Announcements

Abstract

The shape and size of nanoparticles play a crucial role in determining the properties of materials, particularly in applications where atomic-scale interactions are significant. One of the most widely used methods for predicting equilibrium crystal shapes is the Wulff construction, which is based on the surface energies of crystallographic facets. The Wulff construction method calls for accurate calculation for the energies of the various surfaces of the material. Typically, these energies are obtained by simulating slabs of the material. Recently, an analytical model was proposed by P. Rosakis that yields surface energies from the parameters of the interatomic interactions, eliminating the need for direct simulation. The results reveal excellent agreement between the two approaches, while the analytical model exhibits lower computational cost, suggesting that it constitutes an attractive alternative for surface-energy calculations. Using this idea, computational framework is developed to correlate surface energies with the underlying atomic interactions of metallic nanoparticles, aiming at the reliable prediction of nanoparticle shapes. As a primary application, equilibrium Wulff shapes are generated for face-centered cubic (fcc) metallic nanoparticles, including Au, Al, Ag, Ni, Cu, Pd, Pt, and Pb. For each metal, Wulff constructions are obtained using different parameterizations of the Embedded Atom Method (EAM). The obtained results are compared with available experimental data, as well as with calculations based on Density Functional Theory (DFT) and the Modified Embedded Atom Method (MEAM). The comparison indicates that the EAM framework provides a favorable balance between computational efficiency and predictive accuracy, making it an attractive approach for modeling metallic nanoparticles. Finally, the dataset generated from the crystal geometries is utilized in a supervised machine learning study to develop a model capable of identifying the interatomic potential that most accurately describes the crystal under investigation based on its principal geometrical and energetic characteristics.

Abstract

This thesis studies damage-threshold estimation in ultrashort laser interactions with metallic thin films. The damage threshold is the incident laser fluence at which the film reaches melting onset. Estimating this threshold can be computationally expensive, because each evaluation requires two-temperature-model simulations and repeated threshold searches. This motivates surrogate models for faster threshold prediction.

The physical model is the axisymmetric parabolic two-temperature model (pTTM), which describes nonequilibrium energy exchange between laser-excited electrons and the lattice after femtosecond laser deposition. The work develops optical, thermophysical, and numerical components for threshold calculations, including temperature-dependent material response, ballistic and finite-thickness source corrections, implicit and IMEX time-integration schemes, and bracketing with bisection for threshold search. The numerical schemes are supported by empirical convergence and stability analyses.

The study uses one- and two-dimensional pTTM simulations to generate synthetic data for supervised regression. The surrogate target is the maximum lattice temperature, and threshold fluences are recovered by inverting the predicted temperature-fluence response. The surrogate analysis includes a two-dimensional linear pTTM dataset and a nonlinear two-dimensional extension with temperature-dependent optical properties, electron heat capacity, and electron-phonon coupling. LightGBM and multilayer perceptron models are evaluated on synthetic test sets and material-specific fluence cases.

The results show that the trained surrogates reproduce pTTM temperature targets within the sampled ranges and provide computationally efficient threshold estimates. For Ni and Ti, the nonlinear surrogate reduces the discrepancy between surrogate predictions and pTTM fluence-dependent calculations. Threshold estimates for Ag, Ni, and Ti are compared with reference values from the RIANA/PLAN-hub thermal-response calculator.

Abstract

The discovery of graphene in 2004 marked the emergence of a new class of materials known as two-dimensional (2D) materials. Due to their remarkable physical properties and wide range of applications in optoelectronics and photonics, 2D materials, and in particular the transition metal dichalcogenides (TMDs-with the formula MX₂, M standing for metal and X for chalcogenide) and the 2D monochalcogenides (MX), have become the subject of extensive research. This includes the development of optical characterization techniques, capable of providing 2D crystal structural information, at the atomic level.

In this study, we use the ability of 2D MX₂ and 2D MX to produce optical second harmonic generation (SHG) and we present a formalism based on nonlinear optics in crystals to extract the armchair direction, the twist-angle in MX₂ bilayers and probe the in-plane anisotropy in MX monolayers, by performing polarization-resolved SHG (P-SHG) measurements. Finally, we show that by tuning the twist angle in a MX₂/MX heterobilayer, the in-plane anisotropy of the material can be controlled.