Long-lived spin polarisation in the charged InP quantum dots

Long-lived spin polarisation in the charged InP quantum dots

Available online at www.sciencedirect.com Physica E 17 (2003) 361 – 364 www.elsevier.com/locate/physe Long-lived spin polarisation in the charged In...

106KB Sizes 2 Downloads 112 Views

Available online at www.sciencedirect.com

Physica E 17 (2003) 361 – 364 www.elsevier.com/locate/physe

Long-lived spin polarisation in the charged InP quantum dots I.V. Ignatieva; b; c;∗ , I.Ya. Gerlovinc; d , M. Ikezawaa , V.K. Kalevicha; b; e , S.Yu. Verbina; c , Y. Masumotoa a Institute

of Physics, University of Tsukuba, Tsukuba 305-8571, Japan Business Laboratory, University of Tsukuba, Tsukuba 305-8571, Japan c Institute of Physics, St. Petersburg State University, Ulyanovskaya 1 Petrodvorets, St. Petersburg 198504, Russia d Vavilov State Optical Institute, St. Petersburg 190034, Russia e Io+e Physico-Technical Institute, St. Petersburg 194021, Russia b Venture

Abstract This work is aimed at the investigation of the spin relaxation in controllably charged quantum dots (QDs). Measuring the dependence of the spin lifetime on the gate voltage in a heterostructure containing InP QDs, we detected the onset of spin conservation when the QD became negatively charged. The spin lifetime in the charged QDs exceeded 1 ns. Various properties of the long-lived spin polarisation of the QDs in longitudinal magnetic 5eld as well as under various excitation conditions are discussed. ? 2002 Elsevier Science B.V. All rights reserved. PACS: 72.25.Rb; 73.21.La; 78.67.Hc Keywords: Quantum dots; Spin polarisation; Magnetic 5eld; Bias

Localisation of carriers in quantum dots (QDs) eliminates the impact of spatial motion on their spin state via spin–orbit interaction, and, for this reason, greatly reduces the spin relaxation rate. In the case of photogenerated carriers, the spin lifetime is limited not only by spin relaxation but also by the recombination of the carriers with typical times of about 1 ns. This limitation is removed in the doped QDs, where the resident (equilibrium) carriers are permanently present. Therefore, the charged QDs are attractive object for the implementation of the long-lived spin



Corresponding author. Institute of Physics, St. Petersburg State University, Ulyanovskaya 1, Petrodvorets, St. Petersburg 198504, Russia. Fax: +7-812-428-7240. E-mail address: [email protected] (I.V. Ignatiev).

memory. Relatively long spin relaxation times were observed recently in n-doped bulk semiconductors [1,2] as well as in n-doped InAs QDs [3]. Our work is aimed at the investigation of spin relaxation in controllably charged InP QDs. We have found that spin dynamics of the neutral and charged QDs are essentially diIerent. The spin polarisation decay time in neutral dots is of the order of tens of picoseconds, whereas in the presence of resident carriers, the spin lifetime can exceed 1 ns. In order to identify the origin of the long-lived spin polarisation, we studied spin dynamics of the charged QDs as a function of the magnitude and orientation of the external magnetic 5eld. In what follows, we present the main results of this study. We studied the photoluminescence (PL) of a heterostructure with a single layer of InP QDs

1386-9477/03/$ - see front matter ? 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1386-9477(02)00811-1

362

I.V. Ignatiev et al. / Physica E 17 (2003) 361 – 364

Degree of circular polarization

0. 8 0. 3 0. 2

0. 6

0. 1 0. 0 -1

0. 4

0

1

Ubias , V 0V -0.3 V

0. 2

-0.4 V -0.6 V 0. 0 0

200

400

600

800

time, ps Fig. 1. Degree of circular polarisation of the PL at diIerent Ubias shown against each curve. Inset: Bias dependence of amplitude of the long-lived polarisation component (). Solid line is the 5t by a Gaussian. ESt = 14 meV.

sandwiched between In0:5 Ga0:5 P barrier layers. The areal density of the QDs was about 1010 cm−2 . The average base diameter of the QDs was about 40 nm and, height about 5 nm. In order to control the charge of the QDs, the sample was provided with a semitransparent indium tin oxide Shottky contact on the top surface and with an ohmic contact on the n+ GaAs substrate. The thickness of the undoped epitaxial layers, to which the electric voltage was applied, was about 0:5 m. In kinetic studies, the PL was excited by the 3-ps pulses of a Ti:sapphire laser within the PL band (quasi-resonant excitation). The PL kinetics was measured with a time resolution of 6 ps using a 0:25 m subtractive double monochromator and a streak camera. The measurements were made in a cryostat with a superconducting magnet. The PL was detected in the backscattering geometry. A key result of the present work is the detection of the long-lived component in the PL polarisation decay. Fig. 1 shows kinetics of the degree of circular polarisation of the PL, circ , measured at diIerent biases. The degree of polarisation was calculated in a standard manner: circ =(I+ −I− )=(I+ +I− ), where I+ and I− are the PL intensities in the co- and cross-polarised light, respectively. At zero and positive biases, the decay of circ contains a long-lived polarisation component, which remains virtually constant during the

