1: % ======================================================================

2: % based on: revtex4/sample/template.aps

3: % last modified on: 06-26-03

4: % last modified by: Cyrus F. Hirjibehedin

5: % ======================================================================

6: %

7: % Group addresses by affiliation; use superscriptaddress for long

8: % author lists, or if there are many overlapping affiliations.

9: % For Phys. Rev. appearance, change preprint to twocolumn.

10: % Choose pra, prb, prc, prd, pre, prl, prstab, or rmp for journal

11: %  Add 'draft' option to mark overfull boxes with black boxes

12: %  Add 'showplace' option to make PACES codes appear

13: %  Add 'showers' option to make keywords appear

14: %\documentclass[aps,prl,preprint,groupedaddress]{revtex4}

15: %

16: \documentclass[aps,prl,preprint,superscriptaddress,showpacs,amssymb]{revtex4}

17: %\documentclass[aps,prl,twocolumn,superscriptaddress,showpacs,amssymb]{revtex4}

18: %

19: %additional packages to include

20: \usepackage{bm}

21: % ======================================================================

22: % figures should be put into the text as floats.

23: % Use the graphics or graphicx packages (distributed with LaTeX2e)

24: % and the \includegraphics macro defined in those packages.

25: % See the LaTeX Graphics Companion by Michel Goosens, Sebastian Rahtz,

26: % and Frank Mittelbach for instance.

27: %%\usepackage{graphicx}% Include figure files

28: \usepackage{epsfig}

29: %

30: % You should use BibTeX and apsrev.bst for references

31: % Choosing a journal automatically selects the correct APS

32: % BibTeX style file (bst file), so only uncomment the line

33: % below if necessary.

34: %\bibliographystyle{apsrev}

35: %

36: \begin{document}

37: %

38: % Use the \preprint command to place your local institutional report

39: % number in the upper righthand corner of the title page in preprint mode.

40: % Multiple \preprint commands are allowed.

41: % Use the 'preprintnumbers' class option to override journal defaults

42: % to display numbers if necessary

43: %\preprint{unsubmitted}

44: %

45: %Title of paper

46: \title{Spin-Hall effect and spin-coherent excitations in a strongly

47: confined two-dimensional hole gas}

48: %

49: % repeat the \author .. \affiliation  etc. as needed

50: % \email, \thanks, \homepage, \altaffiliation all apply to the current

51: % author. Explanatory text should go in the []'s, actual e-mail

52: % address or url should go in the {}'s for \email and \homepage.

53: % Please use the appropriate macro foreach each type of information

54: %

55: % \affiliation command applies to all authors since the last

56: % \affiliation command. The \affiliation command should follow the

57: % other information

58: % \affiliation can be followed by \email, \homepage, \thanks as well.

59: %\email[cfh@phys.columbia.edu]{Your e-mail address}

60: %\homepage[]{Your web page}

61: %\thanks{}

62: %\altaffiliation{}

63: \author{P. Kleinert}

64: %\email[kl@pdi-berlin.de]{Your e-mail address}

65: \affiliation{Paul-Drude-Intitut f\"ur Festk\"orperelektronik,

66: Hausvogteiplatz 5-7, 10117 Berlin, Germany}

67: %

68: \author{V.V. Bryksin}

69: \affiliation{Physical Technical Institute, Politekhnicheskaya 26,

70: 194021 St. Petersburg, Russia}

71: %

72: \date{\today}

73: %

74: \begin{abstract}

75: Based on a rigorous quantum-kinetic approach, spin-charge coupled

76: drift-diffusion equations are derived for a strongly confined

77: two-dimensional hole gas. An electric field leads to a coupling

78: between the spin and charge degrees of freedom. For weak

79: spin-orbit interaction, this coupling gives rise to the intrinsic

80: spin-Hall effect. There exists a threshold value of the spin-orbit

81: coupling constant that separates spin diffusion from ballistic

82: spin transport. In the latter regime, undamped spin-coherent

83: oscillations are observed. This result is confirmed by an exact

84: microscopic approach valid in the ballistic regime.

