Absorption and photoluminescence spectroscopy on a single self-assembled charge-tunable quantum dot

We have performed detailed photoluminescence (PL) and absorption spectroscopy on the same single self-assembled quantum dot in a charge-tunable device. The transition from neutral to charged exciton in the PL occurs at a more negative voltage than the corresponding transition in absorption. We have developed a model of the Coulomb blockade to account for this observation. At large negative bias, the absorption broadens as a result of electron and hole tunneling. We observe resonant features in this regime whenever the quantum dot hole level is resonant with two-dimensional hole states located at the capping layer-blocking barrier interface in our structure.

Figure 1

Schematic sketches of the heterostructure showing the conduction and valence band edges along the growth direction. A voltage $V_g$ is applied to the gate electrode. An electric field results from both the applied field and the Schottky voltage $V_s$. The QDs are separated from the $n$-doped GaAs by a tunnel barrier, $25\mathrm{nm}$ of undoped GaAs, and from the blocking barrier by $30\mathrm{nm}$ of undoped GaAs. The lever arm of the device is $\lambda =7$. The lower diagram labels the energies used in the electrostatic calculations of the device.

Figure 2

(Color online) (a) PL spectrum of an ${\mathrm{X}}^{1-}$ exciton. The linewidth, $100\mu \mathrm{eV}$, is determined by the resolution of the setup and not by the QD itself. (b) Absorption spectrum on the same QD as (a). The resonance corresponds to excitation of the ${\mathrm{X}}^{1-}$. The resolution is not setup limited and the resonance exhibits a linewidth of $2.3\mu \mathrm{eV}$. Both spectra were measured at the same gate voltage and at a temperature of $4.2\mathrm{K}$. The solid lines are Lorentz curves fitted to the data points.

Figure 3

(Color online) Resonant gate voltage against pump power density for PL and absorption spectroscopy, in both cases keeping the wavelength constant. The data were taken from two different QDs at $4.2\mathrm{K}$ and represent typical behavior for the QDs embedded in the heterostructure of Fig. 1. The line describing the PL data is a guide to the eye.

Figure 4

(a) Behavior of the PL and absorption energies as a function of gate voltage, $V_g$. The PL energies are corrected for space charge effects as described in the text. (b) Energy of the 0, e, 2e, h, ${\mathrm{X}}^0$ and ${\mathrm{X}}^{1-}$ states as a function of $V_g$. The bold lines show the energetically favored states with (upper part) and without (lower part) a hole.

Figure 5

(Color online) Absorption in two different gate voltage regions. The upper graph shows the absorption resonance at a gate voltage of $-0.3\mathrm{V}$ where the linewidth is not influenced by tunneling and the lower graph shows the resonance at $-1.3\mathrm{V}$ where the linewidth is completely dominated by tunneling. The solid lines are Lorentzian fits to the data with linewidths $2.25\mu \mathrm{eV}$ at $V_g=-0.3\mathrm{V}$ and $136\mu \mathrm{eV}$ at $V_g=-1.3\mathrm{V}$.

Figure 6

(Color online) (a) Behavior of the ${\mathrm{X}}^0$ linewidth as function of gate voltage. The dashed line shows the expected linewidth calculated by WKB for electron tunneling and the solid line shows the linewidth for electron and hole tunneling. (b) Plot of the energy of the QD hole state and the energies of the 2D hole states confined at the interface between the capping layer and the blocking barrier versus gate voltage. The 2D states are labeled with an index with 1 describing the ground state. The energy zero lies at the top of the valence band at the interface between the capping layer and the blocking barrier.