Multifunctional Polyoxometalate Platforms for Supramolecular Light‐Driven Hydrogen Evolution

Abstract Multifunctional supramolecular systems are a central research topic in light‐driven solar energy conversion. Here, we report a polyoxometalate (POM)‐based supramolecular dyad, where two platinum‐complex hydrogen evolution catalysts are covalently anchored to an Anderson polyoxomolybdate anion. Supramolecular electrostatic coupling of the system to an iridium photosensitizer enables visible light‐driven hydrogen evolution. Combined theory and experiment demonstrate the multifunctionality of the POM, which acts as photosensitizer/catalyst‐binding‐site[1] and facilitates light‐induced charge‐transfer and catalytic turnover. Chemical modification of the Pt‐catalyst site leads to increased hydrogen evolution reactivity. Mechanistic studies shed light on the role of the individual components and provide a molecular understanding of the interactions which govern stability and reactivity. The system could serve as a blueprint for multifunctional polyoxometalates in energy conversion and storage.


Synthesis of POM-PtCl: (nBu4N)3[MnMo6O18{(OCH2)3CNCH(C11H9N2)PtCl2}2] x 5 DMF.
(nBu4N)3[MnMo6O18{(OCH2)3CNCH(C11H9N2)}2] (100 mg, 0.0446 mmol, 1 eq) and [Pt(DMSO)2Cl2] (37.66 mg, 0.0892 mmol, 2 eq) were dissolved in water-free, de-aerated acetonitrile (10 mL). The orange solution was refluxed under argon atmosphere for 3-4 h. After cooling to room temperature, a slightly turbid solution was obtained which was filtered off and the filtrate was evaporated under vacuum to obtain a yellow precipitate. Orange crystals are formed upon slow diffusion of ethyl acetate into DMF solution. The crystals were collected by filtration and washed with diethyl ether. Yield: 70.2% based on Pt.  (10 mL). The orange solution was refluxed under argon atmosphere for 3-4h. After cooling to room temperature, a slightly turbid solution was obtained which was filtered off and the filtrate was evaporated under vacuum to obtain a yellow precipitate. Orange

Characterization
The FT-IR spectra for both POM-PtCl and POM-PtI exhibit the characteristic bands of Anderson type POMs, i.e. bands between 890-950 cm -1 (terminal Mo=O vibrations) and bands between 660-710 cm -1 , which correspond to bridging Mo-O-Mo vibrations ( Figure S1). [6] Figure S1. FT-IR spectra of POM-PtI (top) and POM-PtCl (bottom) showing the two characteristic bands for Anderson POM as indicated in the graphs. [6] 1 H-NMR showed the aromatic protons broadening of the bands caused by the paramagnetic Manganese center, but the chemical shift and the integrations for the aromatic protons and the aliphatic ones for both POM-PtCl and POM-PtI correspond to two bipyridine units and 3 nBu4N cations. [6] Elemental analysis confirmed the bulk composition and presence of the elements indicated. Further structural characterization was provided by singlecrystal X-ray diffraction (see below).

