhibits no absorption under visible light irradiation owing to the higher redox potential energy level of the Zr6O4(OH)4 SBUs in UiO-66 than the LUMO of the BDC linkers, leading to the low efficiency in LMCT and consequently low efficacy in CO2 photoreduction [94]. Partial substitution of metal cations in MOFs can introduce metal-to-metal charge transfer, which can promote photocatalytic performance especially under visible light irradiation [94,96]. A bimetallic UiO-based MOF, NH2-UiO-66(Zr/Ti), was prepared by Cohen and coworkers by partially substituting Zr in NH2-UiO-66(Zr) with Ti [94]. NH2-UiO-66(Zr/Ti) had a better performance in photocatalytic CO2 reduction under visible light irradiation compared to NH2-UiO-66(Zr). The reason for this is that Ti ions incorporated enable the SBUs to accept the electrons generated via light absorption by the organic linkers. No HCOO- was produced over the parent UiO-66(Zr)-NH2, demonstrating that Ti was critical for photocatalysis.
Sun et al. also prepared Ti-substituted NH2-UiO-66(Zr/Ti) MOFs (NH2-UiO-66(Zr/Ti)-120-16 and NH2-UiO-66(Zr/Ti)-100-4) doped with different amounts of Ti by a post-synthetic exchange method and examined their photocatalytic performance on CO2 reduction under visible light irradiation (Fig. 1(a)) [96]. NH2-UiO-66(Zr/Ti)-120-16 had an enhanced photocatalytic activity toward CO2 conversion with a yield of 5.8 mmol/mol of HCOO- after 10 h visible light irradiation in MeCN solvent with TEOA as a sacrificial agent, which was 1.7 times of that observed over NH2-UiO-66(Zr) (3.4 mmol/mol) under similar conditions. In contrast, NH2-UiO-66(Zr/Ti)-100-4 produced 4.2 mmol mol-1 of HCOO-, which was less than NH2-UiO-66(Zr/Ti)-120-16, but still higher than that over the pristine NH2-UiO-66(Zr). The enhancement in photocatalytic performance over Ti-substituted NH2-UiO-66(Zr/Ti) MOFs is associated with the increase in CO2 adsorption capacity and photocatalytic sites, both of which result from Ti doped into Zr-O clusters of NH2-UiO-66(Zr). Based on the experimental observations and theoretical studies, the mechanism for enhanced photocatalytic reactions over NH2-UiO-66(Zr/Ti) is proposed (see Fig. 1(b)). When Zr4+ centers in Zr6O4(OH)4 are partially substituted by Ti4+ to form (Ti/Zr)6O4(OH)4, the excited NH2-BDC upon visible light irradiation can transfer electrons to either Zr4+ or Ti4+ centers. The theoretical calculations show that there is a higher probability for electrons to be transferred to Ti4+ than that to Zr4+ centers, leading to the formation of (Ti3+/Zr4+)6O4(OH)4 SBUs. The Ti3+ in the excited (Ti3+/Zr4+ )6O4(OH)4 SBUs can play a role of electron donor to donate electrons to Zr4+ , leading to Ti4+–O–Zr3+ formation. As a result, the substituted Ti center in NH2-UiO-66(Zr/Ti) facilitates the interfacial charge transfer from the excited NH2-BDC to Zr–O clusters, which boosts the enhanced photocatalytic reactions over NH2-UiO-66(Zr/Ti).
Recently, Fe-based MOFs as photocatalysts for CO2 reduction have attracted much interest owing to the fact that the Fe–O clusters can be directly photoexcited to induce electron transfer from O2- to Fe3+ to form Fe2+ under visible light irradiation, which drives the photocatalytic reaction [95,118]. Since the Fe sites play a role of photocatalytic centers, Fe-based MOFs are capable of photocatalytically reducing CO2 under visible light irradiation in the absence of LMCT. Wang et al. reported a series of Fe-based MOFs, MIL-101(Fe), MIL-53(Fe), MIL-88B(Fe), all of which had a photocatalytic activity for CO2 reduction to produce HCOO- under visible light irradiation [95]. It showed that the photocatalytic activity of MIL-101(Fe), MIL-53(Fe) and MIL-88B(Fe) toward CO2 conversion into HCOO- in MeCN solvent with TEOA as a sacrificial reactant was 59.0, 29.7, and 9.0 mmol, respectively, after visible light irradiation for 8 h. Of the three investigated Fe-based MOFs, MIL-101(Fe) had the best photocatalytic performance attributing to the existence of the unsaturated Fe sites in its structure that were absent in MIL-53(Fe) and MIL-88B(Fe). MIL-53(Fe) had a better activity than MIL-88B(Fe) attributing to its higher CO2 adsorption capacity (13.5 g/cm3) than MIL-88B(Fe) (10.4 g/cm3) which might be ascribed to its one dimensional framework structure which possessed a better electro-conductivity. The photocatalytic activities of the three Fe-based MOFs were enhanced by amine functionalization. The HCOO- yield of 178, 46.5, and 30 mmol was produced over NH2-MIL-101(Fe), NH2-MIL-53(Fe), and NH2-MIL-88B(Fe), respectively, under the same photocatalytic conditions, which were higher in comparison to the parent Fe-MOFs. This photocatalytic activity enhancement is caused by the existence of the dual excitation pathways: excitation of NH2-functionalized organic ligands to transfer electrons to the Fe center in addition to the direct excitation of Fe-O clusters (see Fig. 2).
Modification of organic linkers
Functionalization of organic linkers has been considered as an effective approach to increase the absorption of light and decrease the recombination rate of the photo-generated charge carriers and consequently to improve the photocatalytic reduction of CO2. To date, several different organic linker functionalization strategies have been developed for enhancing the photocatalytic performance of MOFs toward CO2 reduction under light illumination, including amine functionalization, utilization of porphyrin-based organic linkers, and photosensitizer functionalization.
Amine functionalization
Studies have shown that amine groups incorporated into the organic linkers in MOFs contribute to the CO2 adsorption capacity enhancement, adsorption region broadening, and CO2 photoreduction performance boost. The introduction of amine groups into aromatic polycarboxylates in MOFs can increase the interactions between CO2 molecules and the modified linkers [119]. The CO2 photoreduction improvement is attributed to the lone pair of electrons in amine groups, which can interact with the p*-orbitals of the benzene ring, leading to the donation of electrons to the anti-bonding orbitals. This interaction enables the formation of a new higher energy HOMO level, which leads to a broader optical absorption region and consequently enhanced CO2 photoreduction [69].
In 2012, Li and coworkers [53], for the first time, prepared an amino-functionalized MOF, NH2-MIL-125(Ti), and examined CO2 adsorption and photoreduction under visible light irradiation. NH2-MIL-125(Ti) had a higher CO2 adsorption capacity (132.2 cm3/g) in comparison to that of MIL-125(Ti) (98.6 cm3/g) due to the increased interaction between CO2 molecules and the amine-modified linkers. Moreover, amine functionalization led to a broader adsorption range of 550 nm for NH2-MIL-125(Ti) compared to MIL-125(Ti) with an adsorption range of 350 nm (see Fig. 3(a)). Photoreduction of CO2 over NH2-MIL-125(Ti) was performed in MeCN solvent with TEOA as an electron donor. After visible light irradiation for 10 h, 8.14 mmol of HCOO– was obtained, which was much higher than that of MIL-125(Ti) with no yield. The photocatalytic mechanism is proposed as follows (see Fig. 3(b)): under visible light irradiation, electrons are transferred from organic linkers to Ti4+ cations, and Ti4+ cations are reduced to Ti3+ by TEOA acting as an electron donor, leading to the formation of a long-lived excited charge separation. Ti3+ cations subsequently reduce CO2 to HCOO-. In 2013, the same research group (Li and coworkers) [93] developed another amino-functionalized MOF, NH2-UiO-66(Zr), and found that it had a higher activity for CO2 photoreduction than previously reported NH2-MIL-125(Ti) in the presence of TEOA as a sacrificial agent. Similar to previous reports, substitution of the organic linker in UiO-66(Zr) by NH2-BDC led to a broader optical absorption range of NH2-UiO-66(Zr) due to the increased interaction between the NH2-BDC linker and the Zr-O clusters. CO2 uptake of NH2-UiO-66(Zr) (68 cm3/g) was also improved in comparison to the parent UiO-66(Zr) (53 cm3/g) owing to the enhanced interactions of the NH2 functional groups with the CO2 molecules [119,120]. After visible light irradiation over NH2-UiO-66(Zr) for 10 h, 13.2 mmol of HCOO– was produced. Partial substitution of the organic linker NH2-BDC by (NH2)2-BDC in NH2-UiO-66(Zr) further improved its CO2 photoreduction activity, which gave 20.7 mmol of HCOO- under the same conditions as over NH2-UiO-66(Zr) (Fig. 4). The improvement in CO2 photocatalytic reduction over the mixed NH2-UiO-66(Zr) is attributed to enhanced light absorption in the visible region and increased CO2 adsorption (71 cm3/g). Similar results were reported by Cohen and coworkers [94] who investigated the CO2 photocatalytic reduction performance over (NH2)2-UiO-66(Zr) and (NH2)2-UiO-66(Zr/Ti) prepared by the introduction of diamine-substituted ligands into (NH2)-UiO-66(Zr) and (NH2)-UiO-66(Zr/Ti), respectively. It was found that an increase of the amino group number introduced new energy levels for additional light absorption and charge transfer, leading to enhancement in photocatalytic activities.
