How Structural Defects Affect the Mechanical and Electrical Properties of Single Molecular Wires

We report how individual defects affect single graphene nanoribbons by scanning tunneling and atomic force microscopy pulling experiments simultaneously accessing their electrical and mechanical properties. The on-surface polymerization of the graphene nanoribbons is controlled by cooperative effects as theoretically suggested. Further, we find, with the help of atomistic simulations, that defects substantially vary the molecule-substrate coupling and drastically increase the flexibility of the graphene nanoribbons while keeping their desirable electronic properties intact.

Figure 1

(a) Illustration of the polymerization process on Au(111). Top: DBDA polymerizes at $170°\mathrm{C}$ to anthracene oligomers. Center: Not fully dehydrogenated ribbons are formed at $330°\mathrm{C}$ as exemplified by the chemical structure of a GNR with two defects (upward-pointing anthryl groups indicated in blue, planar anthracene unit indicated in orange). Bottom: At $400°\mathrm{C}$ fully dehydrogenated GNRs are formed. (b) STM image of a GNR adsorbed on Au(111) with several defects. (c) Average number of defects as a function of the annealing temperature. (d) Probability of finding a row of randomly distributed defects as a function of the ribbon length. Vertical line indicates the average length observed in experiments (13 nm and 15 dianthracene units) (e) Experimental distribution (squares) and calculated probability (circles) for a row of randomly distributed defects as a function of the number of defects. We calculated the probability for a ribbon which corresponds to the average experimental length (Supplemental Material [27]). (f) The average bending angle $\alpha$, as indicated in (b), as a function of the number of defects (Supplemental Material [28]). (g) Calculated energy of an adsorbed ribbon for different bending angles $\alpha$ (Supplemental Material [27]).

Figure 2

[(a), (b)] $\mathrm{\Delta }f$ traces with defects as shown in the STM image. The tip height of 0 Å corresponds to a set point of $I=100\mathrm{pA}$ and $U=1\mathrm{V}$, and $I=100\mathrm{pA}$ and $U=-1\mathrm{V}$ for (a) and (b), respectively. The bias voltage during the pulling experiment is 0 and 0.8 V for (a) and (b), respectively.

Figure 3

(a) Simulated force for a pristine GNR (black) and a GNR with four defects (green). (b), (c) Height of the center of mass for the planar adsorbed (b) or the upward-pointing (c) benzene rings during pulling. For comparison, a scaled chemical structure of the defected GNR is plotted between (a) and (b). [(d)–(i)] Molecular snapshots associated with the force peaks and valleys during the pulling of the defected GNR. The planar or upward-pointing anthracene unit that is currently detached is indicated by the arrows in (d), (f), (h) and (e), (g), (i), respectively.

Figure 4

(a) Pulling curve of a GNR with four defects at $+1.8\mathrm{V}$ bias voltage. The blue curve (current divided by its exponential fit) highlights the influence of the defects. (b) Pulling experiment where the current and the $\mathrm{\Delta }f$ are recorded simultaneously. The bias voltage during the pulling experiment is 0.9 V. The ribbon is fully detached at $\sim 140\AA$ when the characteristic feature in the $\mathrm{\Delta }f$ signal occurs. The corresponding drop in current is not visible due to the detection limit of the amplifier. The tip height of 0 Å corresponds to a setpoint of $I=100\mathrm{pA}$ and $U=0.1\mathrm{V}$, and $I=100\mathrm{pA}$ and $U=0.9\mathrm{V}$ for (a) and (b), respectively. (c) Magnification of (b) with insets showing the detachment process. (d) Transmission curves of a GNR when an upward-pointing (blue) or a planar unit (orange) is detached.