arXiv subjects

Stefano Sanvito

Publications and source records attributed to Stefano Sanvito.

At least 37 records · Page 2Linked to original sources

arXiv2024

Quantum-Accurate Machine Learning Potentials for Metal-Organic Frameworks using Temperature Driven Active Learning

Understanding how structural flexibility affects the properties of metal-organic frameworks (MOFs) is crucial for the design of better MOFs for targeted applications. Flexible MOFs can be studied with molecular dynamics simulations, whose accuracy depends on the force-field used to describe the interatomic interactions. Density functional theory (DFT) and quantum-chemistry methods are highly accurate, but the computational overheads limit their use in long time-dependent simulations for large systems. In contrast, classical force fields usually struggle with the description of coordination bonds. In this work we develop a DFT-accurate machine-learning spectral neighbor analysis potential, trained on DFT energies, forces and stress tensors, for two representative MOFs, namely ZIF-8 and MOF-5. Their structural and vibrational properties are then studied as a function of temperature and tightly compared with available experimental data. Most importantly, we demonstrate an active-learning algorithm, based on mapping the relevant internal coordinates, which drastically reduces the number of training data to be computed at the DFT level. Thus, the workflow presented here appears as an efficient strategy for the study of flexible MOFs with DFT accuracy, but at a fraction of the DFT computational cost.

cond-mat.mtrl-sci↗

arXiv2024

Ab-initio transport theory for the intrinsic spin Hall effect applied to 5$d$ metals

We describe how the spin Hall effect (SHE) can be studied from ab-initio by combining density functional theory with the non-equilibrium Green's functions technique for quantum transport into the so-called DFT+NEGF method. After laying down our theoretical approach in particular discussing how to compute charge and spin bond currents, DFT+NEGF calculations are carried out for ideal clean systems. In these the transport is ballistic and the linear response limit is met. The SHE emerges in a central region attached to two leads when we apply a bias voltage so that electrons are accelerated by a uniform electric field. As a result, we obtain a finite spin-Hall current and, by performing a scaling analysis with respect to the system size, we estimate the ballistic spin Hall conductivity (SHC). We consider 5d metals with fcc and bcc crystal structures, finding that the SHC exhibits a rough qualitative dependence on the d-band filling, and comment on these results in relation to existing literature. Finally, within the same DFT+NEGF approach, we also predict the appearance of a current-induced spin dipole moment inside the materials' unit cell and estimate its magnitude.

cond-mat.mes-hall↗

arXiv2024

Half-metallic transport and spin-polarized tunneling through the van der Waals ferromagnet Fe${_4}$GeTe$_{2}$

The recent emergence of van der Waals (vdW) ferromagnets has opened new opportunities for designing spintronic devices. We theoretically investigate the coherent spin-dependent transport properties of the vdW ferromagnet Fe$_4$GeTe$_2$, by using density functional theory combined with the non-equilibrium Green's functions method. We find that the conductance in the direction perpendicular to the layers is half-metallic, namely it is entirely spin-polarized, as a result of the material's electronic structure. This characteristic persists from bulk to single layer, even under significant bias voltages, and it is little affected by spin-orbit coupling and electron correlation. Motivated by this observation, we then investigate the tunnel magnetoresistance (TMR) effect in an magnetic tunnel junction, which comprises two Fe$_4$GeTe$_2$ layers separated by the vdW gap acting as insulating barrier. We predict a TMR ratio of almost 500\%, which can be further boosted by increasing the number of Fe$_4$GeTe$_2$ layers in the junction.

cond-mat.mtrl-sci↗

arXiv2024

A rule-free workflow for the automated generation of databases from scientific literature

In recent times, transformer networks have achieved state-of-the-art performance in a wide range of natural language processing tasks. Here we present a workflow based on the fine-tuning of BERT models for different downstream tasks, which results in the automated extraction of structured information from unstructured natural language in scientific literature. Contrary to existing methods for the automated extraction of structured compound-property relations from similar sources, our workflow does not rely on the definition of intricate grammar rules. Hence, it can be adapted to a new task without requiring extensive implementation efforts and knowledge. We test our data-extraction workflow by automatically generating a database for Curie temperatures and one for band gaps. These are then compared with manually-curated datasets and with those obtained with a state-of-the-art rule-based method. Furthermore, in order to showcase the practical utility of the automatically extracted data in a material-design workflow, we employ them to construct machine-learning models to predict Curie temperatures and band gaps. In general we find that, although more noisy, automatically extracted datasets can grow fast in volume and that such volume partially compensates for the inaccuracy in downstream tasks.

