Computational Modelling of Polyoxometalates: Progress on Accurate NMR Characterization and Reactivity
- Autores: Magdalena Pascual Borràs
- Directores de la Tesis: Xavier López Fernández (dir. tes.), Antonio Rodríguez Fortea (dir. tes.)
- Lectura: En la Universitat Rovira i Virgili ( España ) en 2017
- Idioma: español
- Tribunal Calificador de la Tesis: Jorge Juan Carbó Martín (presid.), Xavier Aparicio Anglès (secret.), Israel-Martyr Mbomekalle (voc.)
- Programa de doctorado: Programa de Doctorado en Ciencia y Tecnología Química por la Universidad Rovira i Virgili
- Materias:
- Enlaces
- Tesis en acceso abierto en: hdl.handle.net TDX
- Resumen
This thesis can be divided in two different parts: i) NMR and ii) Reactivity. The common factor of both sections is Polyoxometalates (POMs). POMs are polymetallic molecular anions formed upon aggregation of octahedral MO6 blocks made of early transition metals (TMs), in their higher oxidation states, and oxygen, that are connected commonly by vertices and edges, forming characteristic structures such as those of the Lindqvist, Keggin or Wells-Dawson anions, which follow the names of the first authors who reported them. The principal features of these POMs are that they are strong oxidizing agents, very acidic and really versatile compounds.
The first part of this thesis is related to the characterization of POMs. In particular, it is based on the calculation of nuclear magnetic resonance (NMR) spectroscopy parameters. NMR has been used for many years and it has become one of the most popular chemical characterization techniques. Despite the fact that oxygen is one of the most important elements, other nuclei such as 1H, 31P, 13C or even 183W have been prevalent in NMR studies of POMs.
First, we focused on 17O nuclei due to their presence in all the POMs structures. We report a theoretical DFT analysis on 17O NMR chemical shifts for a family of prototypical POMs anions. We first made an extensive exploration of density functionals and basis sets to establish a good strategy to determine accurate values. In addition, the diversity of POMs studied provides a unique opportunity to compare chemical shifts (δ) values of nuclei bonded to different transition metals, allowing us to clearly identify and understand factors contributing to these values. Therefore, we have established a DFT-based strategy to accurately compute and rationalize 17O NMR chemical shifts of polyoxoanions. For a set of 75 signals, we show that calculations performed with GGA-type PBE functional, including spin-orbit and scaling corrections, provide a mean absolute error value (MAE) <30 ppm, a small value considering that the range of δ(17O) chemical shifts in these systems is around 1200 ppm (2.5%). We explored the paramagnetic contribution to the shielding, which dominates the 17O chemical shift. It depends on the transition from occupied orbitals to unoccupied ones, which has higher contributions in oxygen atoms. We have shown that the electronegativity of the metal linked to the oxygen is a key factor in order to understand the chemical shift values. The paramagnetic shielding correlates with the M-O distance and also has a linear dependence with the reciprocal of the energy gap of the principal transition, which governs that shielding. Finally, we explored the effect of the protonation on δ(17O) and demonstrated that 17O NMR can be a powerful tool to identify the site(s) of protonation. DFT helps understanding how important protonation is and how it must be properly taken into account to reproduce and correctly interpret δ(17O) in oxometalates. When a molecule has many similar or equivalent oxygen sites that can accept an itinerant proton, the change in δ(17O) signals is almost undetectable. On the other hand, when a molecule has very few distinctively basic sites, the proton gets trapped in one or a few sites and the change in δ(17O) signals is more evident.
NMR of other nuclei which can be part of POMs such as 31P, 29Si, 79Ga and 73Ga offers the possibility to thoroughly study their structure and bonding. One of the most active nuclei used is 31P, which provides straightforward structural information. In the case of POMs, the range of 31P NMR is small (10 ppm), which implies a difficult assignment, but it is considered a powerful technique in structural characterization and monitoring of chemical reactions in the field of POMs. In chapter 4, we present a search for the best density functional strategy for the determination of 31P NMR chemical shifts in POMs. The main computational parameters affecting the quality of such properties are the density functional and the basis set size, as well as the spin-orbit and solvent effects. The influence of the first two parameters on the quality of the 31P NMR chemical shifts was investigated on a large family of compounds based on [XM12O40]n- and [X2M18O62]n- frameworks. The work concludes that using a TZP/PBE for the NMR calculation step and a TZ2P/OPBE for geometry optimization is the best procedure for the accurate determination of 31P NMR chemical shifts. The results obtained with the mentioned methodology presented a MAE of 2.64 ppm. The main variations in δ(31P) come from the paramagnetic contribution to the shielding, which is directly related to occupied-virtual orbital transitions with phosphorous contribution. Thus, the variations in δ(31P) are linked to the energy gap of these transitions.
