Elaboration of II-VI quantum dots

The main fields of study regarding elaboration of II-VI quantum dots (QDs), in our team, in the recent years, are the growth of CdSe/ZnSe QDs for their use as single photon emitters (see part " Quantum optics with II-VI QDs ") and CdTe/ZnTe QDs doped with a single Mn atom for the magneto-optic study of the exciton-Mn spin coupling (see part "Control of a single spin in a QD"). More recently we also started working (i) on CdSe QDs as insertion in ZnSe nanowires, (ii) on type II confining ZnTe/ZnSe QDs and (iii) on QD localization on pre-patterned substrates. Hereafter we present an overview of the most achieved projects, i.e. CdSe/ZnSe self-organized QDs and Mn doped CdTe QDs.

 
People:
Thomas AICHELE, Laurent MAINGAULT, Rita NAJJAR, Ivan-Christophe ROBIN 
 
Results:
 
CdSe/ZnSe self-organized QDs
 
The case of CdSe differs from other strained materials such as Si/Ge or InAs/GaAs: the balance of energies is such that the 2D-3D transition of a strained CdSe layer grown on (001)-ZnSe does not occur spontaneously, in a Stranski-Krastanow growth mode. Here the formation of dislocations is more favorable than the island formation to release strain above the critical thickness. In the literature the formation of CdSe QDs is usually based on annealing processes of the strained layer kept below the critical thickness. We developed a different method to change the surface energy and favor the 2D-3D transition as a possible path to decrease the free energy of the system. It consists in covering a layer of strained CdSe with amorphous selenium by exposing the surface at low temperature (−10 °C) to a Se flux and then to sublimate this amorphous selenium by ramping slowly the sample temperature from −10°C to 280 °C.

To go deeper in the understanding and optimize the process, we have studied the influence of the elastic energy stored in the strained layer by varying the thickness of the CdSe layer by steps of half a monolayer (ML). Figure 1 shows atomic force microscopy (AFM) images obtained ex-situ (left-hand side) and the corresponding RHEED patterns (right-hand side), just after the desorption of the amorphous selenium, for three different thicknesses of CdSe layers and for the reference bare ZnSe surface. The CdSe layers were deposited by atomic layer epitaxy (ALE) at 280 °C to benefit from the well known self regulation growth mode at half a monolayer per cycle and then to obtain a very precise control of the thin layer thickness. We observed that the thickness of the strained CdSe layer plays a key role in the strain relaxation process: CdSe island formation is only observed when 3 ML of CdSe are deposited on ZnSe, whereas the deposition of 2 or 2.5 ML of CdSe leads to strong undulations on the surface, parallel to the [110] direction. The observation of a 2D RHEED pattern is due to the fact that the period of the undulations is two large to be probed at the scale of the RHEED electron beam. The AFM image for 3 ML is drastically different: the formation of small circular islands clearly occurs on the surface ( 3x1010 islands/cm2). This result is in agreement with the spotty RHEED pattern.

Moreover we determined, using the distance between electron diffraction streaks (not shown), that the critical thickness for plastic relaxation occurs between 3 and 3.5 ML/s. Thus, in order to carry out a 2D-3D transition under the best conditions for island formation, it is necessary to keep the CdSe layer thickness in a very narrow window around 3 ML, close to the critical thickness but without exceeding it. The stored elastic energy is then maximum while avoiding the formation of misfit dislocations.

Additionally, two samples were investigated in cross section by TEM (see also part "TEM studies of nanostructures"). Sample A consists of 3 ML of CdSe capped under amorphous selenium whereas sample B was covered in two steps: the amorphous layer was sublimated to reveal the islands and a second deposition of amorphous selenium was performed to protect the surface. In figure 2(a), a high resolution TEM image of sample A, taken along the [110] zone axis, shows a very flat surface. On the other hand, on sample B, well defined islands, about 20 nm wide and 4 nm high are observed (figure 2(b)). The comparison between samples A and B clearly indicates that the island formation occurs during the Se desorption.
 
Mn doped CdTe QDs
 
We also mastered the elaboration of CdTe/ZnTe QDs which is quite similar to the case of selenium based QDs. The specific point we worked on during the last few years is the doping of the CdTe QDs with a single Mn atom for magneto-optical studies. We are not able yet to correlate the presence of a Mn atom and the formation of QD, using for example Mn as nucleation center. The QD doping is then based on statistics, limited by a low occurrence. With a simple quantitative model, we determined the most appropriate Mn density to be incorporated prior to the quantum dot nucleation to enhance the number of QDs doped with exactly one Mn atom. Our criterion for the detection of such a configuration is a unique feature observed in the photoluminescence spectra of a single QD: six narrow lines should appear corresponding to the interaction between an exciton confined in the QD and a Mn ion, whose 5/2 spin has six projections along the growth axis. When more Mn ions interact with the confined exciton, the photoluminescence is broadened by multiple coupling and statistical fluctuations.

A geometric modelization shows that if we optimize the ratio of the Mn density to the QD one, we could get almost half of all the QDs that contain only one Mn. On a surface with only one Mn ion and N QDs, the probability to have this Mn ion into a given QD is p1 =SQD x dQD / N, where dQD is the QD density and SQD is the area of a QD. If they are k>1 atoms on the surface the probability is then pk=k(p1)(1-p1)k-1. This probability is based on a random distribution of both QDs and Mn ions and does not assume any correlation between the QD nucleation and the presence of Mn ions. Then we found that whatever the QD size, an appropriate Mn density do exist to optimized the probability for a QD to contain a single Mn. The maximum probability is 40% (dotted line figure 3).

In practice, a single Mn is optically detectable, with six well separated lines, only if the Mn ion is near the center of the QD and if all the other Mn ions are not in the vicinity of this QD [Phys. Rev. B 73, 045301 (2006)]. Taking into account such additional conditions (solid line on figure 3) induces a strong reduction of the maximum probability to get only one Mn into the QD and also reduces the number of Mn ions needed, with respect to Cd, to get this maximum to a very low value: 0.02%. Such a Mn concentration corresponds to a Mn flux too low to be measured and difficult to control precisely. The first approach we used was to extrapolate the Mn flux measured at higher values using a well calibrated activation energy of an Arrhenius law. Another very specific approach allowed us to precisely control the Mn composition in spite of its low value. The CdTe/ZnTe structure was grown on top of a thin ZnMn6%Te buffer layer. Then we used the segregation effect of Mn ions into ZnTe during growth: the residual amont of Mn in ZnTe is known to be divided by 2 at each additional mono-layer. Then the thickness of the ZnTe spacer between ZnMnTe and the CdSe QDs allows us to precisely adjust the Mn content in the QD layer.

[1] I-C.Robin, R.André, C.Bougerol, T.Aichele, S.Tatarenko, Appl. Phys. Lett. 88, 233103 (2006).

[2] L.Maingault, L.Besombes, Y.Léger, C.Bougerol, H.Mariette, Appl. Phys. Lett. 89, 193109 (2006).

© Institut Néel 2012 l Webdesign chrisgaillard.com l Powered by spip l Last update Sunday 2 February 2020 l