Disorder-enhanced spin polarization of the Zn_1−xCo_xO_1−v concentrated magnetic semiconductor

Abstract

Amorphous concentrated magnetic semiconductor Zn_0.32Co_0.68O_1−v (v refers to oxygen vacancies) thin film was investigated by magnetic and electrical transport measurements as well as Andreev reflection spectroscopy. At a low temperature range, the electrons in the Zn_0.32Co_0.68O_1−v are strongly localized, and electrical transport obeys the Efros variable range hopping law. Spin polarization was measured by Andreev reflection spectroscopy. As high as 64 ± 5% of spin polarization was attained through fitting of the modified Blonder–Tinkham–Klapwijk (BTK) theory. This enhanced spin polarization of Zn_0.32Co_0.68O_1−v likely relates with the structure disorder and high concentration of magnetic cobalt ions, which lead to a spin imbalance impurity band in the tail of the conduction band. Considering room temperature ferromagnetism and high spin polarization, this material appears to be promising for spintronics device applications.

---

Introduction

Oxide magnetic semiconductors that combine both semiconducting and room-temperature ferromagnetic properties are currently one of most attractive research fields in spintronics.¹ For the application of spintronics materials, high spin polarization of the charge carrier is of fundamental importance.² Many techniques, such as superconducting tunnel junctions, spin-resolved photoemission and Andreev reflection (AR), have been used to study spin polarization of magnetic materials.^3–5 Those materials include magnetic metal, half-metal Heusler alloy, III–V group magnetic semiconductors, concentrated magnetic semiconductors (CMS) and semimetals.^5–9 For oxide magnetic semiconductors, many efforts have been made toward understanding the nature of ferromagnetism, spin-dependent electrical transport (magnetoresistance, anomalous Hall effect) and magneto-optical effect.^10,11 However, only a few works have reported on the spin polarization of oxide magnetic semiconductors, especially, the direct measurement of the spin polarization by Andreev reflection.^12,13 Yates et al. found that 55% spin polarization only exists in (Mn, Al) co-doped ZnO thin film. In contrast, Panguluri et al. reported 50% spin polarization in undoped In₂O₃ thin film at helium temperature. To clarify the role of transition metals and yield higher spin polarization in oxide magnetic semiconductors, further work still needs to be done.

Conventionally, it is acknowledged that structure disorder is detrimental to spin polarization due to the nanoscale coherence length of the spin current. Recently, disorder-related spin polarization enhancement has received increasing interest in the field of amorphous ferromagnetic metal and magnetic semiconductors. In metal-based spintronics, disorder-enhanced spin polarization was reported on CoFeB in magnetic tunnelling junctions and Andreev reflection spectroscopy.^14,15 For magnetic semiconductor investigation, Byoughak Lee et al. had theoretically predicted enhanced spin polarization in impurity bands of diluted Ga_1−xMn_xAs magnetic semiconductor due to the formation of Mn clusters, which results in strongly polarized bound states.¹⁶ Hu et al. reported long-range ferromagnetic order in a non-stoichiometric amorphous Co_0.5Zn_0.5O_1−v ternary transition metal oxide semiconductor in theory and experiment.¹⁷ In particular, cobalt-doped ZnO magnetic semiconductor (Zn_1−xCo_xO_1−v) is a notable material for practical application and fundamental physics investigation due to its room-temperature ferromagnetism, visible photoconductivity, large Zeeman splitting by sp–d exchange interaction and small spin–orbital interaction (which is important for high spin polarization).^11,18 Over the past decades, the most attention has been focused on the origin of the long-range ferromagnetism of crystalline-doped ZnO, and a definitive conclusion is forthcoming.¹¹ Although the emerging effect of disorder on spin polarization of magnetic materials is unquestionable, neither has there been an experimental measurement, nor has the impact of interface on the spin polarization of amorphous Zn_1−xCo_xO_1−v magnetic semiconductor been investigated.

