1: %\documentclass[prb,superscriptaddress,showpacs,floatfix,footinbib,fleqn,twocolumn,preprint]{revtex4}

2: 

3: 

4: \documentclass[prl,superscriptaddress,showpacs,floatfix,twocolumn]{revtex4}

5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

6: 

7: \usepackage{amssymb}

8: \usepackage{amsmath}

9: \usepackage{graphicx}

10: %\usepackage{subfigure}

11: \newcommand{\g}[1]{{\bf #1}}

12: %\usepackage[raggedright,raggedleft,ragged]{sidecap}

13: %\setcounter{MaxMatrixCols}{10}

14: %TCIDATA{OutputFilter=LATEX.DLL}

15: %TCIDATA{Version=5.00.0.2606}

16: %TCIDATA{<DEFANGED_META NAME="SaveForMode" CONTENT="1">}

17: %TCIDATA{BibliographyScheme=Manual}

18: %TCIDATA{<DEFANGED_META NAME="GraphicsSave" CONTENT="32">}

19: %TCIDATA{Language=American English}

20: 

21: \DeclareGraphicsRule{.tif}{png}{.png}{`convert #1 `dirname #1`/`basename #1 .tif`.png}

22: 

23: 

24: \begin{document}

25: 

26: \title{Onset of an Insulating Zero-Plateau Quantum Hall State in Graphene}

27: \author{E. Shimshoni}

28: \address{Department of Physics, Bar-Ilan University, Ramat-Gan 52900, Israel}

29: \address{Department of Mathematics–Physics, University of Haifa at Oranim, Tivon 36006, Israel}

30: \author{H.A. Fertig}

31: \address{Department of Physics, Indiana University, Bloomington, Indiana 47405}

32: \address{Department of Physics, Technion, Haifa 32000, Israel}

33: \author{G. Venketeswara Pai}

34: \address{Department of Physics, Technion, Haifa 32000, Israel}

35: \address{Department of Mathematics–Physics, University of Haifa at Oranim, Tivon 36006, Israel}

36: 

37: \pacs{71.10.Pm, 72.10.Fk, 73.20.-r, 73.23.-b, 73.43.Nq}

38: 

39: \begin{abstract}

40: We analyze the dissipative conductance of the zero-plateau quantum

41: Hall state appearing in undoped graphene in strong magnetic fields.

42: Charge transport in this state is assumed to be carried by a

43: magnetic domain wall, which forms by hybridization of two

44: counter--propagating edge states of opposing spin due to

45: interactions. The resulting non--chiral edge mode is a Luttinger

46: liquid of parameter $K$, which enters a gapped, perfectly conducting

47: state below a critical value $K_c\approx 1/2$. Backscattering in

48: this system involves spin flip, so that interaction with localized

49: magnetic moments generates a finite resistivity $R_{xx}$ via a

50: ``chiral Kondo effect''. At finite temperatures $T$, $R_{xx}(T)$

51: exhibits a crossover from metallic to insulating behavior as $K$ is

52: tuned across a threshold $K_{MI}$. For $T\rightarrow 0$, $R_{xx}$ in

53: the intermediate regime $K_{MI}<K<K_c$ is finite, but diverges as

54: $K$ approaches $K_c$. This model provides a natural interpretation

55: of recent experiments.

56: 

57: 

58: \end{abstract}

59: 

60: \date{\today}

61: \maketitle

62: 

63: \textit{Introduction}-- Graphene, a honeycomb network of carbon

64: atoms, is perhaps the most remarkable two-dimensional system to be

65: studied in the last few years \cite{review}.  This system differs

66: from the standard two dimensional electron gas (2DEG) in

67: supporting two Dirac points in its fermion spectrum which are at

68: the Fermi energy when the system is nominally undoped. In the

69: presence of a magnetic field, the Landau level spectrum differs

70: from the standard 2DEG in supporting positive and negative energy

71: (Landau level) states, as well as a zero energy (lowest) Landau

72: level in each of its two valleys.  In the earliest experiments,

73: this last property was understood to account for the absence of a

74: quantized Hall effect at filling factor $\nu=0$

75: \cite{zheng,zhang}.