85: \end{abstract}

86: 

87: % insert suggested PACES numbers in braces on next line

88: \pacs{72.25.-b, 72.10.-d, 72.15.Gd}

89: 

90: % insert suggested keywords - APS authors don't need to do this

91: %\keywords{space charge waves, electron hole system,surface acoustic wave}

92: 

93: %\maketitle must follow title, authors, abstract, \pacs, and \keywords

94: \maketitle

95: 

96: % body of paper here - Use proper section commands

97: % References should be done using the \cite, \ref, and \label commands

98: 

99: The generation and manipulation of a spin polarization by

100: exclusively electronic means in nonmagnetic semiconductors at room

101: temperature is a major challenge of spintronics. Among many

102: interesting phenomena, the intrinsic spin-Hall effect (SHE)

103: \cite{SCIE_1348,PRL_126603} has recently attracted considerable

104: interest. Experimental studies

105: \cite{SCIE_1910,PRL_047204,PRL_096605} reveal an electric-field

106: induced spin accumulation near the edges of a confined

107: two-dimensional electron (hole) gas. Most theoretical

108: interpretations of these experimental data rely on the notion of a

109: spin current oriented transverse to the applied electric

110: field.~\cite{SCIE_1348,PRL_126603} Interestingly, this seemingly

111: clear physical picture still remains the subject of serious

112: debates.\cite{PRL_076604,PRB_165313} The relationship between the

113: spin current and the induced spin polarization seems to be a very

114: subtle issue. The main problem underlying the debates is the

115: notion of a spin current itself because spin is not a conserved

116: quantity in spin-orbit coupled systems. Consequently, any approach

117: that avoids the intricate identification of a more or less

118: suitable spin current is superior. Such an alternative approach

119: not only introduces different calculational techniques, but also

120: suggests alternative interpretations of the effects under

121: consideration. As widely anticipated, a complete physical

122: description of spin-related phenomena is provided by microscopic

123: models based on the spin-density matrix or Keldysh Green functions

124: together with an analysis of its long-wavelength and low-frequency

125: limit. These approaches are more general and free from artefact

126: associated with ambiguous definitions of the spin current. In

127: addition, spin-charge coupled kinetic equations allow the

128: treatment of such interesting phenomena as propagating spin

129: excitations or the relationship between the intrinsic SHE and the

130: zitterbewegung.

131: 

132: In this report, we propose an alternative approach to the SHE by

133: deriving spin-charge coupled drift-diffusion equations for a

134: two-dimensional hole gas (2DHG), which refers to the populated

135: heavy-hole band of thin p-type quantum wells. The related

136: heavy-hole Hamiltonian of the cubic Rashba model has the second

137: quantized form

138: \begin{eqnarray}

139: &&H=\sum_{{\bm{k}},\lambda }a_{\bm{k}\lambda }^{\dag}\left[ \varepsilon_{%

140: \bm{k}}-\varepsilon _{F}\right] a_{%

141: \bm{k}\lambda }-\sum_{\bm{k},\lambda ,\lambda ^{\prime }}\left(

142: \hbar{\bm{\omega}}_{

143: \bm{k}} \cdot {\bm{\sigma }}_{\lambda \lambda ^{\prime }}\right) a_{\bm{k}%

144: \lambda }^{\dag}a_{\bm{k}\lambda ^{\prime }}\nonumber\\

145: &&+ u\sum\limits_{{\bm{k}},{\bm{k}}^{\prime}}

146: \sum\limits_{\lambda}a_{{\bm{k}}\lambda}^{\dag}a_{{\bm{k}}^{\prime}\lambda}

147: -ie{\bm{E}}\sum\limits_{{\bm{k}},\lambda}\nabla_{\bm{\kappa}}

148: a^{\dag}_{{\bm{k}}-{\bm{\kappa}}/2\lambda}a_{{\bm{k}}+{\bm{\kappa}}/2\lambda}\mid_{\bm{\kappa}=0},

