The Hip Laboratory
Kedar Hippalgaonkar's Materials by Design Lab
2022
​
Student: Sharath Ram Kumar (Best Student Award Recipient, TUM Asia Graduation Ceremony 2022)
Mentor: Jayce Cheng, Pawan Kumar
Project: High-Throughput Characterization of Thin-Film Thermoelectric Materials (Under Embargo)
Snippet: High-throughput experimentation is considered to be a promising avenue for the discovery of novel materials, especially in the fields of photovoltaics and thermoelectricity, where there is strong interest in materials such as perovskites and chalcogenides. This study describes a novel apparatus for the automated high-throughput characterization of the thermoelectric properties of thin-film semiconducting materials in an open-air environment, utilising a thermal camera and an electrical probe mounted to a 3-axis CNC machine. The instrument was evaluated against commercially available systems such as the ZEM-3, and the trade-offs in accuracy and ease-of-setup were quantified.
​
​
2022
​
Student: Benedict Anthony
Mentor: Jayce Cheng
Project: Performance Analysis of Object Detection Algorithms Using Small Training Datasets (PDF)
Snippet: Object detection using machine learning approach has seen wide adoption in virtually all known industries in the past decade. Much investment and research has been put into building the most accurate object detection algorithm. However, implementation of these algorithms is only accessible to organizations with vast amount of computing and data procurement resources. In this study, the correlation of overall detection rate, training time and training sample size will be explored.
​
​
Student: James Alexander Rodrigues
Mentor: Jatin Kumar, Andre Low
Project: Machine Learning & Automation to Accelerate Formulations for Personal Care & Cosmetics (PDF)
Snippet: This paper details the use of ML as an optimization tool in the context of personal care & cosmetics formulation by using data collected from physical samples, focusing on viscosity and pH changes, in order to identify possible interactions between ingredients and predict such changes. By leveraging on data science and optimization techniques, we attempt active learning to drive discovery of new products that optimize the aforementioned objectives.
​
​
Student: K Diviyah
Mentor: Balamurugan Ramalingam, Vijila Chellappan
Project: Studies on Surface Modification of ZnO by Natural Therapeutic Agents (PDF)
Snippet: Natural therapeutic ligands have been incorporated with ZnO nanoparticles to form ZnO nanocomposites via surface modification to improve their bioavailability as well as to provide a more localised delivery in vivo. Despite so, the roles of their functional groups in coordinating with ZnO nanoparticles to form such nanocomposites are not well documented. As such, this project narrowed down to several therapeutic ligands containing the functional groups of interest and studied how these functional groups played respective roles in coordinating with Zno nanoparticles.
​
​
Student: Lim Yi Jing Carina
Mentor: Lim Yee Fun, Jonathan Khoo, Albertus Denny Handoko
Project: Surface Effects of Functionalized Cu2O-Derived Cu(0) for C2+-Selective Electrochemical CO2 Reduction (Under Embargo)
Snippet: Selective electrochemical carbon dioxide reduction (CO2RR) to multi-carbon (C2+) products is an attractive method to close the carbon cycle as well as to provide a long-term, large-scale energy storage solution. With copper catalysts, since C2+ product formation is in direct competition with the formation of single-carbon products and hydrogen evolution, methods to modulate the product selectivity are highly desirable. In addition, surface charging effects are not commonly considered in CO2RR experiments. Functionalized cuprous oxide-derived copper synthesized via a simple wet chemistry approach was tested in a H-cell set-up and their surface properties were then studied with electrochemical impedance spectroscopy (EIS) and pulsed voltammetry (PV).
​
​
Student: Thong Jia Li Janessa
Mentor: Lim Yee Fun, Andre Low
Project: Optimizing Material Synthesis – Forward Surrogate Model Development and Bayesian Optimization (Under Embargo)
Snippet: Machine learning has become a popular technique to aid in materials synthesis, with Bayesian Optimization (BO) being a top choice for the exploration of materials design due to its sample efficient approach in evaluating expensive black-box functions. Real world optimization problems may be subject to constraints and consequently the optimization strategies should take that into account. However, the best methods of implementing constrained BO in open source BO libraries have not been studied widely. This study thus investigates the different methods of implementing constraints in Bayesian Optimization, and evaluates the performances of different Python libraries.
​
Student: Lee Wen Yang
Mentor: Beatrice Soh, Jatin Kumar, Aniket Chitre and Daniil Bash
Project: Machine learning and Automation to Accelerate Polymer Property Measurements (Under Embargo)
Snippet: In this study, a high throughput, proxy, experiment has been designed for viscosity screening. Accuracy of this experiment has been cross referenced with high fidelity data obtained from a research standard viscometer. The designed experiment has proven to be high-throughput, scalable, automated and capable of producing accurate viscosity readings.