In this paper, amorphous concentrated magnetic semiconductor Zn_0.32Co_0.68O_1−v thin film was studied by magnetic and electrical transport measurements as well as Andreev reflection spectroscopy. Through Andreev reflection spectroscopy, it was revealed that spin polarization of Zn_0.32Co_0.68O_1−v is as large as 64 ± 5%, which is larger than that reported for crystalline Zn_1−xCo_xO_1−v magnetic semiconductor and metal cobalt.^12,13 Such enhanced spin polarization can be attributed to the strong exchange interaction that is correlated to the structure disorder and high concentration of magnetic cobalt ions in amorphous Zn_0.32Co_0.68O_1−v.

Experimental methods

The concentrated magnetic semiconductor Zn_0.32Co_0.68O_1−v thin film was deposited by alternately sputtering 5 Å Co layers and 5 Å ZnO layers for 30 periods at 20 °C on a glass substrate in pure Ar atmosphere with 0.5 Pa sputtering pressure. The relatively low temperature growth and alternating deposition are a thermal non-equilibrium process, which guarantees the high solubility of cobalt in CMS and prevents the formation of metallic cobalt clusters. A similar preparation process was described for Ti_1−xCo_xO_2−v and Zn_1−xCo_xO_1−v CMS thin film deposition in ref. 19 and 20. Transmission electron microscopy (TEM), energy dispersive X-ray spectroscopy (EDS), and X-ray photoelectron spectroscopy (XPS) studies in these films indicate that the composition of Zn_1−xCo_xO_1−v and Ti_1−xCo_xO_2−v CMS thin film is uniform at nanometer scale but may be inhomogeneous at the sub-nanometer scale. On the other hand, as the thickness of ZnO is less than 1 nm, X-ray diffraction (XRD) and TEM results suggest that the microstructure of Zn_1−xCo_xO_1−v CMS thin film is in an amorphous state but contains small crystal grains with an average size of 4–6 nm.^19–21 However, no metallic Co clusters, CoO insulator clusters or ZnO tunnelling boundary were detected within the entire film, which is different from tunnelling granular film produced using a similar deposition process.²² The magnetic and electrical measurements of Zn_0.32Co_0.68O_1−v films were carried out by alternating gradient force magnetometer (AGM) and van der Pauw method, respectively.

In order to probe the spin polarization information of Zn_0.32Co_0.68O_1−v CMS and exclude the existence of metallic Co or CoO clusters, planar junctions of Zn_0.32Co_0.68O_1−v (30 nm)/Pb (500 nm) were fabricated on glass substrates defined by shadow mask, and the spin polarization was obtained through Andreev reflection spectroscopy. At first, a 30 nm thick Ag stripe was sputtered as the bottom electrode. Then, Zn_0.32Co_0.68O_1−v thin film with 30 nm thickness was deposited. Finally, the sample was transferred into another ultra-high vacuum chamber without air exposure, and a crossed Pb stripe with 500 nm thickness was grown by thermal evaporation. The growth rate of Pb film was higher than 3 nm s⁻¹, and chamber pressure was kept below 1 × 10⁻⁴ Pa during Pb film deposition. The effective junction area is 0.1 × 0.1 mm², and junction resistance varied from several ohms to tens of ohms. Two points deserve to be mentioned regarding the junction fabrication process. The first is that for Zn_0.32Co_0.68O_1−v CMS thin film growth, the last layer is not ZnO, but 5 Å Co, in order to get good ohmic contact with superconducting Pb. The second is that the resistivity of Zn_0.32Co_0.68O_1−v film is of the order of mOhm cm; it serves as conducting electrode rather than insulating barrier. The differential conductance spectra G(V) = dI(V)/dV were measured by standard lock-in technology at liquid helium temperature, and the amplitude of the ac modulation was kept sufficiently small in order to avoid heating and other spurious effects. All differential conductance spectra were normalized by the conductance above gap voltage. We also fabricated thicker Zn_0.32Co_0.68O_1−v (60 nm)/Pb (500 nm) junctions, and consistent measurement results were obtained, suggesting that spin polarization is not an interface effect and is independent of thickness.