Utilization of porphyrin-based organic linkers
Porphyrins have complex cyclic structures consisting of four pyrrole rings linked to each other by methine groups. Due to their strong interactions with CO2, high light adsorption efficiency and catalytic performance, porphyrin-based organic linkers have been incorporated into MOFs for CO2 photoreduction.
Su and coworkers [97] prepared a rhodium(III)-porphyrin zirconium MOF (Rh-PMOF-1(Zr)) using a Rh-based metalloporphyrin tetracarboxylic ligand Rh(TCPP)Cl (TCPP= tetrakis(4-carboxyphenyl) porphyrin) and ZrCl4. The CO2 adsorption and photoreduction over Rh-PMOF-1(Zr) were examined. Rh-PMOF-1(Zr) had a high CO2 adsorption capacity of 53 cm3/gat 298 K. After visible-light irradiation for 18 h, the yield of HCOO– reached 6.1 mmol/mmolcat. This showed that Rh-PMOF-1(Zr) had a long-lived excited-state under vacuum at 298 K with the lifetime of 207 ms, which contributed to the improvement in the photocatalytic activity of Rh-PMOF-1(Zr). Two catalytic reactions contribute to the CO2 photocatalytic reduction to the formation of HCOO– over Rh-PMOF-1(Zr): ① metalloporphyrin ligand plays a role of an antenna to harvest light, generate electrons, and transfer electrons to the zirconium oxo clusters, reducing CO2 to HCOO–, and ② the rhodium-porphyrin ligands serve as photocatalytic centers toward CO2 photoreduction.
Sharifnia and coworkers [98] used TCPP as ligand and Zn(NO3)2·6H2O to prepare a porphyrin-based MOF (Zn/PMOF) and performed photocatalytic reduction of CO2 over Zn/PMOF in the presence of H2O vapor under UV-visible light. After 4 h irradiation, Zn/PMOF had a CO2 photoreduction activity with CH4 formation of 10.43mmol. Only a small amount of CH4 (10.77 mmol) was produced by extending the irradiation time to 24 h, which may have been caused by the fact that the intermediate product and/or byproduct formed after 4 h have saturated the active sites.
Huang et al. [99] synthesized two types of porphyrin-based MOFs, Al/PMOF and Cu-Al/PMOF, and compared their CO2 adsorption capacities and photoreduction of CO2 toward methanol. Al/PMOF was synthesized using TCPP as organic linker and AlCl3·6H2O while Cu-Al/PMOF was produced by doping Cu2+ into Al/PMOF. It showed that Cu2+ in Cu-Al/PMOF contributed to the enhanced photocatalytic reduction of CO2. Cu-Al/PMOF had a higher CO2 adsorption capacity (277.4 mg/g) than that of Al/PMOF (153.1 mg/g). Moreover, the methanol formation rate over Cu-Al/PMOF (262.6 ppm/(g·h)) was 7 times as high as that of Al/PMOF (37.5 ppm/(g·h)). As demonstrated by in situ Fourier transform infrared (FT-IR) spectra, CO2 could be chemically adsorbed on the Cu site in Cu-Al/PMOF, where the linear CO2 molecules would bend, thus lowering the reaction barrier and improving the photocatalytic efficiency.
Ye and coworkers [100] prepared MOF-525 (Zr6O4(OH)4(TCPP-H2)3) by integrating Zr6 clusters with porphyrin-based organic linkers. MOF-525-Co and MOF–525-Zn were also developed by the introduction of coordinatively unsaturated Co sites and Zn sites into the porphyrin units of MOF-525, respectively. As shown in Fig. 5, both MOF-525-Co (33.6 cm3/g) and MOF-525-Zn (28.1 cm3/g) had a higher CO2 adsorption capacity than that of pristine MOF-525 (25.3 cm3/g) due to the enhanced interaction between CO2 molecules and the introduced open Co and Zn metal sites. CO2 photoreduction over MOF-525, MOF-525-Co and MOF-525-Zn was performed in the presence of MeCN solvent and TEOA as an electron donor under visible light irradiation for 6 h, and two products, CO and CH4 were produced. MOF-525-Co had the highest CO evolution rate of 200.6 mmol/(g·h) (yield: 2.42 mmol), and a CH4 evolution rate of 36.76 mmol/(g·h) (yield: 0.42 mmol), followed by MOF-525-Zn (CO, 111.7 mmol/(g·h); CH4, 11.635 mmol/(g·h)) and MOF-525 (CO, 64.02 mmol/(g·h); CH4, 6.2 mmol/(g·h)). The photocatalytic activity enhancement for MOF-525-Co and MOF–525-Zn is partially ascribed to the difference in their charge separation efficiencies. MOF-525-Co and MOF-525-Zn had an enhanced efficiency in charge separation because of the lower energy at Co or Zn site, to which the electrons can be transferred from porphyrin units with a higher easiness. Furthermore, MOF-525-Co had a higher electron transfer efficiency (62.1 %) than that of MOF-525-Zn (24.9 %). These partially account for the better catalytic activity of MOF-525-Co than that of MOF-525-Zn. In addition, MOF-525-Co had a good stability and a good reproducible photocatalytic activity after three cycles (Fig. 5).
Xu et al. [101] prepared a zirconium-porphyrin MOF, PCN-222 (also named as MOF-545 or MMPF-6), by employing zirconium (IV) chloride and TCPP as ligand. PCN-222 had a CO2 uptake of 35 cm3/gat 298 K and 1 atm. Photocatalytic reduction of CO2 over PCN-222 was performed in MeCN as solvent and TEOA as a sacrificial agent. After visible light irradiation for 10 h, HCOO- was produced with a yield of 30 mmol, which was much higher than that observed over the TCPP ligand alone (2.4 mmol), as shown in Fig. 6. It is proposed that upon irradiation, the TCPP in PCN-222 acts as an antenna to harvest visible light, generate electron–hole pairs, and transfer electrons to the Zr-oxo clusters toward CO2 reduction to HCOO- in the presence of TEOA as the electron donor. Two factors contribute to the enhancement in photocatalytic performance of PCN-222: ① the high CO2 adsorption capacity of PCN-222 might enable higher interaction with CO2 in MeCN, thereby promoting the photocatalytic reaction, and ② the ultrafast transient absorption and photoluminescence spectroscopy reveals that the emergence of an extremely long-lived electron trap state in PCN-222 significantly suppresses the electron-hole recombination, thus enhancing the CO2 photoreduction efficiency.
Fig.6 Left: Amount of HCOO- produced as a function of visible light irradiation time over PCN-222 (a), H2TCPP (b), no PCN-222 (c), no TEOA (d), and no CO2 (e). Right: 13C Nuclear magnetic resonance (NMR) spectra for the product obtained from reaction with 13CO2 (a) or 12CO2 (b) (Reproduced with permission from Ref. [101]. Copyright 2015, American Chemical Society) |
Photosensitizer functionalization
Photosensitizers are able to harvest light to generate electron-hole pairs and act as catalyst sites for CO2 photoreduction. Photoactive metal complexes, such as Ru, Re, and Ir-based polypyridine units, have been widely used to functionalize MOFs to enhance their CO2 photocatalytic reaction activities by introducing catalytic active centers and photosensitive sites for visible light harvesting [121–123].
Yan et al. [102] produced an Eu-Ru(phen)3-MOF (phen= phenanthroline) by integrating the triangular Ru(phen)3-derived tricarboxylate ligand as photosensitizer into Eu-MOF with Eu(III)2(m2-H2O) SBUs. Transient absorption results and theoretical calculations showed that photo-excitation of the Ru metalloligands in Eu-Ru(phen)3-MOF initiated electron transfer into the nodes to generate dinuclear [Eu(II)]2 active sites, which selectively converted CO2 to HCOO- with a yield of 47 mmol in 10 h in the presence of MeCN and TEOA under visible light irradiation. This was higher than that observed over the Ru(phen)3-derived tricarboxylate acid metalloligandalone under the same condition.
