High-Acquisition-Rate Single-Shot Pump-Probe Measurements Using Time-Stretching Method

Abstract

Recent advances of ultrafast spectroscopy allow the capture of an entire ultrafast signal waveform in a single probe shot, which greatly reduces the measurement time and opens the door for the spectroscopy of unrepeatable phenomena. However, most single-shot detection schemes rely on two-dimensional detectors, which limit the repetition rate of the measurement and can hinder real-time visualization and manipulation of signal waveforms. Here, we demonstrate a new method to circumvent these difficulties and to greatly simplify the detection setup by using a long, single-mode optical fiber and a fast photodiode. Initially, a probe pulse is linearly chirped (the optical frequency varies linearly across the pulse in time), and the temporal profile of an ultrafast signal is then encoded in the probe spectrum. The probe pulse and encoded temporal dynamics are further chirped to nanosecond time scales using the dispersion in the optical fiber, thus, slowing down the ultrafast signal to time scales easily recorded with fast detectors and high-bandwidth electronics. We apply this method to three distinct ultrafast experiments: investigating the power dependence of the Kerr signal in LiNbO₃, observing an irreversible transmission change of a phase change material, and capturing terahertz waveforms.

Figure 1. Experimental setups for the single-shot ultrafast measurements.

For the Kerr measurement, we placed crossed polarizers (Pol.) before and after the sample to detect the polarization rotated signal. We rotated the half waveplate (HWP) during the measurement, and the intensity of the probe laser was then tuned from 2.3 mJ/cm² to 31 mJ/cm². For the GST measurement, we monitored the transmission change induced by the pump pulse, whose intensity was tuned with the similar setup for Kerr measurement. For the THz demonstration, we generated terahertz waves with an OH1 organic nonlinear crystal, and the electric field was detected using a conventional EO sampling method with a GaP crystal, a quarter waveplate (QWP) and a polarizer. In all the measurements, probe pulses were pre-chirped using SF11 glass rod. The output was sent to a 3-km optical fiber that was connected to a fast photodiode. The signal traces were recorded using a real-time oscilloscope with sufficient bandwidth.

Figure 2. An intensity dependence of a Kerr measurement in LiNbO₃.

(a) Slowed-down signal traces with and without a pump pulse observed directly at the photodiode. Small modulation of the spectrum may be due to an unintentional echo pulses generated in some optics in the setup, which does not severely affect the signal profiles. (b) 500 signal traces (−ΔI/I) captured while changing the pump intensity. The time axis is calibrated by measuring two waveforms with changing the pump delay time. Every waveform is normalized by the average of two probe profiles without the pump as described in the text. (c) Slices of the Kerr waveforms from (b). (d) The power dependence of the peak intensity at the time origin extracted from the data shown in (b). The black solid line indicates the quadratic fitting expected for the Kerr nonlinearity.

Figure 3. Fast acquisition of pump and single-shot probe data on the phase change material GST.

(a) Intensity profiles of the probe pulses (left) and the normalized transmission change (−ΔI/I) (middle). The right-most trace indicates the pump intensity simultaneously monitored using the same oscilloscope. The slices of the ultrafast waveforms are shown in red (frame 400 in the amorphous phase) and blue (frame 150 in the crystalline phase) at the right side of the figure. The blue and orange dashed lines show the crystalline-to-amorphous phase change and the relaxation dynamics in the amorphous phase, respectively, which are taken from the data in refs 4 and 29. (b) The transmission change at 1 ps (top), pump (middle), and probe (bottom) intensities as a function of the frame number. Solid lines in the figures indicate the average of 10 frames around. Since the measurement was made at 500 Hz, the horizontal axis corresponds to the time of the measurement.

Figure 4. The results of the single-shot terahertz measurements.

(a) The probe profiles with and without the terahertz waves as the relative delay between THz and probe pulses was changed. (b) THz time traces compared to the step scan method. The black line indicates the time traces captured using the traditional step-scan method. The data with five different relative delay times are shown to demonstrate the precise measurement of the terahertz waveforms.