Results and discussion

Fig. 1(a) shows the magnetization hysteresis loop of Zn_0.32Co_0.68O_1−v thin film measured at 300 K. The sample shows room temperature ferromagnetism where magnetization saturates at H = 3000 Oe and H_c is about 100 Oe (inset of Fig. 1(a)). Fig. 1(b) shows negative temperature dependence of resistivity at magnetic field 0 Oe and 6 × 10⁵ Oe, respectively. The increasing R with decreasing temperature suggests the semiconductor conducting property, which has been proved by theory and experiment.¹⁷ ln R–T^−1/2 fitting of resistance–temperature (inset of Fig. 1(b)) suggests that the electrical transport obeys Efros variable range hopping (VRH) in the low-temperature range.²² The conducting electrons are strongly localized, and transport is dominated by electron hopping from the initial localized occupied state i to the final vacant state j due to thermal activation. It should be noted that negative magnetoresistance of Zn_0.32Co_0.68O_1−v indicates the spin-polarized carriers due to strong coupling between them and the localized magnetic moments.²⁰

Fig. 1 (a) The magnetization hysteresis loop of Zn_0.32Co_0.68O_1−v; the inset shows magnification of the low field loop. (b) The temperature dependence of Zn_0.32Co_0.68O_1−v resistivity; the inset is ln R–T^−1/2 fitted.

Moreover, no anisotropic magnetization or magnetoresistance was observed in Zn_0.32Co_0.68O_1−v thin film for magnetic field applied parallel or perpendicular to the plane. The isotropy in magnetization and magnetoresistance are consistent with amorphous microstructure, while anisotropy had been reported in Co nanoparticle-embedded ZnO or granular thin films.²³

For spintronics device application and fundamental physics investigation, it is very important to get the spin polarization information directly, especially for carriers that are localized at low temperature. Fig. 2(a) shows a representative normalized differential conductance spectroscopy and fitting of the modified Blonder–Tinkham–Klapwijk (MBTK) theory of the Zn_0.32Co_0.68O_1−v/Pb junction measured at temperature T = 2 K.^24,25 The obvious suppression of subgap conductance suggests a moderated spin polarization in Zn_0.32Co_0.68O_1−v thin film. Through best fitting to the spin-polarized BTK theory, P = 64% with a certainty better than 5% was obtained. This 64% spin polarization of Zn_0.32Co_0.68O_1−v thin film definitively excludes the possibility of cobalt metal clusters at the interface of Zn_0.32Co_0.68O_1−v/Pb because it is larger than that of metal Co (35%).^5,12 More important, this result reveals that amorphous Zn_0.32Co_0.68O_1−v with a high concentration of cobalt ions does enhance the spin polarization, which may be related to the enhanced strength of exchange interaction between cobalt ions.

Fig. 2 (a) Normalized differential conductance spectroscopy of Zn_0.32Co_0.68O_1−v/Pb junction measured at T = 2 K. The junction resistance is 16 Ω. The red line is the fitting of the modified Blonder–Tinkham–Klapwijk (MBTK) theory. (b) Normalized differential conductance spectra (symbols) and MBTK fitting (solid lines) between T = 2 K and 7.2 K. The inset shows temperature dependence of spin polarization with an average P = 64%.

