arXiv subjects

Urvesh Patil

Publications and source records attributed to Urvesh Patil.

9 recordsLinked to original sources

arXiv2026

Machine-learning approach for the phase stability and mechanical properties of disordered alloys at finite temperature

The prediction of stable alloys forming solid-state solutions across large portions of the composition space is a serious theoretical challenge, since one has to evaluate the Gibbs free energy, including both configurational and vibrational contributions. This requires an energy theory capable of extremely high throughput. By taking the Ni-Pd system as prototype, we construct an efficient Jacobi-Legendre machine-learning potential based on density-functional-theory data, which provides accurate energies and forces across the entire composition space. Based on a cluster expansion up to three-body terms and only 873 trainable parameters, this allows us to compute the partition function by directly integrating all accessible microstates, differing for composition, atomic configuration and thermal agitation. We confirm that Ni and Pd are fully miscible, forming an $fcc$ solid-state solution. This is only metastable at room temperature, while becomes thermodynamically stable at around 600~K, with the stability achieved first at the Pd-rich end of the composition range. Interestingly, entropy and heat capacity analysis reveal a competition between the solid-state solution and two intermetallic phases with long-period L1$_0$ structure for NiPd and NiPd$_3$. All in all, our approach offers a powerful and high-throughput workflow for the study of disordered alloys, an approach that can be extended to multi-component systems such as high-entropy alloys.

cond-mat.mtrl-sci↗

arXiv2025

A charge-density machine-learning workflow for computing the infrared spectrum of molecules

We present a machine-learning workflow for the calculation of the infrared spectrum of molecules, and more generally of other temperature-dependent electronic observables. The main idea is to use the Jacobi-Legendre cluster expansion to predict the real-space charge density of a converged density-functional-theory calculation. This gives us access to both energy and forces, and to electronic observables such as the dipole moment or the electronic gap. Thus, the same model can simultaneously drive a molecular dynamics simulation and evaluate electronic quantities along the trajectory, namely it has access to the same information of ab-initio molecular dynamics. A similar approach within the framework of machine-learning force fields would require the training of multiple models, one for the molecular dynamics and others for predicting the electronic quantities. The scheme is implemented here within the numerical framework of the PySCF code and applied to the infrared spectrum of the uracil molecule in the gas phase.

physics.chem-ph↗

arXiv2024

Covariant Jacobi-Legendre expansion for total energy calculations within the projector-augmented-wave formalism

Machine-learning models can be trained to predict the converged electron charge density of a density functional theory (DFT) calculation. In general, the value of the density at a given point in space is invariant under global translations and rotations having that point as a centre. Hence, one can construct locally invariant machine-learning density predictors. However, the widely used projector augmented wave (PAW) implementation of DFT requires the evaluation of the one-center augmentation contributions, that are not rotationally invariant. Building on our recently proposed Jacobi-Legendre charge-density scheme, we construct a covariant Jacobi-Legendre model capable of predicting the local occupancies needed to compose the augmentation charge density. Our formalism is then applied to the prediction of the energy barrier for the 1H-to-1T phase transition of two-dimensional MoS$_2$. With extremely modest training, the model is capable of performing a non-self-consistent nudged elastic band calculation at virtually the same accuracy as a fully DFT-converged one, thus saving thousands of self-consistent DFT steps. Furthermore, at variance with machine-learning force fields, the charge density is here available for any nudged elastic band image, so that we can trace the evolution of the electronic structure across the phase transition.

cond-mat.mtrl-sci↗

arXiv2023

The role of solvent interfacial structural ordering in maintaining stable graphene dispersions

