ISSN 1062-8738, Bulletin of the Russian Academy of Sciences: Physics, 2007, Vol. 71, No. 11, pp. 1610–1612. © Allerton Press, Inc., 2007. Original Russian Text © V.A. Ivanov, 2007, published in Izvestiya Rossiiskoi Akademii Nauk. Seriya Fizicheskaya, 2007, Vol. 71, No. 11, pp. 1651–1653.

Diluted Magnetic Semiconductors and Spintronics V. A. Ivanov Kurnakov Institute of General and Inorganic Chemistry, Russian Academy of Sciences, Leninskii pr. 31, Moscow, 119991 Russia e-mail: [email protected] Abstract—Classification of concentrated and diluted magnetic semiconductors is given and their physicochemical properties that are interesting for spintronics are characterized. The electronic structure of magnetic impurities in semiconductors and the nature of indirect exchange interactions between impurity spins in diluted magnetic semiconductors are considered. On the basis of the proposed theory of kinematic exchange, the Curie temperatures TC for bulk diluted magnetic semiconductors (In,Mn)Sb are estimated. DOI: 10.3103/S1062873807110433

Concerning the questions of solid-state physics in the list of urgent scientific problems, V.L. Ginzburg noted the “appearance of spintronics. The matter is that until recently electronics (specifically, semiconductor electronics) has dealt only with the motion of charges (electrons and holes), whereas the spin variables were ignored or did not play a key role. Currently, much attention is paid to spins of charge carriers, and their behavior is being studied” [1]. The metallic behavior of ferromagnetic alloys, which are widely used, for example, in computer engineering, does not make it possible to use them in semiconductor spintronics. Currently,

the key problem of spintronics is the absence of new ferromagnets possessing (at temperatures above room temperature) a combination of semiconducting, magnetic, and optical properties that are necessary for microelectronics. Compatibility of ferromagnetic and semiconducting properties was revealed for the first time (in 1960) in the magnetic semiconductor CrBr3with the Curie temperature 37 K. This compound became the starting point for the materials listed in the table: concentrated magnetic semiconductors, semimetals, semimagnetic materials, and diluted magnetic semiconductors (DMSs).

Table Diluted magnetic semiconductors III–V hosts IV4(s2p2)

↓

↓

V2(s2p3)

↓

↓ III2(s2p1)

↓

2000

II(s2) IV(s2p2) V2(s2p3) (Ga,Mn)As (TC ~ 170 K), (In,Mn)As (TC = 60 K), (Ga,Mn)P (TC = 400 K), (Ga,Mn)N (TC = 940 K) 1990

1980 1960

Semimagnetic semiconductors AIIBVI and AIVBVI hosts (AII = Zn, Cd, Hg; AIV = Pb, Sn; BVI = S, Se, Te) (CdT)Se, (CdT)Te, PbSnMnTe (T = Fe2+, Co2+, Mn2+), Cd1 – xMnxTe, (TC < 10 K); spin glasses, antiferromagnets, paramagnets Concentrated magnetic semiconductors and semimetals CrBr3 (TC = 37 K) (1960), CdCr2S4 (TC = 84.5 K), HgCr2Se4 (TC = 106 K), CdCr2Se4 (TC = 130 K), EuX (X = O, S, Se, Te), Fe3O4, CrO2 (TC = 400 K), BiMnO3 (TC = 105 K), SeCuO3 (TC = 26 K), YTiO3 (TC = 29 K), EuO (TC = 79 K) Semimetals: MnAs, MnSb, CrAs, CrSb; semimetals–double perovskites A2BB’O6: Sr2FeMoO6 (TC = 420 K), Sr2CrReO6 (TC = 635 K), Sr2CrReO6 (TC = 620 K), Sr2FeMoO6 (TC = 416 K). Heusler alloys X2YZ (X, Y are transition metals; Z are III–V group elements) 1610

DILUTED MAGNETIC SEMICONDUCTORS AND SPINTRONICS

In a DMS, the cation sublattice is randomly substituted by d- or f-metal impurities with conservation of the crystal structure of the semiconductor host. DMSs are isoelectronic and isostructural with the main microelectronic materials, Si and GaAs; hence, they can be incorporated in semiconductor schemes for spintronic applications. A distinctive feature of DMSs is the dependence of their properties on the current carrier density, which can be changed by varying the doping level, voltage, and temperature. II–IV–V2 hosts, which are isoelectronic with the III–V compounds are also promising: the average charges of their cationic pairs are (II + IV)/2 = III. It was reported that such DMSs as (CdGe, Mn)ê2, (ZnGe, Mn)ê2, (ZnSn, Mn)As2, and (CdGe, Mn)As2 with TC = 320, 350, 329, and 355 K, respectively, were obtained on their basis (see references in review [2]). Nevertheless, the reports about synthesis of DMSs with high TC should be accepted with care and carefully analyzed for the presence of precipitates, defects, magnetic inclusions, and interactions between sublattices in such DMSs [2–4]. The impurity concentration in DMSs exceeds the hole density; i.e., the Ruderman–Kittel–Kasuya–Yosida exchange, in which localized magnetic moments interact with each other through charge carriers, is impossible. Since DMSs do not have a mixed valence, the exchange by valence states of the Mn3+ ⇔ Mn4+ type, proposed by Zener for La(Mn, Sr)O3, is also absent in them. The consistent theory of exchange interactions in DMSs (see [2]) is based on the structure of chemical bonds between an impurity and the host and the electron–electron correlations inside impurities. Even in the substituted systems, InSb:Mn, GaAs:Mn, GaP:Mn, or (II–IV)V2:Mn, 3d5 p or 3d5–2 p “clusters” ( p is a weakly bound hole) are energetically more favorable than Mn2+ or MnIV. As Mn2+ ions substitute cations of other valence, they supply both spins and holes to the host, thus forming an impurity band, and donor–acceptor exchange interaction arises in the host in view of the cluster overlap. In p-type DMSs, neighboring impurity Mn ions exchange with a pair of electrons, and the tunneling of these electrons, due to the Vpd hybridization without spin flip, occurs through unfilled p states at the top of the valence band. Thus, the kinematic ferromagnetic exchange interaction Jkin arises. In the DMS (more specifically, in the diluted magnetic dielectric) n-(Ga, Mn)N, the impurity band is formed deep in the wide band gap. At partial occupation of the impurity band, Mn2+(3d5) ions exchange with electrons through vacancies in the bandgap. Strict calculation within the generalized Anderson impurity model [5] takes into account the concentration dependence of Jkin, formation of the