cond-mat.mtrl-sci↗

arXiv2023

Machine-Learning-Assisted Construction of Ternary Convex Hull Diagrams

In the search for novel intermetallic ternary alloys, much of the effort goes into performing a large number of ab-initio calculations covering a wide range of compositions and structures. These are essential to build a reliable convex hull diagram. While density functional theory (DFT) provides accurate predictions for many systems, its computational overheads set a throughput limit on the number of hypothetical phases that can be probed. Here, we demonstrate how an ensemble of machine-learning spectral neighbor-analysis potentials (SNAPs) can be integrated into a workflow for the construction of accurate ternary convex hull diagrams, highlighting regions fertile for materials discovery. Our workflow relies on using available binary-alloy data both to train the SNAP models and to create prototypes for ternary phases. From the prototype structures, all unique ternary decorations are created and used to form a pool of candidate compounds. The SNAPs are then used to pre-relax the structures and screen the most favourable prototypes, before using DFT to build the final phase diagram. As constructed, the proposed workflow relies on no extra first-principles data to train the machine-learning surrogate model and yields a DFT-level accurate convex hull. We demonstrate its efficacy by investigating the Cu-Ag-Au and Mo-Ta-W ternary systems.

cond-mat.mtrl-sci↗

arXiv2023

Machine Learning Predictions of High-Curie-Temperature Materials

Technologies that function at room temperature often require magnets with a high Curie temperature, $T_\mathrm{C}$, and can be improved with better materials. Discovering magnetic materials with a substantial $T_\mathrm{C}$ is challenging because of the large number of candidates and the cost of fabricating and testing them. Using the two largest known data sets of experimental Curie temperatures, we develop machine-learning models to make rapid $T_\mathrm{C}$ predictions solely based on the chemical composition of a material. We train a random forest model and a $k$-NN one and predict on an initial dataset of over 2,500 materials and then validate the model on a new dataset containing over 3,000 entries. The accuracy is compared for multiple compounds' representations ("descriptors") and regression approaches. A random forest model provides the most accurate predictions and is not improved by dimensionality reduction or by using more complex descriptors based on atomic properties. A random forest model trained on a combination of both datasets shows that cobalt-rich and iron-rich materials have the highest Curie temperatures for all binary and ternary compounds. An analysis of the model reveals systematic error that causes the model to over-predict low-$T_\mathrm{C}$ materials and under-predict high-$T_\mathrm{C}$ materials. For exhaustive searches to find new high-$T_\mathrm{C}$ materials, analysis of the learning rate suggests either that much more data is needed or that more efficient descriptors are necessary.

cond-mat.mtrl-sci↗

arXiv2023

Cluster expansion constructed over Jacobi-Legendre polynomials for accurate force fields

We introduce a compact cluster expansion method, constructed over Jacobi and Legendre polynomials, to generate highly accurate and flexible machine-learning force fields. The constituent many-body contributions are separated, interpretable and adaptable to replicate the physical knowledge of the system. In fact, the flexibility introduced by the use of the Jacobi polynomials allows us to impose, in a natural way, constrains and symmetries to the cluster expansion. This has the effect of reducing the number of parameters needed for the fit and of enforcing desired behaviours of the potential. For instance, we show that our Jacobi-Legendre cluster expansion can be designed to generate potentials with a repulsive tail at short inter-atomic distances, without the need of imposing any external function. Our method is here continuously compared with available machine-learning potential schemes, such as the atomic cluster expansion and potentials built over the bispectrum. As an example we construct a Jacobi-Legendre potential for carbon, by training a slim and accurate model capable of describing crystalline graphite and diamond, as well as liquid and amorphous elemental carbon.

cond-mat.mtrl-sci↗

arXiv2023

Linear Jacobi-Legendre expansion of the charge density for machine learning-accelerated electronic structure calculations

