Electrical spin injection from a ferromagnetic CoFe/MgO(100) tunnel contact into GaAs was demonstrated at room temperature by reading out the spin polarization of injected electrons through the circular polarization of electroluminescence (light emitted by recombination of the injected electrons with holes) from GaAs/AlGaAs quantum-well (QW) detectors (Jiang et al., Phys. Rev. Lett. 94, 056601 (2005)). The reported injection efficiencies (at least 52% at 100 K and 32% at 290 K after subtraction of an approximately linear high-field background polarization) are lower bounds, because spin relaxation and recombination inside the QW detector reduce the measured electroluminescence polarization, and the microscopic origin of the linear, field-dependent background above about 2 T is not definitively established. The problem is to convert such electroluminescence measurements into a quantitative determination of the true tunneling spin polarization at the metal-semiconductor interface: to identify and quantify the microscopic mechanisms (Zeeman thermalization, field-dependent D'yakonov-Perel' spin relaxation, and field-dependent recombination time) behind the background polarization, to measure the injected spin polarization independently of QW-detector relaxation, and to characterize the detector lifetime physics well enough to predict and control the nonmonotonic temperature behavior and to optimize QW confinement for reliable room-temperature measurement. Here D'yakonov-Perel' relaxation is the spin-orbit-mediated spin-dephasing mechanism that dominates for conduction electrons in GaAs. Later work, notably time-resolved measurements by the same group, resolved the temperature dependence of the detector lifetimes and yielded an injection-efficiency estimate of about 70%, but a quantitative decomposition of the background and a detector-independent measurement of the interface polarization in GaAs-based devices remain open.

Background

Semiconductor spintronics aims to use the spin of carriers in conventional semiconductors for logic, memory, and optoelectronics, which requires injecting highly spin-polarized currents from ferromagnetic metals at room temperature. Two obstacles dominated this field around 2004. First, the conductivity mismatch problem: the large conductivity difference between a ferromagnetic metal and a semiconductor suppresses spin injection at a direct ohmic contact, unless a resistive tunnel barrier is inserted at the interface. Second, readout: injected electron spin polarization in a semiconductor cannot be measured directly at the contact, so a spin light-emitting diode (spin-LED) is used, in which the spin-polarized electrons recombine with holes in a quantum well (QW, a thin semiconductor layer that confines carriers) and emit circularly polarized light. By the optical selection rules, the degree of circular polarization of the heavy-hole exciton emission, defined as PEL=(I+−I−)/(I++I−)P_{EL}=(I^{+}-I^{-})/(I^{+}+I^{-}) with I±I^{\pm} the intensities of right- and left-circular components, equals the spin polarization of the electrons at the moment of recombination (up to a factor of order one depending on heavy-hole/light-hole mixing, i.e., mixing of the two valence-band hole states of different angular momentum). Crystalline MgO(100) tunnel barriers with CoFe electrodes were predicted to act as symmetry-based spin filters that transmit essentially one spin channel, motivating their use as spin injectors. Jiang et al. (Phys. Rev. Lett. 94, 056601 (2005)) demonstrated a sputtered Co70Fe30/MgO(about 3 nm)/GaAs spin-LED with (100)-textured MgO, measured PEL=57%P_{EL}=57\% at 100 K and 47%47\% at 290 K in a 5 T perpendicular field, and, after subtracting an approximately linear high-field background polarization that appears above the about 2 T needed to rotate the CoFe magnetization out of plane, quoted lower bounds on the tunneling spin polarization (the spin polarization of the tunnel current at the metal-semiconductor interface, TSP) of at least 52% at 100 K and 32% at 290 K. Three features of the experiment were left unresolved. (1) The high-field background: its slope varies with temperature (negative at low temperature, crossing zero near 40-50 K, positive at higher temperature), and it was plausibly attributed to a combination of Zeeman thermalization (thermal equilibration of spin populations split by the Zeeman energy in the QW, relevant at low temperature because GaAs has a negative electron g factor), suppression of D'yakonov-Perel' (DP) spin relaxation by the perpendicular field (DP relaxation, the dominant spin-dephasing mechanism for conduction electrons in GaAs QWs, has a relaxation rate scaling as τs−1∝τpT\tau_{s}^{-1}\propto\tau_{p}T, where τp\tau_{p} is the electron momentum relaxation time and TT the temperature), and a field-dependent electron-hole recombination time; the relative weights of these contributions were not established. (2) Detector relaxation: because spin relaxation and recombination in the QW can only reduce the measured polarization, the reported values are lower bounds, and no measurement probed the injected spin before QW relaxation. (3) Detector physics and design: the nonmonotonic temperature dependence of PELP_{EL} was attributed to DP relaxation in the QW, but the temperature and confinement dependences of τp\tau_{p} and of the recombination time in the actual device geometries were not characterized, and shallow QW detectors had poor electroluminescence signal-to-noise at high temperature.

