Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Automated identification of two-dimensional crystals based on neural network
The goal of this project is to develop an automated setup that is able to identify monolayer flakes based on a recently proposed approach using neural networks. A software code will be developed that is capable of processing images and controlling a HW setup scanning the sample on a holder.
Two-dimensional (2D) materials have emerged as a promising class of layered materials
for optoelectronic applications and discovering novel physical phenomena [1]. Different
materials can be combined in heterostructures with tailored optical and electrical properties.
The assembly of such devices relies on the so-called mechanical exfoliation technique
that also led to the discovery of graphene. In this method, micrometer sized flakes
of 2D materials are exfoliated onto silicon/silicon-oxide (Si/SiO2) substrates that enables
flakes of different thickness to be distinguished easily in an optical microscope by their
difference in contrast.
The goal of this project is to develop an automated setup that is able to identify monolayer
flakes (specifically monolayer graphene and TMDCs) based on a recently proposed
approach using neural networks [2]. A software code will be developed that is capable of
processing images from the optical microscope as well as controlling a hardware setup
that scans the stage with the sample holder.
References:
[1] F. Xia et al., “Two-dimensional material nanophotonics”, Nat. Photon. 8, 899-907 (2014).
[2] E. Greplova et al., “Fully Automated Identification of Two-Dimensional Material Samples”,
Phys. Rev. Applied 13, 064017 (2020).
Prerequisites:
Basic knowledge of computer programming and image processing, interest in machine learning.
Two-dimensional (2D) materials have emerged as a promising class of layered materials for optoelectronic applications and discovering novel physical phenomena [1]. Different materials can be combined in heterostructures with tailored optical and electrical properties. The assembly of such devices relies on the so-called mechanical exfoliation technique that also led to the discovery of graphene. In this method, micrometer sized flakes of 2D materials are exfoliated onto silicon/silicon-oxide (Si/SiO2) substrates that enables flakes of different thickness to be distinguished easily in an optical microscope by their difference in contrast. The goal of this project is to develop an automated setup that is able to identify monolayer flakes (specifically monolayer graphene and TMDCs) based on a recently proposed approach using neural networks [2]. A software code will be developed that is capable of processing images from the optical microscope as well as controlling a hardware setup that scans the stage with the sample holder.
References: [1] F. Xia et al., “Two-dimensional material nanophotonics”, Nat. Photon. 8, 899-907 (2014). [2] E. Greplova et al., “Fully Automated Identification of Two-Dimensional Material Samples”, Phys. Rev. Applied 13, 064017 (2020).
Prerequisites: Basic knowledge of computer programming and image processing, interest in machine learning.
Not specified
Supervisor:
Lujun Wang (lujwang@ethz.ch), Lukas Novotny (lnovotny@ethz.ch)
Supervisor: Lujun Wang (lujwang@ethz.ch), Lukas Novotny (lnovotny@ethz.ch)