Limits of Performance Gain of Aligned CNT Over Randomized Network: Theoretical Predictions and Experimental Validation

Nanobundle thin-film transistors (NB-TFTs) that are based on random networks of single-walled carbon nanotubes are often regarded as high performance alternative to amorphous-Si technology for various macroelectronic applications involving sensors and displays. Here, we use stick-percolation model to study the effect of collective (stick) alignment on the performance of NB-TFTs. For long-channel TFT, small degree of alignment improves the drain current due to the reduction of average path length; however, near-parallel alignment degrades the current rapidly, reflecting the decrease in the number of connecting paths bridging the source/drain. In this paper, we 1) use a recently developed alignment technique to fabricate NB-TFT devices with multiple densities D, alignment thetas, stick length L_S, and channel length L_C; 2) interpret the experimental data with a stick- percolation model to develop a comprehensive theory of NB-TFT for arbitrary D,thetas, L_S, and L_C; and 3) demonstrate theoretically and experimentally the feasibility of fivefold enhancement in current gain with optimized transistor structure.