Charge Photo-Accumulation and Photocatalytic Hydrogen Evolution Under Visible Light at an Iridium(III)-Photosensitized Polyoxotungstate

Abstract

Steady-state irradiation under visible light of a covalent Ir(III)-photosensitized polyoxotungstate is reported. In the presence of a sacrificial electron donor, the photolysis leads to the very efficient photoreduction of the polyoxometalate. Successive formation of the one-electron and two-electron reduced species, which are unambiguously identified by comparison with spectroelectrochemical measurements, is observed with a significantly faster rate reaction for the formation of the one-electron reduced species. The kinetics of the photoreduction, which are correlated to the reduction potentials of the polyoxometalate (POM), can be finely tuned by the presence of an acid. Indeed light-driven formation of the two-electron reduced POM is considerably facilitated in the presence of acetic acid. The system is also able to perform photocatalytic hydrogen production under visible light without significant loss of performance over more than 1 week of continuous photolysis and displays higher photocatalytic efficiency than the related multi-component system, outlining the decisive effect of the covalent bonding between the POM and the photosensitizer. This functional and modular system constitutes a promising step for the development of charge photoaccumulation devices and subsequent photoelectrocatalysts for artificial photosynthesis.

Photoconversion of light into chemical fuels is emerging as a major scientific challenge.^1-4 In the past decades, molecular approaches have mostly focused on one hand on the design of photosensitive systems displaying long-lived photo-induced charge separation states to permit further electron transfers^5-10 and, on the other hand, on catalysts able to use these photogenerated charges for achieving either oxygen^11-18 or hydrogen evolution.^19-24 As these two reactions are multi-electronic processes while photosensitizers deliver electrons and holes sequentially, the charges need to be directed to a charge accumulation site.²⁵ However, only a few molecular photoactive systems with a designed charge accumulation site have been described so far.^26-30 Another requirement is crucial for efficient charge accumulation in such systems: when partially filled, the reservoir should not interfere with the photoactive moiety. Indeed, in classical donor-acceptor (D-A) systems, the electron acceptor, once reduced, potentially becomes an electron donor and often displays light-absorbing properties. Thus it may act, in a subsequent light-driven process, as a deleterious quencher of the excited donor D* by reverse charge transfer (equation 1) or energy transfer (equation 2).

$${\mathrm{A}}^{-}\text{electron donor}:\text{electron transfer}{\mathrm{D}}^{\ast }-{\mathrm{A}}^{-}\to {\mathrm{D}}^{-}-\mathrm{A}$$ (1)

$${\mathrm{A}}^{-}\text{light absorber}:\text{energy transfer}{\mathrm{D}}^{\ast }-{\mathrm{A}}^{-}\to \mathrm{D}-{\mathrm{A}}^{\ast -}$$ (2)

Since they can store several electrons with modest difference in potentials between successive reduction steps,^{31, 32} polyoxometallates (POMs) are attractive candidates to further illustrate charge photoaccumulation. While electrocatalytic reduction of protons by POMs has been thoroughly investigated,^33-35 their activity in water photoreduction has been far less developed and only under UV-light excitation,^36-38 with two recent notable exceptions.^{39, 40} This is in striking contrast with recent but extensive studies on light-driven water oxidation mediated by transition metal substituted POMs. In this context, we recently reported the synthesis and photophysical properties of a series of heteroleptic carbocyclometalated iridium(III)-polyoxometalate conjugates^{41, 42} following mild reaction conditions developed earlier by some of us.^43-46 In such complexes, photoinduced charge-separated excited states of various lifetimes (ranging from nanoseconds to hundreds of nanoseconds) were observed by transient absorption spectroscopy. Most importantly, the functionalization of the heteroleptic cyclometalated iridium(III) on the picolinate ligand provides directionality to the photoinduced electron transfer by enhancing charge separation and delaying charge recombination, which is an asset for preventing reverse charge transfer from proceeding (equation 1). The kinetics of charge separation and charge recombination in these reported POM-[Ir] hybrids were correlated to the redox potential of the POM with the fastest electron transfer rates (both charge separation and charge recombination) observed for the POMs that are the most easily reduced. Among the different POM-[Ir] conjugates we selected the Dawson-type organosilyl hybrid [P₂W₁₇O₆₁{O(SiC₃₆H₂₃N₃O₂Ir)₂}]⁶⁻, named D_Si[Ir], since it offers the best compromise between efficient charge separation and long-lived charge-separated state. This hybrid, isolated as a tetrabutyl ammonium salt, contains two heteroleptic cyclometalated iridium(III) units connected to the mono-lacunary site of the Dawson-type α₂-[P₂W₁₇O₆₁]¹⁰⁻ through a Si-O-Si anchorage (Scheme 1). In this hybrid, the POM and the chromophore are poorly coupled electronically.

Scheme 1.

Molecular representation of the photoactive POM-based hybrid D_Si[Ir], reference POM D_Si[I] and reference iridium complex [Ir] described in this study. In the polyhedral representation, the WO₆ octahedra are depicted with oxygen atoms at the vertices and metal cations buried inside. Color code: WO₆ octahedra, blue; PO₄ tetrahedra, green.