149: \label{Hamil}

150: \end{eqnarray}

151: where $a_{\bm{k}\lambda}^{\dag}$ ($a_{\bm{k}\lambda}$) denote the

152: creation (annihilation) operators with in-plane quasi-momentum

153: $\bm{k}=(k_x,k_y,0)$ and spin $\lambda$. The electric field

154: ${\bm{E}}$ is oriented along the $x$ axis. Furthermore,

155: $\varepsilon_F$ denotes the Fermi energy, $\bm{\sigma}$ the vector

156: of Pauli matrices, $\varepsilon_{\bm{k}}=\hbar^2k^2/(2m)$, and $u$

157: the strength of the 'white-noise' elastic impurity scattering,

158: which gives rise to the momentum relaxation time $\tau$. Contrary

159: to a phenomenological approach, we treat elastic scattering on a

160: full microscopic scope. The spin-orbit coupling is given by

161: \begin{equation}

162: \hbar {\bm{\omega}}_{\bm{k}}=\frac{\alpha}{2}

163: \left[i(k_{+}^3-k_{-}^3),(k_{+}^3+k_{-}^3),0 \right],\label{omega}

164: \end{equation}

165: where $k_{\pm}=k_x\pm ik_y$ and $\hbar\omega_k=\alpha k^3$. Based

166: on the Born approximation for elastic impurity scattering, the

167: Laplace-transformed kinetic equations for the physical components

168: of the spin-density matrix have the form~\cite{PRB_165313}

169: \begin{equation}

170: sf-\frac{i\hbar}{m}(\bm{\kappa}\cdot\bm{k})f+i{\bm{\omega}}_{\bm{\kappa}}

171: ({\bm{k}})\cdot{\bm{f}}

172: +\frac{e{\bm{E}}}{\hbar}\nabla_{\bm{k}}f=\frac{1}{\tau}(\overline{f}-f)+f_0,

173: \label{kin1}

174: \end{equation}

175: \begin{eqnarray}

176: && s{\bm{f}}+2({\bm{\omega}}_{\bm{k}}\times{\bm{f}})

