Single III-N quantum dot spectroscopy

People :
Régis André, Joël Bleuse, Le Si Dang, Bruno Daudin, Bruno Gayral, Jean-Michel Gérard, Kuntheak Kheng, Henri Mariette, Marlène Terrier
Benoît Amstatt, Sebastien Founta, Laurent Maingault, Sebastien Moehl, Julien Renard, Fabian Rol, Franck Tinjod, Adrien Tribu 
Overview and results
GaN quantum dots 
GaN/AlN quantum dots (QDs) have been grown for about ten years, but their electronic properties remain little known. This is in particular due to the difficulty to master the growth of such structures, to the complex valence band structure in the wurtzite phase, to the unknowns concerning the residual doping, to the peculiarities of the internal electric field for wurtzite structures and to the difficulty to perform optical spectroscopy in the UV (250-350 nm) range. In order to gain a better understanding of these structures, we decided to build a dedicated UV microphotoluminescence set-up to study the behavior of single GaN/AlN quantum dots. This set-up is based on a UV refractive microscope objective with NA=0.4 allowing to focus a doubled argon laser (exc=244 nm) on a 1 µm spot. This study was focused on non polar [11-20] GaN/AlN quantum dots (see the chapter on “Non-polar and polar nitride quantum dots” for more details). Time-resolved experiments on ensemble of these QDs have shown that a sizeable quantum confined Stark effect is not observed in these structures.
The typical density of QDs ranges between 5 1010 and 2 1011 cm-2 in these samples. In order to isolate a few QDs under the laser spot, a processing of the sample is thus needed. This is done either by etching mesas by plasma etching, or by opening holes in an Al mask deposited on the sample. In this way, sharp luminescence lines can be isolated and are attributed to QD discrete transitions. These sharp lines are only observed in the high energy part of the QD distribution, for reasons that are not clear to us. We checked that these lines stem from single electron-hole pair recombination by power dependent measurements. The observed linewidths range between 0.5 and 2 meV. These linewidths are due to spectral diffusion, which is particularly important in III-N heterostructures. The lines are narrow enough to probe the broadening as a function of temperature. This was done for temperatures ranging between 4 and 140 K. Asymmetric phonon wings due to acoustic phonon coupling are observed (fig. 1). The modeling of these phonon wings allow to deduce a localization length for the exciton which is smaller than what would be expected from the sole AlN barrier confinement. The system thus behaves as if there was a “sub-quantum dot” within the dot.
This localization effect was probed in another way by analyzing the time-resolved photoluminescence. Indeed, the average decay time for an ensemble of such non-polar QDs is around =250 ps. This might seem fast for quantum dots, but due to the 2/ dependency of the oscillator strength, this corresponds only to an oscillator strength of 9. This result is puzzling for the following reason : the typical QD lateral size is 20 nm (and height 2 nm), while the 2D Bohr radius is around 2nm. One would thus expect the confined exciton to be in the center-of-mass confinement regime. The radiative transition should thus be in the giant oscillator strength regime linked with a bound exciton having a large coherence surface. This is here not the case, which is again coherent with a supplementary localization inside the quantum dot. We measured the time-resolved photoluminescence on single non-polar QDs (fig. 2). The decay is mono-exponential for the three measured QDs. The decay times do not vary monotonously as a function of emission energy. This is consistent with the picture where the in-plane localization is independent of the QD size, so that the decay time depends on the in-plane localization for each QD and not on the QD size (and thus not on the QD transition energy) [1]
The prospects for this study is to understand this localization phenomenon, to study other non-polar QDs such as [11-10] QDs, and to also apply the same experimental techniques to polar GaN QDs, and to GaN QDs embedded in nanowires.
II-VI quantum dots
The strong Coulomb interaction in II–VI QDs as compared to most III–Vs (except GaN dots) makes them very attractive for single photon generation and quantum optical experiments. Our previous detail studies of CdTe/ZnTe QDs has allowed us to identify excitonic complexes (charged exciton, biexciton and multi-exciton states etc..) and single photon emission has been evidence by correlation measurement at low temperature. To go further, one has to investigate the feasibility of such devices at higher temperature. Emission at room temperature can not be achieved in the CdTe/ZnTe system due to the very small unstrained valence band offset (VBO) of the system.