As a consequence, in D_Si[Ir] the POM displays redox properties very similar to those of the hybrid precursor bearing two iodoaryl moieties, named D_Si[I].^{47, 48} In the previously mentioned study, transient absorptions measurements only allowed for the characterization of the first photo-induced electron transfer. Charge photo-accumulation studies can be achieved in the presence of an additional electron donor in the solution that can irreversibly quench the charge separation state, regenerate the initial state of the photosensitizer and make a second photo-induced process possible. We herein provide unprecedented evidence of charge accumulation on a polyoxotungstate by visible-light. Furthermore, this system is able to perform direct photocatalytic hydrogen evolution, albeit at low pace, without noticeably decreasing over more than 1 week of continuous photolysis.

The photoreduction reaction of D_Si[Ir] is easily monitored by electronic spectroscopy since, in their reduced forms, POMs display intense d-d and intervalence charge transfer transitions.^49-52 For reference, spectroelectrochemical measurements carried out on the related D_Si[I] are given in the Supporting Information: electrolysis was followed by rotating disk electrode voltamperometry and UV-Vis spectroscopy to provide the spectroscopic signature of the successively formed one- and two-electron reduced species. The evolution of the electronic spectrum of a degassed DMF solution of D_Si[Ir] (0.2 mM) under visible light irradiation (λ > 400 nm) and in the presence of triethylamine (NEt₃, 1 M) acting as sacrificial electron donor is shown in Figure 1. First, a large band with a maximum absorption at 840 nm, corresponding to the one-electron reduced POM (1e⁻-D_Si[Ir]), grows.

Figure 1.

Left: evolution of the differential visible absorption spectrum of a solution of D_Si[Ir] (0.2 mM) in DMF containing NEt₃ (1 M) during photoirradiation under visible light (green t = 130 sec., blue t = 1 hour). For clarity purpose, the blank corresponds to the absorption of the initially yellow solution before photolysis. Right: distribution of the different reduction states of D_Si[Ir] during the photolysis. Legend: non-reduced D_Si[Ir] (colorless circle), 1e⁻-D_Si[Ir] (green circle) and 2e⁻-D_Si[Ir] (blue circles).

Then, under prolonged photolysis, an additional absorption appears at 710 nm, attributed to the two-electron reduced POM (2e⁻-D_Si[Ir]). As luminescence of the iridium complex precursor [Ir] is not affected by the sacrificial donor under similar conditions, we assume that the active species responsible for the oxidation of the triethylamine is more likely the oxidized photosensitizer, in the charge-separated state, rather than the excited form of the photosensitizer itself (Scheme 2). The formation of 1e⁻-D_Si[Ir] is very fast. We measured a half-reaction time constant of τ_1/2 = 43 s with the setup shown in Figure S1 under continuous visible-light irradiation (400 nm < λ < 800 nm) and a quantum yield at 400 nm of ϕ_{400 nm} = 10.5±1%. The second reduction to 2e⁻-D_Si[Ir] is significantly slower (τ_1/2 = 270 s, ϕ_{400 nm} = 2.3±0.5%). This can be attributed to incomplete charge-separation in the already one-electron reduced species. Indeed, the second reduction process of D_Si[Ir] occurs at a potential 400 mV more negative than the first, slowing down the second photoinduced electron transfer process. However the two-electrons photoreduction observed is much more efficient than that of a previously reported electrostatic [Ru(bpy)₃]₂[S₂M₁₈O₆₂] system.⁵³ If the quantitative photoreduction of [S₂Mo₁₈O₆₂]⁴⁻ at 420 nm has been nicely described, the corresponding tungstate [S₂W₁₈O₆₂]⁴⁻ led to reduced species that could only be poorly characterized after at least one hour of photolysis. Concomitantly, a decrease of the chromophore absorption was observed, suggesting a significant degradation of the system. By introducing a covalent link between the photosensitizer and the polyoxotungstate, we strongly favour direct charge injection to the POM, and prevent the system from degradation.

Scheme 2.

Charge photo-accumulation occurring in the D_Si[Ir] dyad.

Multi-electron processes can be facilitated to a large extent in the presence of protons. In particular, the electrochemical behavior of POMs is drastically affected by the presence of acids.³¹ As shown in Figure 2, in the presence of acetic acid the waves of D_Si[I] are shifted to higher potential, thus confering to the POM improved reservoir properties.

Figure 2.

Evolution of cyclic voltammograms of a 10⁻³ M solution of the reference POM hybrid D_Si[I] in DMF containing 0.1 M TBAPF₆ upon addition of 500 equiv AcOH. Working electrode : glassy carbon; reference electrode, SCE.

