Pump-Probe Spectroscopy

In a pump-probe experiment, a pump pulse excites a sample and induces changes that are measured using a subsequent probe pulse. By varying the time between pump and probe pulses, one can retrieve the recovery time scales of the sample. A concrete example is the measurement of the transmittance through a sample. The pump pulse modifies the transmittance of the sample. For small time difference between the pump and probe, the transmittance of the probe will be low and will increase with increasing time delay. By measuring the transmittance as a function of the delay, it is possible to observe the transmittance recovery timescale of the sample. Knowledge of the relaxation of the complete optical gain and index after a sub-picosecond light pulse is essential to assess the potential of current quantum dot materials for photonic systems applications. Pump-probe techniques provide direct, time domain, measurement of gain and refractive index nonlinearities in optical waveguides with sub-picosecond resolution.

Pump Probe Setup

Fig. 3.1: Schematic diagram of the experimental setup for heterodyne pump-probe spectroscopy. Click on the image for a larger view.

Pump-probe movie

Fig. 3.2: Animation showing how the gain recovery following the pump pulse (red) is measured by varying the delay between the pump (red) and probe (blue) pulses.

Pump-probe spectroscopy is necessary since the timescales of carrier processes in materials (typically tens of pico-seconds) are usually much faster than the bandwidth of conventional detectors. Pump-probe experiments are commonly carried out using picosecond or femtosecond lasers that emit pulse trains. In such a case, the light emitted by a laser is separated into pump and probe beams that reach the sample at different times and, after propagation through the sample, the two beams are again separated in order to analyse only the probe beam. The separation of the pump and probe beams after propagation through the sample can be achieved if the two beams have orthogonal polarisation or different frequencies. This latter solution has led to the development of heterodyne pump-probe spectroscopy.

The heterodyne pump-probe technique, in which pump and probe pulses are distinguished by inducing a small frequency shift between them, was demonstrated for the first time by K. Hall in the mid 1990’s. The technique allows separate extraction of the gain and refractive index dynamics in the waveguide and works for orthogonally as well as parallel polarised pump and probe pulses. In the experiment, a high intensity short pump pulse is used to excite the investigated sample by modifying various carrier populations. It leads to changes in the optical properties of the sample, which can be measured as a change of the transmission (gain) or phase (refractive index) of the low intensity short probe pulse.

In order to follow the gain and refractive index recoveries, the delay between pump and probe pulses is changed and the time resolution of the measurement is determined mainly by the duration of the pulses. Both pump and probe pulses follow the reference pulse which is used for detection purposes. The pump-probe technique provides gain and refractive index recovery at true operating conditions which are essential for the assessment of the optical devices used in ultra-fast signal processing.

Phase measurement

A change in the amplitude of the reference-probe beating signal, caused by pump induced changes in the probe transmission, is measured using the R component (magnitude) of the high frequency lock in amplifier. Similarly a pump induced change in the refractive index, will change the phase of the probe beam with respect to the reference beam. This change is measured directly by the in-phase (X) and out-of-phase (Y) components of the high frequency lock-in amplifier.

In order to perform background free measurements, the pump beam is mechanically chopped at low frequency and the analog output of the high frequency lock-in is connected to a low frequency lock-in amplifier. The low frequency lock-in locks onto this signal, thereby eliminating all changes in the signal which do not oscillate at the same frequency as the chopper.