Tag: Machine Learning

Debra J. Bernhardt joins GETCO2

Posted on by Anna Knudsen

We are excited to welcome Professor Debra J. Bernhardt to GETCO2 as a UQ Chief Investigator.

Debra is an ARC Australian Laureate Fellow in the Australian Institute for Bioengineering and Nanotechnology (AIBN) and School of Chemistry and Molecular Biosciences (SCMB) at The University of Queensland. She was the recent recipient of the RACI Margaret Sheil Leadership Award 2023.

Debra is taking over from Suresh Bhati, who is currently an Emeritus Professor at UQ and will convert from Chief Investigator to Associate Investigator with GETCO2. Suresh’s expertise and leadership have been invaluable in advancing GETCO2, and we appreciate his continuing involvement with GETCO2 as an Associate Investigator.

Debra Bernhardt will contribute renowned expertise in theoretical and computational methods for studying new materials and fluids, and she will be part of Research Theme 4 – Characterisation and Computation.

Curious about computation?

In this video, Debra explains how computational models can be used in the development of theory and computational methods for the study of molecular systems;

You can learn more about Debra’s research at the International Symposium on Green Transformation of Carbon Dioxide in Brisbane, 29 Nov – 1 Dec 2023, where she will present a plenary talk on Contributions of Simulation to Green Transformation of Carbon Dioxide.

Fast-tracking CO2 capture technologies with Machine Learning

Posted on by Anna Knudsen

Carbon capture and storage technologies play a crucial role in mitigating greenhouse gas emissions. One of the challenges is the search for the best material to store CO2 in the most efficient and inexpensive way.

GETCO2’s Associate Investigator Dr Babarao and his team at the Royal Melbourne Institute of Technology in collaboration with Dr Aaron Thornton from CSIRO have looked at a new class of materials called Metal-Organic Frameworks (MOFs) to capture CO2.

You can design these nano-materials in infinite ways, by tuning the shape and composition of the structures. However, optimisation can quickly become costly, both in terms of materials used and computational requirements in simulations. This calls for cheaper and faster methods of evaluating the growing list of candidates.

Tuning the building blocks

Using Machine Learning the research team has managed to dramatically reduce the time it takes to evaluate the materials and find the best candidates. One of the building blocks in the machine learning model is a descriptor that predicts which materials are best suited for CO2 capture. Dr Babaro’s team has developed a new descriptor, that significantly outperforms others, being hundreds of thousands of times faster. They call it the “Effective Point Charge (EPoCh)” descriptor.

The team aims to use this fast method to find suitable materials for CO2 capture and storage that the experimentalist can test and then scale up for commercial use.

This work is funded by the CSIRO Permanent Carbon Locking Future Science Platform and is recently published in Nature Communications Chemistry journal.

Meet the experts at ISGTCO2 29 in Brisbane this November

You can learn more about the research and meet Dr Ravichandar who will present at the International Symposium on Green Transformation of Carbon Dioxide.

For more information, please visit the ISGTCO2 symposium website.

Image:
From left: Supervisor Dr Aaron Thornton and Team Leader Dr Cara Doherty from CSIRO, Dr Ravichandar Babarao from RMIT and Director of CarbonLock FSP Dr Andrew Lenton from CSIRO.