Electronic devices based on carbon nanotubes are among the candidates to eventually replace silicon-based devices for logic applications. Before then, however, nanotube-based radiofrequency transistors could become competitive for high-performance analogue components such as low-noise amplifiers and power amplifiers in wireless systems. Single-walled nanotubes are well suited for use in radiofrequency transistors because they demonstrate near-ballistic electron transport and are expected to have high cut-off frequencies. To achieve the best possible performance it is necessary to use dense arrays of semiconducting nanotubes with good alignment between the nanotubes, but techniques that can economically manufacture such arrays are needed to realize this potential. Here we review progress towards nanotube electronics for radiofrequency applications in terms of device physics, circuit design and the manufacturing challenges.

While the potential for high mobility printed semiconducting nanotube inks has been clear for over a decade, a myriad of scientific and technological issues has prevented commercialization and practical use. One of the most challenging scientific problems has been to understand the relationship between the pristine, individual nanotube mobility (known to be in the 10,000 cm(2)/V·s range) and the as-deposited random network mobility (recently demonstrated in the 100 cm(2)/V·s range). An additional significant scientific hurdle has been to understand, manage, and ultimately eliminate the effects of metallic nanotubes on the network performance, specifically the on/off ratio. Additional scientific progress is important in understanding the dependence of nanotube length, diameter, and density on device performance. Finally, the development of ink formulations that are of practical use in manufacturing is of paramount importance, especially with regard to drying time and uniformity, and ultimately, the issue of scalability and cost must be addressed. Many of these issues have recently been investigated from a phenomenological point of view, and a comprehensive understanding is beginning to emerge. In this paper, we present an overview of solution-based printed carbon nanotube devices and discuss long-term technology prospects. While significant technical challenges still remain, it is clear that the prospects for the use of nanotube ink in a myriad of systems is feasible given their unmatched mobility and compatibility with heterogeneous integration into a variety of applications in printed and flexible electronics.

Carbon-nanotube-based semiconducting inks offer great promise for a variety of applications including fl exible, transparent, and printed electronics and optics. A critical drawback of such inks has been the presence of metallic nanotubes, which causes high-mobility inks to suffer from poor on/off ratios, preventing their applications in a wide variety of commercial settings. Here, we report a comprehensive study of the relationship between mobility, density, and on/off ratios of solution-based, deposited semiconducting nanotube ink used as the channel in fi eld effect transistors. A comprehensive spectrum of the density starting from less than 10 tubes μ m − 2 to the high end of more than 100 tubes μ m − 2 have been investigated. These studies indicate a quantitative trend of decreasing on/off ratio with increasing density and mobility, starting with mobilities over 90 cm 2 V − 1 s − 1 (approaching that of p-type Si MOSFETs) but with on/off ratios ∼ 10, and ending with on/off ratios > 10 5 (appropriate for modern digital integrated circuits), but with mobilities ∼ 1 cm 2 V − 1 s − 1 . These studies provide the fi rst important roadmap for the tradeoffs between mobility and on-off ratio in nanotube based semiconducting inks.Single-walled carbon nanotube (SWNT) based semiconducting inks may have a wide variety of applications in printed electronics (such as inkjet printing, [ 1 ] role to role gravure, [ 2 ] and pad/screen printings [ 3 ] ) as well as offering the possibility of heterogeneous integration of different semiconductor technologies such as Si CMOS, III-V, and optical display technologies. Recent progress in purifi cation techniques [ 4 ] has lead to the prospect of all-semiconducting SWNT inks for unsurpassed performance in printed circuits.In general, it is known that the mobility of individual, pristine semiconducting nanotubes can be up to 10 000 cm 2 V − 1 s − 1 . [ 5 ] However, mobilities for random networks of carbon nanotubes has hovered until recently around the 1 cm 2 V − 1 s − 1 limit. [ 6 ] What sets the mobility of a random network of semiconducting nanotubes in relationship to individual nanotubes? Can the mobility be increased by increasing the density? How does this affect the on/off ratio and what are the physical processes that set limits on this scaling?The most obvious reason that networks have lower mobilties than individual nantoubes is that tube-tube crossings limit the current fl ow from source to drain if the channel length is longer than the nanotube length. Increasing the network density can increase the current (and hence potentially the mobility).However, the complexity of such a system, coupled with the presence of metallic nanotubes that can short-circuit the device if the density exceeds the percolation threshold, means that there is no general theory that explains quantitatively the relationship between mobility, density, and on/off ratio, so that phenomenological experimental approaches are necessary for progress in the fi eld.Although solution-based processing technique...

No abstract

To probe the active sites of nitrogen-doped carbon nanostructures (CN x ), the effect of dihydrogen phosphate (H2PO4 –) anion on their oxygen reduction reaction (ORR) performance was investigated by adding increasing concentrations of phosphoric acid in half-cell measurements. A linear decrease in specific kinetic current at 0.7 V was noted with increasing phosphate anion concentration. It was also found that the adsorption of phosphate species on CN x was strong and the corresponding ORR activity was not recovered when the catalyst was reintroduced to a fresh HClO4 solution. Trends similar to those noted upon addition of H3PO4 to the half-cell were observed when CN x catalysts were soaked in phosphoric acid. Adsorption of dihydrogen phosphate ions on the surface of CN x exposed to phosphoric acid was verified by transmission infrared (IR) and Raman spectroscopy as well as X-ray photoelectron spectroscopy (XPS). XPS results also showed a decrease in the surface concentration of pyridinic-N species accompanied by an increase of equal magnitude in the surface fraction of quaternary-N species, which would include the pyridinic-NH sites. A linear correlation was observed between the loss in pyridinic-N site density and that in ORR activity. The observed poisoning phenomenon is consistent with the two possible active site models, i.e., pyridinic-N sites, which would be rendered inactive by protonation, or the C sites neighboring pyridinic-N species. These latter species would be poisoned by a site blocking effect if they strongly adsorb the phosphate ions. Strong adsorption of negatively charged phosphate ions on neighboring C atoms would also stabilize the pyridinic-NH sites. By identifying a poison that can be used as a probe, this study provides a first step toward identification and quantification of active sites in CN x catalysts.