UROP Proceedings 2020-21

School of Science Department of Physics 67 The Geometry and Physics of Wrinkling Supervisor: ALTMAN Michael Scott / PHYS Student: CUI Zhenhao / PHYS-IRE Course: UROP2100, Fall As the surface potential of sample can affect the imaging of low energy electron microscopy (LEEM), we can use it to do the surface potentiometry analysis. Here we use it to investigate the surface potential of graphene on electrode made with copper (Cu), silicon nitride (SiN) and free-standing graphene. We report that the voltage drops faster at free-standing graphene when compare to the graphene on SiN. Besides, the current tends to start getting into the graphene near the boundary of SiN and Cu, but not the boundary of graphene. And the surface potential shows a jump which increases linearly with the applied transverse current at that boundary. Evidence for Isotropic s-Wave Superconductivity in High-Entropy Alloys Supervisor: JAECK Berthold / PHYS Student: LEUNG Ka Wun Casey / PHYS-IRE Course: UROP1000, Summer High-entropy alloys (HEA) form through the random arrangement of five or more chemical elements on a crystalline lattice. Despite the significant amount of resulting compositional disorder, a subset of HEAs enters a superconducting state below critical temperatures Tc<10 K. The superconducting properties of the known HEAs seem to suffice a Bardeen-Cooper-Schrieffer (BCS) description, but little is known about their superconducting order parameter and the microscopic role of disorder. We report on magnetic susceptibility measurements on films of the superconducting HEA (TaNb)1−x(ZrHfTi)x for characterizing the lower and upper critical fields Hc1(T) and Hc2(T), respectively as a function of temperature T. Our resulting analysis of the Ginzburg-Landau coherence length and penetration depth demonstrates that HEAs of this type are single-band isotropic s-wave superconductors in the dirty limit. Despite a significant difference in the elemental composition between the x=0.35 and x=0.71 films, we find that the observed Tc variations cannot be explained by disorder effects [1]. [1] Casey K.W. Leung, Xiaofu Zhang, Fabian von Rohr, Rolf Lortz, and Berthold Jäck, arXiv:2111.11013 [condmat.supr-con] (2021). Machine Learning Aided Detection of the Quantum State in an Atomic Quantum Simulator Supervisor: JO Gyu Boong / PHYS Student: MAK Ting Hin / PHYS-IRE Course: UROP3100, Fall A convolutional neural network (CNN) is applied to identify the topological quantum phase transition in a system of 173Yb atoms trapped in an one-dimensional optical Raman lattice, based on the insights provided in the previous reports. The CNN is able to identify both experimental transition points by plotting its predictions on a test dataset. Visualizing the filters of the convolution layers suggests that the trained CNN classifies by letting each filter Fk convert the datapoint to a single quantity Gk, which reflects the topological state of the system. The results suggests that it is possible to understand how a trained CNN works by properly analyzing the model.