​
​
Student: Wielly Halim
Mentor: Maheswar Repaka, Zhai Wenhao
Project: Designing High Performance Thermoelectric Materials (Under Embargo)
Snippet: Thermoelectric properties of CuIn1-xSbxTe2 (x=0, 0.03, 0.05, 0.07, 0.1) have been studied. The presence of Sb results in a rise in the carrier concentration from the presence of acceptor that create higher electrical conductivity and power factor. The thermoelectric efficiency is related to figure of merit, where it is proportional with Seebeck coefficient and electrical conductivity and inversely proportional with thermal conductivity. In this experiment, we want to decrease the thermal conductivity by increasing the anharmonicity by doping CuInTe2 system with Sb that will increase the phonon scattering. The phonon scattering will decrease the thermal conductivity of our materials. By doping Sb into CuInTe2, improved figure of merit will be observed for a certain doping concentration of Sb.
​
2021
​
Student: Andre Low (currently employed at Nanyang Technological University)
Mentor: Lim Yee Fun
Project: Applying Machine Learning and Optimization to High-Throughput Experimentation (PDF)
Snippet: This study investigates the effects of different hyperparameters and settings in performance of Bayesian and Particle Swarm Optimization, and in doing so develop best practices for their practical application.
​
​
Student: Frederick Hubert Chenardy (currently employed at Shopee Pte Ltd)
Mentor: Jayce Cheng
Project: Automated Probe Technique for Robust and High-Throughput Electrical Characterization and Measurements (PDF)
Snippet: This study seeks to explore and implement computer vision to automate an electrical probe technique to enhance the robustness and throughput of electrical characterization and measurements, such as the four-point probe sheet resistance and conductivity measurement and the field-effect transistor (FET) measurement.
Demo Videos: (1) (2) (3) (4) (5) (6)
​
​
Student: Chua Zhong Zhe
Mentor: Jatin Kumar
Project: Accelerated Viscosity Measurements of Polymer Solutions Using High Throughput Experimentation and Machine Learning (PDF)
Snippet: In this study, machine learning is used in conjunction with a proxy experiment to relate to high fidelity empirical data, to measure and predict the viscosity of materials in a rapid and high throughput fashion.
​
​
Student: Saif Ali Khan (currently employed at DBS Bank Ltd)
Mentor: Vijila Chellappan
Project: Analysing Hyperspectral Image features to Optimize Degree of Crystallinity in Molecularly- doped P3HT Thin Films (PDF)
Snippet: In this study, we show the correlation of the crystallinity of conjugated polymer films to its electrical performance. We use hyperspectral imaging (HSI) as a proxy tool to investigate the crystallinity of the material from its absorbance spectra. Hyperspectral imaging is a new spectroscopic tool that offers vast amount spectral and spatial information that are being exploited in many fields ranging from medical to food technology.
​
​
Student: ​Denish Kumar (currently employed at The Univac Group)
Mentor: Maheswar Repaka
Project: Thermoelectric Properties of CuIn1-x AxTe2 (A- Al, Sn, Sb and Bi) (PDF)
Snippet: We synthesized the CuIn1-x AxTe2 compounds through solid state reaction followed by spark plasma sintering. The electrical and thermal properties were studied using ZEM-3 and PPMS. We achieved the high power factor with Sb doping among the prepared samples.
​
​
Student: Na Shi Ching
Mentor: Maheswar Repaka
Project: Electrical and thermal transport in thermoelectric, skutterudites, CoSb3 (PDF)
Snippet: Thermoelectric energy harvesting has come into the spotlight in recent years to achieve standards amongst other type of green technology in obtaining sustainable energy. In-filled and Te-doped Co4Sb12 skutterudites were produced via solid-state synthesis and their thermoelectric properties were investigated from 300K to 800K.
​
​
Student: Abdul Rauf Bin Mohamed Raffi
Mentor: Pawan Kumar
Project: Thermoelectric Transport in Conducting Polymers (PDF)
Snippet: In this project, different dopants with varying concentration are used to study the thermoelectric properties of P3HT and PBTTT. Sequential doping is used to control the doping level of the polymers. The effectiveness of the doping is highlighted by doing a UV-vis test. The seebeck coefficient and electrical conductivity were measured using a new lithography-free resistance thermometry-based technique.
2021
​
Student: Bariq Hibatullah Nurlis
Mentor: Lim Yee Fun, Vijila Chellappan
Project: Machine Learning-assisted Hyperspectral Image Analysis (PDF)
Snippet: Hyperspectral imaging is a process to collect and process information from across the electromagnetic spectrum. Increasingly, researchers are using the spectral image to identify material’s properties. However, a direct measurement of these data to the property maybe unclear due to numerous parameters that are affecting the formation of a spectral image. The current study examines the spectral data using machine learning to calculate material’s electrical conductivity.
​
​
Student: Arutselvan Pranav Ganesh​
Mentor: Kedar Hippalgaonkar
Project: Colour Matching Using Bayesian Optimization (PDF)
Snippet: The project aimed to achieve a target colour using the Opentrons OT-2 auto-pipetting instrument to mix food colouring solutions of various volumes. These volumes were dictated by the Bayesian Optimization algorithm, which efficiently searches for the optimal input volumes of dye solutions such that the colours match.
​