Lin and coworkers [103] incorporated ReI(CO)3(5,5′-dcbpy)Cl (at 4 wt% doping level, 5,5′‐dcbpy= 2,2’-bipyridine-5,5′-dicarboxylic acid) into the UiO-67 framework built from Zr6(µ3-O)4(µ3-OH)4 SBUs and BPDC ligands, and tested the photocatalytic reduction of CO2 in MeCN solvent and trimethylamine (TEA) as a sacrificial agent under visible light. The total turnover number (TON) of CO and H2 over the as-doped UiO-67 reached 5.0 and 0.5, respectively, under visible light irradiation of 6 h. After 20 h, CO-TON and H2-TON reached 10.9 and 2.5, respectively, which were higher than that observed for the bare ReI(CO)3(5,5′-dcbpy)Cl (CO-TON= 7.0, H2-TON= 1.0) under the same condition because of the decomposition of ReI(CO)3(5,5′-dcbpy)Cl under irradiation. This enhanced TON is proposed to be associated with the stabilization effect of the active-site isolation of the immobilized Re-based catalysts in the framework. Lately, the same group produced another photosensitizer-functionalized MOF, Zr6(O)4(OH)4[Re(CO)3Cl(bpydb)]6(MOF-1), by integrating the elongated linear (bpy)Re(CO)3Cl-containing dicarboxylate ligand (bpydb= 4,4’-(2,2’-bipyridine-5,5′-diyl)dibenzoate; bpy= 2,2’-bipyridine) with Zr6(µ3-O)4(µ3-OH)4 SBUs (Fig. 7) [104]. Under the same experimental condition as described above, CO-TON and H2-TON reached 6.44 and 0.4 in 6 h, respectively, which were higher than those of the homogeneous counterpart (CO-TON= 1.12 and H2-TON 0.18). Ryu et al. [52] embedded ReI(CO)3(5,5′-dcbpy)Cl and amine (-NH2) functional group within UiO-67 (dented as Re-MOF-NH2) and varied the ratio of the-NH2 functional groups from 0 to 80 mol%. Photocatalytic CO2 conversion was performed in the presence of TEA under visible light. The results showed that Re-MOF-NH2 incorporated with 33 mol% of-NH2 functional groups had the highest photocatalytic CO2 conversion rate (1.5 mmol/(g·h)) to CO, which was three times as that observed for Re-MOF without-NH2 incorporation (0.5 mmol/(g·h)). The enhancement in photocatalytic activity is caused by the induced different bond lengths for Re-CO in ReI(CO)3(5,5′-dcbpy)Cl by the incorporation of-NH2 functional groups, which endows the intermolecular stabilization of carbamate with CO2, thus boosting the photocatalytic activity. Fontecave and coworkers [105] functionalized UiO-67 MOF by replacing 5%–35% of BPDC linkers with Cp*Rh(5,5′-bpydb)Cl2 (Cp* = pentamethylcyclopentadiene) (named as Cp*Rh@UiO-67). Under visible light irradiation for 10 h in the presence of acetonitrile (ACN) and TEOA, HCOO--TON and H2-TON of 10%-Cp*Rh@UiO-67 reached 47 and 36, respectively, with [Ru(bpy)3]Cl2 as a photosensitizer.
Li and coworkers [106] incorporated Ru(CO)2Cl2 into MOF-253 (Al(OH)(5,5′‐dcbpy)), named as MOF-253-Ru(CO)2Cl2, and enhanced the CO2 photoreduction by producing 0.67 mmol of HCOO–, 1.86 mmol of CO as well as 0.09 mmol H2 after irradiation under visible light for 8 h in MeCN and TEOA, whereas no products were detected over pristine MOF-253 under the same condition. The performance of MOF-253-Ru(CO)2Cl2 was further improved via photosensitizer (Ru(bpy)2Cl2) functionalization, by enhancing the light absorption in the visible light region. The HCOO–, CO, and H2 produced in 8 h over sensitized MOF-253-Ru(CO)2Cl2 (Ru(bpy)2Cl2/Ru-complex was 1: 2) was 4.84, 1.85, and 0.72 mmol, respectively, which was much higher than those observed over the non-sensitized MOF-253-Ru(CO)2Cl2 under the same condition. In the sensitized MOF-253-Ru(CO)2Cl2, Ru(CO)2Cl2 reacted with the surface N,N-chelated sites to form MOF-253-supported Ru(bpy)2(X2bpy)2+, which extended the light absorption edge to 630 nm, wider than that of MOF-253-Ru(CO)2Cl2 (470 nm), thus promoting the photocatalytic CO2 reduction. However, a decrease in the photo-reactivity of the sensitized MOF-253-Ru(CO)2Cl2 was observed with increasing the amount of the photosensitizer Ru(bpy)2Cl2, which might be attributed to the pore blocking of MOF-253 by the Ru(bpy)2Cl2.
Luo and coworkers [107] developed a photocatalytic MOF (Y[Ir(ppy)2(4,4’-dcbpy)]2[OH]) (Ir-CP, ppy: 2-phenylpyridine, 4,4’-dcbpy: 2,2’-bipyridine-4,4’-dicarboxylate) using Y(NO3)3 and Ir(ppy)2(Hdcbpy). Photoreduction of CO2 over Ir-CP was performed in MeCN and TEOA under visible light irradiation. It showed that 38.0 mmol HCOO– was produced in 6 h, with a product formation rate of 118.8 mmol /(g·h). The photocatalytic mechanism is proposed as follows: the [Ir(ppy)2(dcbpy)]– unit in Ir-CP is excited under visible light and is reductively quenched by TEOA, and the CO2 molecules are reduced to HCOO– by getting electrons from [Ir(ppy)2(dcbpy)]2– units. Lately, the same group [108,109] incorporated Ru-polypyridine complexes into MOF structures as metalloligands and produced a series of photocatalytic MOFs. Ru-MOF, with a chemical formula of [Cd2[Ru(4,4’-dcbpy)3]·12H2O]n, was constructed from [Cd2(CO2)6] SBUs and [Ru(4,4’-dcbpy)3]4–metalloligands [108]. The morphologies of Ru-MOF can be tuned to form nanoflowers, microflakes, and bulk crystals structures by controlling the reactants concentration, as shown in Fig. 8. In the mixture of MeCN/TEOA, Ru-MOF nanoflowers had the highest HCOO– formation rate of 77.2 mmol/(g·h) under visible light irradiation for 8 h, followed by microflakes (52.7 mmol/(g·h)) and bulk crystals (30.6 mmol/(g·h)). The Ru-MOF nanoflowers had a better photocatalytic activity than their bulk counterparts because of their high visible light harvesting and long-lasting excited-state originated from their large surface area (8.08 m2/g; microflake: 1.33 m2/g) and high energy transfer efficiency. After this work, Luo and coworkers [109] synthesized two Ru-polypyridine-functionalized MOFs, [Cd3[Ru(5,5′-dcbpy)3]2·2(Me2NH2)]n and [Cd[Ru(bpy)(4,4′-dcbpy)2]·3H2O]n, with non-interpenetrated and interpenetrated structures, respectively.CO2 photoreduction over the two MOFs was conducted in the presence of MeCN and TEOA under visible light irradiation. It showed that the HCOO- production over [Cd3[Ru(5,5′-dcbpy)3]2·2(Me2NH2)]n and [Cd[Ru(bpy)(4,4′-dcbpy)2]·3H2O]n reached 16.1 and 17.2 mmol in 6 h (with a production rate of 67.5 and 71.7 mmol/(g·h)), respectively. The difference in photocatalytic performance of the two MOFs may have been associated with their structural stabilities. [Cd3[Ru(5,5′udcbpy)3]2·2(Me2NH2)]n has a non-interpenetrated structure in which the porous framework is supported by the coordination between metal ions and organic linkers as well as the interactions between the framework and the guest molecules in the pores. In the photocatalytic process, the exchange of guest molecules with the solvent molecules may cause the collapse of structure, leading to the decrease in photocatalytic activity. In contrast, [Cd[Ru(bpy)(4,4′-dcbpy)2]·3H2O]n with an interpenetrated structure has a higher structural stability, which endows its better photocatalytic activity than [Cd3[Ru(5,5′-dcbpy)3]2·2(Me2NH2)]n.
Fig.8 Amount of HCOO– produced as a function of irradiation time of visible light over (a) nanoflowers, (b) microcrystals, and (c) bulk crystals of the Ru-MOF ([Cd2[Ru(4,4’-dcbpy)3]·12H2O]n). (d) visible light irradiation without a sample (Inset images (from top to bottom) show nanoflowers, microcrystals and bulk crystals of the Ru-MOF, respectively. Reproduced with permission from reference [108]. Copyright 2015, Royal Society of Chemistry. ) |
MOF composite photocatalysts
MOFs have shown much potential on CO2 photoreduction thanks to their large surface areas, high CO2 uptake, as well as tunable structures and optical properties. To further improve the photocatalytic performance of MOFs, photosensitizers, semiconductors, metals and carbon materials have been incorporated into MOF structures to promote photo-generated electrons transfer and charge separation processes. The performances of MOF composites as photocatalysts for CO2 reduction are summarized in Table 3.
