Electromechanical Properties of Metallic, Quasimetallic, and Semiconducting Carbon Nanotubes under Stretching

An electromechanical system is constructed to explore the electrical properties of various types of suspended single-walled carbon nanotubes under the influence of tensile stretching. Small band-gap semiconducting (or quasimetallic) nanotubes exhibit the largest resistance changes and piezoresistive gauge factors ($\sim 600$ to 1000) under axial strains. Metallic nanotubes exhibit much weaker but nonzero sensitivity. Comparison between experiments and theoretical predictions and potential applications of nanotube electromechanical systems for physical sensors (e.g., strain gauges, pressure sensors, etc.) are discussed.

Figure 1

(a) A schematic device structure. The device fabrication started with a $p$-type Si with a 200 nm-thick thermally grown oxide. A ${\mathrm{S}\mathrm{i}}_3{\mathrm{N}}_4$ (200 nm-thick, HF etch-stop for beam release), ${\mathrm{S}\mathrm{i}\mathrm{O}}_2$ ($2\mu \mathrm{m}$ thick by low pressure chemical vapor deposition, sacrificial layer for cantilever release), poly-Si ($1\mu \mathrm{m}$, for cantilever beam), ${\mathrm{S}\mathrm{i}}_3{\mathrm{N}}_4$ (10  nm layer separating Mo from poly-Si), and Mo (50 nm) films were successively deposited, before catalyst patterning, defining the shapes of cantilevers and terraces by photolithography, cantilever release, and SWNT growth. (b) An SEM image of a device. Inset: Zoom-in view of a suspended SWNT. (c) Scheme for stretching a nanotube.

Figure 2

(a) $I\mathrm{\text{−}}{V}_g$ characteristics for an SGS-SWNT. (b) Real time data of AFM cantilever deflection (top panel) and current (bottom) through the SWNT (under $\mathrm{\text{bias}}=10\mathrm{m}\mathrm{V}$) during multiple manipulation cycles. (c) Beam deflection (top) vs $Z$ and current (bottom panel, $\mathrm{\text{bias}}=10\mathrm{m}\mathrm{V}$) vs $Z$ for one cycle. (d) Normalized resistance change vs strain for the nanotube. Inset: Data for another SGS-SWNT. Lines are drawn to guide the eye in $(\delta R/R_0)\mathrm{\text{−}}\sigma$ plots. The solid squares are calculated data points based on the band-gap change theory described in the text. Note that the data shown here were representative of $\sim 5$ SGS tube devices. On occasion, we observed conductance increase under strain with a device. This could be consistent with the prediction of narrowing of $E_g$ for certain types of SWNTs under strain [7, 8, 10]. Nevertheless, this behavior is rare and needs to be reproduced with more samples.

Figure 3

(a) EM data for a S-SWNT. Inset: FET-like $I\mathrm{\text{−}}{V}_g$ characteristics. Bias $\mathrm{\text{voltage}}=10\mathrm{m}\mathrm{V}$. (b) Normalized resistance change vs strain recorded under $V_g=-20\mathrm{V}$.

Figure 4

(a) EM data for a M-SWNT. Inset: $I\mathrm{\text{−}}{V}_g$ characteristics. Bias $\mathrm{\text{voltage}}=10\mathrm{m}\mathrm{V}$. (b) Normalized resistance change vs strain for the M-SWNT. Inset: Data for another M-SWNT.