Problem Statement

Determine, quantitatively and separately, (i) the tunneling spin polarization of electrons injected from a CoFe/MgO(100) tunnel contact into GaAs at the metal-semiconductor interface, and (ii) the detector-side processes in GaAs/AlGaAs quantum-well (QW) spin light-emitting diodes (spin-LEDs, devices in which the polarization of injected spins is read out through the circular polarization of the emitted light) that mask it in electroluminescence-polarization measurements. Concretely, for the CoFe/MgO(100)/GaAs(AlGaAs) QW spin-LED system of Jiang et al., Phys. Rev. Lett. 94, 056601 (2005): (1) identify the microscopic origin and quantitative contributions of the approximately linear, field-dependent background polarization observed for perpendicular fields above the roughly 2 T needed to saturate the out-of-plane CoFe magnetization, establishing the relative weights and temperature dependences of the proposed mechanisms: Zeeman thermalization (thermal equilibration of the QW spin population across the Zeeman-split levels), suppression of D'yakonov-Perel' spin relaxation (the spin-orbit-mediated dephasing mechanism dominant for conduction electrons in GaAs QWs) by the perpendicular field, and a field-dependent electron-hole recombination time; (2) measure or rigorously bound the injected spin polarization independently of quantum-well relaxation and recombination, so that the interface tunneling spin polarization is known beyond the reported lower bounds (at least 52% at 100 K and 32% at 290 K, and beyond the time-resolved estimate of about 70%); (3) characterize the temperature and quantum-well confinement dependence of the detector momentum relaxation time and recombination time in these device geometries, so that the nonmonotonic temperature behavior of the electroluminescence polarization is predictable rather than only qualitatively attributed to D'yakonov-Perel' relaxation; and (4) establish how the quantum-well detector design (confinement depth, growth conditions) should be optimized so that reliable polarization measurements, and hence inferred injector efficiency, are obtained at room temperature. These four deliverables serve one objective: making electroluminescence-polarization readout a quantitative, artifact-corrected metrology of spin injection for ferromagnet/MgO/semiconductor tunnel contacts. They are jointly posed as one objective, not independent sub-problems. The analysis should be self-consistent and quantitative, and consistent with the field-, temperature-, and confinement-resolved data in the cited literature; newly performed in-device measurements are not required. In deliverable (2), "known beyond ... about 70%" means a detector-independent determination that supersedes the roughly 70% time-resolved estimate, not merely a reported polarization larger than 70%.