Single-crystal X-ray diffraction (scXRD) analysis of POM-PtCl and POM-PtI
Suitable single-crystals of POM-PtCl or POM-PtI were mounted onto a microloop using Fomblin oil. X-ray diffraction intensity data were measured at 150 K on a Bruker D8 QUEST diffractometer (λ(MoKα) = 0.71073 Å) equipped with a graphite monochromator. Structure solution was carried out using SHELX-2013 [7] package through OLEX2 [8] Corrections for incident and diffracted beam absorption effects were applied using empirical methods. [9] Structures were solved by a combination of direct methods and difference Fourier syntheses and refined against F 2 by the full matrix least-squares technique. Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were added to carbon atoms using a riding model. Restraints were used to model the disordered nBu4N + cations. No restraints were used on the actual POM-anions. Corresponding CIF files can be obtained free of charge from the CCDC, reference numbers 2099459 (POM-PtCl) and 2099460 (POM-PtI).    [5,[10][11][12] Spectro-electrochemistry. Insights into the redox activity of POM-PtI were obtained from spectroelectrochemical UV-Vis absorption measurements in water-free, de-aerated DMF containing 0.1 M nBu4NBF4. Under reductive conditions (cyclic voltammetry, scan rate: 100 mVs -1 ) two redox waves between -1.0 and -1.8 V vs. Fc 0/+ are observed. The UV-Vis absorption spectra monitored at the first and second reduction process, i.e., at reduction potentials of -1.0 and -1.8 V vs. Fc 0/+ . The respective spectra show slightly increased molar absorptivity between 250 and 600 nm compared to the spectrum collected without applying a potential (opencircuit-potential, 0.0 V). The difference spectrum at -1.8 V vs. Fc + /Fc shows three major peaks at ca. 300, 390 and 510 nm, which are tentatively associated with a Pt II/I reduction, causing an increase of electron density in the bpy sphere. Due to the minor spectral changes in the investigated spectral window, the first single-electron reduction processes can be associated with the POM (i.e., Mn III/II ).

Light-driven H2 evolution catalysis
General procedure for visible-light-driven catalytic experiments. In a typical experiment, a Schlenk tube (21 cm 3 ) was filled with the photosensitizer (125 µM) catalyst (12.5 µM for POM-catalysts, 25 µM for the reference catalysts), water-free, de-aerated DMF (5.7 ml), electron donor (triethylamine, 1.12 mL, 1.0 M) and a proton source (acetic acid, 0.092 ml, 0.2 M). The Schlenk tubes were irradiated with a LED light-source ((λmax = 470 nm, P ~ 40 mW cm -2 )). After irradiation, head-space gas samples (100 μL) were successively taken using a gas-tight syringe and analyzed by headspace gas chromatography. All catalytic experiments were performed in triplicate, reported values are averaged over the three runs. All catalysis experiments were performed under inert atmosphere. All solvents used for the catalysis experiments were de-gassed thoroughly by bubbling argon through the solution. To study the changes in the emission lifetime of PS upon stepwise addition of TEA or POM-PtI, the emission kinetics of PS (in aerated DMF) were monitored at 420 nm excitation between 540 and 680 nm in 20 nm steps (see Figure S6, left). The respective emission lifetimes were obtained from a global single-exponential fit of the data collected at eight emission wavelengths. The quenching rate constants were calculated using the Stern-Volmer relation,  . The corresponding quenching rate constants are 3.5 × 10 10 M -1 s -1 (POM-PtI) and 5.9 × 10 7 M -1 s -1 (TEA), respectively. Note that a non-linear (i.e. quadratic) fit of the curves did not lead to meaningful values (due to negative curvature), hence a linear approximation has been chosen for quenching rate constant determination.

Determination of diffusion rate constants.
According to the static quenching Collins-Kimball diffusion model a reaction occurs when the sacrificial agent reaches the immediate vicinity of the photoexcited PS. [14,15] The region around the PS is generally treated to be spherical with the radius σ obtained from the volume of the PS. Within the Collins-Kimball model the radius of the reaction volume R is equal to σ. Within this model the timeindependent diffusion rate constants for the Ir-complex [Ir(ppy)2(bpy)] + (PS) and triethylamine (TEA) or POM-PtI were calculated from = 4 · · · ( + ) 2 6 · · · , using the Avogadro-constant (NA), Boltzmann-constant (kB), temperature (T = 298 K), the viscosity of DMF (η = 9.2×10 -4 Pa·s) and the gyration radii of the quencher molecules RQ (RTEA =2.5×10 -10 m, RPOM-PtI = 8.7×10 -10 m) and the Ir-photosensitizer (RPS = 4.2×10 -10 m). The gyration radii of the Ir-complex and TEA are taken from literature, [16] whereas the gyration radii of POM-PtI was estimated from the molecular volume (1681.98 cm 3 mol -1 ) obtained from DFT calculations and using van-der-Waals radii from literature. [17] The resulting diffusion rate constants for PS/TEA and PS/POM-PtI in DMF are 7.7 × 10 9 M -1 s -1 and 8.2 × 10 9 M -1 s -1 , respectively.