The work presented in chapter 5 describes the first thallium containing polyoxometalate, which was synthetized and structurally determined. This project has been done in collaboration with two research groups: Dr. Kortz group and Dr. Tóth group. Our role in this project was to study the stability of the polyanion [Tl2{B-β-SiW8O30(OH)}2]12-. Moreover, a detailed study of 203/205Tl NMR has been performed in order to reproduce the experimental values. The optimized geometry reproduces correctly the X-ray geometry. We studied the most likely positions of the two protons of the molecule in solution. The results showed that there are relatively low energy differences obtained for different di-protonated anions, implying the possibility of mobile protons in solution. In solid state, the protons are linked to OC due to the inaccessibility to the other oxygen types, which are blocked by countercations. The δ(205Tl) are really well reproduced with DFT methodology. However, in the spin-spin coupling, there are large differences between calculated and experimental 2J(205Tl-203Tl). 2J(205Tl-203Tl) depends so much on the protonation and we would need to consider all the species present in solution in order to accurate reproduce 2J(205Tl-203Tl). We have also searched for the best methodology to reproduce 2J(205Tl-203Tl). All the results obtained so far show large deviations for the 2J(205Tl-203Tl) compared to experiment and we are still working on it.
The second part of this thesis includes studies of POMs reactivity. The first project has been done in collaboration with Prof. Anna Proust and Dr. Guillaume Izzet. It is well known that many POMs have the ability to capture electrons with minor structural changes, and then in some media these electrons can be released in presence of an acid generating new molecules. The efficient solar light-driven photochemical generation of fuels such as H2 is a highly interesting topic and still a scientific challenge after decades of research. Matt et. al. (Energy Environ. Sci. 2013, 6, 1504) designed a covalent Ir(III)-photosensitized polyoxotungstate, capable of efficiently evolving H2 in presence of an excess of acid and a sacrificial electron donor. The stepwise mechanism of visible light harvesting, formation of charge-separated species and the proton reduction has been analysed by DFT and time-dependent-DFT (TD-DFT). We characterise the POM-based photosensitised reactant and intermediate species formed during the reaction mechanism, discussing photo- and electrochemical aspects. Electronic structure calculations show that the orbitals involved in the processes described here can be clearly classified as antenna- or POM-like upon their nature. Photochemistry, studied with the TD-DFT formalism, confirm that, upon absorption of UV-vis light by the reactant, the most probable electron excitation involves the [Ir] chromophore localized in the antenna. Partial relaxation of the excited stated generated might populate a d(W)-like empty orbital of the POM fragment, producing the charge-separated species. Triethylamine (NEt3) restores iridium to its initial oxidation state. After two excitation + reduction steps, the electron-rich species has extra electrons located in the POM framework. We have revealed that the two-electron reduced POM would be protonated in the reaction conditions. From the electrochemical point of view of the reaction, we have identified the orbitals of the POM that are found at considerably more positive energies in DMF with respect to aqueous medium, whereas the energy of the empty orbital of H+ is not as dependent on the solvent. The energy of the hydrogen evolution process was computed in water and DMF media to check if the previous arguments are solid, revealing a more favourable process in DMF than in water. Such dissimilar reaction energies bring to light one of the greatest advantages of using DMF instead of water as solvent for H2 evolution. However, the intricacy of the electron transfer from the two-reduced POM to H+ is still now an open issue.
Finally, we wish to remark that reduced hybrid POMs have been used for CO2 reduction, where a Re-organic decorated phosphotungstate Keggin cluster acts as photo-sensitizer, electron reservoir and electron donor. We cannot discard that in parallel to the outer-sphere hydrogen evolution on the surface of the electron-rich POM, a fraction of excited electrons return to the Ir centres and then a hydride is formed. Many organometallic complexes, which contain transition metals, can form hydrides to produce molecular hydrogen. Some studies focus on iridium (III) as catalyst. One of the most discussed mechanisms is the formation of a metal hydride that can react with a proton, resulting in H2. We have checked that this species can evolve H2 with relatively low energy barriers although this hypothesis needs experimental confirmation. So, more studies are needed in this direction.
The last part of this thesis is done in collaboration with Dr. Errington’s group. Heterometallic POMs are of particular interest and they provide a unique opportunity to study the reactivity of isolated heterometal site linked to a matrix of another metal at the molecular level and, hence, to establish a basis of understanding for the activity of a mixed oxide catalyst. The motivation of thiscollaboration comes from the difficulty to explain the behaviour of different heterometallic POMs in front of the protonolysis reaction. In chapter 7, we report a computational study to investigate (a) the effects of different lacunary oxide ligands {M5O18}6- on the protonolysis of the M’-OR bond and (b) the effects of different M’ on the hydrolysis of heterometallic tungstates, [(RO)M’W5O18]n-. We have found a consistent explanation based on DFT calculations of why and how the reactions take place from different heterometallic molecules. The mechanism of protonolysis of different heterometallic Lindqvist and Keggin anions was thoroughly studied. An understanding of the comparative differences in behaviour between polyoxomolybdates and -ungstates is not complete, since our calculations predict the same mechanism in both cases. These results prompted further investigations of these systems and it is worth noting that recent 2D EXSY 1H NMR studies reinforce the conclusions of the computational studies described in this thesis. Moreover, substituted Ti- and Sn- substituted Lindqvist and Keggin polyoxotungstates were also compared. The experimental trends have been reproduced by calculations, and relative stabilities and energy barriers are in complete agreement with measurements.
In the last section of the chapter an experimental study about the protonation of heterometallic tungstate dimers is reported. Two different pathways to the di-protonated form of (TBA)6[(μ-O)(TiW5O18)2] were studied. We have found that the protonated form is stable and it is easily synthesized. This leads us to predict the behaviour of electrophiles when they interact with this class of oxo-bridged dimer. This study provides an understanding about the behaviour towards electrophiles in reactions that are producing a new family of functionalized POMs, where they act as chelating or pincer ligands.