Kohn-Sham density functional theory (KS-DFT) is a powerful method to obtain key materials' properties, but the iterative solution of the KS equations is a numerically intensive task, which limits its application to complex systems. To address this issue, machine learning (ML) models can be used as surrogates to find the ground-state charge density and reduce the computational overheads. We develop a grid-centred structural representation, based on Jacobi and Legendre polynomials combined with a linear regression, to accurately learn the converged DFT charge density. This integrates into a ML pipeline that can return any density-dependent observable, including energy and forces, at the quality of a converged DFT calculation, but at a fraction of the computational cost. Fast scanning of energy landscapes and producing starting densities for the DFT self-consistent cycle are among the applications of our scheme.

cond-mat.mtrl-sci↗

arXiv2023

Machine-Learning Surrogate Model for Accelerating the Search of Stable Ternary Alloys

The prediction of phase diagrams in the search for new phases is a complex and computationally intensive task. Density functional theory provides, in many situations, the desired accuracy, but its throughput becomes prohibitively limited as the number of species involved grows, even when used with local and semi-local functionals. Here, we explore the possibility of integrating machine-learning models in the workflow for the construction of ternary convex hull diagrams. In particular, we train a set of spectral neighbour-analysis potentials (SNAPs) over readily available binary phases and we establish whether this is good enough to predict the energies of novel ternaries. Such a strategy does not require any new calculations specific for the construction of the model, but just avails of data stored in binary-phase-diagram repositories. We find that a so-constructed SNAP is capable of accurate total-energy estimates for ternary phases close to the equilibrium geometry but, in general, is not able to perform atomic relaxation. This is because during a typical relaxation path a given phase traverses regions in the parameter space poorly represented by the training set. Different metrics are then investigated to assess how an unknown structure is well described by a given SNAP model, and we find that the standard deviation of an ensemble of SNAPs provides a fast and non-specie-specific metric.

cond-mat.mtrl-sci↗

arXiv2023

A theoretical perspective on the modification of the magnetocrystalline anisotropy at molecule-cobalt interfaces

We study the modification of the magnetocrystalline anisotropy (MCA) of Co slabs induced by several different conjugated molecular overlayers, i.e., benzene, cyclooctatetraene, naphthalene, pyrene and coronene. We perform first-principles calculations based on Density Functional Theory and the magnetic force theorem. Our results indicate that molecular adsorption tends to favour a perpendicular MCA at surfaces. A detailed analysis of various atom-resolved quantities, accompanied by an elementary model, demonstrates that the underlying physical mechanism is related to the metal-molecule interfacial hybridization and, in particular, to the chemical bonding between the molecular C $p_z$ and the out-of-plane Co $d_{z^2}$ orbitals. This effect can be estimated from the orbital magnetic moment of the surface Co atoms, a microscopic observable accessible to both theory and experiments. As such, we suggest a way to directly assess the MCA modifications at molecule-decorated surfaces, overcoming the limitations of experimental studies that rely on fits of magnetization hysteresis loops. Finally, we also study the interface between Co and both C$_{60}$ and Alq$_3$, two molecules that find widespread use in organic spintronics. We show that the modification of the surface Co MCA is similar upon adsorption of these two molecules, thereby confirming the results of recent experiments.

cond-mat.mtrl-sci↗

arXiv2023

3D-imaging of Printed Nanostructured Networks using High-resolution FIB-SEM Nanotomography

Networks of solution-processed nanomaterials are important for multiple applications in electronics, sensing and energy storage/generation. While it is known that network morphology plays a dominant role in determining the physical properties of printed networks, it remains difficult to quantify network structure. Here, we utilise FIB-SEM nanotomography to characterise the morphology of nanostructured networks. Nanometer-resolution 3D-images were obtained from printed networks of graphene nanosheets of various sizes, as well as networks of WS2 nanosheets, silver nanosheets and silver nanowires. Important morphological characteristics, including network porosity, tortuosity, pore dimensions and nanosheet orientation were extracted and linked to network resistivity. By extending this technique to interrogate the structure and interfaces within vertical printed heterostacks, we demonstrate the potential of this technique for device characterisation and optimisation.

cond-mat.mes-hall↗

arXiv2022

A spectral-neighbour representation for vector fields: machine-learning potentials including spin

