Gap-size dependence of optical near fields in a variable nanoscale two-tip junction

When nanometric metal structures are illuminated with light, large optical near fields arise. These fields are even further enhanced when two nanometer-sized structures are brought in nanometric distance. Then the individual near fields of the structures couple. Measuring the coupled optical near fields directly is hard. Here we demonstrate an scanning tunneling microscope–based setup which allows us to probe the coupling of optical near fields from two nanometer-sharp metal needles facing each other. By measuring the emitted photocurrent as a function of the gap distance, we retrieve the peak fields at the tip apices. The comparison to optical near-field simulations shows that large field enhancement of 15.9 is present at the tip apex. This is an enhancement 4-times larger compared to the individual nanotip. The demonstrated sensitivity of the coupling on the gap size is of large interest for future applications in ultrafast near-field microscopy.

Figure 1

(a) Sketch of the experimental setup. Near-infrared femtosecond laser pulses are focused with an off-axis parabolic mirror (OAP) onto a nanoscale junction, formed by two metal needles. The alignment of the tips, facing each other, is achieved by two-tip tunneling without laser illumination. During the scan the upper tip probes the lower one, moving on a half-circle of radius ${\mathrm{R}}_{\mathrm{conv}}$. (b) Scanning the lower tip with the upper tip in standard STM operation reveals a 3D tiplike shape and demonstrates stable relative positioning of the two tips in 3D on the nanometer scale. (c) From the cross section of the height profile we extract a joint radius of curvature of 30 nm. (d) Example of an interferometric autocorrelation trace (IAC) measured by the emission from one gold tip in a $\sim 200$-nm gap junction. Compared to IACs recorded with beta barium borate (BBO) almost no broadening is visible.

Figure 2

(a) Photocurrent as function of average laser power shows multiphoton photoemission from a single gold tip when tips are separated by about $3\mu \mathrm{m}$ (bias voltage of 1 V). The fitted slope of $2.5\pm 0.1$ indicates the multiphoton order. (b) Current recorded as a function of the gap size variation. With 1.8 mW average laser power the current modulus shows a small decay constant of two orders of magnitude over 12 nm as a function of the relative gap distance $\mathrm{\Delta }z$, in contrast to the current without laser illumination (black with linear fit). Here the current drops by two orders over 2 nm. Note that $\mathrm{\Delta }z=0$ corresponds to a tip-tip distance of approximately 6–8 nm for the curves with laser illumination and of $\sim 1–2$ nm for the data without. The inset shows the influence of the bias voltage on a linear scale. The bias voltage allows us to steer the current direction and thus select which tip photoemits mainly.

Figure 3

(a) Simulated near-field distribution for a gap size of 10 nm for two gold tips of radii 60 and 70 nm, representing the experimental situation. For a small gap size ($\lesssim 60$ nm), the optical near fields of two tips couple, enhancing the combined near field. (b) The simulation of field distributions as a function of gap distance allows us to extract field enhancement factors at the tip apex. For decreasing gap distance $d$, the field enhancement factors of both tips rise dramatically. Both data sets, almost indistinguishable, follow a fit of the form $f\left(d\right)=a{d}^b+c$ (dotted orange line), which approaches for large distance the almost same field enhancement of a single tip (green line). (c) We retrieve the optical near field at the tip apex as the fifth root of the photocurrent derived from Fig. 2. The near-field simulation shows the same decay as observed for the retrieved field enhancement, when the experimental curves are shifted by an offset $z_0$. Consequently, we obtain a quantitative measure for the field enhancement and demonstrate the gap-size dependent coupling of the optical near fields.

Figure 4

(a) Response of the STM feedback loop while the laser focus (red circle) is scanned near the two-tip junction (tip sketches not to scale). Parameters are 12.5-mW average power, equivalent to a peak intensity of $3.5\times {10}^{11}\mathrm{W}/{\mathrm{cm}}^2$, and 0.8-V bias voltage at a constant-current set point of 0.3 nA. Because of large photoemitted current in the junction, the upper tip is retracted by the feedback to hold the current constant. At the shank mainly thermal expansion leads to a response of the STM feedback. (b) Cross section through the maximum retraction along the $z$ axis. The abrupt retraction indicates the coupling of the optical near fields and the large photoemission.