Nano-TCAD (Luisier)Open OpportunitiesWe are looking for motivated students to develop and implement algorithms within a novel, machine-learning (ML) enhanced, atomistic simulation methodology for electrochemical phenomena in memristors. These algorithms include, but are not limited to: (1) calculation of the atomically-resolved electrostatic potential, (2) efficient integration of ML predictions with molecular dynamics (MD), or (3) high-throughput screening of material stacks. - Electrical and Electronic Engineering, Interdisciplinary Engineering, Physics
- Bachelor Thesis, Master Thesis, Semester Project
| In a typical quantum transport simulation, a large system of linear equations has to be solved for a great number of energy points. Especially in our newly developed ab initio NEGF+self-consistent GW model, not only a small grid resolution but also using a wide range of energy values is critical. Using a uniform energy grid with the resolution dependent on the critical part of the bandstructure can make calculations extremely expensive. It is therefore desirable to tune a local resolution of the energy grid with an adaptive algorithm, which is equivalent to a one-dimensional mesh refinement problem. Firstly, different adaptive algorithms should be investigated and implemented. An additional challenge is performing a non-uniform Fast-Fourier Transform (NuFFT) as well as redistributing the computational load in a shared- or distributed memory system. - Electrical Engineering
- Bachelor Thesis, ETH Zurich (ETHZ), Master Thesis, Semester Project
| Electrostatics are at the heart of the operating principle of nanoscale devices, like Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs). By modulating a gate potential, the channel of the device can be switched between an off- and an on-state. It is therefore critical to accurately describe device electrostatics, governed by Poisson’s equation. Due to the irregular geometry of many devices, the finite element method (FEM) is the preferred numerical approach. In our group we are collaboratively developing a next-generation quantum transport (QT) simulator (QuaTrEx) with emphasis on modern high performance computing (HPC) principles. Including a Poisson solver in this novel code will open up many more possibilities for device and material simulation. The development will start with a general approach to set up and solve a system of linear equations. Afterwards, iterative and machine-learning enhanced methods may be investigated. - Electrical Engineering
- Bachelor Thesis, ETH Zurich (ETHZ), Master Thesis, Semester Project, Studies on Micro and Nano Systems (ETHZ)
|
|