Even with the impressive advances in chip- or wafer-scale growth of two-dimensional (2D) crystals, device research on 2D materials still currently significantly relies on the mechanical exfoliation technique to realize flakes of 2D materials for building discrete electronic devices. This is primarily due to the superior quality of the exfoliated flakes, i.e. single crystals, room temperature process, ability to transfer flakes on arbitrary substrates, and create 2D heterostructures, etc. A bottleneck of this process is the manual effort needed to search and identify nanometer-scale thick flakes on the transferred substrate. There have been considerable advances in understanding the thickness-dependent optical properties of 2D flakes, which pave the way for developing a technique to automate the 2D material identification process, which is the objective of this proposal.
Methodology and implementation: The major steps of the automation process are the following: (a) capture optical micrographs of the substrate with flakes as image files; (b) develop an algorithm to scan the images and detect 2D flakes based on based on color, contrast, and/or brightness, etc.; (c) locate the position of the flake by identifying pre-defined location markers on the substrate in a process similar to (b); (d) Using the location information, overlay the image of the flakes or their outline on a CAD layout that can consequently be used for designing electrical devices on the flakes. Step (a) can be either automated, dependent on equipment-availability, or manual. The algorithm in step (b) can be constructed based on either physics-based models, or, look-up tables based on calibration standards from previous measurements.
The student will work with the 2D materials group, primarily with experimentalists in the ED and NAME teams in a closed loop process to develop the automation. The physical model and calibration data for the flake detection algorithm will be provided to the student who is then expected to build a baseline program. Verifications and refinements of the algorithm and the program will be sought periodically with new experimental data.
Necessary skills, and development: Good programming skills in any of the high-level languages such as Python, R, Matlab, etc. Some experience in image processing is desirable. At the end of the internship, apart from a strengthened skill set in algorithms/programming, the student can expect to build a good understanding of the procedures in nanofabrication of solid-state devices.
Impact: The proposed automation will benefit a multiple of IMEC projects on 2D materials, especially on device engineering such as 2D-based TFETs, doping and contact studies on 2D-MOSFETs, etc. Devices with 2D materials is an active international research topic, and publication of a sufficiently accurate and robust flake identification program is likely to be of great interest to researchers worldwide. Also, the scope of the project is not limited to 2D materials alone, but extends to all research on novel materials that rely on exfoliation or the "bottoms-up" approach to device fabrication on such materials.
 Novoselov, K.S., Geim, A.K., Morozov, S.V., Jiang, D., Zhang, Y., Dubonos, S.V., Grigorieva, I.V. and Firsov, A.A., 2004. Electric field effect in atomically thin carbon films. science, 306(5696), pp.666-669.
Type of Project: Internship; Combination of internship and thesis
Master's degree: Master of Engineering Technology; Master of Engineering Science
Duration: 3 to 6 months
Master program: Computer Science; Electrotechnics/Electrical Engineering; Nanoscience & Nanotechnology
Supervising scientist: For further information or for application, please contact Surajit Sutar (Surajit.Sutar@imec.be).
Imec allowance will be provided.