Polarizing nuclear spins in quantum dots by injection of a spin-polarized current Pablo Asshoff, Gunter Wüst, Andreas Merz, Heinz Kalt, and Michael Hetterich Citation: AIP Conference Proceedings 1399, 681 (2011); View online: https://doi.org/10.1063/1.3666560 View Table of Contents: http://aip.scitation.org/toc/apc/1399/1 Published by the American Institute of Physics Polarizing nuclear spins in quantum dots by injection of a spin-polarized current Pablo Asshoff, Gunter Wüst, Andreas Merz, Heinz Kalt and Michael Hetterich Institut für Angewandte Physik, Karlsruhe Institute of Technology (KIT) and DFG Center for Functional Nanostructures (CFN), 76131 Karlsruhe, Germany Abstract. We demonstrate spin-polarization of the nuclei in self-assembled InGaAs quantum dots due to electrical injection of a spinpolarized current. High-resolution optical spectroscopy of the Overhauser shift is used as detection method. The electrically induced spin-polarization is of comparable magnitude as for spin-injection by photoexcitation. The results imply that efﬁcient nuclear spin-polarization in quantum dots can be achieved by purely electrical means. Keywords: Electrical spin-injection, quantum dots, dynamic nuclear polarization, spin-LED PACS: 72.25.Dc, 85.75.-d, 72.25.Hg, 72.25.Rb, 78.67.Hc Quantum dots (QDs) are candidates for a spin-based quantum information processing. The information can be stored either in the spin state of an electron conﬁned within the QD, or in nuclear spin states. These states are coupled to each other by the hyperﬁne interaction: the injected spin-polarized electrons align the nuclear spins via electron-nuclei ﬂip-ﬂop processes. The aligned nuclear spins produce an effective magnetic ﬁeld BN , which in turn affects the electron spins. The electron spins experience a total magnetic ﬁeld of an externally applied ﬁeld B0 plus the nuclear magnetic ﬁeld BN . The difference of the Zeeman splitting between the spin subbands originating from the nuclear ﬁeld results in the so-called Overhauser shift, which can be resolved by spectroscopic methods [1, 2]. Here, we investigate if efﬁcient spin-polarization of the nuclei is possible by electrical injection of spinpolarized electrons. A convenient choice for this is to use single quantum dots embedded in a spin light emitting diode (spin-LED) , which allows for analyzing the Overhauser shift from the light emission from the QDs. We compare the result with an optically induced spinpolarization of the nuclei. These two injection methods are relevant for approaches involving the use of spin states [4, 5, 6]. The investigated quantum dots are incorporated in an all-semiconductor spin-LED (Fig. 1). For electrical spin injection, electrons are injected through the diluted magnetic semiconductor ZnMnSe, which acts as a spin aligner when a magnetic ﬁeld is applied. The layers were deposited in two molecular-beam epitaxy (MBE) systems on a GaAs:Zn(001) substrate (p ∼ 1 × 1019 cm−3 ). In the ﬁrst system, designed for the growth of III–V materials, on the substrate a ∼ 500 nm layer of GaAs:Be (p ∼ 1 × 1019 cm−3 ) was grown, followed by 100 nm FIGURE 1. Cross-sectional schematical view (not to scale) of the spin-LED. For electrical operation, electrons are injected from the top In contact pad, traverse the Zn0.95 Mn0.05 Se:Cl layer and leave it with a deﬁned spin-polarization. The optically active InGaAs QDs and the wetting layer are sandwiched between thin GaAs layers, which renders a suitable band structure for the injection process. Unpolarized holes are ﬂowing into the InGaAs QDs and the WL from the bottom part of the spinLED, leading to radiative recombination. For optical excitation, electron-hole pairs are generated in the WL, both carrier types are captured in the QD states, and recombine. The gold mask on top of the structure, which contains the apertures, is indicated. i-GaAs, the InGaAs QDs / wetting layer (WL) (see  for details) and a 25 nm thick i-GaAs spacer. Plan-view transmission electron microscopy revealed a QD sheet density of ∼ 5 × 1010 cm−2 . The heterostructure was then transferred to the second MBE facility, designed for the growth of II–VI materials, where 750 nm of the spin aligner material Zn0.95 Mn0.05 Se:Cl (n ∼ 1018 cm−3 ) followed by a 200 nm layer of ZnSe:Cl (n = 5 × 1018 cm−3 ) were deposited. The latter layer improves the ohmic contact to the subsequently evaporated In contact pad. Then, gold apertures were fabricated with electron beam lithog- Physics of Semiconductors AIP Conf. Proc. 1399, 681-682 (2011); doi: 10.1063/1.3666560 © 2011 American Institute of Physics 978-0-7354-1002-2/$30.00 681 FIGURE 2. CPD of the emitted light from a speciﬁc QD plotted against the Overhauser shift (B = 6 T, T = 5 K), error bars are obtained from the ﬁts. Blue