Tab.3 Performances of recent photocatalytic MOF composites for CO2photoreduction |
| Strategy | MOF composite | Irradiation | Solvent/ sacrificial agent | Main product | Photocatalytic reactivity | Reaction time/h | Ref. |
|---|---|---|---|---|---|---|---|
| Photosensitizer incorporation | Co-ZIF-9/[Ru(bpy)3]Cl2·6H2O | Visible | MeCN/H2O/ TEOA | CO | 41.8 mmol | 0.5 | [57] |
| H2 | 29.9 mmol | ||||||
| Co-MOF-74/[Ru(bpy)3]Cl2·6H2O | CO | 11.7 mmol | |||||
| H2 | 7.3 mmol | ||||||
| Mn-MOF-74/[Ru(bpy)3]Cl2·6H2O | CO | 1.5 mmol | |||||
| H2 | 2.9 mmol | ||||||
| Zn-ZIF-8/[Ru(bpy)3]Cl2·6H2O | CO | 2.1 mmol | |||||
| H2 | 2.4 mmol | ||||||
| Zr-UiO-66-NH2/[Ru(bpy)3]Cl2·6H2O | CO | 1.2 mmol | |||||
| H2 | 2.2 mmol | ||||||
| Co-ZIF-67 /[Ru(bpy)3]Cl2·6H2O | Visible | MeCN/H2O/TEOA | CO | 29.6 mmol | 0.5 | [59] | |
| H2 | 14.8 mmol | ||||||
| Zn-ZIF-9/[Ru(bpy)3]Cl2·6H2O | CO | 1.8 mmol | |||||
| H2 | 2.0 mmol | ||||||
| Cu-HKUST-1/[Ru(bpy)3]Cl2·6H2O | CO | 1.2 mmol | |||||
| H2 | 1.5 mmol | ||||||
| Fe-MIL-101-NH2/[Ru(bpy)3]Cl2·6H2O | CO | 4.7 mmol | |||||
| H2 | 2.1 mmol | ||||||
| Zr-UiO-66-NH2/[Ru(bpy)3]Cl2·6H2O | CO | 0.9 mmol | |||||
| H2 | 1.2 mmol | ||||||
| UiO-67-Mn(5,5′‐dcbpy) (CO)3Br)/Ru(dmb)3(PF6)2 | Visible | DMF/TEOA/ BNAH | HCOO– | TON= 50 | 4 | [58] | |
| TON= 110 | 18 | ||||||
| UiO-67-Mn(5,5′‐dcbpy)(CO)3Br) (without a photosensitizer) | TON= 18 | 18 | |||||
| Mn(5,5′‐dcbpy)(CO)3Br)/ Ru(dmb)3(PF6)2 | TON= 32 | 4 | |||||
| TON= 57 | 18 | ||||||
| Mn(bpy)(CO)3Br)/Ru(dmb)3(PF6)2 | TON= 35 | 4 | |||||
| TON= 70 | 18 | ||||||
| UiO-67-5,5′‐dcbpy)/Ru(dmb)3(PF6)2 | TON= 38 | 18 | |||||
| [Ru(dmb)3]2+ | HCOO– | TON= 33 | 18 | ||||
| Semiconductor incorporation | ZIF-8/TiO2 (ZIF-8 growth step on TiO2 film was repeated twice) | UV | H2O vapor | CO | 0.53 mmol/(g·h) | 5 | [124] |
| CH4 | 0.18 mmol/(g·h) | ||||||
| ZIF-8/Ti/TiO2 nanotube | UV-visible | Na2SO4 (0.1 mol L-1) | C2H5OH | 10 mmol/L | 3 | [125] | |
| CH3OH | 0.7 mmol/L | ||||||
| Co-ZIF-9/TiO2 (mass ratio of Co-ZIF-9 in composite is 0.03) | UV-visible | H2O vapor | CO | 8.79 mmol | 10 | [54] | |
| CH4 | 0.99 mmol | ||||||
| H2 | 1.30 mmol | ||||||
| TiO2 | CO | 3.58 mmol | |||||
| CH4 | 0.60 mmol | ||||||
| H2 | 0.63 mmol | ||||||
| Co-ZIF-9 | CO | 0 | |||||
| CH4 | 0 | ||||||
| H2 | 0 | ||||||
| Physical mixture of TiO2 and Co-ZIF-9 with the mass ratio of 0.03:0.97 | CO | 3.86 mmol | |||||
| CH4 | 0.42 mmol | ||||||
| H2 | 0.56 mmol | ||||||
| Cu-BTC/TiO2 | UV | H2O vapor | CH4 | 2.64 mmol/(g TiO2·h) | 4 | [126] | |
| TiO2 | CH4 | 0.52 mmol/(g TiO2·h) | |||||
| H2 | 2.29 mmol/(g TiO2·h) | ||||||
| Cu-BTC | CH4 | 0 | |||||
| H2 | 0 | ||||||
| Cu-BTC/TiO2 (molar ratio of Cu-BTC to TiO2 is 3.33) | N/A | CO2/H2O vapor | CO | 256.38 mmol/(g TiO2·h) | 8 | [127] | |
| TiO2 | 11.48 mmol/(gTiO2·h) | ||||||
| Cu-BTC | 0 | ||||||
| CPO-27-Mg/TiO2 | UV | H2O vapor | CO | 40.9 mmol/g | 10 | [128] | |
| CH4 | 23.5 mmol/g | ||||||
| TiO2 | CO | 22.5 mmol/g | |||||
| CH4 | 13.7 mmol/g | ||||||
| CPO-27-Mg | CO | 0 | |||||
| CH4 | 0 | ||||||
| Physical mixture of TiO2 and CPO-27-Mg with the ratio of 6:4 | H2 | 8.5 mmol/g | |||||
| CO | 18.9 mmol/g | ||||||
| CH4 | 7.1 mmol/g | ||||||
| NH2-UiO-66/TiO2 (with 19%(wt) NH2-UiO-66) | UV-visible | CO2/H2 | CO | 3.74 mmol/(g·h) | [129] | ||
| NH2-UiO-66/TiO2 (19.5%(wt)NH2-UiO-66) | 4.24 mmol/(g·h) | ||||||
| NH2-UiO-66/TiO2 (24.5%(wt) NH2-UiO-66) | 3.37 mmol/(g·h) | ||||||
| NH2-UiO-66/TiO2 (36.8%(wt) NH2-UiO-66) | 2.85 mmol/(g·h) | ||||||
| TiO2 | 2.85 mmol/(g·h) | ||||||
| NH2-UiO-66 | 1.50 mmol/(g·h) | ||||||
| Co-ZIF-9/CdS | Visible | MeCN/H2O /TEOA/bpy | CO | 50.4 mmol | 1 5 | [130] | |
| H2 | 11.1 mmol | ||||||
| Co-MOF-74/CdS | CO | 39.6 mmol | |||||
| H2 | 7.7 mmol | ||||||
| Mn-MOF-74/CdS | CO | 1.0 mmol | |||||
| H2 | 2.0 mmol | ||||||
| Zn-ZIF-8/CdS | CO | 0.6 mmol | |||||
| H2 | 0.6 mmol | ||||||
| Zr-UiO-66-NH2/CdS | CO | 0.4 mmol | |||||
| H2 | 0.3 mmol | ||||||
| CdS | CO | 0.5 mmol | |||||
| H2 | 0.6 mmol | ||||||
| Co-ZIF-9 | CO | 0 | |||||
| H2 | 0 | ||||||
| UiO-66-NH2/Cd0.2Zn0.8S (10%(wt)UiO-66-NH2) | Visible | Na2S/Na2SO3 | H2 | 4591.6 mmol/(g·h) | [55] | ||
| CH3OH | 4.1 mmol/(g·h) | ||||||
| UiO-66-NH2/Cd0.2Zn0.8S (20%(wt)UiO-66-NH2) | H2 | 5846.5 mmol/(g·h) | |||||
| CH3OH | 6.8 mmol/(g·h) | ||||||
| UiO-66-NH2/Cd0.2Zn0.8S (30%(wt)UiO-66-NH2) | H2 | 5235.9 mmol/(g·h) | |||||
| CH3OH | 5.9 mmol/(g·h) | ||||||
| UiO-66-NH2/Cd0.2Zn0.8S (40%(wt)UiO-66-NH2) | H2 | 4922.7 mmol/(g·h) | |||||
| CH3OH | 5.3 mmol/(g·h) | ||||||
| Cd0.2Zn0.8S | H2 | 2804.2 mmol/(g·h) | |||||
| CH3OH | 2.0 mmol/(g·h) | ||||||
| UiO-66-NH2 | H2 | 0 | |||||
| CH3OH | 0 | ||||||
| Co-ZIF-9/mesoporous g-C3N4 | Visible | MeCN/H2O /TEOA/bpy | CO | 20.8 mmol | 2 | [131] | |
| H2 | 3.3 mmol | ||||||
| Co-ZIF-9 | CO | 0 | |||||
| H2 | 0 | ||||||
| g-C3N4 | CO | 0 | |||||
| H2 | 0 | ||||||
| ZIF-8/g-C3N4 nanotubes (molar ratio of g-C3N4 nanotubes to ZIF-8 is 10) | UV-visible | CO2/H2O vapor | CH3OH | 0.64 mmol/(g·h) | 1 | [56] | |
| ZIF-8/g-C3N4 nanotubes (molar ratio of g-C3N4 nanotubes to ZIF-8 is 8) | 0.75 mmol/(g·h) | ||||||
| ZIF-8/g-C3N4 nanotubes (molar ratio of g-C3N4 nanotubes to ZIF-8 is 5) | 0.45 mmol/(g·h) | ||||||
| ZIF-8/g-C3N4 nanotubes (molar ratio of g-C3N4 nanotubes to ZIF-8 is 2) | 0.31 mmol/(g·h) | ||||||
