Seminar by Dr. Le Dai Nam and MSc. Phan Anh Luan

On July 19, 2024, the lectures from Dr. Le Dai Nam and MSc. Phan Anh Luan takes place on the 5th Floor of the Library with detailed content as follows

Dr. Le Dai Nam presents about "Phonon-assisted Casimir interactions: the case of piezoelectric materials"

Abstract:

The strong coupling between electromagnetic field and lattice oscillation in piezoelectric materials gives rise to phonon polariton excitations. The coupling between such quasiparticles and the electromagnetic field opens up new directions to modulate the ubiquitous Casimir force. Here by utilizing the generalized Born-Huang hydrodynamics model, three types of phonons in piezoelectrics are studied: longitudinal optical phonon, transverse optical phonon and interface phonon polariton. As a result of the coupling between these phonons and the electromagnetic field, the electromagnetic and elastic boundary conditions result in a complex set of Fresnel reflection matrices which prevents the utilization of the standard Lifshitz approach for calculating Casimir forces in the imaginary frequency domain. Our calculations are based on an approach within real frequency and finite temperatures, through which various regimes of the Casimir interactions are examined. Our study shows that piezoelectrics emerge as a set of materials where the ubiquitous Casimir force can be controlled via phonon properties for the first time.

MSc. Phan Anh Luan presents about " Empirical Tight-Binding simulations for disordered semiconductor alloys"

Abstract:

The semiconductor alloys are one of the important approaches to open up more design space for novel electronic and optoelectronic devices, thus understanding the impacts of disorder in alloyed configurations is crucial. From the theoretical side, this involves the need of theoretical models that are able to properly incorporate the alloy disorder's effects into calculations of the electronic and optical properties of the materials/devices. I will try to convince the audience that the empirical tight-binding (ETB) method is a good tool for this task because it can well balance between the computational demand and the required accuracy. Considering group-IV and III-VI semiconductor materials only, one of the most widely used schemes was proposed by Jancu et al. [1] in 1998. However, Jancu's scheme and its variants/modifications fail to treat strain effects in random alloys and thus wrongly reproduce the alloy’s band gap in some cases. I will discuss two approaches in our group to resolve this limitation. The first one employs the power of Machine Learning to add the proper corrections to the onsite energies for each ion by learning from reference DFT targets. On the other hand, a physics-based approach backed by a recently proposed scheme of Tan et al. [2] addresses the limitation of Jancu’s scheme by adding parameters accounting for the corrections from the multipole expansion analysis of the atomic potentials to both onsite and hopping energies. We will analyze the pros and cons of each approach and then demonstrate practical application of ETB in large-supercell simulations for GaAsSb alloy as an example.