Interactions between catalyst and photosensitizer
To gain insights into the interactions between catalyst and photosensitizer, the interaction energy of PS with the catalyst was evaluated by DFT calculations. To assess the stability of different binding sites, the interaction of the PS with isolated model systems was determined first. These model systems represent separate components of the POM-PtX catalysts ( Figure S8) and serve as a reference for the interaction at the individual regions of  Table S2. Table S3. Interaction energies of reference and catalyst systems with the photosensitizer. The interaction energies are given in kJ/mol .

Charge analysis and distribution of electron density
To enable an estimation of how additional negative charge can be distributed over the catalysts, we performed single-point calculations with an additional electron for the reference and catalyst systems. The Hirshfeld net charges of the additionally negatively charged systems was then set in reference to the molecule's initial density. This procedure is analogous to the determination of condensed Fukui functions. While condensed Fukui functions are normally used as descriptors for the chemical reactivity of a molecule (with respect to its electrophilic and nucleophilic regions), these functions will in this case, however, serve as a measure for the uptake of an additional electron. Here we use the condensed Fukui function  + for the addition of an electron: In this case, ( ) describes the atomic charges at the respective atomic center of the original system and ( + 1) those of the system with an additional electron. In the following, the normalized values for  + are given in percent and serve as a measure of the portion of the additional electron, which a certain region of the molecule has accepted. The corresponding values for reference and catalytic systems are shown in Table S3.
As shown in Table S3,    UV-Vis studies on the catalyst stability. The stability of molecular catalysts under catalytic conditions is a general concern. In this context, the (photo)stability of POM-PtI in our system has been examined using multiple physicochemical methods. We first tested the stability of the catalyst in DMF solvent by measuring the UV-Vis spectrum of the POM-PtI at different time intervals. The measurements showed minor changes in the 400 -450 nm region, which might indicate slow degradation or structural changes of the catalyst under prolonged irradiation. The exact details are still under investigation. Micro-filtration analysis. To exclude electrostatic aggregation and colloid formation between [Ir(ppy)2(bpy)]PF6/ POM-PtI (3-) in our reaction system, a literature-known colloid detection procedure using micro-filtration was performed: [18] A DMF solution containing [Ir(ppy)2(bpy)]PF6 and POM-PtI (molar ratio: 10/1) was prepared. The UV-Vis spectrum of the solution was recorded, the solution was filtered through a 0.2 µm pore size PTFE syringe filter and the UV-Vis spectrum was recorded again. UV-Vis spectroscopic analysis of the solution before and after filtration indicates no significant changes ( Figure S12), suggesting that no significant number of particles have been removed from solution. Figure S12. UV-Vis spectroscopic analysis of the reaction solution before and after micro-filtration. No changes are observed after filtration.
In-situ exchange of Cl-with I-ligands. The formation of {PtI2} species starting from {PtCl2} in the presence of excess iodide is known from literature as a means to access more active HER catalysts. [19,20] Here, we used this concept to verify whether this reaction is still possible for the new POM-PtCl cluster. To this end, the standard light-driven HER catalysis was prepared using POM-PtCl as catalyst, and an excess of nBu4NI (1000 fold with respect to the catalyst) was added. The standard HER catalysis was performed, and we observed a TON increase of ca. 20 %, indicating (at least partial) in-situ formation of POM-PtI. Further insights into this ligand exchange were obtained by UV-Vis spectroscopy: a DMF solution of POM-PtCl (10 -5 M) was prepared and an excess of nBu4NI (0.5 M) was added. UV-Vis spectroscopy of the solution after stirring for 48 h clearly demonstrated the conversion of POM-PtCl to the more active POM-PtI species, see below.