ReviewHybrid polyoxometalate materials for photo(electro-) chemical applications
Graphical abstract
Introduction
Design of effective materials for solar photo(electro)catalysis is important across diverse applications including energy generation, water purification, chemical synthesis and environmental waste remediation. Semiconductors such as TiO2, WO3 and ZnO have been the primary focus in the pursuit of catalytic materials for heterogeneous photocatalysis. However, the photoactive metal cluster complexes, the polyoxometalates (POMs), have been growing in significance in this field due to their synthetic versatility, well behaved redox properties including light driven redox chemistry, their low cost and low environmental impact. Polyoxometalates have been widely explored over many years across a breadth of applications ranging from catalysis through medicine and sensing [1], [2], [3], [4], [5], [6], [7]. This review focuses on the photochemical and photophysical properties of POMs, especially the Keggin and Dawson structures of Mo and W, and overviews some of the recent developments in the application of such POMs in heterogeneous photo(electro)catalysis. Assembly of POM building blocks into functioning devices has typically relied on electrostatic binding but has more recently been extended to covalent systems and the particular advantages of these two approaches is discussed, along with the photophysics of inter-species charge transfer in both systems. POM-based water oxidation catalysis using transition-metal substituted sandwich POM compounds (see Fig. 1(f)) is also briefly considered.
The polyoxometalates are transition metal oxide clusters of d0 or d1 metal ions bridged via oxygen atoms. Literally hundreds of POM complexes have been prepared, most commonly from Mo, W, V and Nb. In many structures, the cluster framework includes a heteropolyanion comprising for example phosphorous or sulfur, e.g. in the form of tetrahedral SO42− [8]. The structures of POMs are diverse, but the most common are the Keggin, Dawson, Lindqvist, Anderson, sandwich and Preyssler structures, which are illustrated in Fig. 1. The Keggin and Dawson structures have been most widely studied [9]. In addition, more recent covalent functionalization of traditional POM architectures has led to new classes of hybrid species [10], [11], [12].
The Dawson-type POMs are useful components in interfacial supramolecular devices due to their well-behaved redox chemistry. They are capable of undergoing multiple proton and electron transfer steps reversibly and without substantial decomposition under irradiation and/or electrochemical bias. In addition, both the Keggin and Dawson structures exhibit excellent chemical stability relative to some other POMs: for example the well-known and commercially available phosphotungstate K3[PW12O40] is stable towards thermal decomposition up to 580 °C [16].
The conventional Dawson structure; α-[S2Mo18O62]4− is shown in Fig. 2(c). This species is sufficiently stable under redox conditions that both the one and two-electron reduced products have been electro-synthesized by bulk electrolysis and isolated as a stable crystalline species [17], [18]. Interestingly, efforts to isolate a crystal structure of the two-electron reduced form yield a protonated form of the two-electron reduced material, i.e., the reduced forms tend to be very basic and easily protonated. The Dawson structures in their oxidized states do not typically absorb strongly in the visible spectral region. However, their reduced analogues usually have intense visible metal to metal charge transfer (MMCT) transitions and are thus deeply coloured.
The Dawson polyanion exhibits a horizontal mirror plane which separates two Mo9 clusters with the principle C3 axis of rotation connecting the two sulfur atoms. However, there is diversity even amongst conventional Dawson and Keggin structures. For example non-conventional sulfite-containing Dawson clusters α/β-[S2Mo18O60]4− (Fig. 2(a) and (b)) exhibit a structure which is distorted inward at the equator. This “peanut” configuration arises from the charge distribution of the encapsulated SO32− groups, which are not isostructural with true tetrahedral ions such as PO43− and SO42− [19]. The structural isomers vary in the orientation of the encapsulated sulfite moieties relative to each other as viewed down the principle axis of rotation. In the α case the sulfite groups are staggered and in the β case they are eclipsed. Since the staggered conformation is thermodynamically favoured, the α isomer is the more stable. This difference is considered negligible in the case of the fully oxidized POM ions, but in their reduced forms this energy difference becomes increasingly significant, and from electron paramagnetic resonance (EPR) spectroscopic studies the β → α isomerization occurs in solution over a period of 30 days, both under light and dark conditions (see Fig. 2).
