Arxiv:0804.1400V2 [Cond-Mat.Mes-Hall] 18 Oct 2008 Oeiti Aiu Ytm,Icuigsemiconductor Including Systems, Various in Heterostructures Exist to U11 Ufcsa Omtemperature
Total Page:16
File Type:pdf, Size:1020Kb
Nonequilibrium spin transport on Au(111) surfaces Ming-Hao Liu,1, ∗ Son-Hsien Chen,1, 2 and Ching-Ray Chang1 1Department of Physics, National Taiwan University, Taipei 10617, Taiwan 2Department of Physics and Astronomy, University of Delaware, Newark, DE 19716-2570, USA (Dated: October 18, 2008) The well-known experimentally observed sp-derived Au(111) Shockley surface states with Rashba spin splitting are perfectly fit by an effective tight-binding model, considering a two-dimensional hexagonal lattice with pz-orbital and nearest neighbor hopping only. The extracted realistic band parameters are then imported to perform the Landauer-Keldysh formalism to calculate nonequi- librium spin transport in a two-terminal setup sandwiching a Au(111) surface channel. Obtained results show strong spin density on the Au(111) surface and demonstrate (i) intrinsic spin-Hall ef- fect, (ii) current-induced spin polarization, and (iii) Rashba spin precession, all of which have been experimentally observed in semiconductor heterostructures, but not in metallic surface states. We therefore urge experiments in the latter for these spin phenomena. PACS numbers: 73.20.At, 73.23.-b, 71.70.Ej I. INTRODUCTION ter of the Au(111) surface states is about α =0.36 eV A.˚ Subsequent findings of giant Rashba spin-orbit coupling is also claimed in Bi(111) surfaces17 with α 0.83 eV A˚ Two-dimensional electron gas (2DEG) is known and in Bi/Ag(111) surface alloy18 with α ≈3.05 eV A.˚ to exist in various systems, including semiconductor ≈ heterostructures1 and metallic surface states.2 Due to It is therefore legitimate to expect the previously men- the lack of inversion symmetry introduced by the in- tioned spin-dependent phenomena to be observed on terface or surface, the spin degeneracy, the combining those metallic surfaces with strong Rashba coupling. In consequence of the time reversal symmetry (Kramers de- this paper we theoretically study nonequilibrium spin generacy) and the inversion symmetry, is removed and transport in 2DEG held by Au(111) surface states, which exhibit not only strong Rashba coupling but also sim- the energy dispersion becomes spin-split. In semiconduc- 19 tor heterostructures, one of the underlying mechanisms ple parabola-like dispersions. The latter characteristic leading to such spin splitting is known as the Rashba enables successful description of the band structure us- 3 ing the simplest tight-binding model (TBM), which then spin-orbit coupling, which stimulates a series of discus- 20,21,22 sion on plenty of intriguing spin-dependent phenomena. provides the Landauer-Keldysh formalism (LKF) Well studied phenomena include spin precession,4,5 spin- with reasonable or even realistic band parameters. Hall effect (SHE),6,7,8 and current-induced spin polar- This paper is organized as follows. In Sec. II we de- ization (CISP),7,8,9,10 all of which have been experimen- tally observed in semiconductor heterostructures. Con- _ _ _ Total Density of States (103 eV−1) trarily, none of these in metallic surface states is reported, ← → M Γ M 0 1 2 3 even though the Rashba effect has been shown to exist 2.5 E Band Top therein.11,12 0 F _ 2 To the lowest order in the inplane wave vector kk, the _ M K _ two spin-split energy branches are expressed as E± = −0.1 M ~2 2 ⋆ ⋆ _ 1.5 kk/2m αkk (m the electron effective mass), so that Γ ± TDOS the Rashba spin splitting ∆E = E+ E− = 2αkk is − −0.2 1 linear in kk. Here the proportional constant α is com- E (eV) E (eV) monly referred to as the Rashba coupling constant or ETBM + 0.5 arXiv:0804.1400v2 [cond-mat.mes-hall] 18 Oct 2008 the Rashba parameter. Typical values of α in semicon- ETBM −0.3 − −2 ETBM ductor heterostructures are at most of the order of 10 + TBM E 0 7,10,13 E F eV A,˚ while in metallic surface states α can be one − Eexp (a) (b) + Band Bottom or two orders larger. −0.4 Eexp − −0.5 _ _ _ −0.2 −0.1 0 0.1 0.2 The first evidence of spin splitting in metallic sur- ° Γ K M k (A−1) face states was pioneered by LaShell et al. on || Au(111) surfaces at room temperature.11 The origin of their observed spin splitting was later recognized FIG. 1: (Color online) (a) Tight-binding energy dispersion TBM exp as the Rashba effect by performing the first-principles E± and the experimentally measured binding energy E± electronic-structure and photoemission calculations,14,15 of Ref. 11, along the Γ¯M¯ direction. The surface Brillouin which are in good agreement with the spin-resolved pho- zone is sketched in the inset. (b) Total density of states and TBM ¯ ¯ ¯ ¯ toemission experiments.15,16 Concluded Rashba parame- E± along ΓK and KM directions. 