Mastering carbon nanotube dispersion: A simplified model for industrial and environmental innovation

Carbon nanotubes (CNTs) are celebrated for their extraordinary mechanical, electrical, and thermal properties, yet their industrial adoption remains hindered by aggregation issues. Achieving stable dispersion in organic solvents is critical for unlocking their potential in advanced composites, flexible electronics, energy storage, and environmental remediation. Current quantitative structure-property relationship (QSPR) models for predicting CNT dispersibility rely on computationally intensive descriptors, such as quantum-chemical or topological parameters, which limit their practical accessibility. This study introduces a streamlined predictive model that uses only three intuitive solvent descriptors—hydrogen-bonding capacity, hydrophobicity, and a novel π-π interaction parameter—to achieve exceptional accuracy (training r² = 0.917, external validation r² = 0.963) and precision (RMSE = 0.236 vs. 0.337 for prior models). Innovations include leveraging amine/amide functional groups for stabilization and eliminating dependence on complex computational tools. The model’s robustness is validated through rigorous statistical testing (leave-many-out cross-validation q² = 0.823) and applicability domain analysis. By prioritizing simplicity without compromising performance, this work bridges the gap between lab-scale nanotechnology research and scalable industrial applications, such as water purification and pollution remediation, offering a user-friendly alternative to traditional QSPR frameworks.