PL lifetime of about 300 ps. The amplitude of this component, slow , lies in the range of tens of percents for small Stokes shift (about 15 meV). The value slow rapidly drops with increasing negative voltage and vanishes at Ubias ¡ − 0:5 V. It has been shown in Refs. [4,5] that the InP QDs in the structure under study contain resident carriers (electrons) that arrived from the doped substrate. The negative voltage removes these electrons, so that the QDs become neutral under the voltage Ubias ¡−0:5 V. Therefore, one can conclude that the long-lived circular polarisation of the PL is observed only for the QDs containing resident electrons. As has been shown earlier [6], the main reason for the fast decay of the degree of circular polarisation in the neutral InP QDs is related to the reversible dephasing resulting from a large spread of anisotropic exchange splittings in the ensemble of QDs. The presence of a long-lived spin polarisation in the negatively charged QDs indicates that the anisotropic component of the exchange coupling, in these dots, is, to a considerable extent, suppressed. The reason for this suppression in QDs containing odd numbers of carriers can be readily explained by the fact that such QDs have a half-integer spin and the anisotropic exchange does not split the spin states due to the Kramers theorem [5,7]. In the QDs with even number of carriers, such a mechanism cannot be realised, and we can suggest that these dots, like the neutral ones, should show a fast decay of the degree of polarisation. Indeed, as can be seen from Fig. 1, the decay of circ for the charged dots displays, with an approximately equal weight, both the fast and slow components. In the framework of the model proposed, we can suggest that the dots with even and odd number of the carriers are responsible, respectively, for the fast and slow components. The analogy, noted above, between spin dynamics of neutral dots and of charged dots with even number of carriers is, however, far from complete. The diIerence between them is revealed distinctly in kinetics of circular polarisation of the PL in the longitudinal magnetic 5eld (the Faraday con5guration) shown in Figs. 2 and 3. A characteristic feature of the kinetics for the charged dots is its dependence on the magnetic 5eld direction for a speci5ed sign of circular polarisation of the exciting light. For one-5eld direction, the amplitude of the slow component increases with increasing

I.V. Ignatiev et al. / Physica E 17 (2003) 361 – 364

363

Ubias=0.1 V 0.8

B = +6 T

0.4

B = -6 T

0.2

B=0T

Ubias=-0.75 V

0. 2

B=6T

ρcirc

ρcirc

0.6

0. 4

B=0T 0.0

0

200

400

600

∆ρcirc

0. 0 0

(a)

200

400

6 00

time, ps Fig. 3. Degree of circular polarisation of the PL, circ , for neutral InP QDs at zero magnetic 5eld and B = ±6 T. Note that the polarisation kinetics at B = +6 T is not distinguished from that at B = −6 T. Smooth solid curve is the 5t by a two-exponential function with 1 = 50 ps, 2 = 800 ps.

0.2

0.1

0.0 0

(b)

200

400

600

time, ps

Fig. 2. (a) Degree of circular polarisation of the PL, circ , for charged InP QDs at zero magnetic 5eld and B = ±6 T. Smooth solid curve is 5t by a two-exponential function: circ = 1 exp(−t= 1 ) + 2 exp(−t= 2 ) with 1 = 33 ps, 2 = 2900 ps. (b) DiIerence of the circular polarisations, Lcirc = circ (+6T) − circ (−6T). Smooth solid curve is the 5t by function: Lcirc = 1 (1 − exp(−t= )) with = 150 ps.

magnetic 5eld strength, whereas for the opposite direction, it remains practically constant or even slightly decreases (Fig. 2a). Note that the diIerence between the values of the degree of polarisation for two directions of the magnetic 5eld increases in time with the characteristic constant of about 150 ps (see Fig. 2b). Since we suppose that PL of the QDs with odd number of carriers is completely polarised, the increase of circ in magnetic 5eld can be caused only by changes of spin states of the QDs with even number of carriers. These changes can be naturally attributed to the temperature redistribution of populations of the 5ne-structure components split by the magnetic 5eld. The measured value of should be considered, in this case, as the time of establishment of thermal equilibrium in the spin system. By comparing the data

