Search

International audience; Cooling of nuclear spins in doped semiconductors via their dynamic polarization by optical pumping is a powerful method for harnessing ubiquitous fluctuations of nuclear spin, which constitute a well-known decoherence channel for the electronic spins. The idea of spin cooling is based on the hypothesis of spin temperature, which states that nuclear spin system (NSS) reaches an internal thermal equilibrium long before it comes to equilibrium with the external bath (crystal lattice). Although thermodynamic framework has been successfully employed for the description of a variety of the experimental data, a rigorous check of this concept in semiconductors was impossible until recently, in particular at low magnetic field. The reason for that is the lack of experimental techniques allowing nonperturbative optical control over adiabatic transformation of the NSS.We have recently developed such methods, based on off-resonant Faraday rotation and spin noise spectroscopy [1, 2]. Using these techniques, combined with dark-interval photoluminescence spectroscopy, we established a comprehensive picture of the nuclear spin relaxation efficiency, its magnetic field, temperature, and carrier concentration dependence in doped GaAs, a model system in the field of nuclear spin physics in semiconductors [3-5]. We also analyzed the interplay between four relevant relaxation mechanisms: hyperfine interaction, quadrupole interaction, spin diffusion and Korringa mechanisms.Understanding of field dependence of NSS dynamics allowed us to obtain a new insight into the NSS thermodynamics, and verify the spin temperature concept in GaAs bulk material and microcavities [6]. We have demonstrated, that NSS exactly follows the predictions of the spin temperature theory, despite the quadrupole interaction that was earlier reported to disrupt nuclear spin thermalization in quantum dots [7]. Our results open a way for the deep cooling of nuclear spins in semiconductor structures, with the prospect of realizing nuclear spin-ordered states for high-fidelity spin-photon interfaces.This work is supported by a joint grant of the Russian Foundation for Basic Research (RFBR, Grant No. 16-52-041301-4 and National Center for Scientific Research (CNRS, PRC SPINCOOL No. 148362)References:[1] R. Giri et al, Physical Review Letters, 111, 087603 (2013)[2] I. I. Ryzhov et al, Applied Physics Letters, 106, 242405 (2015)[3] M. Kotur et al, Physical Review B, 94(8), 081201 (2016)[4] M. Vladimirova et al, Physical Review B, 95, 125312 (2017)[5] M. Kotur et al, eprint arXiv:1802.05013 (2018)[6] M. Vladimirova et al, Physical Review B, 97, 041301 (2018)[7] P. Maletinsky et al, Nature Physics, 5, 407 (2009)

International audience; spin fluctuations in atomic vapors and in semiconductors. As such it is quite attractive for probing non-perturbatively electrons or nuclear spins in atomic vapours or in semiconductors, particularly when the laser is detuned from optical resonances, an aspect which has triggered considerable interest to spin noise spectroscopy in recent years. The signal in spin noise spectroscopy can be considered as arising from intensity fluctuations due to the interference between the light scattered by Raman spin-flip excitations and the probe laser. This suggests that homodyne or heterodyne detection of spin noise by mixing of the Raman signal with a local oscillator is feasible, which we have demonstrated recently [1]. This opportunity is essential to spin noise spectroscopy because it allows us to increase its sensitivity and accessible frequency range.Homodyne and/or heterodyne detection of spin noise for scattered light directions different from that of the probe should give access to spatio-temporal spin correlations [2], and therefore could considerably extend the application range of spin noise spectroscopy. Here we demonstrate non-collinear homodyne detection of spin noise in n-doped CdTe epilayers. This allows us to exploit different scattering geometries based on spin-flip Raman selection rules. We are able to measure simultaneously different fluctuating spin components by homodyne mixing in different space directions, and also for different local oscillator polarizations. [1] S. Cronenberger and D. Scalbert, Review of Scientific Instruments 87, 093111 (2016).[2] G. G. Kozlov, I. I. Ryzhov, and V. S. Zapasskii, Phys. Rev. A 95, 043810 (2017).

