Carbon Nanotube Networks: A New Electronic Material

Background: Single-wall carbon nanotubes (SWNTs) are unique electronic structures. They are one-dimensional wires composed entirely of surface atoms yet exhibit transport properties superior to bulk single-crystalline silicon (Si). This high electron mobility makes them an ideal candidate for electronic device applications, while their virtually infinite surface-to-volume ratio offers extraordinary sensitivity for chemical and biological sensor applications. However, a major obstacle presently preventing their commercial and/or military implementation in new classes of electronic devices is the lack of a technique for the controlled assembly of large numbers of SWNTs with precisely controlled position and orientation. Until this obstacle is overcome or circumvented, SWNT-based devices and sensors will remain in the realm of impressive laboratory curiosities without real-world applications.

At NRL we have developed an approach to realizing SWNT electronic applications that circumvents the assembly problem stated above by using a random network of SWNTs (Fig. 10). We have recently discovered that such carbon nantoube networks (CNNs) form an electrically conducting thin film, and that this network is conducting because of the high-quality electrical contact formed between intersecting nanotubes.1 We have found that the electronic properties of CNNs can range from semiconducting to metallic, depending on the network density. Devices fabricated from such networks circumvent the requirement of precise position and structural control because the devices display the averaged properties of many randomly distributed interconnected SWNTs.

FIGURE 10
Optical image of a CNN device; inset shows an atomic force microscope image of the CNN that electrically connects the two electrodes.

An additional feature of CNNs is that they can be processed into devices using conventional fabrication technology without affecting their electrical properties. Thus, CNNs retain the interesting electronic properties of individual SWNTs while providing the processing capabilities of conventional electronic materials. This combination of features opens up a wide range of electronic and sensor applications, two of which (macroelectronics and chemical detection) are described below.

Chemical Detection: Because they are composed entirely of surface atoms, the electrical resistance of SWNTs is extremely sensitive to the physisorption of certain chemical vapors. In addition, the otherwise chemical inertness of SWNTs makes them very robust for operating in harsh environments. Consequently, SWNTs offer the potential for a new class of chemical sensors.

In such sensors, the adsorption of an analyte molecule with electron donor or acceptor properties results in a charge transfer between the analyte and the SWNT that changes its electrical resistance. Because many chemical warfare agents and toxic industrial chemicals have strong electron transfer properties, CNN-based devices can serve as highly sensitive chemical sensors for defense and homeland security applications.

At NRL we have performed preliminary experiments to test the viability of CNN chemical detectors, with a focus on achieving high levels of sensitivity and chemical specificity. Our initial experiments have established that SWNT network devices are capable of detecting nerve agents with sub-ppb sensitivity.2 In these experiments, we fabricated CNN-based chemical sensors and used these devices to detect dimethyl methylphosphonate (DMMP), a nerve agent simulant (Fig. 11). In order to obtain a chemically specific response, we have combined the CNNs with chemoselective polymers recently developed at NRL. This combination has the potential to provide highly sensitive, low-power chemical detectors for chemical warfare agents and toxic industrial chemicals.

FIGURE 11
Response of a CNN sensor to doses of DMMP, a chemical simulant for nerve agents. The resistance of the CNN increases upon exposure and is refreshed by heating the device.

Macroelectronics: An important emerging area of technology is macroelectronics. Macroelectronics has the goal of providing inexpensive electronics on polymeric substrates. Such electronics will be "printed" onto large-area polymeric films by using fabrication techniques more analogous to text and imaging printing than conventional semiconductor microfabrication technology. Applications of such macroelectronics include lightweight flexible displays, inexpensive RF identification, and smart materials or clothing. Conventional semiconductors are not suitable for such applications, because such materials require a crystalline substrate and are too expensive. Consequently, a large investment has been made in developing organic semiconductors that can be deposited at room temperature onto a wide range of substrates. This effort has achieved only moderate success because of the low-quality electron transport found in organic semiconductors.

CNNs provide a promising new material for macorelectronic applications because of the high-quality electron transport intrinsic to SWNTs and because CNNs can be deposited at room temperature onto polymeric and many other substrates. We have demonstrated thin-film transistors fabricated using CNNs that have an electron mobility more than an order of magnitude larger than the mobility found in organic semiconductors (Fig. 12). Consequently, with further development CNNs should provide a low-cost electronic material for flexible, unbreakable displays and inexpensive macroelectronics.

FIGURE 12
Transistor characteristics for a CNN thin-film transistor; the inset is an optical image of the source, drain, and gate electrodes of the device

Future Work: The promising results of our initial research indicate that CNNs can serve as a new electronic material for a variety of applications such as macroelectronics and chemical sensing where submicron device size is not required. In such applications, the larger device size averages the properties of many individual SWNTs, yielding good device-to-device reproducibility. Future work includes improving the transistor performance by using chemical separation techniques to purify SWNTs based on electronic type, chemical functionalization of the SWNTs to further enhance the sensitivity and chemical specificity of the sensors, and exploring additional CNN applications such as biosensing.

[Sponsored by ONR]