and red data points correspond to experiments with optically generated and electrically injected charge carriers, respectively. The dashed line is deduced from the PL experiments. Since for photoexcitation with linearly polarized light the polarization is very sensitive to small misalignments of the optical components, the CPD slightly deviates from zero for this conﬁguration. raphy to access the luminescence of single QDs. Finally, optical lithography was employed to deﬁne squareshaped spin-LEDs, one of which with a surface area of (400 μ m)2 is investigated here. The sample was placed in Faraday geometry in a magneto-optical cryostat at a temperature of T = 5 K and with a ﬁeld B0 = 6 T. To investigate a single QD, one of the gold apertures on top of the spin-LED was placed in the focus of a 60× microscope objective mounted inside the cryostat with piezoelectric actuators, providing an optical path from the sample surface to the outside of the cryostat and vice versa, used for detection and laser excitation, respectively. The emission of the light was guided by a ﬁber to a 0.85 m double monochromator (Horiba, Spex 1402) with two 1200 grooves/mm gratings. The dispersed light was then detected by a CCD (Andor, iDus DU420A-BR-DD). By inserting a quarterwave plate and a linear polarizer in the optical path, we could additionally determine the circular polarization degree (CPD) of the emitted light, which corresponds to the spin-polarization inside the quantum dot , and is deﬁned as (Iσ + − Iσ − )/(Iσ + + Iσ − ), with Iσ +(−) denoting the intensity of σ +(−) -polarized light. In a ﬁrst experiment, we determined the Overhauser shift and the CPD from the spectra when the quantum dots were electrically excited. We applied a voltage across the structure and generated a spin-polarized current, leaving the quantum dots ﬁlled with spin-polarized electrons . When resident electrons recombine with holes ﬂowing into the device from the p-doped side of the spin-LED, new electrons can populate the quantum dot, and transfer spin-polarization to the nuclei, leading to a quasi-equilibrium. A constant current I = 8 mA 682 (U ∼ 4 V) was ﬂowing through the device, resulting in a high current density, with the Overhauser shift being saturated. For comparison, in a second experiment we generated a spin-population in the QDs by CW photoexcitation, and measured the Overhauser shift and CPD. As for the electroluminescence measurements, high enough excitation intensities were used to reach a regime where the Overhauser shift is saturated. For the photoexcitation, we changed the helicity of the impinging light in order to obtain Overhauser shifts for varying average electron spinpolarizations. In Fig. 2 the magnitude of the Overhauser shift in relation to the spin-injection efﬁciency (corresponding to the CPD) is shown for a single QD. Results obtained from the spectra for the cases of electrical excitation and photoexcitation are plotted. Data obtained from photoexcitation are shown as blue data points. A linear approximation  is indicated by the dashed line in the graph. As can be seen, a similar relationship between CPD and Overhauser shift as for the photoexcitation occurs. Indeed, the Overhauser shift is of comparable magnitude as deduced from the photoexcitation experiments. This implies that electrical excitation leads to efﬁcient spinpolarization of the nuclei which corresponds to the case of photoexcitation. This work has been performed within project A2 of the DFG Research Center for Functional Nanostructures (CFN). It has been further funded by a grant from the Ministry of Science, Research and the Arts of BadenWürttemberg (Az.: 7713.14-300). P.A. acknowledges ﬁnancial support from the Karlsruhe School of Optics and Photonics (KSOP) and the Karlsruhe House of Young Scientists (KHYS). REFERENCES 1. A. W. Overhauser, Phys. Rev. 92, 411 (1953). 2. T. Yokoi, S. Adachi, H. Sasakura, S. Muto, H. Z. Song, T. Usuki, and S. Hirose, Phys. Rev. B 71, 041307(R) (2005). 3. W. Löfﬂer, M. Hetterich, C. Mauser, S. Li, T. Passow, and H. Kalt, Appl. Phys. Lett. 90, 232105 (2007). 4. M. Ghali, T. Kümmell, J. Wenisch, K. Brunner, and G. Bacher, Appl. Phys. Lett. 93, 073107 (2008). 5. A. Greilich, S. E. Economou, S. Spatzek, D. R. Yakovlev, D. Reuter, A. D. Wieck, T. L. Reinecke, and M. Bayer, Nature Physics 5, 262 (2009). 6. S. Datta, and B. Das, Appl. Phys. Lett. 56, 665 (1990). 7. T. Passow, S. Li, P. Feinäugle, T. Vallaitis, J. Leuthold, D. Litvinov, D. Gerthsen, and M. Hetterich, J. Appl. Phys. 102, 073511 (2007). 8. M. Hetterich, W. Löfﬂer, P. Asshoff, T. Passow, D. Litvinov, D. Gerthsen, and H. Kalt, “Electrical Spin Injection into Single InGaAs Quantum Dots,” in Advances in Solid State Physics, edited by R. Haug, Springer, 2009, vol. 48, p. 103.