We introduce a translational and rotational invariant local representation for vector fields, which can be employed in the construction of machine-learning energy models of solids and molecules. This allows us to describe, on the same footing, the energy fluctuations due to the atomic motion, the longitudinal and transverse excitations of the vector field, and their mutual interplay. The formalism can then be applied to physical systems where the total energy is determined by a vector density, as in the case of magnetism. Our representation is constructed over the power spectrum of the combined angular momentum describing the local atomic positions and the vector field, and can be used in conjunction with different machine-learning schemes and data taken from accurate ab initio electronic structure theories. We demonstrate the descriptive power of our representation for a range of classical spin Hamiltonian and machine-learning algorithms. In particular, we construct energy models based on both linear Ridge regression, as in conventional spectral neighbour analysis potentials, and gaussian approximation. These are both built to represent a Heisenberg-type Hamiltonian including a longitudinal energy term and spin-lattice coupling.

cond-mat.mtrl-sci↗

arXiv2022

Transition between large and small electron polaron at neutral ferroelectric domain walls in BiFeO$_3$

Ferroelectric domain walls are planes within an insulating material that can accumulate and conduct charge carriers, hence the interaction of the domain walls with the charge carriers can be important for photovoltaic and other electronic applications. By means of first principles calculations we predict a transition from a large two-dimensional electron polaron to a small polaron at the domain walls at a critical electron density, with polaron signatures in optical absorption and photoluminescence. We find that large and small polarons at the domain walls create different absorption peaks within the band gap that are not present in the case of pristine domain walls. These are an extended Drude peak in the case of large electron or hole polarons and a narrow mid-gap peak in the case of the small electron polaron.

cond-mat.mtrl-sci↗

arXiv2022

Inversion of the chemical environment representations

Machine-learning generative methods for material design are constructed by representing a given chemical structure, either a solid or a molecule, over appropriate atomic features, generally called structural descriptors. These must be fully descriptive of the system, must facilitate the training process and must be invertible, so that one can extract the atomic configurations corresponding to the output of the model. In general, this last requirement is not automatically satisfied by the most efficient structural descriptors, namely the representation is not directly invertible. Such drawback severely limits our freedom of choice in selecting the most appropriate descriptors for the problem, and thus our flexibility to construct generative models. In this work, we present a general optimization method capable of inverting any local many-body descriptor of the chemical environment, back to a cartesian representation. The algorithm is then implemented together with the bispectrum representation of the local structure and demonstrated for a number of molecules. The scheme presented here, thus, represents a general approach to the inversion of structural descriptors, enabling the construction of efficient structural generative models.

cond-mat.mtrl-sci↗

arXiv2022

Data-Driven Time Propagation of Quantum Systems with Neural Networks

We investigate the potential of supervised machine learning to propagate a quantum system in time. While Markovian dynamics can be learned easily, given a sufficient amount of data, non-Markovian systems are non-trivial and their description requires the memory knowledge of past states. Here we analyse the feature of such memory by taking a simple 1D Heisenberg model as many-body Hamiltonian, and construct a non-Markovian description by representing the system over the single-particle reduced density matrix. The number of past states required for this representation to reproduce the time-dependent dynamics is found to grow exponentially with the number of spins and with the density of the system spectrum. Most importantly, we demonstrate that neural networks can work as time propagators at any time in the future and that they can be concatenated in time forming an autoregression. Such neural-network autoregression can be used to generate long-time and arbitrary dense time trajectories. Finally, we investigate the time resolution needed to represent the system memory. We find two regimes: for fine memory samplings the memory needed remains constant, while longer memories are required for coarse samplings, although the total number of time steps remains constant. The boundary between these two regimes is set by the period corresponding to the highest frequency in the system spectrum, demonstrating that neural network can overcome the limitation set by the Shannon-Nyquist sampling theorem.

