KNU Department of Physics Education Team led by Sooran Kim, Discovers Machine-Learning-Guided Prediction Models of Critical Temperature of Cuprates
- Date
- 2021/08/02
- Writer
- Oh
- Hit
- 1120
KNU Department of Physics Education Team led by Professor Sooran Kim collaborated
in joint research with Harvard University School of Engineering and Applied
Sciences Team to develop a model for predicting the critical temperature of
cuprate superconductors via machine-learning and Ab initio quantum chemistry
methods, and to guide the experimental search for new cuprate superconductors. The
paper on these results was published on the front page of the highly renowned
journal in physics, “The Journal of Physical Chemistry Letters” on July 8. The
first author for the paper was Dongeon Lee, KNU Department of Physics Education.
Cuprate superconductors are known to be substances with the highest
experimental maximum superconducting transition temperature (Tc,max), although
their underlying mechanism of superconductivity remains elusive. Thus, in this
study, KNU Department of Physics Education Team develops ML models for
predicting Tc,max of hole-doped cuprates by using material-dependent parameters
beyond simple composition features. Additionally in the search for new cuprate
candidates, the team has found that cuprates with Ga as an apical cation
exhibited Tc,max comparable to that of cuprates with Hg as an apical cation, suggesting
that machine learning could guide the design of high-Tc superconductors in the
future.
Professor Sooran Kim states, “This research shows significant evidence of
high accuracy for prediction models of the critical temperature of cuprates via
machine learning and Ab initio quantum chemistry methods. Moreover, it assists
in developing a quantitative understanding of the underlying mechanism of
cuprates as well as in guiding the search for new cuprate superconductors. In
fact, we are undertaking a follow-up study with machine learning on a different
superconductor at present.”