Research data supporting publication "Optical Control of Single Atom Dynamics in Plasmonic Nanogaps"

Figure 1: (E) 633 nm laser irradiation intensity over the course of a typical write-read experiment (note y-scale). (F) Corresponding SERS time scans for (E). Initial ‘check’ scan collected at 5 μW μm-2 with 8 s integration time per spectrum, ‘write’ at 3 mW μm-2 for 170 ms irradiation time, ‘read’ at 10 μW μm-2 every 4 s (further scans in Fig. S1). Figure 2: Picocavity writing and reading dynamics vs laser power density. (A) Empirical cumulative density distribution functions (ECDF) of picocavity write time distributions along with exponential CDF fits (F_W=1-e^(-Γ_W t)) for 1 mW μm-2 and 5 mW μm-2 incident laser power densities (dashed). Shaded regions denote 95% confidence intervals. (B) Histogram of picocavity write times at 1mW μm-2 write intensity, with exponential probability distribution function (dashed line, f_W=Γ_W e^(-Γ_W t)) derived from the ECDF in (A). (C) Picocavity write rate Γ_W vs laser intensity with fit to Eq.(1) model assuming 1 eV activation barrier (28); ◼ marks critical intensity I_c. Error bars derived from CDF least squares fits are too small to be seen. (D) Percentage of picocavities that survive for 200 s vs read illumination intensity (at write intensity 3 mW μm-2). Upper and lower limit survival percentages extracted from criteria detailed in Methods and Supplementary Note S1. (E) Picocavity % survival during read vs laser intensity in the preceding write (read intensity 40 μW μm-2). Error bars in (D-E) are derived from binomial distribution variance. Figure 3: (A) 785 nm excitation SERS spectra resulting from a write-read-erase experiment with a 20 h storage period at the read stage. Only the final write spectrum is shown. All spectra are normalised to the 290 cm-1 nanocavity peak and smoothed with a 3rd order 5 pt Savitzky-Golay filter. (C-E) 633 nm excitation SERS spectra resulting from write-read experiments on three example picocavities over the course of a week, with increasing periods of dark storage between read spectra. Only the final write spectra (red) are shown. (F) 633 nm excitation SERS spectra resulting from read experiments on a control NPoM with no prior write stage, over the course of 76 h. All spectra in (C-F) are normalised to the 1050 cm-1 nanocavity peak and smoothed with a 3rd order 11 pt Savitzky-Golay filter. Figure 4: (A) Example histogram of picocavity state evolution times, with exponential probability distribution (dashed line, f_S=Γ_S e^(-Γ_S t)) derived from fitting 80 μW μm-2 ECDF in (B). (B) Example ECDFs of picocavity state evolution time distributions along with exponential CDF fits (F_S=1-e^(-Γ_S t)). Shaded regions denote 95% confidence intervals. (C) Picocavity state change rate Γ_S dependence on laser intensity (line is guide to eye). Error bars are derived from the CDF least squares fits. Figure 5B: Migration energy of the adatom and surface area of the pit in (A) vs simulation trajectory over time.