quant-ph↗

arXiv2022

Fe- and Co-based magnetic tunnel junctions with AlN and ZnO spacers

AlN and ZnO, two wide band-gap semiconductors extensively used in the display industry, crystallise in the wurtzite structure, which can favour the formation of epitaxial interfaces to close-packed common ferromagnets. Here we explore these semiconductors as material for insulating barriers in magnetic tunnel junctions. In particular, the {\it ab initio} quantum transport code {\it Smeagol} is used to model the $X$[111]/$Y$[0001]/$X$[111] ($X=$ Co and Fe, $Y=$ AlN and ZnO) family of junctions. Both semiconductors display a valance-band top with $p$-orbital character, while the conduction band bottom exhibits $s$-type symmetry. The smallest complex-band decay coefficient in the forbidden energy-gap along the [0001] direction is associated with the $Δ_1$ symmetry, and connects across the band gap at the $Γ$ point in 2D Brillouin zones. This feature enables spin filtering and may result in a large tunnelling magnetoresistance. In general, we find that Co-based junctions present limited spin filtering and little magnetoresistance at low bias, since both spin sub-bands cross the Fermi level with $Δ_1$ symmetry. This contrasts the situation of Fe, where only the minority $Δ_1$ band is available. However, even in the case of Fe the magnitude of the magnetoresistance at low bias remains relatively small, mostly due to conduction away from the $Γ$ point and through complex bands with symmetry different than $Δ_1$. The only exception is for the Fe/AlN/Fe junction, where we predict a magnetoresitance of around 1,000\% at low bias.

cond-mat.mes-hall↗

arXiv2022

Spin transfer torque in Mn$_3$Ga-based ferrimagnetic tunnel junctions from first principles

We report on first-principles calculations of spin-transfer torque (STT) in epitaxial magnetic tunnel junctions (MTJs) based on ferrimagnetic tetragonal Mn$_3$Ga electrodes, both as analyzer in an Fe/MgO stack, and also in an analogous stack with a second Mn$_3$Ga electrode (instead of Fe) as polarizer. Solving the ballistic transport problem (NEGF + DFT) for the nonequilibrium spin density in a scattering region extended to over 7.6 nm into the Mn$_3$Ga electrode, we find long-range spatial oscillations of the STT decaying on a length scale of a few tens of angstroms, both in the linear response regime and for finite bias. The oscillatory behavior of the STT in Mn$_3$Ga is robust against variations in the stack geometry and the applied bias voltage, which may affect the phase and the amplitude of the spatial oscillation, but the wave number is only responsive to variations in the longitudinal lattice constant of Mn$_3$Ga (for fixed in-plane geometry) without being commensurate with the lattice. Our interpretation of the long-range STT oscillations is based on the bulk electronic structure of Mn$_3$Ga, taking also into account the spin-filtering properties of the MgO barrier. Comparison to a fully Mn$_3$Ga-based stack shows similar STT oscillations, but a significant enhancement of both the TMR effect at the Fermi level and the STT at the interface, due to resonant tunneling for the mirror-symmetric junction with thinner barrier (three monoatomic layers). From the calculated energy dependence of the spin-polarized transmissions at 0 V, we anticipate asymmetric or symmetric TMR as a function of the applied bias voltage for the Fe-based and the all-Mn$_3$Ga stacks, respectively, which also both exhibit a sign change below 1 V. In the latter (symmetric) case we expect a TMR peak at zero, which is larger for the thinner barriers because of a spin-polarized resonant tunneling contribution.

cond-mat.mes-hall↗

arXiv2021

Reactivity of transition-metal alloys to oxygen and sulphur

Oxidation and tarnishing are the two most common initial steps in the corrosive process of metals at ambient conditions. These are always initiated with O and S binding to a metallic surface, so that one can use the binding energy as a rough proxy for the metal reactivity. With this in mind, we present a systematic study of the binding energy of O and S across the entire transition-metals composition space, namely we explore the binding energy of {\bf 88} single-phase transition metals and of {\bf 646} transition-metal binary alloys. The analysis is performed by defining a suitable descriptor for the binding energy. This is here obtained by fitting several schemes, based on the original Newns-Anderson model, against density-functional-theory data for the 4$d$ transition metal series. Such descriptor is then applied to a vast database of electronic structures of transition-metal alloys, for which we are able to predict the range of binding energies across both the compositional and the structural space. Finally, we extend our analysis to ternary transition-metal alloys and identify the most resilient compounds to O and S binding.

cond-mat.mtrl-sci↗