Chiral DNTs with methyl groups in a nanocomposite, i.e. DNT_C-CH₃, exhibit the highest interfacial thermal conductance of 0.115 GW m⁻² K⁻¹, representing a 140% improvement over CNTs.

Concrete is a highly heterogeneous composite material, and accurately predicting its elastic modulus remains a major challenge in mechanical analysis. To address this, this study systematically investigates the predictive performance of several classical homogenization methods for estimating the effective elastic modulus of concrete, including the dilute approximation, self-consistent method, generalized self-consistent method, Mori–Tanaka model, differential method, as well as the Voigt and Reuss models. To enhance prediction accuracy, an improved computational framework is proposed based on an iterative strategy that enables dynamic updating of model parameters. This approach combines principles of mesomechanics with numerical simulation techniques and is implemented using Mathematica for both symbolic and numerical computations. The performance of the models is evaluated under varying aggregate volume fractions and aggregate-matrix stiffness combinations, and validated using multiple experimental datasets from the literature. The results show that the iterative strategy significantly improves the predictive accuracy of several models, reducing the maximum error by up to 30%. Further analysis indicates that the dilute method performs best at low aggregate volume fractions, the Mori–Tanaka model yields the most accurate results when the aggregates are stiff and moderately concentrated, and the generalized self-consistent method outperforms the standard version when the elastic moduli of the aggregate and matrix are similar.

CNT-modified asphalt nanocomposites exhibit superior pullout performance compared to those reinforced with DNTs, primarily due to π–π stacking interactions that facilitate the wrapping of asphalt molecules around the CNT surfaces.