Structure Property Correlation in Rutile Related Materials

To fight global warming, electrifying end use energy sources like electric cars or industrial processes to reduce the dependence of fossil fuel is considered one of the main solutions. To achieve this, materials with applications in renewable energy have drawn a lot of attention in the past decades, whether for production or storage of electrical energy. Concerning energy storage, research has focused on improving performances of batteries, finding new kinds of capacitors, or new types of energy storage systems, in particular, nonlinear dielectrics, i.e. materials which exhibit ferroelectric, antiferroelectric or relaxor ferroelectric behaviours. The vast majority of the work in this area has tended to focus on materials which crystallize in the perovskite structure (due to compositional flexibility and well-understood chemical toolboxes). However, some materials crystallizing in other systems such as tetragonal tungsten bronzes, binary systems (e.g. ZrO2/HfO2) and rutile-related structures have exhibited non-linear dielectric behaviours. In particular, rutile-based materials offer chemical and structural flexibility and the potential to host exciting electrical properties but have received considerably less attention. Thus, the main goal of this thesis was to investigate rutile or rutile related structures to understand their dielectric behaviour.

This thesis first investigates the structural, magnetic and electronic properties of MFeNbO6 materials (where M = Ti, Zr or Hf). TiFeNbO6 crystallized with the disordered rutile structure with the P42/mnm space group. Increasing the ionic radii of the M4+ cation from Ti4+ (0.605 Å) to Zr4+ (0.720 Å) and Hf4+ (0.710 Å) resulted in a buckling of the octahedral chains with these materials crystallising with the disordered α-PbO2 structure (space group Pbcn). In contrast, with previous reports for TiFeNbO6 we find no evidence of relaxor ferroelectric behaviour with all three materials instead exhibiting Maxwell-Wagner-like relaxation. Impedance spectroscopy confirmed semiconductor behaviour. Magnetically, all three materials could be described by infinite irregular antiferromagnetic S = 5/2 chains with weak ferromagnetism below 10 K. These chains had random propagation directions between nearest Fe-Fe neighbours consistent with the cation-disordered nature of these materials. Observations of a negative imaginary part of the ac susceptibility can be related to the defect magnetic topology. The thesis then focused on the series of materials, MFeTaO6 (where M = Ti, Zr or Hf). TiFeTaO6 was previously reported to be isostructural with the TiFeNbO6 sample and also exhibit relaxor ferroelectric behaviour. The work presented here confirms the disordered rutile structure (P42/mnm) for TiFeTaO6. Likewise, similar structural results were obtained when the M4+ cation was replaced by those with larger ionic radii, i.e the Hf and Zr materials adopt the α-PbO2 structure. No evidence was found for relaxor ferroelectric behaviours for the samples TiFeTaO6 and HfFeTaO6. The antiferromagnetic chain-like model was used to fit the FC/ZFC magnetometry measurements for both samples. ZrFeTaO6, however, gave a different result. First, the dielectric measurements did not show any relaxation behaviour, with low value and showing an evolution that could be interpreted by either a transition from a FE phase to a PE phase, or a Debye relaxation. P-E loop measurement showed a loop that could interpreted as a FE hysteresis loop. However, the value of the polarization, several times higher to what to expect to a FE measurement, combined with the I-E graph showing a behaviour close to the other two sample. Also, the DC conductivity showed the same kind of behaviour observed to the other two samples and the FC/ZFC magnetometry measurements were again consistent with the disordered nature of the sample.

Next, the iron niobate, FeNbO4, and iron columbite, FeNb2O6, materials were studied. The aim of this chapter was to investigate the influence of the ordering in samples with the same composition, since FeNbO4 can either crystallises in a disordered phase, called ixiolite, or in a layered A-B-A-B stacked phase, called wolframite. Iron columbite (FeNb2O6) crystallizes in a layered structure with A-B-B-A-B-B stacking. Unfortunately, during these studies a completely disordered phase was not obtained. Instead, a slightly richer Nb5+ wolframite was observed, giving a slightly disordered phase. The wolframite sample showed a Maxwell-Wagner relaxation and a resistive P-E loop suggesting that cation ordering does not promote ferroelectric behaviour. However, this may also arise due to the poor density of the pellet used for these measurements. In contrast, the Nb5+-rich FeNbO4 composition, showed a relaxation that could be interpreted as a relaxor-like ferroelectric behaviour suggesting that the excess Nb5+ gives rise to the formation of polar nano-regions. Both samples exhibited antiferromagnetic behaviour. The iron-columbite, FeNb2O6, material did not show any relaxation behaviour over the temperature range studied, with P-E measurements providing no evidence for any loop that could be interpreted as a ferroelectric hysteresis loop. Contrary to the wolframite samples, the magnetic measurements for FeNb2O6 confirmed that the material exhibits ferromagnetic order.

Finally, the material, Sr2(TiO)(PO4)2 was studied. This material is known to exhibit a layered structure, with each layer comprised of chains of TiO6 octahedra, bound by the top and linked by PO4 tetrahedra. Each layer is separated by Sr2+ cations. Within this structure the Ti cations are off-centred within the octahedra, leading to a chain of short/long Ti-O bonds. The aim of this chapter was to investigate if the Ti cations within the octahedra could be switched by applying an electric field, leading to ferroelectric or antiferroelectric behaviour, and if by optimising the cooling rate if it was possible to make the different chains cooperative. Samples with different cooling rates were made with all samples containing various amounts of secondary phases. To observe if the chains were cooperative, 31P NMR measurements were made. These measurements demonstrated that the cooling rate did have an influence on the amount of cooperation between chains. P-E loop measurements were taken of one of the samples with the lowest amount of secondary phase, and a ferroelectric hysteresis was observed. However, there is not enough evidence to conclusively determine the origin of the ferroelectric behaviour i.e. is it an intrinsic property of Sr2(TiO)(PO4)2, or if it arises as a result of the presence of ferroelectric secondary phases.