Solution-Processed Carbon Nanotube and Chemically Synthesized Graphene Nanoribbon Field Effect Transistors.
- Author(s): Bennett, Patrick B.
- Advisor(s): Bokor, Jeffrey
- et al.
Carbon nanotubes (CNTs) possess great potential as high performance semiconducting channels due to their one-dimensional nature, extremely high mobility, and their demonstrated ability to transport electrons ballistically in transistors. However, the presence of metallic CNTs in CNT films and arrays represents a major impediment towards large-scale integration. Methods of solution purification have demonstrated partial success in metallic CNT removal, although their effects on device performance are unknown. While this problem may be solvable, new synthesis techniques have recently resulted in the creation of high-density films of graphene nanoribbons (GNRs) with atomically smooth edges, uniform widths, and uniform band structure. These may ultimately supplant CNTs in device applications due to their theoretically similar, but uniform electronic properties.
This work aims to study the effects of purification of semiconducting CNTs in thin film transistors (TFTs) and to develop methods to increase device performance when metallic CNTs are present. Devices consisting of large networks of CNTs as well as short channel, single CNT devices are characterized to determine the effects of solution processing on CNTs themselves. Short channel, bottom-up GNR devices are fabricated to compare their performance to CNT transistors.
The first half of this dissertation describes the methods of integrating CNTs from various sources into transistors. Growth and transfer are described, as well as methods of creating aqueous suspensions for solution processing. Development of novel surfactant materials based on biomimetic polymers used to suspend CNTs in solution are reported and characterized. Methods of deposition out of solution and onto insulating substrates are covered. Device fabrication from start to finish is detailed, with the subtleties of processing required to produce sub 10-nm channel length devices explained.
The second half reports devices produced via these techniques in order to study the performance of solution-processed CNT devices. TFT performance is limited by metallic CNTs that can short channels, but can be improved by structuring the CNT film, either through patterning or induced alignment. Increasing semiconducting CNT purity does not necessarily increase device performance because of the decreased lengths of the purified CNTs. Extremely high purity semiconducting CNT solutions, however, are not subject to these same limitations, with transistors exhibiting improved mobilities while also scaling to sub-µm channel lengths. Short channel devices down to 15 nm are then presented, demonstrating ballistic transport in solution-processed CNTs, despite their inferior electronic performance at µm-scale lengths. Finally, short channel devices utilizing chemically synthesized GNRs as channels are presented and characterized to directly probe the mechanisms of electron transport in these materials for the first time.