177: -\frac{i\hbar}{m}(\bm{\kappa}\cdot\bm{k}){\bm{f}}

178: +i{\bm{\omega}}_{\bm{\kappa}}

179: ({\bm{k}})f+\frac{e{\bm{E}}}{\hbar}\nabla_{\bm{k}}

180: {\bm{f}}\nonumber\\

181: &&=\frac{1}{\tau}(\overline{{\bm{f}}}-{\bm{f}})+\frac{1}{\tau}

182: \frac{\partial}{\partial\varepsilon_{\bm{k}}}

183: \overline{f\hbar{\bm{\omega}}_{\bm{k}}}-

184: \frac{\hbar{\bm{\omega}}_{\bm{k}}}{\tau}

185: \frac{\partial}{\partial\varepsilon_{\bm{k}}} \overline{f}

186: +{\bm{f}}_0, \label{kin2}

187: \end{eqnarray}

188: where an additional frequency appears

189: \begin{eqnarray}

190: {\bm{\omega}}_{\bm{\kappa}}({\bm{k}})&=&\frac{3\alpha}{\hbar}

191: \left[(k_y^2-k_x^2)\kappa_y-2k_xk_y\kappa_x,\right. \nonumber\\

192: &&\left. (k_x^2-k_y^2)\kappa_x-2k_xk_y\kappa_y,0 \right],

193: \end{eqnarray}

194: which depends on ${\bm{\kappa}}$. The cross line over

195: ${\bm{k}}$-dependent quantities denotes an integration over the

196: polar angle $\varphi$ of the in-plane vector ${\bm{k}}$. $f_0$ and

197: ${\bm{f}}_0$ denote the initial charge and spin density

198: components, respectively. The quantum Boltzmann equations

199: (\ref{kin1}) and (\ref{kin2}) are treated in the long-wavelength

200: limit in order to derive spin-charge coupled drift-diffusion

201: equations. To this end, the kinetic equations are written in a

202: matrix form $A\widehat{f}+{\cal

203: E}\widehat{f}=B\overline{\widehat{f}}+ \widehat{\delta}$, where

204: the matrix $A$ collects all contributions that are independent of

205: the electric field ${\bm{E}}$ and not integrated over the angle

206: $\varphi$ [$\widehat{f}$ denotes the four component vector

207: $(f,{\bm{f}})$]. To calculate the matrix $B$ on the right-hand

208: side of this equation, we assume $\alpha k_F^3\ll

209: \varepsilon_{k_F}$ and restrict $\kappa$ contributions up to

210: $\kappa^2$. The matrix equation is solved iteratively in the case

211: of weak electric field contributions ${\cal E}$ (the matrix ${\cal

212: E}$ contains first-order derivatives and $\widehat{f}$ is

213: decomposed according to $\widehat{f}_0+\widehat{f}_1$ with

214: $\widehat{f}_0\sim E^0$ and $\widehat{f}_1\sim E$). The solution

215: of the equation is written in the form

216: \begin{equation}

217: (1+\overline{C}_1^{-1}\overline{C}_2)^{-1}\overline{C}_1^{-1}

218: \overline{\widehat{f}}=\widehat{\delta},

219: \end{equation}

220: where $\overline{\widehat{f}}_0=\overline{C}_1\widehat{\delta}$

221: and $\overline{\widehat{f}}_1=\overline{C}_2\widehat{\delta}$. The

222: general expressions for the transport coefficients are very

223: cumbersome but simplify considerably in the low-field case and

224: under the condition $\alpha k_F^3\ll \varepsilon_{k_F}$.

225: 

226: As we are mainly interested in the SHE, let us focus on the

227: coupling between the spin and charge degrees of freedom. By

228: applying the outlined schema, we obtain our main analytical result

229: \begin{equation}

230: (s+i\frac{v_d}{\sigma_0}\kappa_x+D_0\kappa^2)\overline{f}+i\Gamma_z\kappa_y\overline{f}_z=f_0,\label{e1}

231: \end{equation}

232: \begin{equation}

233: (s+\frac{1}{\tau_{sz}}+i\frac{v_d}{\sigma_0}\kappa_x+D_z\kappa^2)\overline{f}_z

234: +i\Gamma_0\kappa_y\overline{f}=0,\label{e2}

235: \end{equation}

236: with the transport coefficients

237: \begin{equation}

238: D_0=\frac{D}{\sigma_0^2},\quad

239: D_z=D\frac{\sigma_0^2-12\Omega^2}{(\sigma_0^2+4\Omega^2)^2},\label{e10}

240: \end{equation}

241: \begin{equation}

242: \Gamma_0=v_d\frac{9\Omega^2}{2\gamma}\frac{3\sigma_0^2-4\Omega^2}{(\sigma_0^2+4\Omega^2)^2},

243: \quad \frac{1}{\tau_{sz}}=\frac{4\Omega^2}{\sigma_0\tau},

244: \end{equation}

245: \begin{equation}

246: \Gamma_z=v_d\frac{9\Omega^2}{2\gamma}

247: \sigma_0^2\frac{4\sigma_0\Omega^2+8\Omega^2-3\sigma_0^2s\tau}

248: {(\sigma_0s\tau +4\Omega^2)(\sigma_0^2+4\Omega^2)}.

249: \end{equation}

250: The parameters introduced in these equations are given by:

251: $v_d=eE\tau/m$, $\sigma_0=s\tau+1$, $D=v^2\tau/2$,

252: $\gamma=\varepsilon_k\tau/\hbar$, and $\Omega=\omega_k\tau=K(k)l$

253: (with $K(k)=\alpha mk^2/\hbar^2$ and the mean-free path

254: $l=v\tau$). As the contributions $\sim \kappa_x$ do not affect our

255: analysis, we took them to lowest order in $\Omega$. The

256: Eqs.~(\ref{e1}) and (\ref{e2}) have been derived for

257: $\Omega/\gamma\ll 1$ but unrestricted values $\Omega$. In the

258: absence of the electric field ($E=0$), the Eqs.~(\ref{e1}) and

259: (\ref{e2}) completely decouple. This decoupling, which applies to

260: all components of the spin-density matrix, represents a speciality

261: of the cubic Rashba model.~\cite{PRB_193316,PRB_195330} The

262: situation is completely different for electrons (linear Rashba

263: model), for which at zero external fields the charge density

264: $\overline{f}$ couples to the transverse spin component

265: $\overline{f}_r=i({\bm{\kappa}}\times{\overline{\bm{f}}})_z$,

266: whereas $\overline{f}_z$ is connected with the longitudinal

267: component $\overline{f}_d=i{\bm{\kappa}}\cdot{\overline{\bm{f}}}$.

268: However, both for the linear and cubic Rashba model additional

269: couplings arise due to an applied electric field. The related

270: magnetization gives rise to a magnetoelectric effect that has been

271: thoroughly investigated in the literature for semiconductors with

272: spin-orbit interaction.

