Photoluminescence Dynamics Spectroscopy
Time Resolved Photo-Luminescence (TRPL) is an experimental technique that provides the spectral and temporal evolution of the emission of a sample following its illumination by a short pulse of light. More precisely, the short pulse of light generates electron-hole pairs that decay to lower energy levels of the sample. These electron-hole pairs can subsequently recombine and emit light. The emitted light is composed of a set of wavelengths corresponding to transition energies of the sample and, as a result, the measurement of the optical spectrum as a function of time provides a means to measure the transition energies and their lifetimes. Since these decay times are on the order of picoseconds or nanoseconds, and the intensity of light emitted can be very weak, a conventional spectrum analyser cannot provide the resolution required. Instead, it is necessary to use a device known as a streak camera.
|Fig. 1: Schematic diagram of a streak camera.|
|Fig. 2: Schematic diagram of the experimental setup for TRPL spectroscopy.|
The streak camera is an ultra high-speed detector which captures light emission phenomena occurring in extremely short time periods. The operating principle is shown in Fig. 1. The light pulse to be measured is projected onto the slit and is focused by a lens into an optical image on the photocathode, able to cover a wavelength range between 300 nm and 1500 nm. Here, the photons are converted into a stream of electrons proportional to the intensity of the incident light. As the electron stream created from the light pulse passes between a pair of sweep electrodes, a time-varying voltage is applied to the electrodes, resulting in a high-speed sweep. This means that the early part of the pulse is deflected less than the later part of the pulse, so that different parts of the pulse strike the micro channel plate (MCP) at different positions. Thus the temporal structure of the pulse is converted into a spatial distribution, or ‘streak’, pattern. As the electrons pass the MCP, they are multiplied several thousands of times and are then bombarded against the phosphor screen, where they are converted back into light. The fluorescence image corresponding to the early part of the incident light pulse is positioned at the top of the phosphor screen, with later parts positioned in descending order; in other words, the axis in the perpendicular direction on the phosphor screen serves as the temporal axis. The brightness of the fluorescence image is proportional to the intensity of the corresponding incident light pulses and the position in the horizontal direction on the phosphor screen corresponds to the wavelength of the incident light.
To perform TRPL experiments, the sample under investigation is mounted on a cold finger in the vacuum cell of a closed cycle helium cryostat (operating from 7 K to 300 K), and is illuminated with very fast (picosecond or femtosecond) laser pulses. The optical response is then collected by a microscope objective and is sent to the entrance slit of an imaging spectrograph, directly connected with the input optic of a streak camera. An overall picture of the schematic experimental setup is depicted in Fig. 2.
|Fig. 3: (a) Schematic of carrier dynamics in an InGaAs/GaAs quantum well. (b) The resultant streak image.|
The typical spectrum collected by a TRPL experiment is called a streak image. It depicts the intensity of the optical response of the sample (colour scale) as a function of the wavelength (horizontal axis) and the time (vertical axis). The example of Fig. 3(a) concerns the emission dynamics of an InGaAs/GaAs quantum well (QW) at 8 K. A 780 nm laser pulse excites the InGaAs QW above the band gap of the material (~890 nm at this temperature), generating photo-carriers in the GaAs barrier. From here they flow into the well, and recombine to emit light. As the number of carriers is depleted, the intensity of the emitted light decays, as is evident in the streak image (Fig. 3(b)). The carrier dynamics occurring during these processes provides not only information about the position of the peak emission, as in a normal PL experiment, but also the lifetime of the excitons involved in the recombination. In fact, considering vertical (Fig. 4(a)) or horizontal (Fig. 4(b)) cuts of the streak image, the trends of the intensity of the emission as a function of time or wavelength, respectively, can be extracted. In this particular example, for InGaAs/GaAs QW, the peak position is measured at 888 nm and the decay time, after a single exponential fitting, is estimated to be 0.7 ns.
|Fig. 4: Plots extracted from the streak image of Fig 3; (a) Intensity versus time at 888nm, (b) Intensity versus wavelength at 0.5ns.|
|Fig. 5: TRPL streak image of Ga(As)Sb QDs with no capping QW.|
|Fig. 6: TRPL streak image of Ga(As)Sb QDs with an In0.2Ga0.8As capping quantum well.|
The mid- and far-infrared wavelength range is of great interest for a number of applications, including medical imaging, free space communications, and gas sensing. For example, many common pollutants have easily identifiable signature absorptions in the infrared, such as hydrocarbons (3.3 μm), HCl (3.55 μm), CO2 (4.25 μm), CO (4.73 μm), NO2 (3.45 μm), N2O (4.5 μm ), NO (5.24 μm) and SO2 (4.0 μm). These long wavelengths are inaccessible to conventional InAs/GaAs devices, and so alternative material systems must be employed. Antimonide-based compounds are a promising candidate for this. In the PDD group, time-resolved photoluminescence (TRPL) measurements were performed on molecular-beam epitaxy (MBE) grown GaAsSb/GaAs quantum dot structures: one with an InGaAs capping quantum well and one without. These structures exhibit a type II band alignment, meaning that the holes are confined in the dots, while the electrons are confined outside the dots, principally via Coulomb attractions. The addition of the QW in the first structure thus serves to increase the confinement of the electrons and facilitate greater electron-hole overlap and recombination rates. TRPL was used to determine the structures' ground state transition energies and their dependence on the number of photo-generated carriers.
The samples were excited using a PicoQuant pulsed laser diode emitting 780 nm, 60ps pulses with a repetition rate of 1MHz. Streak images were recorded with the Hamamatsu streak camera. In these type-II Ga(As)Sb QDs, the photoluminescence peak was seen to red-shift over time (Figs. 5 and 6). In combination with theoretical modelling, this was shown to be connected to Coulomb effects caused by the decreasing number of photo-generated carriers in the dots.
The group also has a vacuum Fourier Transform Infrared (FTIR) spectrometer capable of measuring emissions in the far infrared and THz regions, in particular for observing intraband transitions at low temperature under electrical and optical pumping. This opens up fresh research avenues for the group, with the possibility of exploring and developing optical sensors for pharmaceutical applications, for example.
Dr. Tomasz Ochalski
Tel: +353 21 490 4862