Machine Learning Helps To Discover New Polymers Which Can Be Used For 5G Connectivity

There was always a need of such a robust polymer which contains tremendous thermal conductivity, a property that would be instrumental in heat management, for instance, in the fifth-generation mobile communication technologies that are the 5G.

But due to many reason development of this kind of material was not feasible. Now a study published by Japanese researchers demonstrates the successful discovery of new polymers having very high thermal conductivity, inspired by machine-learning-assisted polymer chemistry.

The study points out how successful machine learning models that are trained to “learn” from data provided to find new materials can be over traditional ways of hunting for high-performance materials.

This ML model is trained on a substantially limited amount of polymeric properties data, expertise from laboratory synthesis and advanced technologies for thermophysical property measurements.

The researcher said, “ML for polymer design is a challenging but promising field as these materials have properties that differ from metals and ceramics, and are not yet fully predicted by the existing theories.”

The team of researchers includes Ryo Yoshida, Junko Morikawa, and Yibin Xu.

Ryo Yoshida is a Professor and Director of the Data Science Center for Creative Design, Junko Morikawa is a Professor at the School of Materials and Chemical Technology and  Yibin Xu is a Group Leader of Thermal Management and Thermoelectric Materials Group.

The research was conducted by Statistical Mathematics (ISM), Research Organization of Information and Systems, Tokyo Institute of Technology and Center for Materials Research by Information Integration, Research and Services Division of MaDIS.

To create this ML system the researchers used Bayesian molecular design algorithm trained to recognize quantitative structure-property relationships with respect to thermal conductivity and other targeted polymeric properties.

The Bayesian molecular design process generated a library of virtual chemical structures, that was produced by targeting the glass transition and melting temperatures and then proceeded with laboratory synthesis and experimental characterization of the thermophysical properties.

The polymers with higher glass transition temperatures tend to be achieved by rigid structures, which result in higher thermal conductivity.

The team used the transferred thermal conductivity model to screen promising candidates over the virtual library. An ML framework referred to as “transfer learning” was introduced to obtain a thermal conductivity model with the given small data set.

For the given target property to be predicted from the limited supply of data, models on physically related proxy properties were pre-trained using an adequate amount of data, which captured common features relevant to the target task of predicting thermal conductivity.

The team integrated this model with a particularly designed ML algorithm for computational molecular design, which is referred to as the iQSPR algorithm formerly developed by Yoshida and his colleagues.

Applying this method allowed the identification of numerous promising “virtual” polymers. The model was able to identify new molecules and was also helping to mitigate the issue of limited data.

The model is based on a data set of polymeric properties from PoLyInfo, the leading database of polymers in the world located at NIMS. Regardless of its massive size, PoLyInfo has a limited quantity of data on the heat transfer properties of polymers.

Finally, three chemical structures were selected from a list of 1000 designed candidates on the basis of criteria involving synthetic accessibility and ease of processing.

Then, the monomers of these candidates were synthesized and polymerized using retrosynthetic routes designed by synthetic chemists. The synthesized polymers exhibited a glassy state, and two of them were crystallized by annealing.

During the test, the new polymers possess a high thermal conductivity of up to 0.41 W/mK.

This number is 80% higher than that of common polyimides, a group of frequently used polymers that have been mass-produced since the 1950s for applications spanning from cookware to fuel cells.

Junko Morikawa, Professor, School of Materials and Chemical Technology, Tokyo Institute of Technology said, we would like to try to create an ML-driven high-throughput computational system to design next-generation soft materials for applications going beyond the 5G era.

Through our project, we aim to pursue not only the development of materials informatics but also contribute to fundamental advancement of materials science, especially in the field of phonon engineering, he further added.

(You can read the full paper here.)

More in AI

SoftBank Launches New $108 Billion Vision Fund To Invest in AI

Python 3.8 (B3) to release on July 29 | Here’s new features & what you can expect?

FaceApp That Can Make You Old, How Actually Does It Work?

DeepNude An AI App That “Undressed” Women Show How Harmful DeepFakes Is

Leave a Reply

Your email address will not be published.