In the recent years the spectroscopy of spin noise has won its spurs for spin dynamics related studies in semiconductors, largely because it exhibits several quite attractive features. Noticeably it enables almost perturbation-free detection of spontaneous electron spin fluctuations. In addition as the amplitude of spin noise grows when the size of the probed region is reduced, it is well adapted for spatially resolved studies. Also the spin noise spectrum is quite sensitive to internal effective fields, which allows to probe locally the existence of nuclear fields. Finally, the combination of spin noise spectroscopy and optical heterodyning has been demonstrated, which permits enhanced sensitivity and broadband detection, while keeping a high spectral resolution [1].In this poster we will present results obtained by heterodyne detection of spin noise in an n-doped CdTe epilayer with donor density n∼3×〖10〗^17cm-3. Thanks to the enhanced sensitivity gained by heterodyne detection we could detect the spin noise of electrons bound to neutral donors for probe powers as low as 5 µW. In these conditions we observe an extremely slow hopping rate W0 of electron spin between neighbouring donors, and extremely slow electron spin relaxation rate s (see Figure 1). The observed noise spectrum at zero field is characteristic of electron spin precession in the frozen nuclear field acting on the electrons bound to the donors and exhibits two components. A central lorentzian line corresponding to the electron spin component along the nuclear field, and two satellites gaussian lines corresponding to the spin precession in the nuclear field. We analyze our results in the framework of a theoretical model, which takes into account both the electron spin precession, and the hopping between donors, but slightly modified to take into account an eventual non-zero nuclear spin polarization [2]. In agreement with the theory we can see that the satellites merge with the central line as the hopping rate increases (at the highest probe power). Surprisingly W0 and s become quite small at the lowest probe power. We find that the electron spin relaxation time becomes longer than 1 µs in this regime.

International audience; Cooling of nuclear spin system (NSS) in doped semiconductors via dynamic polarization by optical pumping is a powerful method for harnessing ubiquitous fluctuations of nuclear spin. The idea of NSS cooling is based on the hypothesis of spin temperature, which states that NSS reaches an internal thermal equilibrium long before it comes to equilibrium with the external bath (crystal lattice). Although thermodynamic framework has been successfully employed for the description of a variety of the experimental data, a rigorous check of this concept in semiconductors was impossible until recently, in particular at low magnetic field. The reason for that is the lack of experimental techniques allowing nonperturbative optical control over adiabatic transformation of the NSS. We have recently developed such methods, based on off-resonant Faraday rotation and spin noise spectroscopy [1,2]. Using these techniques, combined with photoluminescence spectroscopy, we established a comprehensive picture of the nuclear spin relaxation efficiency, its magnetic field, temperature, and carrier concentration dependence in both n- and p-doped GaAs, a model system in the field of nuclear spin physics in semiconductors [3-5]. We also analyzed the interplay between four relevant relaxation mechanisms: hyperfine interaction, quadrupole interaction, spin diffusion towards paramagnetic impurities, and Korringa mechanisms. Figure 1 illustrates these processes in a nuclei-electron coupled system. Understanding of field dependence of NSS dynamics allowed us to obtain a new insight into the NSS thermodynamics, and verify the spin temperature concept in GaAs bulk material and microcavities [6]. We have demonstrated that NSS exactly follows the predictions of the spin temperature theory, despite the quadrupole interaction that was earlier reported to disrupt nuclear spin thermalization in quantum dots [7]. These results open a way for the deep cooling of nuclear spins in semiconductor structures, with the prospect of realizing nuclear spin-ordered states for high-fidelity spin-photon interfaces. References[1] R. Giri et al, Physical Review Letters, 111, 087603 (2013)[2] I. I. Ryzhov et al, Applied Physics Letters, 106, 242405 (2015)[3] M. Kotur et al, Physical Review B, 94, 081201(R) (2016)[4] M. Vladimirova et al, Physical Review B, 95, 125312 (2017)[5] M. Kotur et al, Physical Review B, 97, 165206 (2018)[6] M. Vladimirova et al, Physical Review B, 97, 041301 (2018)[7] P. Maletinsky et al, Nature Physics, 5, 407 (2009)

