Full text loading...
(a) SMP jack fixed to a modified standard tip carrier. Short tip is glued directly to the jack as marked. (b) SMP plug terminating the RF coaxial cable as used for applying the voltage pulses. Plug is glued to a fixing plate which is screwed to the STM body. (c) Schematic cross-section through the STM with main components marked.
Setup for pulsed STM measurements: PPG generates “pump” (CH1) and delayed “probe” (CH2) voltage pulses, which are summed by a power combiner (+). The time delay of CH2 is modulated with a reference frequency to enable lock-in detection. The second power combiner adds the bias voltage provided by the STM electronics. The RF coaxial cable guides the signal to the tip of the microscope which is unshielded on the last 2 mm. The sample is connected via low frequency wiring to an I–V-converter withbandwidth 1 kHz and gain 109 V/A. For frequencies exceeding the bandwidth of the preamplifier, the circuit is shunted by the parasitic sample-ground capacitance (labeled “shunt capacitance”). The signal of the I–V-converter is fed back into the STM electronics and simultaneously into the input channel of the lock-in amplifier. For pulse superposition measurements (Fig. 3 ), the lock-in output is displayed as a function of pump-probe delay.
(a) Voltage pulse train generated by pulse pattern generator (nominal amplitude 2 V, nominal width 100 ps) and combined with DC bias as measured by an oscilloscope, connected at 50 Ω input impedance to the output of the second power combiner. The combined attenuation of both power combiners (21.6 dB) results in 121 mV pulse amplitude. The pulse train is repeated at 166 MHz, while the “probe” pulse delay is modulated between the delay of interest Δt and a time delay of 3 ns at 1403 Hz. Δt is swept during measurements. The figure shows ns, i.e., the probe precedes the pump pulse. (b) Lock-in output Vlock-in as a function of for an HOPG sample and a PtIr tip. Each measurement point is recorded after a waiting time of three lock-in time constants . Three measurement runs are averaged. . Solid line: Simulation of the lock-in output using the measured I(V)-characteristic of the tunneling junction (inset) and the pulse train shown in (a).
(a) 8.6 nm × 8.6 nm constant-current image of HOPG with atomic resolution. . (b) Corresponding current image. (c) Simultaneously acquired lock-in data at 1403 Hz reference frequency with a 3 ms time constant exhibiting a 61 mV RMS noise. (d)–(f) Lock-in data acquired the same way as in (c) but with enabled PPG using time delays as marked. Deduced peak voltage at the junction: 307 mV. Measured pulse width: 133 ps FWHM (150 ps nominally). Repetition rate: 1.67 GHz. The insets in (c)–(f) show corresponding autocorrelation images scaled to an identical contrast.
Article metrics loading...