InGaAs/GaAs 量子ドットのナノインデンテーションによる発光停止と位置推定

指導教員: 埼玉大学大学院理工学研究科教授 荒居善雄In this thesis, the low-temperature (10 K) photoluminescence (PL) of self-assembled InGaAs/GaAs quantum dots (QDs) was studied under the elastic indentation of a flat cylindrical or dome-shaped nanoprobe that generates localized strain fields around itself. As the indentation force increases with the flat cylindrical nanoprobe indented at a fixed position, the intensity of the enhanced fine PL peak from a single QD was observed to firstly increase, followed by a decrease, and be finally quenched, while the energy of the PL peak was linearly blueshifted with the force. The PL peak energy shift per unit force, i.e., the blueshift rate, was measured in the range from 22.7 to 87.8 meV/mN. The observed force at which a PL peak disappears, i.e., the quenching force varies from QD to QD. This variation is ascribed to the diversely distributed strain fields in and around each QD and therefore can be related to the QD location with respect to the nanoprobe center. In the case of using the dome-shaped nanoprobe for the indentation with a horizontal scan, the PL emissions from some of observed QDs were firstly blueshifted then symmetrically redshifted with the movement of the probe position during the scan. This symmetric blue- and red-shift was not observed in the indentation of scan of the flat cylindrical probe. In order to clarify the mechanism of PL quenching and the influence of probe tip shape on the PL emissions from QDs, a numerical simulation of the strain distribution is carried out by a 3-dimensional finite element method, depending on the shape of nanoprobe used in the experiments. The modification of the energy band structure resulting from strain is then calculated based on the deformation potential theory. With a quantitative relationship between the blueshift rate and the QD’s distance from the probe center, derived from the simulation, we proposed a method to estimate the location of the embedded QDs from the experimentally measured blueshift rate. The observed QDs were determined to be located around the nanoprobe edge, which coincides with the observation that only the PL emitted from the QDs located under the probe edge is enhanced and observable under nanoprobe indentation. The estimation method presented in this study provides a reliable and simple way to estimate the location of the embedded QDs with respect to the probe center by nanoprobe indentation. Based on the estimation of the locations of QDs under the indentation of the flat cylindrical nanoprobe, the observed PL quenching was analyzed quantitatively to be ascribed to two possible mechanisms, the crossover between the Γ and X or L band of InGaAs and the electron-repulsion resulting from the strain-induced potential gradient. The crossover between the Γ and X or L band of InGaAs was deduced to be possible only on the basis of assuming low X or L band gap energy of InAs and high bowing parameter for the X or L band gap energy of InGaAs. On the other hand, if the electron-repulsion resulting from the strain-induced potential gradient was responsible for the PL quenching, an indentation-induced potential height at quenching was derived to be from 43.5 to 133.5 meV. Calculations in the case of dome-shaped nanoprobe have shown that the change in the probe tip shape modified the generated strain field in and around single QD, thus resulting in the different behavior of PL emissions observed in the indentation experiments. With the PL results from the indentation of scan of dome-shaped probe, the locations of those QD, whose PL emissions exhibited symmetric blue- and red-shift, were directly obtained from the experiment at which the PL emissions from those QDs reached the maximum values. This is an advantageous improvement in the estimation of the location of embedded QDs achieved by using the dome-shaped probe, which can not be realized in the case of flat cylindrical probe