| ZIF-8/g-C3N4 nanotubes (molar ratio of g-C3N4 nanotubes to ZIF-8 is 1) | 0.16 mmol/(g·h) | ||||||
| g-C3N4 nanotubes | 0.49 mmol/(g·h) | ||||||
| Bulk g-C3N4 | 0.24 mmol/(g·h) | ||||||
| ZIF-8 nanocrystals | 0 | ||||||
| UiO-66/g-C3N4 nanosheets | Visible | MeCN/TEOA | CO | 9.9 mmol/(g g-C3N4·h) | 6 | [132] | |
| UiO-66/bulk g-C3N4 | 3.2 mmol/(g g-C3N4·h) | ||||||
| g-C3N4 nanosheets | 2.9 mmol/(g g-C3N4·h) | ||||||
| bulk g-C3N4 | 2.0 mmol/(g g-C3N4·h) | ||||||
| UiO-66 | 0 | ||||||
| BIF-20/g-C3N4 nanosheets (10%(wt) g-C3N4 nanosheets) | Visible | MeCN/TEOA | CO | 3.42 mmol | 6 | [133] | |
| CH4 | 1.12 mmol | ||||||
| BIF-20/g-C3N4 nanosheets (15%(wt) g-C3N4 nanosheets) | CO | 4.86 mmol | |||||
| CH4 | 1.45 mmol | ||||||
| BIF-20/g-C3N4 nanosheets (20%(wt) g-C3N4 nanosheets) | CO | 6.12 mmol | |||||
| CH4 | 1.76 mmol | ||||||
| BIF-20/g-C3N4 nanosheets (25%(wt) g-C3N4 nanosheets) | CO | 5.14 mmol | |||||
| CH4 | 1.51 mmol | ||||||
| ZIF-8/Zn2GeO4 (25%(wt) ZIF-8) | N/A | Na2SO3 | CH3OH | 0.22 mmol/(g·h) | 11 | [134] | |
| Metal incorporation | Pt/NH2-MIL-125(Ti) | Visible | MeCN/TEOA | HCOO– | 12.96 mmol | 8 | [135] |
| H2 | 235 mmol | ||||||
| Au/NH2-MIL-125(Ti) | HCOO– | 9.06 mmol | |||||
| H2 | 40.2 mmol | ||||||
| NH2-MIL-125(Ti) | HCOO– | 10.75 mmol | |||||
| H2 | 0 | ||||||
| 1%(wt) Co/NH2-MIL-125(Ti) | Visible | MeCN/TEOA | HCOO– | 384.2 mmol | 10 | [136] | |
| 2%(wt) Co/NH2-MIL-125(Ti) | 321.8 mmol | ||||||
| 3%(wt) Co/NH2-MIL-125(Ti) | 239.4 mmol | ||||||
| NH2-MIL-125(Ti) | 162.8 mmol | ||||||
| Ag⊂Re3-MOF (16 nm thick Re3-MOF) | Visible | MeCN/TEOA | CO | TON ≈ 2.8 | 48 | [137] | |
| ReI(CO)3(5,5′‐dcbpy)Cl | TON ≈ 1.7 | ||||||
| Carbon materials incorporation | 1%(wt) UiO-66-NH2/graphene | Visible | DMF/TEOA/H2O | HCOO– | 12.3 mmol | 4 | [138] |
| CH4 | 0.25 mmol | ||||||
| H2 | 15.2 mmol | ||||||
| 1.5%(wt) UiO-66-NH2/graphene | HCOO– | 21.2 mmol | |||||
| CH4 | 0.59 mmol | ||||||
| H2 | 13.9 mmol | ||||||
| 2%(wt) UiO-66-NH2/graphene | HCOO– | 33.5 mmol | |||||
| CH4 | 0.90 mmol | ||||||
| H2 | 13.2 mmol | ||||||
| 2.5%(wt) UiO-66-NH2/graphene | HCOO– | 14.9 mmol | |||||
| CH4 | 0.51 mmol | ||||||
| H2 | 15.1 mmol | ||||||
| 3%(wt) UiO-66-NH2/graphene | HCOO– | 8.6 mmol | |||||
| CH4 | 0.19 mmol | ||||||
| H2 | 16.8 mmol | ||||||
| UiO-66-NH2 | HCOO– | 3.1 mmol | |||||
| CH4 | 0.11 mmol | ||||||
| H2 | 16.9 mmol | ||||||
| Graphene | HCOO– | 0 | |||||
| CH4 | 0 | ||||||
| H2 | 0 | ||||||
| UiO-66-NH2/graphene (hydrothermal synthesis) | HCOO | 16.1 mmol | |||||
| H2 | 20.4 mmol | ||||||
| Al-PMOF/5%(wt) NH2-rGO | Visible | MeCN/TEOA | HCOO– | 685.6 mmol/(g·h) | 6 | [139] | |
| Al-PMOF/15%(wt) NH2-rGO | 479.8 mmol/(g·h) | ||||||
| Al-PMOF/25%(wt) NH2-rGO | 476.4 mmol/(g·h) | ||||||
| Al-PMOF | 165.3 mmol/(g·h) |
Notes: CAN–acetonitrile; BNAH–1-benzyl-1,4-dihydronicotinamide; bpy–2,2’-bipyridine; BTC–benzene-1,3,5-tricarboxylate; Cp*–pentamethylcyclopentadiene; DMF–N, N-dimethylformamide; MeCN–acetonitrile; phen–phenanthroline; TEA–trimethylamine; TEOA–triethanolamine; TON–total turnover number; 5,5′‐dcbpy–2,2’-bipyridine-5,5′-dicarboxylic acid; bpydb –4,4’-(2,2’-bipyridine-5,5′-diyl)dibenzoate; ppy–2-phenylpyridine; 4,4′‐dcbpy–2,2’-bipyridine-4,4’-dicarboxylate; dmb–4,4′-dimethyl-2,2′-bipyridine |
MOF composites incorporated with photosensitizers
As discussed in Section 2.2.3, metal-complex photosensitizers have been used to functionalize MOFs to enhance the photocatalytic activity. Photosensitizers can also be incorporated into MOFs to produce MOF composites for CO2 photoreduction.
Ru-based photosensitizer has been incorporated into several MOFs to enhance CO2 photocatalytic reduction [57–59]. Wang and coworkers [57] investigated the performance of a series of MOFs (Co-ZIF-9, Co-MOF-74, Mn-MOF-74, Zn-ZIF-8, Zr-UiO-66-NH2) in conjunction with [Ru(bpy)3]Cl2·6H2O toward CO2 photocatalytic reduction, where the MOF acts as a co-catalyst and [Ru(bpy)3]Cl2·6H2O acts as a photosensitizer. Photoreduction of CO2 was performed in MeCN and H2O solvent with TEOA as a sacrificial agent. After visible light irradiation for 0.5 h, 41.8 mmol of CO and 29.9 mmol of H2 were obtained over Co-ZIF-9 incorporated with [Ru(bpy)3]Cl2·6H2O, which outperformed the counterparts with a lower yield of CO and H2 under the same condition (as shown in Table 3). Co-ZIF-9 is a microporous cobalt-containing benzimidazolate MOF. The superior performance of Co-ZIF-9 with regard to CO2 photocatalytic reduction is a result of the synergetic effect of the imidazolate-based ligand which has a strong interaction with CO2 molecules for CO2 adsorption and cobalt with electron-mediating functions. Lately, the same group incorporated [Ru(bpy)3]Cl2·6H2O as a photosensitizer into Co-ZIF-67 acting as a co-catalyst [59]. Under the similar reaction condition, 37 mmol of CO and 13 mmol of H2 were produced over Co-ZIF-67/[Ru(bpy)3]Cl2·6H2O in 0.5 h under visible light irradiation. This was higher than that observed over Zn-ZIF-8/[Ru(bpy)3]Cl2·6H2O (1.8 mmol of CO and 2.0 mmol of H2 production in 0.5 h under the same condition), where the co-catalyst Zn-ZIF-8 has the same organic linker (2-methylimidazole) as Co-ZIF-67 but different metal site (Zn-based). This further confirmed the effect ofcobalton the photocatalytic reaction.