of Figs. 1 and 2, one can readily note that the time of establishment of equilibrium in the dots with even number of carriers is much longer than the reversible dephasing time. The polarisation dynamics, in these dots, cannot be elementary. First, the dephasing rapidly destroys the light-induced polarisation and only after that the thermal equilibrium between the spin sublevels starts to be established, which is observed experimentally (see Fig. 2). Under the temperature equilibrium, the PL polarisation is determined only by the order of arrangement of the Zeeman sublevels, i.e., by the magnetic 5eld direction. Thus, to explain the experimental data, we have to assume that, in the QDs with even number of electrons, the irreversible relaxation processes occur that lead to establishment of thermal equilibrium in the spin subsystem during the time of the order of 150 ps. As was already mentioned above, we ascribe the long-lived polarisation observed in the absence of the magnetic 5eld to the spin orientation in the QDs with odd number of carriers. The irreversible phase relaxation between spin sublevels of these dots would have led to a decay of the degree of circular polarisation with the same characteristic time = 150 ps. However, as can be seen from Fig. 1, the slow polarisation

364

I.V. Ignatiev et al. / Physica E 17 (2003) 361 – 364

component of PL of the charged dots virtually does not decay during the measuring time (1000 ps). We can conclude from this that the processes of irreversible dephasing in QDs with odd number of carriers are suppressed. For the neutral QDs, the eIect of longitudinal magnetic 5eld on the PL kinetics is revealed in appearance of a slowly varying circularly polarised component, whose amplitude increases with the 5eld strength as shown in Fig. 3. Similar long-lived component were observed by Paillard et al. [8] in PL of the InAs QDs. The analysis shows that the slow component arises in magnetic 5elds for which the Zeeman splitting exceeds the anisotropic one that results in suppression of the reversible dephasing of the spin states. It is important that the PL kinetics of neutral QDs, unlike that of the charged ones, does not show any appreciable dependence of the slow component amplitude on the direction of the magnetic 5eld as seen from Fig. 3. This means that the irreversible spin relaxation rate in neutral dots is low, and lifetime of the electron–hole pair appears to be too small for the temperature equilibrium between the Zeeman components to be established. In summary, the main result of this paper consists in the detection of a nearly undamped component in kinetics of the degree of circular polarisation of PL of the InP QDs with resident electrons. We come to the conclusion that the long-lived polarisation is related to spin orientation in the QDs containing odd number of carriers. The spin orientation in the dots with even number of carriers is destroyed by the reversible dephasing resulting from a spread of the anisotropic exchange splittings. In addition, we have found a dependence of kinetics of the degree of polarisation of the charged QDs on the magnetic 5eld direction. It is concluded that this dependence is related to irreversible spin relaxation leading to temperature

redistribution of populations of the Zeeman sublevels. It is shown that this relaxation can occur only in the QDs with even number of carriers. The reason for the dependence of the irreversible spin relaxation on the number of carriers in the QDs calls for further studies. Acknowledgements The authors are grateful to Dr. T. Okuno for assistance in the experiments, to Dr. K. V. Kavokin and Dr. I. A. Yugova for fruitful discussions, and to Dr. Zapasskii for a critical reading of the manuscript. This work is supported by program “Research for the Future” from JSPS-RFTF97P00106, by INTAS, Grant No. 1B 2167, and by the Russian Foundation for Basic Research. References [1] R.I. Dzioev, B.P. Zakharchenya, V.L. Korenev, M.N. Stepanova, Fiz. Tverd. Tela 39 (1997) 1975 (Phys. Solid State 39 (1997) 1765). [2] J.M. Kikkawa, D.D. Awschalom, Phys. Rev. Lett. 80 (1998) 4313. [3] S. Cortez, O. Krebs, S. Laurent, M. Senes, X. Marie, P. Voisin, R. Ferreira, G. Bastard, J.-M. Gerard, T. Amand, Phys. Rev. Lett. 89 (2002) 207401. [4] Y. Masumoto, I.V. Ignatiev, I.E. Kozin, V.G. Davydov, S.V. Nair, H.-W. Ren, J.-S. Lee, S. Sugou, Jpn. J. Appl. Phys. 40 (2001) 1947. [5] I.E. Kozin, V.G. Davydov, I.V. Ignatiev, A.V. Kavokin, K.V. Kavokin, G. Malpuech, H.W. Ren, M. Sugisaki, S. Sugou, Y. Masumoto, Phys. Rev. B 65 (2002) 241312(R). [6] I.A. Yugova, I.Ya. Gerlovin, I.V. Ignatiev, S. Yu. Verbin, Y. Masumoto, in: Proceedings of the 14th Indium Phosphide and Related Materials Conference, Stockholm, Sweden, May 12–16, 2002, pp. 71–74. [7] K.V. Kavokin, Phys. Stat. Sol. (a) 190 (2002) 221. [8] M. Paillard, X. Marie, P. Renucci, T. Amand, A. Jbeli, J.M. GSerard, Phys. Rev. Lett. 86 (2001) 1634.