To overcome this limitation, we have grown and study CdTe QDs embedded in ZnMgTe barriers. The incorporation of Mg into ZnTe barriers is expected to increase the potential confinement for the heavy hole, and so to significantly strengthen the excitonic recombination. This is indeed revealed in our PL studies by both an extension of the radiative recombination regime up to higher temperatures (to 150 K for 30% Mg, instead of 60K for pure ZnTe barriers) and an increase of the activation energy of nonradiative recombination. However, the in-plane confinement is less enhanced by Mg, which allows observation of interdot carrier transfer with increasing temperature (thermally activated process via the 2D wetting layer states), as evidenced directly by analysis of the PL intensities of different single dots [2]. We have shown that the confinement of carriers in CdTe QDs can be significantly enhanced by inserting thin MgTe layers surrounding the QD plane [3]. Furthermore, the careful analysis of the excitonic luminescence lineshape as a function of temperature has proven that the use of the more confining ZnMgTe barriers surrounding CdTe quantum dots is an efficient way to reduce two different kinds of exciton-phonon interaction, namely anti-Stokes scattering with acoustic and optical phonons as well as mixing of excitonic states with the acoustic phonon continuum [4].
Room temperature operation of single photon source requires that the separation of the exciton and the biexciton remain larger than the thermal broadening of the optical linewidths. From the study of the thermal broadening of single dot emission lines (fig. 3) we have shown that the possibility to distinguish exciton from biexciton in CdTe (with ZnMgTe barriers) was limited to a temperature of about 200K. To go further, we have to use CdSe/ZnSe QDs for which single photon emission should be possible up to room temperature [5]. The study of CdSe/ZnSe QDs grown in our group reveals that the QDs are naturally n doped. Magneto-optical studies of negative trions in Voigt geometry have shown that the hole g-factor is not only nonzero, but can reach values as high as the electron g-factor. Moreover, we find a correlation between the hole g-factors and the intensity differences within the trion quadruplet [6] that appears in Voigt geometry .These facts reveal strong heavy hole–light hole mixing which is particularly favorable for weakly strained QDs. Such a weak strain may arise from a large size of the QDs, a certain Zn content inside the dots and/or defects generated next to the dots.
The fine-structure splitting (FSS) of the radiative excitonic doublet, which results from the electron-hole exchange interaction in elongated QDs, hampers the generation of entangled photon pairs. Indeed, the two possible cascades for a biexciton X2 to recombine into an exciton X become energetically distinguishable. In collaboration with the group of Prof. Forchel (Physikalisches Institut, Universität Würzburg) we have investigated postgrowth thermal annealing (TA) of our CdSe QDs. We have demonstrated that the FSS can be modified by TA (fig. 4). With increasing temperature of TA, an initial diminution of the structure splitting, followed by its sign reversal and sharp increase in the opposite direction, has been recorded [7]. We assign this observation to an enhanced interdiffusion close to the side-walls of the small mesa due to defects generated during the etching process. Also, the binding energy of the charged exciton shows a maximum when the QD is in the high symmetry configuration, and drastically reduces subsequently when the QD undergoes an elongation. Such an experimental method, of tuning the FSS and controlling the QD anisotropy, is an important step toward the generation of entangled photons.
 References :
[1] Probing exciton localization in nonpolar GaN/AlN quantum dots by single-dot optical spectroscopy, F. Rol, S. Founta, H. Mariette, B. Daudin, Le Si Dang, J. Bleuse, D. Peyrade, J.-M. Gérard, and B. Gayral, Phys. Rev. B 75, 125306 (2007)
[2] CdTe/Z1-xMgxTe self-assembled quantum dots : Towards room temperature emission, F. Tinjod, S. Moehl, K. Kheng, B. Gilles, and H. Mariette, J. Appl. Phys. 95, 102 (2004)
[3] Reduction of exciton-phonon interaction due to stronger confinement in single quantum dots
S. Moehl, F. Tinjod, K. Kheng, H. Mariette, Phys. Rev. B 69, 245318 (2004)
[4] Enhanced carrier confinement in quantum dots by raising wetting layer state energy
S. Moehl, L. Maingault, K. Kheng, and H. Mariette, Appl. Phys. Lett. 87, 033111 (2005)
[5] Temperature Dependence Of The Exciton Homogeneous Linewidth In CdTe and CdSe Self-assembled Quantum Dots : Limit Of Single Photon Source Operation, K. Kheng, S. Moehl, I. C. Robin, L. Maingault, R. Andre, and H. Mariette, AIP Conference Proceedings ICPS 2006, Volume 893, pp. 917 (2007)
[6] Strong heavy-hole–light-hole mixing in CdZnSe quantum dots, S. Moehl, I. C. Robin, Y. Léger, R. André, L. Besombes, and K. Kheng, Phys. Stat. Sol. (b) 243, No. 4, 849 (2006)
[7] Annealing induced inversion of quantum dot finne-structure splitting, E. Margapoti, A. Tribu, T. Aichele, L. Worschech, A. Forchel, R. André, and K. Kheng, Appl. Phys. Lett 90, 181927 (2007)
In collaboration with :
David Peyrade1, Emanuela Margapoti2
(1) Laboratoire des Techniques de la Microélectronique, (2) Université Würzburg (Allemagne)
© Institut Néel 2012 l Webdesign l Propulsé par spip l Dernière mise à jour : mercredi 5 février 2020 l