Mechano-switching devices from carbon wire-carbon nanotube junctions

Well-known conductive molecular wires, such as cumulene or polyyne, provide a model for interconnecting molecular electronics circuits. In recent experiments, the appearance of carbon wire bridging between two-dimensional electrodes, i.e., graphene sheets, was observed [C. Jin et al., Phys. Rev. Lett. 102, 205501 (2009)], thus demonstrating a mechanical way of producing cumulene. In this work, we studied the structure and conductance of carbon wire suspended between carbon nanotubes (CNTs) of different chiralities (zigzag and armchair), and corresponding conductance variation upon stretching. We found that the geometric structure of the carbon wire bridging CNTs was similar to the experimentally observed structures in carbon wire obtained between graphene electrodes. We show a way to modulate conductance by changing bridging sites between carbon wire and CNTs without breaking the wire. Observed current modulation via cumulene wire stretching or elongation together with CNT junction stability makes this a promising candidate for use in mechano-switching devices for molecular nanoelectronics.

Figure 1

Two-probe systems for measuring the conductance of carbon wires connected to semi-infinite zigzag (4,0)CNT (left panel) and armchair (4,4)CNT (right panel).

Figure 2

Various structures of the short carbon wires (C${}_4$ and C${}_5$) bridged between zigzag (4,0)CNTs and armchair (4,4)CNTs, in which the gap width is varied. Different geometric structures and bridging sites of the carbon wires are represented by I, II, III, IV. The linear carbon wire can transform into either cumulene “C” or polyyne “P”.

Figure 3

Binding energies vs gap width of C${}_4$ and C${}_5$ wires connected to (a) (4,0)CNT and (b) (4,4)CNT electrodes, and corresponding zero-bias transmission on the energy-length plane $T(E,L)$ (bottom). The location of a transition state of structures is marked by red dots (shown in the upper panels). The length-dependent transmission is correlated to the binding energies and wire structures.

Figure 4

Zero-bias transmission on the energy-length plane $T(E,L)$ and corresponding projected density of $p_{m=1}$ and $p_{m=-1}$ states for carbon wire stretched between armchair (4,4)CNT electrodes at the two lengths corresponding to the II(C) and III(P) configurations as indicated in Fig. 3.

Figure 5

Zero-bias conductance of (4,0)@C${}_n$-II, $n=4$–9.

Figure 6

$I$-$V$ characteristics of (a) (4,0)@C${}_5$ and (b) (4,0)@C${}_4$ varying the gap width.

Figure 7

(a) C${}_5$ wire bridging between (4,0)CNTs. By twisting the CNT, we can change bridging sites and bonding at the junction. (b) Bridge sites of the left and right (4,0)CNT electrodes (left panel) and the corresponding conductance values of each structure (right panel). The blue circle represents the carbon atoms at the edge and the bridge sites between each (4,0)CNT and the C${}_5$ wire are marked by red dots, the carbon atoms next to the edge atoms are marked with a cross symbol.

Figure 8

Zero-bias transmission function and the PDOS of $p_x$ and $p_y$ orbitals of cumulene C${}_5$ wire, corresponding to the wire configuration in Fig. 7.