Carrier distribution processes in Quantum Dot ensembles

• Main Author: Hutchings, Matthew D.
• Published: Cardiff University 2012
• Subjects: 537.6; QC Physics
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.567533

In this thesis the development of new analysis methods that study the carrier distributions in quantum dots (QDs) directly from experimental measurement of spontaneous emission and gain spectra are described. These were applied to three InAs QD structures that are nominally identical except for the doping type in the active region, one p-doped, one n-doped and one left un-doped for comparison. The effect that carrier localisation within individual dots had on this temperature dependence of the carrier distribution under injection was studied and related to key aspects associated with laser device performance. The nature of QD occupation in the three samples was determined through measurement of the carrier temperature (TC) of the electrons populating the QD states. It was found that the un-doped samples QDs were in thermal equilibrium with the bulk lattice down to 200 K. Below this temperature the sample’s QD states become decoupled from the lattice and at 60 K QD occupation was shown to be random. The p-doped sample was shown to be non-thermal between 300 K and 200 K where at 150 K the occupation of QDs became random. The TC was observed to decrease for this sample below 200 K and this was attributed to fewer dopants ionising as the temperature decreased. The n-doped sample was also shown to be non-thermal between 300 K and 200 K with the QD occupation becoming random at 100 K. In all three samples, above 300 K, the measured TC was lower than that predicted by a Fermi-Dirac distribution and this was attributed to the these QDs having a large number of closely spaced hole states leading to a size dependence of the number of these states. This means an individual ΔEf exists for a given set of dot sizes. So emission from an ensemble of dots is “smoothed” across different ΔEf levels leading to a reduction in the apparent TC. These results have a significant effect on the threshold current densities of these samples and suggest that the differences observed due to doping will not be reproduced by calculations assuming a quasi-thermal equilibrium across the QD structure. The temperature dependence of the shift in gain peak energy was determined for the un and p-doped samples. This showed that the blue-shift of the gain peak due to state-filling in un-doped QD structures is independent of temperature, at a given value of peak gain, over the temperature range studied (200 K to 350 K). In the pdoped sample however, the state filling is temperature dependent at any fixed gain with a shift of 8meV observed between 200 K and 350 K. This was attributed to the wide electron state distribution and the lowering of the electronic quasi-Fermi level by the p-doping. This renders p-doped materials unsuitable for any technology application where gain peak wavelength temperature stability is required for efficient operation.