Defective Dawson structures, known as lacunary POMs, can be prepared under basic conditions from their parent ion via the loss of an MO unit and the synthesis of unmodified parent POM structures is generally straightforward. The lacunary anion has been used as the basis for functionalization via covalent bond formation or co-ordination of metal cations to the Dawson structure [21], [22]. Higher polyoxo architectures self-assemble from simple starting materials such as Na2WO4 in acidic media and since almost all POMs are anionic (with the exception of polyoxotitanates, which are often charge neutral [23], [24]) isolation of the product is generally achieved by addition of an excess of cation such as K+ to yield water-soluble POMs or an organic ion such as [(But)4N]+ to yield POMs soluble in non-aqueous solvent. The anionic charge on the polyoxometalate provides a useful synthetic route to the preparation of electrostatic constructs by ion-pairing (see Section 2.1). Purification of polyoxometalates is usually achieved by recrystallisation or fractional recrystallisation; however the giant polyoxomolybdates “Mo-brown ball” (a Mo132 cluster [25]) and “Mo blue wheel” (a Mo154 cluster [26]) can be separated using gel electrophoresis, which opens up a new path to POM purification [27].
In their oxidized forms, Dawson and Keggin polyoxometalates exhibit oxygen pπ to metal dπ, ligand to metal charge transfer (LMCT) transitions with absorbance maxima typically in the near UV spectral region between 250 and 400 nm. The red-tail of these transitions can extend into the visible spectral region, although this is less true of the tungstates than the molybdate or vanadate complexes and, overall, visible absorbance is generally weak for fully oxidized POM complexes.
Exciting into the LMCT, the resulting CT excited state nominally comprises a reduced metal centre and cation radical at the oxygen donor. The reactivity of the latter is regarded to be the origin of the photocatalytic ability of the polyoxometalates. POM LMCT excited states can lead to photocatalytic oxidation of organic substrate through direct electron transfer or through electron coupled atom transfer leading to reduction of the POM. Frequently, the polyoxometalate reduction is a multi-electron process which may also be accompanied, in the presence of organic substrates with labile protons or sometimes water, by proton transfer to the polyoxometalate. Such photochemically induced oxidation of the substrate appears to occur principally via a pre-association of POM and substrate, typically, through H-bonding with water or a labile proton in an organic substrate [28]. Where water is present, either as the primary substrate or as a supporting solvent with an organic substrate, photoexcitation of the POM-H2O complex leads to H2O H-abstraction and formation of a hydroxyl radical [29], [30], [31]. The OH radical is a powerful oxidant capable of undergoing non-selective reactions with most organic substrates. For polyoxometalate-substrate systems which do not involve the participation of water, formation of a pre-associated equilibrium complex between catalyst and substrate via H-bonding can occur leading to direct electron and H transfer [32] or indeed Oxygen transfer [33].
As the substrate and POM pre-associate, this leads to ultrafast electron transfer, outside of the diffusion controlled rates anticipated for conventional bimolecular reactions. For example, Hill et al. reported photocatalytic oxidation of the polar organic substrates N-methylpyrrolidinone, 1,1,3,3-tetramethylurea and 1,3-dimethyl-2-imidazolidinone by hydrated Keggin POMs. The photoproduct signal evolved on the sub-nanosecond timescale, and similar results have been reported for other POM systems [32], [34]. POM oxidative photochemistry has been shown across several studies to conform to the Langmuir–Hinshelwood kinetics, [35], [36], [37]. This model indicates that at constant catalyst concentration and relatively low concentrations of substrate the photochemical reaction is first order, but as the substrate concentration approaches saturation the reaction becomes zero-order in the substrate. Such behaviour is consistent with a pre-association complex formed between catalyst and substrate.
