Insights into the mechanism of photocatalytic water reduction by DFT-supported in situ EPR/Raman spectroscopy.
D. Hollmann,F. Gärtner,R. Ludwig,Enrico Barsch,H. Junge,Matthias Blug,S. Hoch,M. Beller,A. Brückner
DOI: https://doi.org/10.1002/anie.201103710
2011-10-17
Angewandte Chemie
Abstract:Considering the foreseeable shortage of fossil resources and global warming, the development of sustainable-energy technologies is of vital interest. An attractive option for the production of more benign energy vectors is the generation of hydrogen by photocatalytic water reduction. This concept facilitates the transformation of sunlight as the ultimate energy source into transportable energy carriers such as hydrogen. Hence, significant efforts are currently being undertaken to increase the activity and stability of suitable water-splitting catalysts. 3] The overall process can be divided into the two half reactions: water oxidation and water reduction. Studying these half reactions in detail, in particular the formation, operation, and decomposition of the catalyst, provides essential information for the development of new more efficient and environmentally benign catalysts. Recently, the Beller group disclosed an efficient water-reduction catalyst system consisting of [Ir(ppy)2(bpy)]PF6 (ppy = 2-phenylpyridine, bpy = 2,2’-bipyridine) as photosenzitizer (IrPS), [Fe3(CO)12] as water-reduction catalyst (WRC), and triethylamine (TEA) as sacrificial reductant (SR; Scheme 1). It is supposed that the catalytic cycle starts by photoexcitation of IrPS and charge separation, and subsequent reduction of its excited state by TEA (SR, cycle I). From the reduced state IrPS an electron is transferred to the WRC, which subsequently reduces aqueous protons to H2 (cycle II). To date, the only intermediate that has been experimentally identified by in situ IR spectroscopy in the water-reduction cascade (Scheme 1) is the anion [HFe3(CO)11] , which is considered to be the catalytically active species. However, the preceding steps leading to its formation as well as pathways responsible for the observed deactivation with time are still not known. Thus, more comprehensive in situ studies using additional methods are highly desired. It is probable that the one-electron-transfer processes in the catalytic cycles I and II (Scheme 1) lead to paramagnetic radical intermediates. Such species are accessible by EPR spectroscopy, while the diamagnetic [HFe3(CO)11] anion is EPR-silent but can be observed by vibrational in situ spectroscopic methods. To gain a more detailed insight into catalytic cycles I and II and to identify possible deactivation processes, we have monitored the reaction simultaneously by in situ EPR/Raman spectroscopy. To the best of our knowledge, photocatalytic water-splitting reactions have never been studied by these coupled techniques. The interpretation of our experimental data is supported by DFT calculations and additional in situ IR studies. First, catalytic cycle I was investigated. As expected, the IrPS complex (low-spin d, diamagnetic) showed no EPR signal in a solution containing THF/TEA/H2O (8:2:1) in the absence of [Fe3(CO)12] without light irradiation. However, if this solution is irradiated at 300 K, an intense isotropic signal at g = 1.9840 is observed (Figure 1). This signal corresponds to the reduced form of the iridium photosensitizer (IrPS ), which is formed by reductive quenching of the excited state (IrPS*) by TEA. A similar signal was formed neither in pure THF nor in THF/H2O, suggesting that 1) TEA is needed as a reducing agent and 2) excitation by light is essential to initiate the electron transfer. However, it must also be mentioned that the signal rapidly declines with time, probably because of ligand dissociation from IrPS (for additional information see Figure SI1 in the Supporting Information). In a reaction mixture containing all the necessary components of the waterreduction system (THF, H2O, TEA, IrPS, and Fe-WRC), no Scheme 1. General principle of H2 formation through the photocatalytic water-reduction cascade.