Abstract:It has been a great challenge to optimize the growth conditions toward structure-controlled growth of single-wall carbon nanotubes (SWCNTs). Here, a high-throughput method combined with machine learning is reported that efficiently screens the growth conditions for the synthesis of high-quality SWCNTs. Patterned cobalt (Co) nanoparticles were deposited on a numerically marked silicon wafer as catalysts, and parameters of temperature, reduction time and carbon precursor were optimized. The crystallinity of the SWCNTs was characterized by Raman spectroscopy where the featured G/D peak intensity (IG/ID) was extracted automatically and mapped to the growth parameters to build a database. 1,280 data were collected to train machine learning models. Random forest regression (RFR) showed high precision in predicting the growth conditions for high-quality SWCNTs, as validated by further chemical vapor deposition (CVD) growth. This method shows great potential in structure-controlled growth of SWCNTs.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve The paper aims to optimize the growth conditions of single-walled carbon nanotubes (SWCNTs) to achieve structurally controlled high-quality SWCNT synthesis. Specifically, the study efficiently screens the growth conditions required for synthesizing high-quality SWCNTs by combining high-throughput screening methods and machine learning techniques. In the traditional chemical vapor deposition (CVD) process, there are at least 12 growth parameters, which together form a vast parameter space (10^12 possible combinations). The traditional trial-and-error method is time-consuming and inefficient. Therefore, this study proposes a new method to optimize these parameters, thereby improving the quality and yield of SWCNTs. ### Main Content Summary 1. **High-Throughput Growth Experiments**: SWCNT growth experiments were conducted by depositing cobalt catalyst nanoparticles of different thicknesses on silicon wafers and varying growth temperature, reduction time, and carrier gas flow rate. 2. **Raman Spectroscopy Characterization**: Raman spectroscopy was used to measure the quality of SWCNTs (through IG/ID values), and a database containing 1,280 sets of data was established. 3. **Machine Learning Modeling**: A random forest regression (RFR) model was used to model the relationship between growth conditions and SWCNT quality. The results showed that the RFR model had the highest prediction accuracy (R²=0.88). 4. **Experimental Validation**: Further CVD growth experiments were conducted to verify the predicted optimal growth conditions, confirming their effectiveness. ### Conclusion The combination of high-throughput screening and machine learning methods significantly improved the efficiency of optimizing high-quality SWCNT growth conditions. This method demonstrates potential in the controlled synthesis of SWCNTs with specific structures and properties.

High-throughput screening and machine learning for the efficient growth of high-quality single-wall carbon nanotubes

Water-Assisted Preparation of High-Purity Semiconducting (14,4) Carbon Nanotubes

Selective growth of semiconducting single-walled carbon nanotubes solely from carbon monoxide

Growth Of High-Quality Single-Wall Carbon Nanotubes Without Amorphous Carbon Formation

Dynamics of growing carbon nanotube interfaces probed by machine learning-enabled molecular simulations

Optimizing the growth of vertically aligned carbon nanotubes by literature mining and high-throughput experiments

Growth of high-density single-wall carbon nanotubes with a uniform structure using a CoRu catalyst

Carbon fiber-promoted activation of catalyst for efficient growth of single-walled carbon nanotubes

Thin-Film Metal Catalyst For The Production Of Multi-Wall And Single-Wall Carbon Nanotubes

Statistical patterns in high-throughput growth of single-wall carbon nanotubes from Co/Pt/Mo ternary catalysts

Growth Of High-Density-Aligned And Semiconducting-Enriched Single-Walled Carbon Nanotubes: Decoupling The Conflict Between Density And Selectivity

Metal-catalyst-free growth of single-walled carbon nanotubes.

Growth of Single-Walled Carbon Nanotubes from Solid Carbon Nanoparticle Seeds via Cap Formation Engineering with a Two-Step Growth Process and Water Vapor Supply

Scalable growth of single-walled carbon nanotubes with a highly uniform structure

Self-Aligned Growth and Optimization of Carbon Nanotube

Growth of High-Density Horizontal SWNT Arrays by Pre-Cracking of Carbon Source

CVD Growth of Carbon Nanotube Forest with Selective Wall-Number from Fe–Cu Catalyst

Growth of Single-Walled Carbon Nanotubes with Controlled Structure: Floating Carbide Solid Catalysts

Growth of Single-Walled Carbon Nanotubes from Ag-15 Cluster Catalysts

Direct Growth of Single-Walled Carbon Nanotubes without Metallic Residues by Using Lead as a Catalyst

Nearly Exclusive Growth of Small Diameter Semiconducting Single-Wall Carbon Nanotubes from Organic Chemistry Synthetic End-Cap Molecules