In order to get the best fitting to the BTK theory, four physical parameters, including spin polarization P, interfacial scattering barrier strength Z, superconducting gap Δ and inelastic broadening parameter Γ, are used.^24,26,27 We emphasize that the fitting is always performed in a straightforward manner with actual measurement temperature. For MBTK fitting of differential conductance spectra at T = 2 K, Z = 0.05, Δ ∼ 0.93 meV and Γ = 0.80 meV are obtained. Consistent with low junction resistance, the small interfacial scattering barrier strength Z can be qualitatively attributed to the enhanced junction transparence due to a moderated spin polarization.^7,8 The fitting yields the superconducting gap Δ ∼ 0.93 meV, which is smaller than the bulk BCS gap of Pb ∼ 1.3 meV, possibly attributed to the suppression of superconductivity due to interface intermixing or the existence of magnetic impurity.^25,28 Here, the conductance peaks at ±Δ are completely absent, since the AR process does not require available quasiparticle states in the superconductor.⁷ Based on Zutic and Das Sarma's theoretical results, the spin polarization actually enhances junction transparency of the ferromagnet/superconductor junction.²⁹ Specially, the conductance peaks at ±Δ can be completely suppressed by a moderated spin polarization at the ferromagnet/superconductor interface.²⁹ Similar conductance characteristics have been confirmed in the Andreev reflection spectra of CrO₂ half-metal.^6,30 In order to better understand the nature of spectroscopic broadening in our junction, the inelastic scattering parameter Γ is introduced into the energy term (E + iΓ) for fitting.²⁷ This inelastic term Γ is equivalent to the effective temperature () that is often used in point contact Andreev reflection spectra fitting due to heating effect.^25,31 A large inelastic scattering term Γ = 0.80 meV is obtained, which is nearly comparable to the magnitude of the energy gap. Considering the disordered interface of the Zn_0.32Co_0.68O_1−v/Pb superconducting junction, cobalt oxide and a large amount of magnetic cobalt ions would result in high inelastic scattering strength at the interface, which is also the reason for the suppressed superconductivity. The high inelastic scattering leads to broadening and flattening conductance spectroscopy peaks at ±Δ. Furthermore, high inelastic scattering strength can cause spin-mixing effect (spin flips) and dilute the intrinsic spin polarization. Therefore, 64% spin polarization is underestimated, and further cleaning of the interface is necessary for intrinsic spin polarization measurement and device application.

We also studied the effect of temperature on the conductance spectra as shown in Fig. 2(b). As the superconducting transition temperature of Pb is approached, the subgap conductance increases gradually. At the critical temperature of Pb, T_c = 7.2 K, the conductance is a constant that is consistent with good ohmic contact of the Zn_0.32Co_0.68O_1−v/Pb junction. As temperature increases, the spin polarization remains constant at around 64 ± 5% (inset of Fig. 2(b)). The corresponding fitting parameters Z, Δ and Γ are summarized in Fig. 3(a)–(c). The interfacial scattering barrier strength Z is small, while inelastic scattering Γ increases as temperature increases. The superconducting gap value decreases as temperature increases and can be well fitted by the BCS gap law, which is consistent with independent measurement of the superconducting Pb thin film.

Fig. 3 The temperature dependence of parameters extracted from fitting of the data in Fig. 2(b). (a) Interfacial scattering barrier strength Z, (b) inelastic broadening parameter Γ, (c) superconducting gap of Pb. The red line in (c) shows the best fit to the BCS gap law.

The Andreev reflection spectra show quite clearly that within the temperature range measured, the spin polarization of Zn_0.32Co_0.68O_1−v can be enhanced by amorphous structure and the high concentration of Co ions. The origin of spin polarization of Zn_0.32Co_0.68O_1−v, especially in variable range hopping regime, is a matter of considerable importance in physics. Recent theoretical research in amorphous transition metal oxides found that high spin polarization and magnetization is probably associated with the accommodation of concentrated transition metal atoms.¹⁷ Compared to crystalline structure, there are much more oxygen deficiencies in the amorphous thin film Zn_0.32Co_0.68O_1−v. These oxygen deficiencies offer charge carriers, which mediate the ferromagnetic exchange interaction of cobalt magnetic moments. On the other hand, the high concentration of magnetic cobalt ions reduces the distance and thus increases the strength of exchange interaction of local cobalt magnetic moments. Oxygen deficiency and the high concentration of cobalt ions both result in long-range ferromagnetism in the cobalt-rich region. Furthermore, the enhanced coupling of electrons and localized magnetic moment leads to a spin imbalance density of states in the tail of the conduction band.¹⁷ This spin-polarized band structure also is indicated by the negative magnetoresistance in our Zn_0.32Co_0.68O_1−v thin film. However, to clarify the influence of structure disorder on spin polarization of Zn_0.32Co_0.68O_1−v, further element-specific electronic structure study, such as X-ray absorption fine structure (XAFS), X-ray magnetic dichroism (XMCD) et al., is needed.