Cohen and coworkers [58] incorporated a photosensitizer [Ru(dmb)3]2+ (dmb= 4,4′-dimethyl-2,2′-bipyridine) into UiO-67 which was functionalized with a Mn+bipyridine complex, Mn(5,5′‐5,5′‐dcbpy)-(CO)3Br for CO2 photocatalytic reduction. In the mixture of N, N-dimethylformamide (DMF), TEOA and BNAH, HCOO- was produced with the TON of 110 under visible light irradiation for 18 h, which outperformed the UiO-67-Mn(5,5′‐dcbpy)-(CO)3Br without a photosensitizer [Ru(dmb)3]2+ and the homogeneous reference systems (Fig. 9). The superior photocatalytic performance is attributed to the isolated active sites in MOF framework, which endows the high stability of the catalyst and impedes the dimerization of the singly reduced Mn complex.
Fig.9 TON of HCOO– as a function of reaction time over UiO-67-Mn(bpy)(CO)3Br (red), Mn(bpy)(CO)3Br (green), Mn(bpydc)(CO)3Br (blue), UiO-67-bpydc (black), no added Mncomplex or MOF (only Ru2+ , brown), and UiO-67-Mn(bpy)(CO)3Br without added Ru2+ (gray) (Reproduced with permission from Ref. [58]. Copyright 2015, American Chemical Society) |
MOF composites incorporated with semiconductors
Coupling MOFs with semiconductors has been shown to be another approach to reducing the recombination rate of the photo-generated charge carries and consequently enhancing the photocatalytic performance. Several types of semiconductors such as TiO2, CdS and graphitic C3N4 have been embedded into MOFs to form heterogeneous structures which retain the properties of both the semiconductors and MOFs that are beneficial to CO2 photocatalytic reactions.
MOF-TiO2 composites
Several studies reported the development of ZIF-8/TiO2 composites for CO2 photoreduction. Zhang and coworkers [124] integrated ZIF-8 into the TiO2 film grid and conducted CO2 photocatalytic reactions in H2O as an electron donor without the use of a sacrificial agent under UV irradiation for 5 h. The amount of ZIF-8 in ZIF-8/TiO2 composite was varied by repeating the ZIF-8 growth step on the TiO2 film from one to three times, which are denoted as TiMOF-1, TiMOF-2, and TiMOF-3, respectively. TiMOF-3 with the highest amount of ZIF-8 had the highest CO2 adsorption uptake, followed by TiMOF-2, TiMOF-1, and TiO2. TiMOF-2 had the best performance with a CO yield of 0.53 mmol/(g·h) and a CH4 yield of 0.18 mmol/(g·h), which was 38% and 157% higher than that observed over pure TiO2 film under the same condition. TiMOF-2 outperformed TiMOF-3 which had the highest CO2 adsorption capacity due to the fact that the large amount of ZIF-8 in TiMOF-3 covered the TiO2 film, impeding the photo excitation process of TiO2. It is proposed that both TiO2 and ZIF-8 can be activated to generate photo-induced charge carries under UV irradiation, while TiO2 is more effective in the photoexcitation process than ZIF-8 because of its higher photoactivity and narrower band gap (-0.45 eV vs. -0.5 eV for ZIF-8). Electrons can also transfer from TiO2 to ZIF-8 and be involved in the photoreduction process over ZIF-8.
Cardoso and coworkers [125] developed a MOF-based Ti/TiO2 composite photocatalyst by growing ZIF-8 thin films on Ti/TiO2 nanotube (NT) electrodes using a layer-by-layer process. Spectroscopic and voltammetric assays revealed that the CO2 adsorbed on ZIF-8 formed stable carbamates. The photoelectrocatalytic reduction of CO2 over Ti/TiO2NT-ZIF-8 electrodes was performed in Na2SO4 (pH 4.5) saturated with CO2 at a constant potential of+0.1 V under UV-visible light irradiation. 10 mmol/L of ethanol and 0.7 mmol/L of methanol were produced in 3 h, which increased around 20 and 430 times, respectively, compared to the values observed over Ti/TiO2NT due to the low CO2 adsorption capacity of Ti4+ species in the absence of ZIF-8.
Co-ZIF-9/TiO2 composites with different mass ratios of Co-ZIF-9 (named as ZIFx/T; where x represents the mass ratios of Co-ZIF-9 in the composite, which equals 0.01, 0.03, 0.10, 0.20, 0.30, 0.40, and 0.60) were synthesized via an in situ synthetic method by Ye and coworkers for CO2 photoreduction [54]. Of the ZIFx/T composites, ZIF0.03/T had the best photocatalytic performance with a CO yield of 8.79 mmol, CH4 of 0.99 mmol and H2 of 1.30 mmol under UV-visible light irradiation for 10 h, which was higher than that of the pure TiO2 (3.58 mmol of CO, 0.60 mmol of CH4 and 0.63 mmol of H2) and Co-ZIF-9 (no CO, CH4 or H2 was detected). ZIF0.03/T also had a better photocatalytic performance than the physical mixture of TiO2 and Co-ZIF-9 at the same mass ratio of 0.03:0.97 (where 3.86 mmol of CO, 0.42 mmol of CH4, and 0.56 mmol of H2 were produced) due to the better charge separation. It was found that when the mass ratios of Co-ZIF-9 was higher than 0.1, the photocatalytic activity decreased with increasing the Co-ZIF-9 content in ZIFx/T, which might be attributed to the heavier charge recombination.
Pipelzadeh et al. [140] reported the CO2 photoreduction to CH4 and CO over a ZIF-8/TiO2 composite with core-shell structure in a photoreactor equipped with simulated sunlight at a constant pressure (CP, 5 bar) and an intentionally controlled pressure swing (PS) at 50 mL/min (PS-50) and 100 mL/min (PS-100). A high CO yield was achieved in the PS mode at a production rate of 13.2 mmol/(g·h) (PS-50, 80% increase than CP-50 operation) and 15.6 mmol/(g·h) (PS-100, 30% increase than CP-100). The PS mode had a better promotion effect on CO production than CH4. Continuous alteration of the reactants and product adsorption/desorption over the photocatalyst in the PS mode was beneficial to the regeneration of Ti3+ active sites, which contributed to the enhancement of photoreduction activity [17]. In addition, a higher gas flow rate facilitated CO removal to avoid catalyst poisoning. The calcination of ZIF-8/TiO2 composite at 300°C further increased the yield of CO with 45.16 mmol/(g·h) under PS-100 condition, which was higher than that observed in the CP-100 mode (33.46 mmol/(g·h)) because of the higher structural stability of the ZIF-8/TiO2 composite.
Composite photocatalysts composed of Cu-BTC (also named as HKUST-1, BTC= benzene-1,3,5-tricarboxylate) and TiO2 were also developed for CO2 photocatalytic reduction. Ye and coworkers [126] synthesized a Cu-BTC/TiO2 composite with a core-shell structure (see morphologies in Fig. 10) and evaluated the CO2 photocatalytic performance. Under UV irradiation for 4 h, the production rate of CH4 and H2 from CO2 over the bare TiO2 reached 0.52 and 2.29 mmol/(gTiO2·h), respectively, whereas the formation rate of CH4 over the Cu-BTC/TiO2 composite reached 2.64 mmol/(gTiO2·h) with no H2 detected. No product was produced over Cu-BTC, because its conjugated structure did not favor charge separation. This indicates that the CH4 yield and the selectivity of CH4 to H2 over Cu-BTC/TiO2 composite are significantly improved compared to the bare TiO2 and Cu-BTC in photocatalytic reduction. Under UV irradiation, the TiO2 in Cu-BTC/TiO2 composite is photoexcited to generate charge pairs and the electrons can be effectively transferred to Cu-BTC, as demonstrated by the ultrafast spectroscopy. This facilitates the charge separation in TiO2 and supplies active electrons to CO2 molecules adsorbed on Cu-BTC, leading to an enhancement in photocatalytic activity of Cu-BTC/TiO2 composite. Theoretical simulations demonstrate that the activation-energy barrier for CO2 on the Cu sites in Cu-BTC will be lowered upon receiving the photo-excited electrons from TiO2, enabling the CO2 reduction occurring on the Cu sites of Cu-BTC and the enhanced selectivity of CH4 to H2 production. Lately, Wang and coworkers [127] synthesized Cu-BTC/TiO2 composites in microdroplets via an aerosol route. Similarly, Cu-BTC/TiO2 composites had a higher photocatalytic performance than the pure TiO2 and Cu-BTC. The yield of CO from CO2 conversion over Cu-BTC/TiO2 composites increased as a function of the molar ratio of Cu-BTC to TiO2 ranging from 0 (i.e., bare TiO2) to 3.33 (see Fig. 11), where the production rate of CO over 3.33Cu-BTC/TiO2 increased to 256.35 mmol/(gTiO2·h) as compared to the bare TiO2 with a CO formation rate of 11.48 mmol/(gTiO2·h). As demonstrated by the in situ diffuse reflectance infrared Fourier transform spectrometer (DRIFTS) analysis, the enhancement in the photoreduction performance of the Cu-BTC/TiO2 composites may have been caused by the improved adsorption of reactants on the catalyst.
