Keywords: carbon nanotubes, CNTs, CNTFETs, ultralong, field–effect transistors, FETs, mobility changes, inversion regions, deoxyribo nucleic acid, DNA, classifiers, support vector machine, SVM, multilayer perceptron, MLP, nanoelectronics, nanotechnology, ethanol, chemical vapour deposition, CVD
Identification of mobility changes induced by deoxyribo nucleic acid in the inversion regions of ultralong single walled carbon nanotube field–effect transistor using support vector machine and multi–layer perceptron
Using ethanol based chemical vapour deposition (CVD), ultralong single walled carbon nanotubes (SWCNTs) were consistently grown in the centimetre scale, and a number of long–channel carbon nanotube field–effect transistors (CNTFETs) were fabricated. Here, we extracted the field–effect mobility of millimetre–long CNTFETs from the strong inversion region and near–threshold region respectively before and after the addition of deoxyribo nucleic acid (DNA). Support vector machine (SVM) and multi–layer perceptron (MLP) classification algorithms were used to discriminate between the bare device mobility and the DNA induced mobility data in the two regions. SVM classification yielded an average accuracy, sensitivity, and specificity of 97.81%, 97.66%, and 95.38% respectively. On the other hand, MLP based classification resulted in an average accuracy, sensitivity, and specificity of 92.10%, 95.07%, and 90.70%, respectively.