273: 

274: The time dependence of the coupled spin-charge transport is

275: calculated by an inverse Laplace transformation of the solution of

276: Eqs.~(\ref{e1}) and (\ref{e2}). Due to the complicated $s$

277: dependence of all transport coefficients, a non-Markovian temporal

278: evolution is expected. However, the straightforward determination

279: of this complicated time dependence of charge and spin densities

280: becomes problematic. Eqs.~(\ref{e1}) and (\ref{e2}) are only

281: solvable by inverse Fourier- and Laplace transformations under

282: appropriate additional approximations concerning the $s$

283: dependence. This delicate mathematical problem will be accounted

284: for in a forthcoming paper. Here, we focus on steady-state

285: solutions ($\sigma_0=1$).

286: 

287: The character of the coupled spin-charge transport strongly

288: depends on the strength of the spin-orbit coupling, which is

289: expressed by the dimensionless parameter $\Omega=K(k)l$. It is the

290: most striking feature of the drift-diffusion Eqs.~(\ref{e1}) and

291: (\ref{e2}) that the character of the spin transport changes

292: radically with increasing coupling strength $\Omega$. The

293: appearance of such a crossover is due to the unusual expression

294: for the diffusion coefficient $D_z$ in Eq.~(\ref{e10}), which has

295: recently been derived by an alternative

296: approach.~\cite{PRB_193316} Moreover, the very same result is also

297: obtained for the linear Rashba model, when the frequency

298: $\omega_k$ is appropriately redefined. With increasing spin-orbit

299: coupling $K(k)$ or relaxation time $\tau$, the diffusion

300: coefficient $D_z$ changes its sign. A negative diffusion

301: coefficient indicates an instability of the spin system. Under

302: this condition, spin diffusion has the tendency to strengthen

303: initial spin fluctuations. The competition between this self

304: strengthening and spin relaxation processes results in a spatial

305: oscillatory spin pattern. Going from weak ($\Omega<1/\sqrt{12}$)

306: to strong ($\Omega>1/\sqrt{12}$) spin-orbit coupling, we observe a

307: transition in the spin system from a diffusive behavior to a

308: ballistic regime. We shall show that ballistic spin transport is

309: characterized by undamped spin oscillations.

310: 

311: To be specific, let us solve Eqs.~(\ref{e1}) and (\ref{e2}) for a

312: stripe geometry ($-L_0\leq y\leq L_0$). The inverse Fourier

313: transformation is accomplished by the replacement

314: $\kappa_y\rightarrow i\partial/\partial y$, whereas along the $x$

315: axis all quantities are constant. The resulting set of

316: differential equations is easily solved. For the boundary

317: condition $\overline{f}(-L_0)=\overline{f}(L_0)=f_0$ and

318: $\overline{f}_z(-L_0)=\overline{f}_z(L_0)=0$, we obtain

319: \begin{equation}

320: \overline{f}_z(k,y)=f_0\frac{\tau_{sz}\Gamma_0}{D_0}\,\frac{y\sinh(L_0/l_0)-L_0\sinh(y/l_0)}

321: {\sinh(L_0/l_0)},\label{e13}

322: \end{equation}

323: where $l_0=\sqrt{D_z\tau_{sz}}$.

324: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

325: \begin{figure}

326: %%\vspace*{5cm}

327: \centerline{\epsfig{file=figure1.eps,width=7.0cm}} \caption{Out-of

328: plane spin polarization induced by an electric field applied

329: parallel to the stripe of a 2DHG (with $\gamma=0.5$). The thick

330: and thin lines were calculated with $\Omega=0.25$ and

331: $\Omega=0.5$, respectively.} \label{fig1}

332: \end{figure}

333: % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%

334: An interesting effect, which we do not follow up in this paper,

335: results from the $y$ dependence of the charge density

336: $\overline{f}$ that is strongly affected by the boundary condition

337: and that gives rise to a self-consistent electric field oriented

338: along the $y$ direction. In Fig.~1, the thick line illustrates the

339: result for the spin polarization in the diffusive regime

340: ($D_z>0$), when the spin-orbit coupling is weak

341: $\Omega<1/\sqrt{12}$. The electric field aligned along the $x$

342: axis induces a spin polarization at the edges of the stripe. This

343: SHE has received a great deal of attention. Most popular is the

344: description by means of a spin current oriented perpendicular to

345: the electric field. Many theoretical studies (see, for instance,

346: Ref.[\onlinecite{PRB_085308,PRB_193316}]) introduced the spin

347: current $\widehat{J}_{\mu}^{i}$ by a symmetrized product of spin

348: and velocity operators

349: $(\widehat{v}_{\mu}\sigma^{i}+\sigma^{i}\widehat{v}_{\mu})/2$. It

350: was claimed that at least for the cubic Rashba model, the SHE

351: introduced in such a manner is robust against disorder.

