Advanced machine learning decision policies for diameter control of carbon nanotubes

Rahul Rao,Pavel Nikolaev,Michael A Susner,Jennifer Carpena-Núñez,Kristofer G Reyes,Benji Maruyama

doi:10.1038/s41524-021-00629-y

Abstract

The diameters of single-walled carbon nanotubes (SWCNTs) are directly related to their electronic properties, making diameter control highly desirable for a number of applications. Here we utilized a machine learning planner based on the Expected Improvement decision policy that mapped regions where growth was feasible vs. not feasible and further optimized synthesis conditions to selectively grow SWCNTs within a narrow diameter range. We maximized two ranges corresponding to Raman radial breathing mode frequencies around 265 and 225 cm−1 (SWCNT diameters around 0.92 and 1.06 nm, respectively), and our planner found optimal synthesis conditions within a hundred experiments. Extensive post-growth characterization showed high selectivity in the optimized growth experiments compared to the unoptimized growth experiments. Remarkably, our planner revealed significantly different synthesis conditions for maximizing the two diameter ranges in spite of their relative closeness. Our study shows the promise for machine learning-driven diameter optimization and paves the way towards chirality-controlled SWCNT growth.

Highlights

The utilization of single-walled carbon nanotubes (SWCNTs) in a wide range of applications, including “Beyond Moore’s Law” computing[1], hinges on our ability to control their physical properties[2]
An machine learning (ML) planner based on the Expected Improvement (EI) decision policy was used to map the growth parameter phase spaces for maximizing the intensities of two radial breathing modes (RBMs) at 225 and 265 cm−1, which correspond to SWCNT diameters around 1.06 and 0.91 nm, respectively
The SWCNT growth experiments were performed in ARESTM

Summary

Introduction

The utilization of single-walled carbon nanotubes (SWCNTs) in a wide range of applications, including “Beyond Moore’s Law” computing[1], hinges on our ability to control their physical properties[2]. Control over SWCNT diameters is desirable for applications such as molecular transport[5], membranes[6], field emission[7], and sensing[8]. There are many published reports on diameter-controlled growth of SWCNTs by chemical vapor deposition (CVD), the most popular synthesis method. As the SWCNT diameter is defined primarily by the size of the catalyst particle[9,10], most of these reports focus on synthesis from catalysts with defined sizes or compositions[11,12,13,14,15,16]. Other reported alternatives for controlling SWCNT diameters involve careful selection of the growth substrates[17,18] and manipulation of the type and amount of hydrocarbon feedstocks[19,20,21,22]

Methods

Results

Conclusion