Preprint Technical Note Version 1 Preserved in Portico This version is not peer-reviewed

Data-Driven Inverse Design of Low-Dimensional Nanocarbons: Revealing Hidden Growth-Properties Relationships and Identifying Universal Descriptors

Version 1 : Received: 25 December 2023 / Approved: 27 December 2023 / Online: 27 December 2023 (05:53:15 CET)

How to cite: Lukin, A.; Gülseren, O. Data-Driven Inverse Design of Low-Dimensional Nanocarbons: Revealing Hidden Growth-Properties Relationships and Identifying Universal Descriptors. Preprints 2023, 2023122057. https://doi.org/10.20944/preprints202312.2057.v1 Lukin, A.; Gülseren, O. Data-Driven Inverse Design of Low-Dimensional Nanocarbons: Revealing Hidden Growth-Properties Relationships and Identifying Universal Descriptors. Preprints 2023, 2023122057. https://doi.org/10.20944/preprints202312.2057.v1

Abstract

Recent advances in nanomaterials have been heavily influenced by low-dimensional nanocarbon allotropes. In particular, carbyne has attracted attention for its potential as a true one-dimensional carbon chain with sp1 hybridization. To maximize the capabilities of this material, we employ a focused data-driven inverse design approach based on the carbon nanomaterials genome concept. This involves using deep learning neural network models to identify key descriptors tied to desired properties, enabling property prediction and reverse engineering of nanocarbons. Our iterative approach entails: (i) gathering growth/property data on nanostructures; (ii) identifying informative numerical/categorical predictors; (iii) developing deep learning models mapping descriptors to properties; (iv) refining models with new insights; (v) determining required descriptors/conditions for target properties via inverse mapping; (vi) validating models by synthesizing predicted nanostructures; and (vii) enhancing models with validation data. This allows uncovering hidden growth-property connections, precisely tuning nanocarbons for desired attributes. We introduce new methodologies including exciting synergistic effects, synchronizing atomic vibrations, active screen plasma, energy-driven transformations, surface acoustic micro/nano-manipulation, doping and directed self-assembly to expose relationships and integrate insights into the inverse design flow. This research promises to accelerate discovery of next-generation low-dimensional nanocarbons with exceptional properties and applications.

Keywords

low-dimensional nanocarbons; machine learning-powered inverse design; data-driven carbon nanomaterials genome approach; deep learning neural networks; multifactorial neural network predictive models; growth-property relationships; universal descriptors; collective atomic vibrations; nano-enhanced interfaces

Subject

Chemistry and Materials Science, Nanotechnology

Comments (0)

We encourage comments and feedback from a broad range of readers. See criteria for comments and our Diversity statement.

Leave a public comment
Send a private comment to the author(s)
* All users must log in before leaving a comment
Views 0
Downloads 0
Comments 0


×
Alerts
Notify me about updates to this article or when a peer-reviewed version is published.
We use cookies on our website to ensure you get the best experience.
Read more about our cookies here.