352: Experimental results \cite{PRL_047204} seem to confirm this

353: physical picture. However, there is a principal difficulty with

354: such an approach. The above mention definition of the current is

355: not sufficiently general. It would completely fail for any hopping

356: transport problem, for which the eigenstates have no dispersion.

357: This definition only applies, whenever the Hamiltonian commutes

358: with the dipole operator. Obviously, this is not the case for the

359: Rashba Hamiltonian. Consequently, it is necessary to go back to

360: the more general definition, which expresses the current by the

361: time derivative of the dipole

362: operator.~\cite{PRL_076604,PRB_165313} This more general treatment

363: of the spin current reveals a close relationship between the

364: field-induced spin accumulation and the spin current expressed by

365: a quasi-chemical potential. From a technical point of view, the

366: current that applies in its most general form to a homogeneous

367: system is calculated from the quantity

368: $\nabla_{\bm{\kappa}}\widehat{f}({\bm{k}},{\bm{\kappa}})$ at

369: ${\bm{\kappa}}={\bm{0}}$ and not only from the density matrix

370: $\widehat{f}({\bm{k}})$. Based on this general definition of the

371: spin current, it was concluded that an electric-field-induced

372: steady-state spin-Hall current does not exist in the cubic Rashba

373: model.~\cite{JPCM_7497} On the other hand, a SHE was demonstrated

374: by a recent experiment.~\cite{PRL_047204} This calamity indicates

375: that the notion of a spin current is not useful for studying the

376: SHE. An alternative, which is proposed in this paper, determines

377: the spin accumulation from quantum kinetic equations or the

378: associated spin-charge coupled drift-diffusion equations. The

379: specific difficulty of the latter approach is the formulation of

380: appropriate boundary conditions.~\cite{PRB_113309,PRB_115331}

381: 

382: The character of the SHE dramatically changes in strongly

383: spin-orbit coupled systems ($\Omega>1/\sqrt{12}$ so that $D_z<0$).

384: The result is illustrated in Fig.~1 by the thin line. A

385: spin-coherent standing wave travels through the stripe. It is

386: remarkable that these oscillations are not damped although a

387: finite elastic scattering is present. The occurrence of a periodic

388: spin pattern is not unusual and has been investigated in the

389: literature.~\cite{PRB_155317,PRB_033316,PRB_195308,PRB_205307} The

390: novelty here is that such an oscillatory spin pattern can be

391: induced by an electric field. The rapid variation of the out-of

392: plane spin polarization induces a magnetic field that leads to

393: circulating microscopic currents. The retroaction of these

394: currents on spin may result in a finite damping of spin

395: oscillations.

396: 

397: In the strong-coupling limit ($\Omega=K(k)l\gg 1$), the

398: oscillatory spin pattern changes on a length scale $K^{-1}(k)$

399: that is much smaller than the mean-free path $l$. This fact

400: conflicts with basic assumptions of the drift-diffusion approach,

401: which is only applicable for diffusion lengths much smaller than

402: the mean-free path. Although macroscopic transport equations were

403: found to be valid even when the spin-diffusion length is somewhat

404: less than the mean-free path \cite{PRB_212410}, it is

405: indispensable to treat this point in detail. Large values for

406: $\Omega$ give rise to spin-relaxation times $\tau_{sz}$, which are

407: much smaller than the elastic scattering time $\tau$. This

408: condition characterizes the ballistic spin regime. Therefore, we

409: go back to the kinetic Eqs.~(\ref{kin1}) and (\ref{kin2}) and

410: solve them under the condition $\tau\rightarrow \infty$ and to

411: first order in the electric field. For the out-of plane spin

412: polarization, we obtain

413: \begin{equation}

414: (\sigma^2+4\omega_k^2)f_z+\frac{eE\sigma}{\hbar}\frac{\partial}{\partial

415: k_x}f_z=\frac{2i}{\sigma}({\bm{\omega}}_{{\bm{k}}x}{\bm{\omega}}_{{\bm{\kappa}}y}