Conclusions

In conclusion, spin polarization of the concentrated magnetic semiconductor Zn_0.32Co_0.68O_1−v is probed by Andreev reflection spectroscopy. The spin polarization is 64 ± 5% at low temperature, in which charge carriers are strongly localized, as reflected by Efros variable range hopping transport. We believe the spin polarization of Zn_0.32Co_0.68O_1−v is underestimated due to strong inelastic scattering strength at the interface of Zn_0.32Co_0.68O_1−v/Pb. To elucidate the nature of high spin polarization of localized carriers in concentrated magnetic oxide semiconductors, the correlation of microstructure and electronic structure may shed some light on spintronics material and device applications.

Acknowledgements

We acknowledge the financial support from He'nan Educational Committee Grant No. 15B140001, NSF Nos. 11547176 and U1504517, the National Basic Research Program of China No. 2013CB922303 and 111 project B13029.

References

1. T. Dietl, Science, 2000, 287(5455), 1019.
2. G. M. Muller, J. Walowski, M. Djordjevic, G.-X. Miao, A. Gupta, A. V. Ramos, K. Gehrke, V. Moshnyaga, K. Samwer, J. Schmalhorst, A. Thomas, A. Hutten, G. Reiss, J. S. Moodera and M. Munzenberg, Nat. Mater., 2009, 8(1), 56.
3. Polarized electrons in surface physics, ed. R. Feder, World Scientific, Singapore, 1985.
4. R. Meservey and P. M. Tedrow, Phys. Rep., 1994, 1973, 238.
5. R. J. Soulen, J. M. Byers, M. S. Osofsky, B. Nadgorny, T. Ambrose, S. F. Cheng, P. R. Broussard, C. T. Tanaka, J. Nowak, J. S. Moodera, A. Barry and J. M. D. Coey, Science, 1998, 282(5386), 85.
6. Y. Ji, G. J. Strijkers, F. Y. Yang, C. L. Chien, J. M. Byers, A. Anguelouch, G. Xiao and A. Gupta, Phys. Rev. Lett., 2001, 86(24), 5585.
7. J. G. Braden, J. S. Parker, P. Xiong, S. H. Chun and N. Samarth, Phys. Rev. Lett., 2003, 91, 056602.
8. C. Ren, J. Trbovic, R. L. Kallaher, J. G. Braden, J. S. Parker, S. von Molnár and P. Xiong, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 205208.
9. T. Guan, C. Lin, C. Yang, Y. Shi, C. Ren, Y. Li, H. Weng, X. Dai, Z. Fang, S. Yan and P. Xiong, Phys. Rev. Lett., 2015, 115, 087002.
10. M. Bibes and A. Barthelemy, IEEE Trans. Electron Devices, 2007, 54, 1003.
11. D. Zhu, G. Liu, Y. Tian, S. Yan, Q. Cao, L. Cai, M. Fu, J. Zhang, K. Zhang, S. Ye, H. Li, S. Hu, Y. Chen, S. Kang, Y. Dai and L. Mei, Appl. Phys. Lett., 2014, 105(7), 072404.
12. K. A. Yates, A. J. Behan, J. R. Neal, D. S. Score, H. J. Blythe, G. A. Gehring, S. M. Heald, W. R. Branford and L. F. Cohen, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80(24), 245207.
13. R. P. Panguluri, P. Kharel, C. Sudakar, R. Naik, R. Suryanarayanan, V. M. Naik, A. G. Petukhov, B. Nadgorny and G. Lawes, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79(16), 165208.
