Breaking the aristotype: featurisation of polyhedral distortions in perovskite crystals

While traditional crystallographic representations of structure play an important role in materials science, they are unsuitable for efficient machine learning. A range of effective numerical descriptors have been developed for molecular and crystal structures. We are interested in a special case, where distortions emerge relative to an ideal high-symmetry parent structure. We demonstrate that irreducible representations form an efficient basis for the featurisation of polyhedral deformations with respect to such an aristotype. Applied to dataset of 552 octahedra in ABO3 perovskite-type materials, we use unsupervised machine learning with irreducible representation descriptors to identify four distinct classes of behaviour, associated with predominately corner, edge, face, and mixed connectivity between neighbouring octahedral units. Through this analysis, we identify SrCrO3 as a material with tuneable multiferroic behaviour. We further show, through supervised machine learning, that thermally activated structural distortions of CsPbI3 are well described by this approach.

X-RAY studies of ferroelectric lithium niobate crystals doped with rare earth elements

The lattice structural distortions of LiNbO3 crystals doped with samarium with a concentration of 1.0, 1.9, 2.5 mol. % were studied by X-ray diffraction methods. It was found that the samarium atoms and some of the niobium atoms occupied the vacant positions of lithium in the crystal lattice, and some of the niobium atoms were located in the empty octahedron. The octahedra of SmLiO6 was distorted more strongly than octahedra of NbLiO6 when niobium atoms entered the vacant positions of lithium. The smallest changes in the bond lengths in the octahedra of the main motif and in the defect region as well as along the polar axis in the lithium niobate lattice were observed in the sample with a samarium concentration of 2.5 mol. %.

New Modular Organic Platform For Understanding The Effect Of Structural Changes On Slow Magnetic Relaxation In Mononuclear Octahedral Copper(II) Complexes

Current advances in molecular magnetism are aimed at the construction of molecular nanomagnets and spin qubits for their utilization as high-density data storage materials and quantum computers. Mononuclear coordination compounds with low spin values of S=½ are excellent candidates for this endeavour, but their construction via rational design is limited. This particularly applies to the single copper(II) spin center, having been only recently demonstrated to exhibit slow relaxation of magnetisation in the appropriate octahedral environment. We have thus prepared a novel, modular organic scaffold that would allow one to gain in-depth insight into how purposeful structural differences affect the slow magnetic relaxation in monometallic, transition metal complexes. As a proof-of-principle, we demonstrate how one can construct two, structurally very similar complexes with isolated Cu(II) ions in an octahedral ligand environment, the magnetic properties of which differ significantly. The differences in structural symmetry effects and in magnetic relaxation are corroborated with a series of experimental and theoretical techniques, showing how symmetry distortions and crystal packing affect the relaxation behaviour in these isolated Cu(II) systems. Our highly modular organic platform can be efficiently utilized for the construction of various transition-metal ion systems in the future, effectively providing a model system for investigation of magnetic relaxation via targeted structural distortions.

Quasi-particle interference of the van Hove singularity in Sr2RuO4

The single-layered ruthenate Sr2RuO4 is one of the most enigmatic unconventional superconductors. While for many years it was thought to be the best candidate for a chiral p-wave superconducting ground state, desirable for topological quantum computations, recent experiments suggest a singlet state, ruling out the original p-wave scenario. The superconductivity as well as the properties of the multi-layered compounds of the ruthenate perovskites are strongly influenced by a van Hove singularity in proximity of the Fermi energy. Tiny structural distortions move the van Hove singularity across the Fermi energy with dramatic consequences for the physical properties. Here, we determine the electronic structure of the van Hove singularity in the surface layer of Sr2RuO4 by quasi-particle interference imaging. We trace its dispersion and demonstrate from a model calculation accounting for the full vacuum overlap of the wave functions that its detection is facilitated through the octahedral rotations in the surface layer.

High-temperature superconductivity on the verge of a structural instability in lanthanum superhydride

The possibility of high, room-temperature superconductivity was predicted for metallic hydrogen in the 1960s. However, metallization and superconductivity of hydrogen are yet to be unambiguously demonstrated and may require pressures as high as 5 million atmospheres. Rare earth based “superhydrides”, such as LaH10, can be considered as a close approximation of metallic hydrogen even though they form at moderately lower pressures. In superhydrides the predominance of H-H metallic bonds and high superconducting transition temperatures bear the hallmarks of metallic hydrogen. Still, experimental studies revealing the key factors controlling their superconductivity are scarce. Here, we report the pressure and magnetic field dependence of the superconducting order observed in LaH10. We determine that the high-symmetry high-temperature superconducting Fm-3m phase of LaH10 can be stabilized at substantially lower pressures than previously thought. We find a remarkable correlation between superconductivity and a structural instability indicating that lattice vibrations, responsible for the monoclinic structural distortions in LaH10, strongly affect the superconducting coupling.