416: -{\bm{\omega}}_{{\bm{k}}y}{\bm{\omega}}_{{\bm{\kappa}}x})f_0,

417: \end{equation}

418: where $\sigma=s-i{\bm{\kappa}}\cdot{\bm{v}}$. Calculating the

419: inverse Laplace and Fourier transformation and integrating over

420: the angle $\varphi$, we arrive at the analytical solution

421: \begin{eqnarray}

422: &&f_z(k,{\bm{r}},t)=\frac{3f_0}{16\pi r}\frac{\partial}{\partial

423: k}\left[\frac{eE}{\varepsilon_k}\left(\frac{\sin 2K(k)r}{2K(k)r}

424: -\cos 2K(k)r \right) \right]\nonumber\\

425: && \times\left\{

426: \begin{array}{ccc}

427:   \delta(y-\sqrt{(vt)^2-x^2}) & {\rm for} & 0<y\leq vt \\

428:   -\delta(y+\sqrt{(vt)^2-x^2}) & {\rm for} & vt\leq y<0 \\

429: \end{array}\right.\,,

430: \end{eqnarray}

431: which describes the field-induced spin polarization that occurs,

432: when initially a drop of charge carriers is injected into the 2DHG

433: at the position ${\bm{r}}={\bm{0}}$.

434: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

435: \begin{figure}

436: %%\vspace*{5cm}

437: \centerline{\epsfig{file=figure2.eps,width=7.0cm}} \caption{The

438: amplitude $F(r)=\eta x{\rm d}((\sin x/x-\cos x)/x)/{\rm d}x$ in

439: dependence on $r$ with $x=2K(k_F)r$ and $K(k_F)=0.5$. For the

440: factor $\eta$, we have: $\eta = \pm 1$ for $y \gtrless 0$.}

441: \label{fig2}

442: \end{figure}

443: % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%

444: A $\delta$-like wave front of spin polarization travels through

445: the homogeneous 2DHG. The amplitude of this narrow wave front

446: oscillates as illustrated in Fig.~2. As expected, the relief is

447: antisymmetric with respect to the $y$ axis. Most interesting for

448: our comparison with the above drift-diffusion approach is the

449: observation that the wavelength of the spatial spin pattern is of

450: the order of $K^{-1}(k_F)$. Therefore, this exact analytical

451: result confirms the existence of a field-induced long-living spin

452: pattern in strongly spin-orbit coupled systems as predicted by the

453: drift-diffusion equations.

454: 

455: Our study of electric-field induced spin phenomena revealed a

456: close relationship between the SHE and spin-coherent oscillations.

457: We compare this conclusion with recent results that demonstrated

458: that the intrinsic SHE and the zitterbewegung are essentially the

459: same kind of phenomena.~\cite{PRB_205307} Consequently, the

460: question arises whether the above treated spin-coherent waves have

461: to be identified with the zitterbewegung, which is a purely

462: relativistic effect. Indeed, both kinds of oscillatory spin

463: excitations exhibit common features. The characteristic wavelength

464: of both types of oscillations amounts about $100$~nm (calculated

465: by adopting the typical values $\alpha m/\hbar^{2}\sim 2$~nm and

466: $k_F\sim 0.1$~nm). Moreover, both the zitterbewegung

467: \cite{PRL_206801} and the spin-coherent waves [cf.

468: Eq.~(\ref{e13})] are resonantly enhanced, whenever the width of

469: the stripe matches a characteristic wavelength of the spin

470: excitation. However, some features of spin-coherent waves are not

471: compatible with such an identification with the zitterbewegung.

472: The dispersion relation of coupled spin-charge excitations is

473: calculated from the vanishing determinant of Eqs.~(\ref{e1}) and

474: (\ref{e2}). In general, one obtains not only spin-coherent

475: solutions but also damped excitations, whereas the transition

476: between them could be driven by the electric field. In addition,

477: the spin-coherent waves that appear at strong spin-orbit coupling

478: are separated from the SHE by a sharp threshold. For the

479: zitterbewegung such a threshold is not expected as its existence

480: is solely due to at least two energy bands separated by a nonzero

481: gap. We think that the interesting study of the relationship

482: between spin-coherent waves, the zitterbewegung, and long-living

483: spin-coherent states \cite{PRB_155317} will continue in the near

484: future.