14. P. V. Paluskar, J. J. Attema, G. A. de Wijs, S. Fiddy, E. Snoeck, J. T. Kohlhepp, H. J. M. Swagten, R. A. de Groot and B. Koopmans, Phys. Rev. Lett., 2008, 100(5), 057205.
15. S. X. Huang, T. Y. Chen and C. L. Chien, Appl. Phys. Lett., 2008, 92(24), 242509.
16. B. Lee, X. Cartoixà, N. Trivedi and R. M. Martin, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76(15), 155208.
17. S.-J. Hu, S.-S. Yan, Y.-P. Zhang, M.-W. Zhao, S.-S. Kang and L.-M. Mei, J. Appl. Phys., 2014, 116(4), 043711.
18. Z. Yang, Appl. Phys. A: Mater. Sci. Process., 2013, 112(2), 241.
19. S.-S. Yan, C. Ren, X. Wang, Y. Xin, Z. X. Zhou, L. M. Mei, M. J. Ren, Y. X. Chen, Y. H. Liu and H. Garmestani, Appl. Phys. Lett., 2004, 84(13), 2376; H.-Q. Song, L.-M. Mei, S.-S. Yan, X.-L. Ma, J.-P. Liu, Y. Wang and Z. Zhang, J. Appl. Phys., 2006, 99(12), 123903.
20. S. S. Yan, J. P. Liu, L. M. Mei, Y. F. Tian, H. Q. Song, Y. X. Chen and G. L. Liu, J. Phys.: Condens. Matter, 2006, 18(46), 10469; C. Yan-Xue, Y. Shi-Shen, L. Guo-Lei, M. Liang-Mo and R. Miao-Juan, Chin. Phys. Lett., 2007, 24(1), 217.
21. Q. Li, T. T. Shen, Z. K. Dai, Y. L. Cao, S. S. Yan, S. S. Kang, Y. Y. Dai, Y. X. Chen, G. L. Liu and L. M. Mei, Appl. Phys. Lett., 2012, 101(17), 172405.
22. Z.-Y. Quan, L. Zhang, W. Liu, H. Zeng and X.-H. Xu, Nanoscale Res. Lett., 2014, 9, 6; Z.-Y. Quan, X.-H. Xu, X.-L. Li, Q. Feng and G. A. Gehring, J. Appl. Phys., 2010, 108(10), 103912.
23. D. Y. Li, Y. J. Zeng, L. M. C. Pereira, D. Batuk, J. Hadermann, Y. Z. Zhang, Z. Z. Ye, K. Temst, A. Vantomme, M. J. Van Bael and C. Van Haesendonck, J. Appl. Phys., 2013, 114, 033909; S. Zhou, K. Potzger, J. von Borany, R. Grötzschel, W. Skorupa, M. Helm and J. Fassbender, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77(3), 035209.
24. I. I. Mazin, A. A. Golubov and B. Nadgorny, J. Appl. Phys., 2001, 89(11), 7576.
25. Y. Bugoslavsky, Y. Miyoshi, S. K. Clowes, W. R. Branford, M. Lake, I. Brown, A. D. Caplin and L. F. Cohen, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71(10), 104523.
26. G. E. Blonder, M. Tinkham and T. M. Klapwijk, Phys. Rev. B: Condens. Matter Mater. Phys., 1982, 25(7), 4515.
27. R. C. Dynes, V. Narayanamurti and J. P. Garno, Phys. Rev. Lett., 1978, 41(21), 1509.
28. P. Chalsani, S. K. Upadhyay, O. Ozatay and R. A. Buhrman, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 094417.
29. I. Žutić and S. Das Sarma, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 60(24), R16322.
30. J. S. Parker, S. M. Watts, P. G. Ivanov and P. Xiong, Phys. Rev. Lett., 2002, 88, 196601.
31. R. P. Panguluri, B. Nadgorny, T. Wojtowicz, X. Liu and J. K. Furdyna, Appl. Phys. Lett., 2007, 91, 252502.

---