Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Supplementary on-line information Sample technology: The micropillar fabrication is based on a microcavity structure grown by molecular beam epitaxy (MBE). Figure 1 shows the design of the MBE layer sequence. The structure provides strong optical confinement in the growth direction by Bragg reflector stacks composed of 20 and 23 periods of λ/4 GaAs and AlAs layers (69 nm /82 nm) in the top and bottom mirrors, respectively. Here (quantum dots) in the cavity materials. The actual cavity is inserted between the Bragg reflectors and has a thickness of 296 nm (single λ ). At the center of the GaAs cavity an InGaAs QD layer is embedded. Due to the use of a single λ cavity the QDs are positioned at the antinode of the fundamental mode in the growth direction. The investigation of strong coupling effects for individual quantum dots in the micropillars requires us to design the samples appropriately so that the cavity resonance for the pillar diameter of interest is located in a spectral range in which clearly distinct single dot lines occur. Therefore the samples are grown in such a way that in the planar structures the cavity mode is located about 40 meV below the maximum of the dot emission band. The lateral optical confinement in the 1.5 µm micropillars shifts the fundamental mode by about 25 meV to higher energies into the spectral range of the onset of the dot emission band. Dot linewidth: The measured quantum dot linewidth is usually of the order of several tens of µeV. It is due to inhomogeneous broadening e.g. by random charge fluctuations created by the nonresonant excitation around the dot which modify the dot exciton energy during the long data recording times (seconds to minutes). As strong coupling occurs via the emission and reabsorption of a photon within the photon lifetime in the cavity (about 10 ps, much smaller than e.g. recombination time scales) the influence of statistical charge fluctuations is negligible. For the criterion for the onset of strong coupling the use of the homogeneous quantum dot exciton linewidths is appropriate. At low temperatures this linewidths is on the order of a few to about 10 µeV and therefore much smaller than the cavity linewidth. Control of in-plane dot position: We would like to point out, that the in-plane dot position in the micropillar cannot be controlled in the present epitaxial growth and patterning processes. Therefore the in-plane dot location may not coincide with the antinode of the in-plane cavity field. For the estimate of the oscillator strength of the dot this implies that we derive a lower limit. Single dot identification: We have chosen the relative spectral positions of the cavity mode and the quantum dot emission band appropriately that the cavity is placed at the low energy side of the dot emission. In this energy range only few dots emit, facilitating single dot studies. As can be seen in Fig. 2b of the manuscript, there is typically one dot per meV in the energy range around the cavity mode for a 1.5 µm diameter micropillar. In principle it is possible that there are two dots interacting simultaneously with the cavity mode. However, when the dots are tuned out of resonance with the cavity mode a two component spectrum would be expected as long as the dots are separated by the spectral resolution of the set-up which approximately coincides with the spectral resolution used in the experiments (both about 50 µeV). We observe no indication for this in the strong or weak coupling cases investigated. The probability to have two dots within an energy range of 50 µeV can be estimated to about a few times 10-3. In addition, in order to obtain strong coupling both dots would have to be located close to the antinode of the field, i.e. at the center of the cross section of the pillar which reduces the probability further. Therefore it is very unlikely that we observe an interaction of the cavity mode with more than one quantum dot exciton. The oscillator strength of 50 estimated in the manuscript assuming a single quantum dot exciton in the cavity showing strong coupling is also consistent with values observed for studies on natural quantum dots in the literature. Guest et al. have obtained oscillator strength values ranging from 45 to 180 for natural GaAs/AlGaAs quantum dots. Our estimate is at the lower end of this range. This is consistent with the fact, that the dot is most likely not located at the exact maximum of the field. Due to this our evaluation gives only a lower limit for the oscillator strength. Number of photons in a cavity at any given time: We use a laser power of 2 µW for the strong coupling study. As a worst case model we assume that the laser is entirely focussed on the micropillar and we neglect reflection at the cryostat window and the sample surface. As we are using 532 nm excitation the main loss of photons which has to be considered for the population of the cavity is absorption in the thick upper Bragg reflector. Using an absorption coefficient of GaAs of 6x104/cm (M.D. Sturge, Phys. Rev. 127, 768 (1962)) at 532 nm we obtain that only about 0.57 nW reach the cavity. Using again a worst case estimate (i.e. giving us the highest number of photons) we assume that all the photons reaching the cavity are absorbed there and the resulting carrier pairs all reach the dot under consideration for the strong coupling, we obtain 1.5x109 excitons which are created in that dot per second. The photon life time in the present cavities is about 10 ps. As can be seen from the product of the number of excitons per time with the photon lifetime the average population probability of the cavity by photons is 0.015. Therefore effects due to the population of the cavity with more than one photon at a time are negligible.