ABSTRACT

Two-dimensional single-walled (carbon) nanotube (SWNT) networks-also referred to as thin films-for large area electronics have been a rapidly growing area of research since 2003. SWNT thin films provide an alternative to the lithographically intensive fabrication process for individual SWNT devices and can be deposited at room temperature over large areas on inexpensive substrates. SWNT thin films display unique and tunable optoelectronic properties and thus can be utilized in applications ranging from electrodes for organic solar cells and light-emitting diodes to thinfilm transistors. Here we describe the different fabrication routes for the assembly of SWNT thin films primarily including solution

deposition approaches-starting from dispersion related issuesand then briefly describing dry transfer strategies. Thereafter, we concentrate on the optoelectronic properties of SWNT thin films and postdeposition strategies for engineering their optoelectronic properties. Finally, the use of SWNT thin films in various devices will be presented in detail. 4.1 IntroductionCarbon nanotubes (CNTs) are graphene sheets rolled up to form seamless cylindrical tubes.1-3 Their ends can be capped with bisected fullerene molecules, or they can be open. Nanotubes can be tens of micrometers long and are either single walled (one shell) or multiwalled nanotubes (MWNTs) (many shells).1-3 Double-walled (carbon) nanotubes (DWNTs), which possess characteristics of both SWNTs and MWNTs, have also been isolated and synthesized in large quantities.4,5 It is now widely known that the electronic properties of SWNTs depend on the arrangement of carbon atoms around the circumference of the tubes. The chirality (orientation of the graphene lattice relative to the tube axis) and diameter of SWNTs determine the band structure and hence the electronic properties. As a result, SWNTs can be metallic (~0 eV bandgap) or semiconducting (0.4-0.7 eV bandgap).1-3 The bandgap of the semiconducting SWNT is strain induced due to curvature effects, and it decreases with the SWNT diameter. One-third of SWNTs with all possible chiralities are metallic, the other two-thirds being semiconducting. 1-3 SWNTs can be synthesized using three primary methods: arc discharge6-8, laser ablation9-11, and chemical vapor deposition (CVD).12-15 The vapor-liquid-solid (VLS) growth mechanism leads to the formation of CNTs where diffusion of carbon occurs through a liquid-phase catalyst, followed by precipitation of graphitic filaments. The carbon source for nanotubes is vaporized carbon atoms from a solid target in arc discharge and laser ablation and hydrocarbon gaseous species in CVD. Arc discharge and laser ablation can be classified as high-temperature (>3,000 K near the discharge) and short-reaction-time (ms-ms) processes, whereas CVD is a moderatetemperature (700-1,400 K) and long-reaction-time (typically minutes to hours) process. Arc discharge and laser ablation produce SWNTs as powder samples in bundles, while CVD allows synthesis of SWNTs

on substrates16-19 as well as in powder form.20,21 Additionally, CVD allows control over the diameter22,23, length, and orientation24,25 of nanotubes. One important point to note is that although there are promising efforts on the preferential growth of semiconducting/ metallic SWNTs, it is presently not possible to grow all metallic or all semiconducting SWNTs.23,26-28 Recently, the catalyst-conditioning ambient investigated by Harutyunyan et al. was shown to affect the relative abundance of metallic and semiconducting tubes.29Successful separation of SWNTs by diameter and electrical properties to enhance semiconducting/metallic behavior in device applications will be discussed in Section 4.5. The realization of MWNTs at NEC Laboratories by Sumio Iijima in 199130 followed by the discovery of SWNTs in 19936,31 started a new era of research in nanotechnology. The unique one-dimensional structure and cylindrical symmetry of SWNTs leads to appealing mechanical and electrical properties, which have received a great deal of attention and investigation. Numerous potential applications have been proposed for CNTs due to their extraordinary characteristics. A list of applications includes but is not limited to conductive and highstrength composites, energy storage devices, sensors, field emission displays, and semiconductor devices such as field-effect transistors (FETs). Nanoelectronics utilizing CNTs has been considered as the most promising application of nanotechnology. In principle CNTs can be useful for downsizing circuit dimensions and providing a corresponding increase in computational power (Moore’s law32). Additional effects such as ballistic transport also make them interesting for high-frequency and interconnect applications.33,34 The unique quantum wire-like properties make them useful in novel devices such as a spin transport medium for spintronics.35,36Experiments have shown that metallic SWNTs can carry currents up to 109 A/cm2 (compared to ~105 A/cm2 for metals), which make them particularly useful in high-power electronic circuits.37,38 On the other hand, semiconducting SWNTs connected to two metal electrodes can function as FETs. Semiconducting SWNTs can be switched from conducting to insulating state at room temperature by modulating the gate voltage.39 However, fabrication methods for individual SWNT devices such as transistors are not integrated and sometimes involve crude techniques such as dragging an SWNT over predefined electrodes using an atomic force microscope (AFM)

tip40,41 or drop-casting a dilute SWNT solution onto prepatterned electrodes39 or directly growing SWNTs between electrodes.42-44Although outstanding and record-beating results have been obtained using individual SWNTs, several inherent variabilities in SWNT properties, such as chirality and diameter, make it challenging to fabricate reproducible and uniform devices. In addition, the limited drive current per SWNT makes it almost impossible to match the current levels required for applications such as microwave circuits or display drivers. The practical fabrication and integration challenges along with fundamental limitations such as low drive currents and inadequate density of tubes for interconnects have limited the implementation of SWNTs into electronics. Therefore, despite the intense efforts and substantial investment in CNT research over the last two decades, individual SWNT devices are not likely to be incorporated into mainstream electronics applications in the foreseeable future. SWNT networks on the other hand offer advantages by alleviating some of the device integration challenges associated with individual SWNT devices.45 That is, by creating stable SWNT inks, networks of SWNTs can be deposited onto substrates over large areas using solution-based techniques such as inkjet printing, spin coating, spraying, or roll-to-roll printing. Although the electronic properties of the SWNT thin films are substantially lower than those of individual nanotubes, the ease of processing and device fabrication makes them useful for large-area electronics where devices with moderate performance on inexpensive and flexible platforms are required. In addition, SWNT thin films possess unique properties, which have opened new applications. For example, SWNT thin films are transparent and conducting and can be deposited on flexible substrates such as plastics, making them useful as a potential replacement for indium tin oxide (ITO) in organic electronic devices. The recent advancements in postgrowth separation of metallic SWNTs from semiconducting ones using a centrifugal density gradient by Hersam et al.46-49 has allowed fabrication of high-performance optoelectronic devices from SWNT networks. In this chapter, we review the progress in the field of SWNT thin-film devices. The field is relatively new, with the first report on SWNT network devices by Snow et al.50 appearing in 2003, but has grown rapidly, in part, due to advancements from Rinzler et al.51-53, Gruner et al.54-56, Rogers et al.57-60, and others.