Governing Factors for Carbon Nanotube Dispersion in Organic Solvents Estimated by Machine Learning

Abstract

The insolubility of single‐walled carbon nanotubes (SWCNTs) in most common organic solvents has been the cause of a bottleneck in their practical utilization. Aqueous SWCNT inks containing amphiphilic surfactants are widely used for processing including coatings and composite fabrication. Most practical processes are, however, designed to be compatible with organic solvents, generating a technological mismatch between production and utilization. This work reports on the surfactant‐assisted dispersion of SWCNTs in useful organic solvents, at up to quantitative yields. A feature extraction based on machine learning offers seemingly important, highly intuitive physicochemical factors that lead to efficient dispersion. These elucidated factors are associated not only with solvents–surfactant, but also solvent–SWCNT interactions. The organic SWCNT dispersion as well as its research methodology developed here may find widespread applications ranging from nanofluidics to functional materials design.

Dye adsorption-assisted colloidal dispersion of single-walled carbon nanotubes in polar solvents

“…Cellulosic polymers and some types of vinyl polymers have also been reported to aid in the colloidal dispersion of SWCNTs in arbitrary organic solvents. 34–36 A recent study based on artificial intelligence-assisted feature extraction suggested that weak interactions such as hydrogen bonding could be a key factor in the design of efficient surfactants. 36…”

“…34–36 A recent study based on artificial intelligence-assisted feature extraction suggested that weak interactions such as hydrogen bonding could be a key factor in the design of efficient surfactants. 36…”

“…Cellulosic polymers and some types of vinyl polymers have also been reported to aid in the colloidal dispersion of SWCNTs in arbitrary organic solvents. [34][35][36] A recent study based on artificial intelligenceassisted feature extraction suggested that weak interactions such as hydrogen bonding could be a key factor in the design of efficient surfactants. 36 Compared to polymer surfactants, there are few studies on surfactants derived from low-molecular-weight molecules for efficient SWCNT dispersion, [24][25][26][27][28]37,38 except for those on well-known micellar solutions.…”

Machine Learning-Assisted Exploration and Identification of Aqueous Dispersants in the Vast Diversity of Organic Chemicals

“…There have only been a few reports addressing complex systems, such as the fabrication of dispersions and interaction with other materials, where the dispersion outcome is inseparably governed by the fabrication factors. For example, Nonoguchi et al reported a ML study of the efficacy of various organic solvents for SWCNT dispersion using ethylcellulose as a dispersant . In addition, Nakada et al reported a ML model to analyze and predict new conditions for the efficient exfoliation of two-dimensional interlayer materials .…”

“…For example, Nonoguchi et al reported a ML study of the efficacy of various organic solvents for SWCNT dispersion using ethylcellulose as a dispersant. 20 In addition, Nakada et al reported a ML model to analyze and predict new conditions for the efficient exfoliation of two-dimensional interlayer materials. 21 Furthermore, Glaubitz et al reported an attempt to determine the relationship between particle size and ultrasonic irradiation conditions in the liquid state using ML to ensure the reproducibility of ultrasonic treatment in nanoparticle dispersion.…”