Although, as described, direct photocatalytic oxidation of organic compounds such as alcohols proceed via a pre-association complex (i.e.: via electron transfer from the substrate) [34], hydroxyl radical generation has also been cited as a reaction mechanism in the presence of water and/or organics, i.e.:POM + H2O → POM (e−) + OH + H+
Evidence for this second mechanism is derived from the observation of several hydroxylated products and intermediates after the irradiation of p-chlorophenol in the presence of [W10O32]4−, and from spin-trapping experiments involving EPR spectroscopy [38], [39]. Nonetheless, photochemistry was also observed in dry, inert solvents to a large degree, indicating that the reaction can occur without radicals and may be attributed to electron transfer under these conditions. Reactions due to the presence of trace amounts of water could be ruled out as the concentration of water present was insufficient to support the photochemistry observed. It has been postulated that the dominant mechanism is electron transfer as in many cases specific products only are formed and hydroxyl radicals tend to be highly unselective [28], [40]; for example the selective oxidation of benzene to phenol was recently reported for the tungstovanadate [PW11VO40]4− under simulated solar irradiation [41].
The transient spectroscopy of POM excited states is, surprisingly, relatively unexplored. However, ultra-fast studies of the decatungstate anion [42], [W10O32]4−, following 355 nm laser excitation showed that the LMCT excited state persists for ∼30 ps [43], followed by relaxation and formation of longer-lived (∼5 ns) transient species which was capable of photo-oxidation of organic substrate [44]. More recently, Hill et al. synthesized a range of Dawson and Krebs-type POMs covalently bound to Re(I) or Mn(I) tricarbonyl ligands and observed Re(I)/Mn(I) to POM MMCT states which were measured to decay in <50 fs using pump-probe TR-FTIR and transient absorption measurements [45], [46]. These papers provide good evidence for direct photochemical electron transfer from chromophore to POM in covalently bonded systems. Regardless of whether intramolecular or intermolecular electron transfer is under study it is clear that POM excited state chemistry is both complex and very fast.
In general, POMs based on molybdenum or tungsten exhibit weak visible spectral absorbance and therefore the overlap with the solar spectrum is typically poor. Reduced forms of both molybdates and tungstates appear dark blue due to the appearance of MMCT bands at visible wavelengths; however excitation into these bands does not result in significantly increased visible/NIR photocatalytic activity. In the case of polycrystalline TiO2, a sensitizer can be successfully employed to extend absorption into the visible region and the field of dye-sensitized solar cells (DSSCs) is now quite advanced [47]. Indeed, the current generation of DSSCs is now well-established commercially with several companies competing for market share. In order for POM-based photochemical technologies to evolve, an analogous approach may be possible through the use of a sensitizer to enhance solar light harvesting efficiencies. Since the advancement of POM-sensitizer photochemistry is a full generation behind that of DSSCs, a greater understanding of POM-sensitizer photophysics is required.
An important challenge in applying POMs to photocatalysis lies in efficiently closing the photocatalytic cycle, i.e. the re-oxidation of the POM catalyst, as this step is typically rate-determining. In some polyoxotungstates the photocatalytic cycle has been closed using molecular oxygen as the re-oxidant, which is present in excess under ambient conditions. For example, Neumann and co-workers have studied in detail the re-oxidation of photoreduced [PV2Mo10O40]7− to [PV2Mo10O40]5− in acidic media by O2 [48]. However, oxidation of reduced forms of POM by molecular oxygen is typically either kinetically very slow or thermodynamically unfeasible for unsubstituted polyoxomolybdates [17], [49]. However, it is also possible to regenerate the oxidized POM by its reaction with protons to yield H2 in the absence of oxygen (see [17] for further details). Fig. 3 shows a POM photocatalytic cycle involving oxidation of an alcohol, RCH2OH, in which [Pn−RCH2OH]* is the pre-associated complex at which electron transfer occurs.
To optimize their application in photocatalysis, the challenge in the medium term is to develop POM-sensitizer systems with good visible spectral coverage, high stability under photocatalytic conditions and fast, efficient heterogeneous interfacial kinetics. A key way in which a sensitizer can be built into the POM structure is through formation of a charge-transfer salt. These can be formed between POMs and a cationic sensitizer species via a facile metathesis reaction, and the resulting salts tend to exhibit high formation constants (frequently 105–107, see Section 2.1). Synthesis of the stoichiometrically precise salts is often achieved, but, particularly in the case of metallopolymers (see Section 3.3), this has proven to be more challenging [51]. An advantage of this approach is that electrostatic assembly can be used to drive facile interfacial self-assembly of multilayer systems.