1611

impurity band, and reconstruction of the heavy-hole band: εt

J kin

4 P 12 ( ε )Γ 12 ( ε ) V -, = ------ dε -------------------------------------------------------------------------------4 π 2 V 2 2 2 εF [ ε – E d – V P 11 ( ε ) ] + ------ π ρ ( ε ) 4

∫

∫

where P12(ε) = dω sin[k(ω)R]ρ(ω)/[k(ω)R(ε – ω)] are, respectively, the real and imaginary parts of the Green’s function of holes, Γ12(ε) = 2πρ(ε)sin[k(ε)R]/k(ε)R; P12(ε) + (i/2)Γ12(ε) integration is performed from εF to the top of the impurity band. In these formulas, k = (−2mε)1/2/
is the wave vector of holes with the density of states ρ(ω); R is the average distance between impurities, whose concentration is x = a3/(4R3) (a is the lattice constant); and Ed is the impurity level in the host. The theory of kinematic exchange in DMSs is consistent with the experimental data and makes it possible to predict new materials and, using the known parameters of the electronic structure, calculate TC = zJkinS(S + 1)/6kB (z is the coordination number of magnetoactive impurities and S is the cluster spin). Synthesis of bulk (In,Mn)Sb DMSs was performed in [4, 6] taking into account the results of the calculations of TC (figure). Unfortunately, experimental values of Ed for InSb:Mn are absent, and their potential spread yields both small (~50–100 K) and large values of TC. The theory developed can be applied to low-dimensional DMSs (heterostructures, quantum wells), which are characterized by enhanced resonant scattering of holes from impurities, and quantum confinement amplifies the exchange and correlation interaction effects [7]. For heterostructures, oscillatory dependence of the Curie temperature on the thickness of DMS layers was predicted, which explains the large values of TC in δ-doped DMSs (digital alloys) [4]. TC 400 300 200 100 0 0.1

0.2

0.3 n, nm–3

Dependence of TC on the hole density in the In0.99Mn0.01Sb DMS (obtained by E.A. Ugolkova). The Mn levels Ed = −1.906 and –2.002 eV (upper and lower curves, respectively) are counted from the top of the valence band. The hybridization parameter Vpd = 1.1 eV.

BULLETIN OF THE RUSSIAN ACADEMY OF SCIENCES: PHYSICS

Vol. 71

No. 11

2007

1612

IVANOV

ACKNOWLEDGMENTS This study was supported by the Russian Foundation for Basic Research, project no. 05-02-17666.

4. 5.

REFERENCES 1. Ginzburg, V.L., Usp. Fiz. Nauk, 1999, vol. 169, p. 419; Usp. Fiz. Nauk, 2002, vol. 172, p. 213. 2. Ivanov, V.A., Aminov, T.G., Novotortsev, V.M., and Kalinnikov, V.T., Izv. Akad. Nauk, Ser. Khim., 2004, no. 11, p. 2255. 3. Ivanov, V.A., in Sovremennye problemy i dostizheniya obshchei i neorganicheskoi khimii, (Modern Problems

6. 7.

and Advances in General and Inorganic Chemistry), Kuznetsov, N.T., Ed., Moscow: Nauka, 2004. Ivanov, V.A., Ugolkova, E.A., Pashkova, O.N., et al., J. Magn. Magn. Mater., 2006, vol. 300, p. 32. Ivanov, V.A., Krstajic, P.M., Peeters, F.M., et al., J. Magn. Magn. Mater., 2003 vol. 258–259, p. 237; Krstajic, P.M., Ivanov, V.A., Peeters, F.M., et al., Europhys. Lett., 2003, vol. 61, p. 235; Krstajic, P.M., Peeters, F.M., Ivanov, V.A., et al., Phys. Rev. B: Condens. Matter Mater. Phys., 2004, vol. 70, 195 215. Pashkova, O.N., Sanygin, V.P., Ivanov, V.A., et al., Neorg. Mater., 2006, vol. 42, p. 1. Ivanov, V.A., Proc. Sbornik trudov XX mezhdunarodnoi shkoly-seminara po NMMM (XX Int. School–Seminar on NMMM, Moscow: Izd-vo MGU, 2006, p. 874.

BULLETIN OF THE RUSSIAN ACADEMY OF SCIENCES: PHYSICS

Vol. 71

No. 11

2007