2 scribe the Au(111) surface band structure by using an In the vicinity of Γ¯, i.e., k a 1, Eqs. (3) and | k| ≪ effective TBM, through which the experimentally mea- (4) are approximated by F (kk) 3VR (kx iky) a and 11 2 ≈−2 − sured energy dispersions can be perfectly reproduced. G(k ) 6Vppπ (3Vppπa /2)k , respectively, and the k ≈ − k Section III is devoted to nonequilibrium spin transport Hamiltonian matrix (2) then takes the form on a finite Au(111) surface channel attached to two exter- 2 nal leads, using the LKF with band parameters extracted 3Vppπa 2 E0 k 3VR (kx iky) a in Sec. II. The intrinsic SHE, the CISP, and the Rashba H − 2 k − − kka≪1 = 2 , spin precession will be shown by directly imaging the lo- 3Vppπa 2 3VR (kx + iky) a E0 kk cal spin densities. We conclude in Sec. IV. − − 2 (6) where E Ep + 6Vppπ. Equation (6) is consistent 0 ≡ II. Au(111) SURFACE BAND STRUCTURE with the pz-resolved effective Hamiltonian of the earlier TBM by Petersen and Hedeg˚ard, who considered all the three p-orbitals, subject to the intra-atomic spin-orbit A. Effective tight-binding model coupling.26 We first demonstrate that the sp-derived Shockley sur- face states on Au(111) from Ref. 11 can be well described B. Extraction of band parameters by an effective TBM [see Fig. 1(a)] for a single sheet of two-dimensional hexagonal lattice, taking into account We now fit our tight-binding dispersions (5) with the only pz-orbital hopping between nearest neighbors, sub- experiment of Ref. 11. This can be done by comparing ject to the Rashba spin-orbit coupling. The Hamiltonian the low-kk expansion of Eq. (5), matrix can be written as23,24 2 3Vppπa k t 2 H i k· I E(kk) Ep +6Vppπ k 3VRa kk , (7) = Ep11+ e [Vppπ11+ VRez (~σ tI )] , (1) ka≪1 ≈ − 2 k ± | | · × XtI with that of the free-electron model, E(kk) = E0 + ~2 ⋆ 2 where 11 is the 2 2 identity matrix, Ep is the p-orbital ( /2m )kk αkk. In addition to the band offset E0 = × ± 2 2 ⋆ energy, tI represents the six nearest neighbor hopping Ep + 6Vppπ, we identity Vppπ = (2/3a )(~ /2m ) − vectors, Vppπ is the band parameter describing the orbital and VR = α/3a. Using the reciprocal vector g1 = −1 integral under the two-center approximation of Slater and (4π/√3a)(1/2, √3/2, 0) and M¯ = 1.26 A˚ from Ref. 25 Koster, V is the Rashba hopping parameter, and ~σ = 2 R 11, we have ~2/2m⋆ 15.2 eV A˚ , α 0.3557 eV A,˚ (σ , σ , σ ) is the Pauli matrix vector. The three terms in x y z and E 0.415 eV.≈ The norm of g≈gives M¯ such Eq. (1) are the energy band offset, the kinetic hopping, 0 1 that the≈ lattice − constant is a = 4π/√3 g = 5. 7581 A.˚ and the Rashba hopping, respectively. Arranging the two 1 Hence we deduce V = 0.3056 eV, V| |= 0.0206 eV, primitive translation vectors for the hexagonal lattice as ppπ R and E =1.4188 eV. Substituting− these parameters into t = (√3/2, 1/2, 0)a and t = ( √3/2, 1/2, 0)a where a p 1 2 Eqs. (3)–(5), a nearly perfect consistency between our is the lattice constant, the six nearest− neighbor hopping effective TBM and the experimentally measured binding vectors are t = t , t , (t + t ), and Eq. (1) then I 1 2 1 2 energy of Ref. 11 can be seen in Fig. 1(a). The ex- takes the explicit± form± of ± perimentally measured Fermi surface of the concentric rings slightly distorted from circles19 can be reproduced H Ep + G(kk) F (kk) (kk)= ∗ (2) as well, but we do not explicitly show. F (kk) Ep + G(kk) with III. NONEQUILIBRIUM SPIN TRANSPORT F (kk)= iVR[(1 + √3i) sin kk t1 · A. Landauer-Keldysh formalism vs tight-binding + (1 √3i) sin k t + 2 sin kya] (3) − k · 2 model √3kxa kya G(k )=2V [2 cos cos + cos(k a)]. (4) 21 k ppπ 2 2 y Next we apply the Landauer-Keldysh formalism, namely the nonequilibrium Keldysh Green’s function Equation (2) can be diagonalized to yield the energy dis- formalism27 applied on Landauer multiterminal ballistic persions nanostructures. For detailed introduction to the LKF, see Refs. 20,22. To make use of the previously extracted E(k )= Ep + G(k ) F (k ) . (5) k k ± | k | band parameters in the LKF calculation, we consider the second-quantized single particle Hamiltonian,24 Noting from Eq. (3) that VR is embedded in F (kk), the † † above dispersion contains the Rashba term to all (odd) = εnc cn + c [t + itR (~σ dmn) ] cn, (8) H n m 0 × z orders in kk.