Mn spin dynamics is measured in very diluted (Cd,Mn)Te crystals by time-resolved Kerr rotation. Spin beats due to the hyperfine interaction between the 3d electrons of the Mn ions and their own nuclei are detected. It is shown that the effect of the crystal field can be strongly suppressed for "magic" orientations of the magnetic field. This particular orientation of the field permits the optical read-out of the Mn nuclear spin state. Manganese ions trapped on a semiconductor lattice have uniform properties and relatively long spin lifetimes, which make them promising for optical manipulation. In particular, Mn2+ ions embedded in a II-VI semiconductor are S-state ions, weakly coupled to the lattice. For this reason quite long Mn electronic spin relaxation times are expected and have been observed [1]. Here we show that the spin coherence time is mainly limited by dipole-dipole interactions at a Mn concentration x=0.001, and reaches up to 15 ns. At this low concentration, fine and hyperfine structures of Mn2+ are resolved in electron spin resonance experiments. In the time-domain it corresponds to low-frequency beats as shown in Fig. 1.

International audience; Nuclear spin dynamics in semiconductors has been studied extensively by optical means for more than forty years. Yet generally the existing techniques require the injection of out of equilibrium carriers in the semiconductor in order to probe the nuclear polarization, which in turn disturbs the nuclei to be measured. This is the case for photoluminescence experiments, but also for time resolved Faraday/Kerr rotation, in which a shift in the electron Larmor frequency is induced by the nuclear field.1 In this work, we demonstrate a non-destructive detection of nuclear magnetization in n-GaAs via cavity enhanced Faraday rotation of an off-resonant light beam. We use a two-step protocol, as proposed by Artemova and Merkulov2, where preparation and measurement are separated in time. The nuclear spins are first cooled down by optical pumping, then the pump beam is switched off and the subsequent evolution of nuclear polarization is detected by the Faraday rotation of a below band gap probe beam. This experimental protocol applied to metallic GaAs just above the metal-insulator transition reveals a non- monotonous evolution of the nuclear induced Faraday rotation. We interpret this behavior as a consequence of the existence of two contributions in the Faraday rotation, which are related to optical transitions involving either localized or delocalized states. Therefore, this detection scheme turns out to be sensitive to nuclear spins situated within the localization radius of bound electron states, as well as to all other nuclear spins in the sample. The contributions to the Faraday rotation of different nuclear species with different localizations in the sample are identified in nuclear magnetic resonance experiments. Our results suggest that the nuclear relaxation in metallic n-GaAs is dominated by diffusion towards fast relaxing centers, rather than by the Korringa mechanism.

International audience; We report time-resolved Kerr rotation experiments on bulk CdMnTe with very low Mn concentration. We observe spin beatings which are the result of a complex spin dynamics of the d-electrons in the hyperfine field of their nucleus and the tetragonal crystal field. We find the transverse Mn spin relaxation time to increase with the increase of the effective temperature of Mn subsystem. At higher temperatures we measure T2 up to 12 ns, which is much longer than any previously measured values.

Faraday rotation is a well-known magneto-optical phenomenon: polarization plane of the light wave is rotated upon transmission through a magnetized medium. It is believed that most efficient Faraday rotators are transparent magnetic materials, such as rare-earth doped glasses and garnets, as well as diluted magnetic semiconductors, where magnetic ions create strong magnetization. In contrast, in conventional non-magnetic semiconductors the rotation due to magnetization of electron gas in the presence of equivalent magnetic field is much smaller. We show, that using optical orientation of electron gas in GaAs confined in a microcavity (Q=50000) it is possible to reach the rotation angles larger than 10° at only 1% spin polarization of the electron gas and in the absence of magnetic field ( Fig. 1 ). This strong rotation exceeds by orders of magnitude the rotation ever measured in bulk n-GaAs samples of similar concentration () [1] [2] . We deduce from these experiments the Verdet constant associated with the electron spin polarization density. This also allows for the direct detection of the extremely slow electron spin polarization decay (~250 ns) in the absence of the pump ( Fig. 2 ). It suggests that spin relaxation time in the dark can be much longer than the relaxation time extracted from Hanle depolarization in either photoluminescence [3] or Faraday rotation ( Fig. 1 ) experiments.