Fig.11 CO2 photoreduction analysis of TiO2 and HKUST-1/TiO2 composites (CO yield over the HKUST-1/TiO2 composites and pure TiO2) Inset image: CO yield peak time of the HKUST-1/TiO2 composites and pure TiO2 (Reproduced with permission from Ref. [127]. Copyright 2017, American Chemical Society) |
In addition to ZIF and Cu-based MOFs, TiO2 was also incorporated into other MOF structures to form MOF-TiO2 composites for CO2 photoreduction. Li and coworkers [128] combined TiO2 with CPO-27-Mg (also named as Mg2(DOBDC), DOBDC= 2,5-dioxido-1,4-benzenedicarboxylate) to produce a CPO-27-Mg/TiO2 composite via a hydrothermal self-assembly method. There exists a high concentration of open alkaline metal sites (Mg2+) in CPO-27-Mg structure, endowing the high CO2 adsorption capacity of CPO-27-Mg. Under UV irradiation for 10 h, 40.9 mmol/g of CO and 23.5 mmol/g of CH4 were produced over CPO-27-Mg/TiO2, which were higher than those observed over pure TiO2 (22.5 mmol/g of CO and 13.7 mmol/g of CH4). The enhanced performance of CPO-27-Mg/TiO2 composite on CO2 photoreduction is attributed to its high CO2 adsorption capacity and the existence of open Mg2+ metal sites. The CO2 photoreduction over a physical mixture of TiO2 and CPO-27-Mg (The ratio of TiO2 to CPO-27-Mg was 6:4.) was also performed under the similar condition, and 8.5 mmol/g of H2, 18.9 mmol/g of CO, and 7.1 mmol/g of CH4 were produced, which were lower than those produced over the CPO-27-Mg/TiO2 composite. This demonstrates the indispensable effect of the strong interaction between CPO-27-Mg and TiO2 in CPO-27-Mg/TiO2 composite for the CO2 photoreduction. TiO2 nanosheets were coupled with NH2-UiO-66 using an in situ growth strategy by Petit and coworkers [129]. The content of NH2-UiO-66 in the composite was varied from 19 wt% to 37 wt%. NH2-UiO-66/TiO2 had a better performance in photo reducing CO2 to CO than their single moiety. This improvement is a result of the enhanced abundance of long-lived charge carriers and high CO2 adsorption capacity in NH2-UiO-66/TiO2 composites.
MOF-CdS composites
CdS semiconductor has been widely used for CO2 photoreduction as a photocatalyst [141–143]. Wang et al. [130] incorporated CdS into Co-ZIF-9 and the as-obtained Co-ZIF-9/CdS composite had a good photocatalytic activity toward CO2 conversion to CO under visible light irradiation. CO2 photoreduction was performed in MeCN and H2O solvent with TEOA as a sacrificial agent and bipyridine (bpy) as an assistant for electron transfer. After visible light irradiation for 1 h, 50.4 mmol of CO and 11.1 mmol of H2 were produced over Co-ZIF-9/CdS composite, which outperformed its counterparts with lower yields of CO and H2 and pure CdS semiconductor (0.5 mmol of CO and 1.6 mmol of H2) under the same condition(as shown in Table 3). The photocatalytic mechanism was proposed as follows: under visible light irradiation, the CdS semiconductor was excited and charge carriers were generated. The photo-generated electrons transferred to Co-ZIF-9 and reduced the CO2 molecules adsorbed on Co-ZIF-9 to CO. Meanwhile the protons existed in the reaction system were also reduced to H2 by the excited electrons. Lately, Su et al. [55] prepared a series of UiO-66-NH2/Cd0.2Zn0.8S composites with different UiO-66-NH2 contents using a solvothermal method. CO2 photoreduction was performed over UiO-66-NH2/Cd0.2Zn0.8S composites under visible light irradiation, which all had an enhanced photocatalytic activity in comparison to their single components. The UiO-66-NH2/Cd0.2Zn0.8S composite with a UiO-66-NH2 content of 20 wt% had the best photocatalytic performance with a H2 production rate of 5846.5 mmol/(g·h) and a CH3OH production rate of 6.8 mmol/(g·h). The efficient charge separation and transfer between Cd0.2Zn0.8S and UiO-66-NH2 contributed to the enhanced photocatalytic activity of UiO-66-NH2/Cd0.2Zn0.8S composites.
MOF-graphitic C3N4 composites
Graphitic carbon nitrides (g-C3N4) with different morphologies and structures have been integrated with MOFs to improve the CO2 photoreduction activity. Wang and coworkers [131] coupled mesoporous g-C3N4 with Co-ZIF-9, which acted as a light harvester and co-catalyst, respectively, to fabricate a Co-ZIF-9/g-C3N4 composite. The Co-ZIF-9/g-C3N4 composite efficiently catalyzed CO2 to CO and H2 under visible light irradiation. 20.8 mmol of CO and 3.3 mmol of H2 were obtained over the Co-ZIF-9/g-C3N4 composite in 2 h, whereas no product was detected over the pristine Co-ZIF-9 and g-C3N4. Liu and coworkers [56] developed a series of ZIF-8/g-C3N4 composites by growing different contents of ZIF-8 nanoclusters on the surface of g-C3N4 nanotubes. The ZIF-8/g-C3N4 composites had an increased CO2 adsorption capacity than g-C3N4 nanotubes without sacrificing the light absorption capacity owing to the incorporation of ZIF-8 nanoclusters. Because of the high CO2 capture capacity of ZIF-8 and the promoted charge separation efficiency from the g-C3N4 nanotubes, the ZIF-8/g-C3N4 composites had an enhanced photocatalytic performance on CO2 reduction, where the highest production rate of methanol reached 0.75 mmol/(g·h) over the ZIF-8/g-C3N4 composite in which the mass ratio of g-C3N4 nanotubes to ZIF-8 was 8 under light irradiation for 1 h. Under similar conditions, g-C3N4 nanotubes and bulk g-C3N4 had a production rate of methanol of 0.49 and 0.24 mmol/(g·h), respectively, whereas no methanol was produced over the pure ZIF-8 nanocrystals. In addition, g-C3N4 nanosheets were combined with UiO-66 [132] and BIF-20 (a zeolite-like porous boron imidazolate framework) [133] MOFs to form MOF/g-C3N4composites, which all had an enhanced performance on CO2 photocatalytic reduction.
In addition to TiO2, CdS and C3N4, another type of semiconductor was also incorporated into MOFs to generate a composite photocatalyst for CO2 photocatalytic reduction. Wang and coworkers [134] developed a ZIF-8/Zn2GeO4 composite by growing ZIF-8 nanoparticles on Zn2GeO4 nanorods. The ZIF-8/Zn2GeO4 composite inherited both the high CO2 adsorption capacity of ZIF-8 nanoparticles and the high crystallinity of Zn2GeO4 nanorods. The ZIF-8/Zn2GeO4 composite with 25 wt% ZIF-8 had a CO2 adsorption capacity of 15.5 cm3/g, which was higher than the pure Zn2GeO4 nanorods (4.9 cm3/g) due to the high CO2 adsorption ability of ZIF-8. After 11 h of light irradiation in Na2SO3, the production of methanol at a rate of 0.22 mmol/(g·h) over the ZIF-8(25wt%)/Zn2GeO4 composite was observed. The yield of methanol over the ZIF-8(25 wt%)/Zn2GeO4 composite had a 62% increase in comparison to the pure Zn2GeO4 nanorods under light irradiation for 10 h. The enhanced CO2 photoreduction performance of ZIF-8/Zn2GeO4 composite may have been resulted from the high CO2 adsorption capacity of ZIF-8 and the higher light response.