We then investigated the photoreduction reaction of D_Si[Ir] in the presence of acetic acid (0.1 M). The study of the stability of the heteroleptic iridium(III) complex covalently linked to the POM showed that in DMF or MeCN solutions, the presence of strong acid (HCl, trifluoracetic acid) leads to an irreversible loss of picolinate ancillary ligands as recently described elsewhere.⁵⁴ However, we checked that, even in large excess (500 equiv) of acetic acid (AcOH), D_Si[Ir] is stable for few days (Figure S4). We note that AcOH (pKa =13.5 in DMF) is unable to protonate NEt₃ (pKa =9.2 in DMF) under these conditions⁵⁵ and thus does not significantly influence the apparent pH value of the reaction medium.⁵⁶ The presence of the acid barely modifies the initial evolution of the absorption spectra. The initial one-electron reduced POM, has an absorption maximum at 845 nm and the absorption of the two-electron reduced POM is centred at 695 nm. While the formation of 1e⁻-D_Si[Ir] is only slightly faster compared to the photoreduction in the absence of AcOH (τ_1/2 = 37 s), formation of 2e⁻-D_Si[Ir] is drastically accelerated (τ_1/2 = 60 s) in the presence of acid. Again, the photoreduction kinetics seems to be correlated to the redox properties of the POM. Indeed, the presence of AcOH affects more significantly the redox potential of the second reduction of the POM than the first one (see Figure 2; ΔE₁ = 60 mV, ΔE₂ = 280 mV). As a consequence, in the presence of acetic acid, the light-driven formation of the two-electron reduced POM is considerably facilitated, while the kinetics of formation of the one-electron reduced POM is almost unchanged. We have previously described how the structure-related redox potentials of the POM impacts the kinetics of the charge separation/recombination steps.⁴¹ Further tunability is thus addressable through the acid dependence of the reduction processes. This makes POMs unique for the design of finely controlled molecular electron reservoirs.

As POMs are known to catalyze proton reduction under certain conditions,^{31, 33, 34, 38} we thus wonder whether the POM moiety in D_Si[Ir] could act both as an electron reservoir and as a H₂-evolving catalyst under photolysis conditions. Gas chromatography monitoring of the photolysis of D_Si[Ir] in the presence of NEt₃ and acetic acid in DMF shows that H₂ continuously accumulates in the headspace of the reaction vessel. After an initial induction period, H₂ evolution was sustained for days with a turnover frequency (TOF) of ca. 0.25 h⁻¹. Up to 41 turnovers (calculated vs the POM) are formed within a 7-days experiment (Figure 3). No H₂ could be detected when NEt₃ was omitted or when D_Si[Ir] was replaced by the reference compound D_Si[I]. Only traces of H₂ were detected when a mononuclear carbocyclometalated picolinate iridium(III) complex, [Ir], was used instead of D_Si[Ir]. Interestingly, the turnover frequency for hydrogen evolution is twice lower for the multicomponent system consisting of D_Si[I] and 2 equiv of [Ir] under the same conditions, emphasizing again the beneficial effect of the covalent tether between the photosensitizer and the POM (See inset in Figure 3).

Figure 3.

Photochemical production of hydrogen from DMF solutions (10 mL) of NEt₃ (1 M) and AcOH (0.1 M) catalyzed by D_Si[Ir] (0.2 mM). Comparison with the multicomponent system consisting of D_Si[I] (0.2 mM) and 2 equiv. of [Ir] (0.4 mM) is shown in the inset (colorless dots). In these experiments, one turnover corresponds to 2 μmol (48 μL) H₂ evolved.

We have thus demonstrated that upon steady-state visible irradiation and in the presence of a sacrificial electron donor, the carbocyclometalated iridium(III)-polyoxometalate conjugate D_Si[Ir] is capable of photo-accumulating two electrons on the POM in a very efficient manner. This all-integrated system offers unique advantages for charge photo-accumulation since it fulfills the previously mentioned requisites.

(1) The iridium complex, because of its heteroleptic character, favors charge transfer to the POM and slows down backward electron transfer. This prevents the reduced POM from acting as a deleterious quencher of the excited iridium complex by reverse charge transfer (i.e. equation 1). (2) The absorbance of both one-electron and two-electron reduced POMs, although important in the lower energy part of the visible spectrum is modest to weak at ca. 510 nm (emission maximum wavelength of the iridium complex)⁵⁷ disfavoring energy transfer from the excited iridium complex to the reduced POM (i.e. equation 2). (3) The presence of acid leads to proton-assisted reduction of the POM, which provides the POM with improved reservoir properties and enhances the rate of photoproduction of the two-electron reduced POM. Finally, the system displays promising photocatalytic activity for hydrogen evolution. Further studies are required to establish the catalytic mechanism at work and eventually identify the catalytic active species formed in situ during the induction period.⁵⁸ While the catalytic rate is modest, the deceleration over seven days of continuous photolysis is very low, which establishes the robustness of the system upon turn-over. Indeed, monitoring of the ³¹P NMR signals of D_Si[Ir] under irradiation indicates that the POM framework of the hybrid is maintained (Figure S4). We are currently working on the improvement of photocatalytic activity for hydrogen evolution of the system. This implies optimization of the photocatalytic conditions (solvent, electron donor…), improvement of the robustness of the photoactive complex towards acid and design of multicomponent POM-[Ir]/electrocatalyst systems.