References

  1. X. Jiang, R. Wang, R. M. Shelby, R. M. Macfarlane, S. R. Bank, J. S. Harris, and S. S. P. Parkin, Highly Spin-Polarized Room-Temperature Tunnel Injector for Semiconductor Spintronics using MgO(100), Phys. Rev. Lett. 94, 056601 (2005). DOI: 10.1103/PhysRevLett.94.056601, https://doi.org/10.1103/PhysRevLett.94.056601
  2. G. Salis, R. Wang, X. Jiang, R. M. Shelby, S. S. P. Parkin, S. R. Bank, and J. S. Harris, Temperature independence of the spin-injection efficiency of a MgO-based tunnel spin injector, Appl. Phys. Lett. 87, 262503 (2005). DOI: 10.1063/1.2149369, https://doi.org/10.1063/1.2149369
  3. R. Wang, X. Jiang, R. M. Shelby, R. M. Macfarlane, S. S. P. Parkin, S. R. Bank, and J. S. Harris, Increase in spin injection efficiency of a CoFe/MgO(100) tunnel spin injector with thermal annealing, Appl. Phys. Lett. 86, 052901 (2005). DOI: 10.1063/1.1787896, https://doi.org/10.1063/1.1787896
  4. X. Jiang, R. Wang, R. M. Shelby, and S. S. P. Parkin, Highly efficient room-temperature tunnel spin injector using CoFe/MgO(001), IBM J. Res. Dev. 50, 111-120 (2006). DOI: 10.1147/rd.501.0111, https://doi.org/10.1147/rd.501.0111
  5. E. Oh, T. K. Lee, J. H. Park, J. H. Choi, Y. J. Park, K. H. Shin, and K. Y. Kim, Carrier lifetime and spin relaxation time study for electrical spin injection into GaAs, J. Appl. Phys. 106, 043515 (2009). DOI: 10.1063/1.3186026, https://doi.org/10.1063/1.3186026
  6. O. Brandt, M. Ramsteiner, T. Flissikowski, J. Herfort, and H. T. Grahn, Rate-equation model of spin dynamics and polarized light emission for spin light-emitting diodes, Phys. Rev. B 81, 115302 (2010). DOI: 10.1103/PhysRevB.81.115302, https://doi.org/10.1103/PhysRevB.81.115302
  7. H. Höpfner, C. Fritsche, A. Ludwig, A. Ludwig, F. Stromberg, H. Wende, W. Keune, D. Reuter, A. D. Wieck, N. C. Gerhardt, and M. R. Hofmann, Magnetic field dependence of the spin relaxation length in spin light-emitting diodes, Appl. Phys. Lett. 101, 112402 (2012). DOI: 10.1063/1.4752162, https://doi.org/10.1063/1.4752162
  8. H. Höpfner, C. Fritsche, A. Ludwig, A. Ludwig, F. Stromberg, H. Wende, W. Keune, D. Reuter, A. D. Wieck, N. C. Gerhardt, and M. R. Hofmann, Spin injection, transport, and relaxation in spin light-emitting diodes: magnetic field effects, Proc. SPIE 8813, 881318 (2013). DOI: 10.1117/12.2023324, https://doi.org/10.1117/12.2023324
  9. C. H. Li, G. Kioseoglou, A. Petrou, M. Korkusinski, P. Hawrylak, and B. T. Jonker, Highly polarized emission from electrical spin injection into an InGaAs quantum well with free carriers, Appl. Phys. Lett. 103, 212403 (2013). DOI: 10.1063/1.4832460, https://doi.org/10.1063/1.4832460
  10. Y. Lu, V. G. Truong, P. Renucci, M. Tran, H. Jaffrès, C. Deranlot, J.-M. George, A. Lemaître, Y. Zheng, D. Demaille, P.-H. Binh, T. Amand, and X. Marie, MgO thickness dependence of spin injection efficiency in spin-light emitting diodes, Appl. Phys. Lett. 93, 152102 (2008). DOI: 10.1063/1.2999631, https://doi.org/10.1063/1.2999631
  11. P. Barate, S. Liang, T. T. Zhang, J. Frougier, M. Vidal, P. Renucci, X. Devaux, B. Xu, H. Jaffrès, J. M. George, X. Marie, M. Hehn, S. Mangin, Y. Zheng, T. Amand, B. Tao, X. F. Han, Z. Wang, and Y. Lu, Electrical spin injection into InGaAs/GaAs quantum wells: A comparison between MgO tunnel barriers grown by sputtering and molecular beam epitaxy methods, Appl. Phys. Lett. 105, 012404 (2014). DOI: 10.1063/1.4887347, https://doi.org/10.1063/1.4887347
  12. P. Barate, S. H. Liang, T. T. Zhang, J. Frougier, B. Xu, P. Schieffer, M. Vidal, H. Jaffrès, B. Lépine, S. Tricot, F. Cadiz, T. Garandel, J. M. George, T. Amand, X. Devaux, M. Hehn, S. Mangin, B. Tao, X. F. Han, Z. G. Wang, X. Marie, Y. Lu, and P. Renucci, Bias Dependence of the Electrical Spin Injection into GaAs from Co-Fe-B/MgO Injectors with Different MgO Growth Processes, Phys. Rev. Applied 8, 054027 (2017). DOI: 10.1103/PhysRevApplied.8.054027, https://doi.org/10.1103/PhysRevApplied.8.054027
  13. S. Iba, H. Saito, K. Watanabe, Y. Ohno, and S. Yuasa, Growth condition dependence of photoluminescence polarization in (100) GaAs/AlGaAs quantum wells at room temperature, J. Appl. Phys. 118, 083901 (2015). DOI: 10.1063/1.4928325, https://doi.org/10.1063/1.4928325
  14. A. Spiesser, H. Saito, S. Yuasa, and R. Jansen, Tunnel spin polarization of Fe/MgO/Si contacts reaching 90% with increasing MgO thickness, Phys. Rev. B 99, 224427 (2019). DOI: 10.1103/PhysRevB.99.224427, https://doi.org/10.1103/PhysRevB.99.224427
  15. Y. Huang, V. Polojärvi, S. Hiura, P. Höjer, A. Aho, R. Isoaho, T. Hakkarainen, M. Guina, S. Sato, J. Takayama, A. Murayama, I. A. Buyanova, and W. M. Chen, Room-temperature electron spin polarization exceeding 90% in an opto-spintronic semiconductor nanostructure via remote spin filtering, Nat. Photonics 15, 475-482 (2021). DOI: 10.1038/s41566-021-00786-y, https://doi.org/10.1038/s41566-021-00786-y
  16. K.-T. Chang, W. Liang, Y. Lu, and H. Lu, Principles and advances in spin light-emitting diodes, npj Spintronics 3, 44 (2025). DOI: 10.1038/s44306-025-00107-x, https://doi.org/10.1038/s44306-025-00107-x