MOF composites incorporated with metals
Because of their high Fermi energy levels, noble metal nanoparticles can effectively separate photo-generated charge pairs of photocatalysts [144,145]. Therefore, another approach for promoting the CO2 photocatalytic reduction is to dope noble metal nanoparticles into MOFs structures to decrease the recombination rate of the photo-generated electrons and holes. Li and coworkers [135] synthesized M-doped NH2-MIL-125(Ti) (M= Pt and Au) and conducted CO2 photocatalytic reaction in saturated CO2 with TEOA as a sacrificial agent under visible light irradiation. Both H2 and HCOO- were produced over M/NH2-MIL-125(Ti), while no H2 but only HCOO- was produced over bare NH2-MIL-125(Ti). The reason for this is that the electron-trapping effect of noble metals promotes the hydrogen evolution. In addition, it is noted that Pt and Au had different photocatalytic performances on the production of HCOO-. Compared to bare NH2-MIL-125(Ti) (yield of HCOO-: 10.75mmol), Pt/NH2-MIL-125(Ti) had a higher production yield of HCOO- (12.96 mmol), while Au/NH2-MIL-125(Ti) had a lower production of HCOO- (9.06 mmol) under visible light irradiation for 8 h. As demonstrated by ESR studies and DFT calculations, hydrogen spillover from Pt to the bridging oxygen linked to Ti atoms occurred in Pt/NH2-MIL-125(Ti), leading to the Ti3+ formation and boosting the hydrogen-assisted CO2 reduction to HCOO-. In contrast, it was difficult to achieve the hydrogen spillover from Au to the NH2-MIL-125(Ti) framework in Au/NH2-MIL-125(Ti), and thus resulting in a lower HCOO- formation over Au/NH2-MIL-125(Ti). Fu et al. [136] doped different contents of Co (from 1 to 3 wt%) into NH2-MIL-125(Ti) and produced Co/NH2-MIL-125(Ti) composites for CO2 photoreduction under visible light irradiation. 1 wt% Co/NH2-MIL-125(Ti) had the best performance on CO2 photoreduction with an HCOO- formation of 384.2 mmol in 10 h, which was 2-fold higher than that observed over pure NH2-MIL-125(Ti) (162.8 mmol) due to the enhanced visible-light harvesting and electron transfer stemmed from the addition of Co.
Yaghi and coworkers [137] functionalized UiO-67 with Re complexes, ReI(CO)3(5,5′‐dcbpy)Cl photosensitizer, of various densities (Ren-MOF, n = 0, 1, 2, 3, 5, 11, 16, and 24 complexes per unit cell) and found that the photocatalytic activity of the MOF system can be controlled by controlling the density of Re complexes in the framework, in which Re3-MOF had the best performance on CO2-to-CO conversion. The photocatalytic activity was further enhanced by coating a 16 nm layer of Re3-MOF onto Ag nanocubes (Ag⊂Re3-MOF, see Fig. 12), where a 7-fold improvement of CO2-to-CO reduction compared to pure Re3-MOF under visible light irradiation was achieved. Both Re complexes and Ag nanocubes contributed to the CO2 photocatalytic activity enhancement of Ag⊂Re3-MOF. Because the quadrupolar localized surface plasmon resonance (LSPR) scattering peak (lmax~ 480 nm) of Ag nanocube overlapped with the absorption range of ReI(CO)3(5,5′‐dcbpy)Cl (400 nm<l<550 nm) in the visible region [146,147] and Ag⊂Re3-MOF structure inherited the LSPR features of Ag cores, the photoactive Re metal sites within MOF shell were spatially localized into a strong electromagnetic field induced by LSPR of the Ag nanocubes for photocatalytic enhancement.
MOF composites incorporated with carbon materials
Still, another important strategy to develop active photocatalysts for CO2 photoreduction is the incorporation of carbon materials such as graphene and its derivatives into MOFs. Graphene can act as a photosensitizer to extend the light adsorption region from the UV to the visible light region. Moreover, the excellent conductivity of graphene can facilitate the rapid transfer of the photo-generated electrons and suppress the electron-hole recombination, and subsequently boost the photocatalytic activity.
Li and coworkers [138] synthesized UiO-66-NH2/graphene composites via microwave-assisted in situ growth of different amounts (1–3 wt%) of UiO-66-NH2 nanocrystals onto graphene. As compared to the pure UiO-66-NH2 (3.1 mmol of HCOO-, 0.11 mmol of CH4, and 16.9 mmol of H2 were produced.) and the UiO-66-NH2/graphene synthesized via a traditional hydrothermal route (16.1 mmol of HCOO- and 20.4mmol of H2 were produced.), the as-obtained UiO-66-NH2/graphene composites had a better performance on CO2 photoreduction activity and selectivity under visible light irradiation. Of the UiO-66-NH2/graphene composites as prepared, 2 wt% UiO-66-NH2/graphene composite had the best performance, where 33.5mol of HCOO-, 0.9 mmol of CH4 and 13.2 mmol of H2 were produced in 4 h. The performance enhancement is attributed to both the fine particle size and high dispersion of UiO-66-NH2 nanocrystals, which enables more light trapping to generate charge pairs and shortens the electron transfer pathway to facilitate electron transfer. Additionally, the strong UiO-66-NH2/ graphene interaction can also effectively accelerate electrons transfer efficiency to enhance photocatalytic activity for CO2 reduction.
Do and coworkers [139] incorporated different contents of amine-functionalized reduced graphene oxide (NH2-rGO) (5–25 wt%) into a TCPP-based MOF (Al/PMOF) to form Al-PMOF/NH2-rGO composites as photocatalysts for CO2 reduction. Al-PMOF/NH2-rGO composites had an enhanced photocatalytic activity for CO2 reduction, where the HCOO- formation rate reached 685.6 mmol /(g·h) over Al-PMOF/5 wt% NH2-rGO under visible light irradiation in 6 h, which was much higher than the HCOO- production rate observed over pure Al-PMOF (165.3 mmol /(g·h)). The photocatalytic mechanism is proposed as follows: upon visible light irradiation, TCPP is responsible for light harvesting and is excited to produce photo-generated electron-hole pairs, and the electrons are transferred from TCPP to graphene which acts as an electron acceptor. The electrons transferred from graphene reduce the adsorbed CO2 to HCOO- in the presence of TEOA that serves as a hydrogen source.
Most recently, MOF with a special microstructure in CO2 photoreduction was reported. Lin and coworkers [148] prepared Ni-based MOF (Ni2(OH)2BDC) monolayers (Ni MOLs) and examined the performance for photoreduction of CO2 under visible light irradiation with [Ru(bpy)3]Cl2⋅6H2O as a photosensitizer and TEOA as an electron donor. After 2 h reaction in pure CO2, Ni MOLs had a CO production rate of 12.5 mmol/h and a H2 production rate of 0.28 mmol/h, whereas bulk Ni MOFs had a lower CO production rate of 7.23 mmol/h. In diluted CO2, Ni MOLs had a CO selectivity of 96.8 %, which outperformed most of the reported systems in diluted CO2. It was proposed that the strong affinity of Ni MOLs to CO2 molecules enabled their high CO2 adsorption ability and stabilized the initial Ni‐CO2 adducts, thus promoting CO2-to-CO conversion. In addition, weak affinity of Ni MOLs to H2O impeded the transfer of protons, thereby reducing the H2 formation.
Conclusions and outlook
In recent years, MOFs have attracted great attention and showed much potential as photocatalysts for CO2 reduction because of their super-high surface areas, tunable structures, and high CO2 adsorption capacity. This review summarizes the recent research progress in the development of MOFs and MOF-based composite photocatalysts for the photocatalytic reduction of CO2. Several strategies in improving light harvesting, CO2 adsorption and charge separation have been discussed, which provides guidelines for rational design of MOF-based photocatalysts with enhanced performance on CO2 reduction under visible light irradiation. Although great progress has been made in the development of MOF-based photocatalysts for CO2 reduction, there still exist some challenges and large potential need to be explored. For instance, ① to date several thousands of MOFs with different structures have been developed, however, only several types of MOFs such as ZIFs, UiO- and MIL-based MOFs are under exploration as photocatalysts for CO2 reduction. More efforts need to be made in investigation of many other MOF systems with active catalytic centers, which may be promising for photocatalysis. ② Most MOFs have a poor stability in aqueous solution and suffer structural collapse, which limit their practical applications in catalysis processes where water is involved. Therefore, it is imperative to develop novel MOF photocatalysts which have an excellent chemical stability in aqueous solution and meanwhile retain the photocatalytic functionality. ③ The current research achievements in photocatalytic reduction of CO2 over MOF-based photocatalysts have not yet met the requirement for large-scale industrial applications, therefore, exploration in further improvement of CO2 photoreduction is required. So far, many novel materials such as two-dimensional (2D) nanomaterials like graphene and 2D semiconductors have been developed; integrating these novel nanomaterials into MOF structures may be considered as a promising strategy to enhance the photocatalytic performance of MOFs. It is believed that with the rapid development progress of MOF materials and other novel photocatalytic materials, MOF-based photocatalysts will have even greater potential to fulfill the requirements for practical applications in heterogeneous photocatalysis in the future.