Section snippets
Photophysics of CT salts between POMs and Ru(II) polypyridyl complexes
Balzani et al. reported the first study into the interaction between POMs and Ru(II) polypyridyl complexes in 1987. They investigated the rates of photoinduced intramolecular electron-transfer from [Ru(bpy)3]2+ (where bpy = 2,2′-bipyridyl) to the polyoxotungstates [Mn(OH)PW11O39]6− and [Co(H2O)SiW11O39]6− and observed that changing the polyoxotungstate and Ru complex allowed them to tune the electron transfer properties of the system in solution. This study concentrated primarily on solution
Limitations of POM photocurrent generation
As POMs are increasingly incorporated into composites and supramolecular architectures, their application across a range of key photonic areas is growing. This review has focused in particular on photocatalysis as we believe this is a key future area for development using POM materials. There are several factors which severely limit photocurrent generation by POM films.
- (a)
Surface area: Unlike nanocrystalline semiconductor films, POM and hybrid POM LBL film photoanodes do not generally produce high
Acknowledgements
J. J. W., T. E. K. and R. J. F. gratefully acknowledge funding from Science Foundation Ireland under the SFI Research Frontiers Programme, Award No. 07/RFP/MASF386 and SFI TIDA programme, Award No. 14/TIDA/2400. AMB also expresses his gratitude to the Australian Research Council for financial support.
References (292)
- et al.
Coord. Chem. Rev.
(2013) - et al.
J. Electroanal. Chem.
(1997) - et al.
Polyhedron
(1996) - et al.
J. Photochem. Photobiol. A
(1996) - et al.
J. Mol. Catal. A: Chem.
(2011) Coord. Chem. Rev.
(1998)- et al.
J. Colloid Interface Sci.
(2014) - et al.
Electrochim. Acta
(1999) - et al.
Inorg. Chim. Acta
(2010) - et al.
Talanta
(2009)
Inorg. Chem. Commun.
J. Solid State Chem.
J. Colloid Interface Sci.
Inorg. Chem. Commun.
Inorg. Chim. Acta
Appl. Catal. A-Gen.
J. Solid State Chem.
Coord. Chem. Rev.
Chem. Commun.
J. Mol. Struct.
Colloid Surf. A
Eur. J. Inorg. Chem.
Polyoxometalate molecular science
NATO Sci. Ser.
J. Mater. Chem. A
Anti-Cancer Drug
Front. Biosci.
Inorg. Chem.
Chem. Soc. Rev.
Chem. Soc. Rev.
Inorg. Chem.
Angew. Chem. Int. Ed.
Dalton Trans.
Inorg. Chem.
J. Am. Chem. Soc.
Inorg. Chem.
J. Am. Chem. Soc.
Inorg. Chem.
Angew. Chem. Int. Ed.
Chem. Eur. J.
Inorg. Chem.
J. Am. Chem. Soc.
J. Chem. Soc. Dalton Trans.
Angew. Chem. Int. Ed.
Nature
Chem. Sci.
Chem. Soc. Rev.
J. Chem. Soc. Dalton Trans.
Eur. J. Inorg. Chem.
J. Am. Chem. Soc.
Cited by (323)
Engineering heteropolyblue hole transfer layer for efficient photoelectrochemical water splitting of BiVO<inf>4</inf> photoanodes
2024, Applied Catalysis B: EnvironmentalRecent advances of polyoxometalate-based materials applied for electron-related devices
2024, Coordination Chemistry ReviewsA comprehensive review on advances in polyoxometalate based materials for electrochemical water splitting
2024, Coordination Chemistry ReviewsConstruction of Co/Ni-modified P<inf>4</inf>Mo<inf>6</inf> compounds for photocatalytic CO<inf>2</inf> conversion
2024, Journal of Molecular Structure
- 1
Current address: Department of Chemistry, Stephenson Institute for Renewable Energy, The University of Liverpool, L69 7ZF, UK.