Title: Statistical analysis of effective electro-mechanical properties and percolation behavior of aligned carbon nanotube/polymer nanocomposites via computational micromechanics
Krishna Talamadupula and Gary D Seidel
Computational Materials Science -- 2021 -- Vol. 197, 110616-32pgs
Abstract
CNT (carbon nanotube) embedded polymer nanocomposites exhibit significant variation in their electromechanical
properties depending upon factors such as CNT volume fraction, CNT dispersion, CNT alignment
and properties of the polymer. Of interest is electrical percolation where the electrical conductivity of the CNT/
polymer nanocomposite increases several orders of magnitude with increase in CNT volume fraction. Estimates
and distributions for the electrical conductivity and piezoresistive coefficients of the CNT/polymer nanocomposite
are obtained and analyzed with increasing CNT volume fraction and varying barrier potential, a
parameter that controls the extent of electrical tunneling. The barrier potential is seen to play a key role in
determining electrical percolation in terms of the CNT volume fractions that correspond to this percolation
transition. The effect of CNT alignment is analyzed by comparing the electro-mechanical properties in the
alignment direction versus the transverse direction, where it is found that the properties in the alignment direction
are larger than that of the transverse direction. Estimates of piezoresistive coefficients are converted into
gage factors and discussed in the context of experimental sources in literature. The methodology for this work
involves generating several semi-random 5 µm × 5 µm computational realizations for different CNT volume
fractions with randomized CNT seeding. These realizations are then analyzed using finite elements to obtain
volume averaged effective values, which are then subsequently used to generate estimates (mean from several
realizations) for mechanical stiffness, electrical conductivity and piezoresistive coefficients. The distribution in
these properties is also studied using measures of coefficient of variation, skewness and kurtosis. It is found that
electrical percolation can be captured by the transition of these measures of variation from low to high and back
down to low values with increase in CNT volume fraction.