Breaking the aristotype: featurisation of polyhedral distortions in perovskite crystals

While traditional crystallographic representations of structure play an important role in materials science, they are unsuitable for efficient machine learning. A range of effective numerical descriptors have been developed for molecular and crystal structures. We are interested in a special case, where distortions emerge relative to an ideal high-symmetry parent structure. We demonstrate that irreducible representations form an efficient basis for the featurisation of polyhedral deformations with respect to such an aristotype. Applied to dataset of 552 octahedra in ABO3 perovskite-type materials, we use unsupervised machine learning with irreducible representation descriptors to identify four distinct classes of behaviour, associated with predominately corner, edge, face, and mixed connectivity between neighbouring octahedral units. Through this analysis, we identify SrCrO3 as a material with tuneable multiferroic behaviour. We further show, through supervised machine learning, that thermally activated structural distortions of CsPbI3 are well described by this approach.

Deep Bayesian local crystallography

The advent of high-resolution electron and scanning probe microscopy imaging has opened the floodgates for acquiring atomically resolved images of bulk materials, 2D materials, and surfaces. This plethora of data contains an immense volume of information on materials structures, structural distortions, and physical functionalities. Harnessing this knowledge regarding local physical phenomena necessitates the development of the mathematical frameworks for extraction of relevant information. However, the analysis of atomically resolved images is often based on the adaptation of concepts from macroscopic physics, notably translational and point group symmetries and symmetry lowering phenomena. Here, we explore the bottom-up definition of structural units and symmetry in atomically resolved data using a Bayesian framework. We demonstrate the need for a Bayesian definition of symmetry using a simple toy model and demonstrate how this definition can be extended to the experimental data using deep learning networks in a Bayesian setting, namely rotationally invariant variational autoencoders.

Particularities of photonic and structural transformations in the polystyrene film during pulsed infrared laser irradiation

Two new kinds of the structural in-homogeneities in the polystyrene films were created by the irradiation of the film with nanosecond pulses of the infrared laser with 1064 nm wavelength. At the beginning of the irradiation dark micro-inclusions having unusual close to rectangular geometries appeared. After about 30 seconds of the irradiation one more kind structural distortions appeared having the geometries of the systems of the concentric micro-rings. Simultaneously with the creation of these concentric micro-rings the film started to emit bright blue luminescence all over its area and significant focusing of the laser beam during its propagation through the film was observed resulting in the generation of the bright white flashes at the screen installed behind the film. The creation of the dark rectangular spots as well as of the concentric micro-rings are ascribed to the non-linear interactions between the laser radiation and the structural distortions produced in the film due to photo-thermal processes excited in the film.

Analysis of changes of cavity volumes in predefined directions of protein motions and cavity flexibility

Dynamics of protein cavities associated with protein fluctuations and conformational plasticity is essential for their biological function. NMR ensembles, Molecular Dynamics (MD) simulations combined with Principal Component Analysis (PCA), and Normal Mode Analysis (NMA) provide appropriate frameworks to explore functionally relevant protein dynamics and cavity changes relationships. Within this context, we have developed ANA (Analysis of Null Areas), an efficient method to calculate cavity volumes. ANA is based on a combination of algorithms that guarantees its robustness against numerical differentiations. This is a unique feature with respect to other methods. Herein, we test ANA as a biophysical and bioinformatic method to analyze different structural and dynamics properties of cavities. In order to address this task, we have developed an updated and improved version of ANA that expands it use to quantify changes in cavity features, like volume and flexibility, due to protein structural distortions performed on predefined biologically relevant directions, e.g, directions of largest contribution to protein fluctuations (PCA modes) obtained by MD simulations or ensembles of NMR structures, collective NMA modes or any other direction of motion associated with specific conformational changes. A web page has been developed and its facilities are explained in detail, making the software available to all users. Firstly, we show that ANA can be useful to explore gradual changes of cavity volume and flexibility associated with protein ligand binding. Secondly, we perform a comparison study of the extent of variability between protein backbone structural distortions, and changes in cavity volumes and flexibilities evaluated for an ensemble of NMR active and inactive conformers of the epidermal growth factor receptor (EGFR) structures. Finally, we compare changes in size and flexibility between sets of NMR structures for different homologous chains of dynein.