Liquid phase exfoliation (LPE) is the most promising method for the low-cost, scalable production of two-dimensional nanosheets from their bulk counterparts. Extensive exfoliation occurs in most solvents due to the huge amount of energy introduced by sonication or shear mixing. However, the subsequent dispersion is not always stable, with extensive reaggregation occurring in some solvents. Identifying the optimal solvent for a particular layered material is difficult and requires a fundamental understanding of the mechanism involved in maintaining a stable dispersion. Here, we use molecular dynamics calculations to show that when graphene is immersed in a solvent, distinct solvation layers are formed irrespective of the choice of solvent and their formation is energetically favourable for all considered solvents. However, energetic considerations such as these do not explain the experimental solvent-dependence of the dispersion concentration. Instead, we find that solvents with high diffusion coefficients parallel to the graphene layer result in the lowest experimental concentration of graphene in solution. This can be explained by the enhanced ease of reaggregation in these solvents. Solvents with smaller diffusion coefficients result in higher experimental graphene concentrations as reaggregation is prevented. In the low diffusion limit, however, this relationship breaks down. We suggest that here the concentration of graphene in solution depends primarily on the separation efficiency of the initial exfoliation step. Based on this, we predict that the concentration of exfoliated graphene in solvents such as benzaldehyde and quinoline, which have low diffusion constants, can be increased dramatically by careful tuning of the experimental sonication parameters.

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↗

arXiv2019

Composition Dependence of the Charge Driven Phase Transition in Group-VI Transition Metal Dichalcogenides

Materials exhibiting multiple stable phases can be used as functional components in electronic and optical applications if the phase transition is controllable. Group-VI transition metal dichalcogenides (TMDs, MX$_2$, where M=Mo, W and X = S, Se) are known to undergo charge induced transitions from semi-conducting H phases to metallic T phases. This occurs, for example, when bulk TMDs are exfoliated with the aid of alkali ion intercalants. However, it is difficult to experimentally decouple the effect of composition-dependent phase transition barriers from indirect effects related to the exfoliation process. Here, using first-principles calculations, we study the energetics of transition between the different structural polytypes of four group-VI TMDs upon lithium adsorption. We find that both the activation barrier from the H phase to the metallic phase in charged monolayers and the reverse barrier in neutral monolayers are required to explain experimental results. We show that the high proportion of metallic phase found in WS$_2$ monolayers after alkali treatment can be explained by high transition barriers to revert back to the H phase once in a neutral state. The calculated barriers however cannot explain the low proportion of metallic phase found in MoS$_2$ monolayers in some experiments and so non-electronic effects must also play a role.

cond-mat.mtrl-sci↗

arXiv2018

Adsorption of common solvent molecules on graphene and MoS$_2$ from first-principles

Solvents are an essential element in the production and processing of two-dimensional (2D) materials. For example, the liquid phase exfoliation of layered materials requires a solvent to prevent the resulting monolayers from re-aggregating, while solutions of functional atoms and molecules are routinely used to modify the properties of the layers. It is generally assumed that these solvents do not interact strongly with the layer and so their effects can be neglected. Yet experimental evidence has suggested that explicit atomic-scale interactions between the solvent and layered material may play a crucial role in exfoliation and cause unintended electronic changes in the layer. Little is known about the precise nature of the interaction between the solvent molecules and the 2D layer. Here, we use density functional theory calculations to determine the adsorption configuration and binding energy of a variety of common solvent molecules, both polar and non-polar, on two of the most popular 2D materials, namely graphene and MoS$_2$. We show that these molecules are physisorbed on the surface with negligible charge transferred between them. We find that the adsorption strength of the different molecules is independent of the polar nature of the solvent. However, we show the molecules induce a significant charge rearrangement at the interface after adsorption as a result of polar bonds in the molecule.

cond-mat.mtrl-sci↗

arXiv2016

Resonant enhancement in nanostructured thermoelectric performance via electronic thermal conductivity engineering

The use of an asymmetric broadening in the transport distribution, a characteristic of resonant structures, is proposed as a route to engineer a decrease in electronic thermal conductivity thereby enhancing the electronic figure of merit in nanostructured thermoelectrics. Using toy models, we first demonstrate that a decrease in thermal conductivity resulting from such an asymmetric broadening may indeed lead to an electronic figure of merit well in excess of 1000 in an idealized situation and in excess of 10 in a realistic situation. We then substantiate with realistic resonant structures designed using graphene nano-ribbons by employing a tight binding framework with edge correction that match density functional theory calculations under the local density approximation. The calculated figure of merit exceeding 10 in such realistic structures further reinforces the concept and sets a promising direction to use nano-ribbon structures to engineer a favorable decrease in the electronic thermal conductivity.

cond-mat.mes-hall↗