485: 

486: The experimental observation of the field-induced spin-coherent

487: waves should be possible by high-resolution scanning-probe

488: microscopy imaging techniques. The direct experimental proof of

489: this effect would fascilitate developments both in spintronics and

490: basic research.

491: 

492: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

493: %\begin{acknowledgments}

494: This work was supported by the Deutsche Forschungsgemeinschaft and

495: the Russian Foundation of Basic Research under the grant number

496: 05-02-04004.

497: %\end{acknowledgments}

498: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

499: 

500: %\bibliographystyle{prsty}

501: %\bibliography{abbrev,spin}

502: \begin{thebibliography}{10}

503: 

504: \bibitem{SCIE_1348}

505: S. Murakami, N. Nagaosa, and S.~C. Zhang, Science {\bf 301},  1348

506: (2003).

507: 

508: \bibitem{PRL_126603}

509: J. Sinova, D. Culcer, Q. Niu, N.~A. Sinitsyn, T. Jungwirth, and

510: A.~H.

511:   MacDonald, Phys. Rev. Lett. {\bf 92},  126603  (2004).

512: 

513: \bibitem{SCIE_1910}

514: Y.~K. Kato, R.~C. Myers, A.~C. Gossard, and D.~D. Awschalom,

515: Science {\bf 306},

516:    1910  (2004).

517: 

518: \bibitem{PRL_096605}

519: V. Sih, W.~H. Lau, R.~C. Myers, V.~R. Horowitz, A.~C. Gossard, and

520: D.~D.

521:   Awschalom, Phys. Rev. Lett. {\bf 97},  096605  (2006).

522: 

523: \bibitem{PRL_047204}

524: J. Wunderlich, B. Kaestner, J. Sinova, and T. Jungwirth, Phys.

525: Rev. Lett. {\bf

526:   94},  047204  (2005).

527: 

528: \bibitem{PRL_076604}

529: J. Shi, P. Zhang, D. Xiao, and Q. Niu, Phys. Rev. Lett. {\bf 96},

530: 076604

531:   (2006).

532: 

533: \bibitem{PRB_165313}

534: V.~V. Bryksin and P. Kleinert, Phys. Rev. B {\bf 73},  165313

535: (2006).

536: 

537: \bibitem{PRB_195330}

538: O. Bleibaum and S. Wachsmut, Phys. Rev. B {\bf 74},  195330

539: (2006).

540: 

541: \bibitem{PRB_193316}

542: T.~L. Hughes, Y.~B. Bazaliy, and B.~A. Bernevig, Phys. Rev. B {\bf

543: 74},  193316

544:    (2006).

545: 

546: \bibitem{PRB_085308}

547: J. Schliemann and D. Loss, Phys. Rev. B {\bf 71},  085308  (2005).

548: 

549: \bibitem{JPCM_7497}

550: P. Kleinert and V.~V. Bryksin, J. Phys.: Condens. Matter {\bf 18},

551: 7497

552:   (2006).

553: 

554: \bibitem{PRB_113309}

555: O. Bleibaum, Phys. Rev. B {\bf 74},  113309  (2006).

556: 

557: \bibitem{PRB_115331}

558: V.~M. Galitski, A.~A. Burkov, and S.~D. Sarma, Phys. Rev. B {\bf

559: 74},  115331

560:   (2006).

561: 

562: \bibitem{PRB_155317}

563: Y.~V. Pershin, Phys. Rev. B {\bf 71},  155317  (2005).

564: 

565: \bibitem{PRB_033316}

566: J. Wang, K.~S. Chan, and D.~Y. Xing, Phys. Rev. B {\bf 73},

567: 033316  (2006).

568: 

569: \bibitem{PRB_195308}

570: Y. Jiang, Phys. Rev. B {\bf 74},  195308  (2006).

571: 

572: \bibitem{PRB_205307}

573: P. Brusheim and H.~Q. Xu, Phys. Rev. B {\bf 74},  205307  (2006).

574: 

575: \bibitem{PRB_212410}

576: D.~R. Penn and M.~D. Stiles, Phys. Rev. B {\bf 72},  212410

577: (2005).

578: 

579: \bibitem{PRL_206801}

580: J. Schliemann, D. Loss, and R.~M. Westervelt, Phys. Rev. Lett.

581: {\bf 94},

582:   206801  (2005).

583: 

584: \end{thebibliography}

585: 

586: \end{document}

587: