Transport and optical response of molecular junctions driven by surface plasmon polaritons

We consider a biased molecular junction subjected to external time-dependent electromagnetic field. The field for two typical junction geometries (bowtie antennas and metal nanospheres) is calculated within finite-difference time-domain technique. Time-dependent transport and optical response of the junctions is calculated within nonequilibrium Green’s-function approach expressed in a form convenient for description of multilevel systems. We present numerical results for a two-level (highest occupied molecular orbital-lowest unoccupied molecular orbital) model and discuss influence of localized surface plasmon-polariton modes on transport.

Figure 1

(Color online) Current on the left, $I_L$, and right, $I_R$, interfaces vs time for single-level model. Numerical results (dashed line, red) are compared to analytical expression (solid line, blue). Also shown is sum of the currents, $I_L+I_R$ at the two interfaces (dotted line, black). See text for parameters.

Figure 2

(Color online) Results of FDTD simulations. Left panel shows intensity enhancement as a function of the incident wavelength (in nanometer) in logarithmic scale for two spheres of 20 nm in diameter with a gap of 10 nm (solid line, black) and bowtie antenna with a gap of 10 nm (dashed line, red), and 5 nm (dashed-dotted line, blue). Top right inset represents steady-state intensity enhancement distribution in logarithmic scale for two spheres system at the resonant wavelength of 368.202 nm. Lower right inset shows intensity distribution for the bowtie antennas with a gap of 5 nm at 602.647 nm.

Figure 3

(Color online) Comparison of exact numerical solution (solid line, red) to adiabatic approximation (dashed line, blue) for the two-level (HOMO-LUMO) model. Shown are (a) levels populations and (b) current at the left interface vs time. See text for parameters.

Figure 4

(Color online) The two-level (HOMO-LUMO) model. Shown are (a) current and (b) total optical response, Eq. (24), vs time for bowtie inanimateness (the strongest signal, blue) and two spheres junction geometries. In the latter case, the response is calculated for two positions of the molecule in the junction: in the middle between the spheres (the weakest signal, red) and closer to one of the spheres [intermediate signal, (a) white silhouette and (b) solid line, black]. (c) shows contour map of optical flux, Eq. (23), for bowtie geometry vs outgoing frequency and time. A sketch of optical transitions at different frequencies is shown in the inset. See text for parameters.

Figure 5

(Color online) Shown are (a) current vs time for the two-level (HOMO-LUMO) model calculated at two different biases, and two sketches of optical pulse influence on current at (b) prearrangement, $V\sim 1.8\text{V}$ and (c) postrecession, $V\sim 2.2\text{V}$ bias. See text for parameters.

Figure 6

(Color online) Role of energy-transfer process. Shown are (a) total optical response vs time with (dotted line, red) and without (solid line, blue) electron-hole excitations and (b) difference, $\Delta I=I_0-I_{e\text{−}h}$, between current calculated without $\left(I_0\right)$ and with $\left(I_{e\text{−}h}\right)$ electron-hole excitations vs time and bias. Calculations are performed within adiabatic approximation scheme. See text for parameters.

Figure 7

(Color online) Effect of induced time-dependent phase on electron transport. Shown are time-dependent current resulting from incident monochromatic laser pulse (dotted line, blue) and local electric field with plasmon excitation induced time-dependent phase (solid line, red). The former is scaled to be comparable with enhanced local-field signal. Calculation is done for bowtie antennas with 5 nm gap. Incident pulse duration is 30 fs centered at $\lambda =602.6\text{nm}$. Other parameters are as in Fig. 4