76: % Moreover, with increasing filling, the first Hall

77: %conductance step turns out to be {\it half} the height of subsequent

78: %steps, confirming the unusual nature of the single particle spectrum

79: %in this system \cite{zhang}.

80: 

81: In stronger fields, and with higher quality samples, a plateau

82: does appear to emerge near $\nu=0$ which is likely associated with

83: a resolution of the spin-splitting of the lowest Landau level

84: \cite{zhang06,abanin07,ong07}. However, the behavior of the

85: longitudinal (dissipative) resistance ($R_{xx}$) at this filling

86: has been difficult to explain: different measurements have

87: demonstrated that it may either decrease \cite{abanin07} or

88: increase \cite{ong07} with falling temperature $T$ in samples that

89: appear to be quite similar \cite{novoselov_pr}. Moreover, in Ref.

90: \cite{ong07} the $\nu=0$ quantized Hall state exhibits a transition to

91: an insulator at a critical magnetic field $H_c$, where $R_{xx}(T\rightarrow 0)$

92: diverges. The scaling of $R_{xx}$

93: as a function of field $H$ below $H_c$ appears to signify a quantum phase

94: transition of the Kosterlitz--Thouless (KT)\cite{KT} type.

95: 

96: In this paper, we demonstrate that such a variety of transport

97: behaviors can be explained quite naturally if the system contains

98: local magnetic moments near its edge, which may occur due to

99: chemical passivation of dangling bonds or imperfections in the

100: lattice \cite{lehtinen04}. Interaction with such magnetic degrees of

101: freedom provides a mechanism for back--scattering in the primary

102: channel for charge conduction at $\nu=0$, a magnetic domain wall

103: (DW) \cite{brey06,abanin06,fertig06} which may carry

104: current that is unaffected by ordinary, {\it non--magnetic}

105: impurities. Transport properties of this mode are dramatically

106: altered by tuning of a Luttinger parameter $K$

107: \cite{giamarchi_book}, which in this system depends on the details

108: of the edge potential and on the magnetic field $H$.

109: 

110: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

111: \begin{figure}[t]

112: %\begin{center}

113: %\includegraphics[scale=.5]{phases.eps}

114: \includegraphics[scale=.5]{phase_d.eps}

115: \caption{(Color online). Phase diagram of the DW coupled to a

116: magnetic impurity via a fixed Kondo interaction. The solid blue

117: curve depicts the gap $\Delta_s(K)$. The dashed lines mark the

118: boundaries between distinct transport regimes (see text).}\label{phases}

119: %from a metallic behavior ($dR_{xx}/dT>0$) to insulating

120: %behavior ($dR_{xx}/dT<0$) at $T>\Delta_s$.}

121: %\end{center}

122: \end{figure}

123: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

124: 

125: Our main findings are summarized in the phase diagram depicted in

126: Fig. \ref{phases}. For $K>K_c$, the ideal system is a Luttinger

127: liquid, with $K$ large enough that a Kondo

128: impurity renders the system insulating at zero temperature

129: [$R_{xx} \rightarrow \infty$].

130: As $K$ is tuned below $K_c$, the impurity-free DW system

131: undergoes a quantum phase transition

132: to a perfect conducting state protected by a gap $\Delta_s(K)$

133: (solid line in Fig. 1). In this case a Kondo impurity generates

134: a {\it finite} resistance at zero temperature, approaching

135: this value from {\em below} ($dR_{xx}/dT < 0$).

136: The unusual thermal behavior of this

137: ``odd metal,''  results from a competition between

138: two relevant perturbations which separately would render the

139: system either perfectly conducting or truly insulating.

140: Finally, for $K<K_{MI}$, a Kondo impurity is an irrelevant

141: perturbation, yielding more typical metallic behavior,

142: $dR_{xx}/dT > 0$.

143: This variety of behavior is consistent with existing

144: published data \cite{abanin07,ong07}.

145: 

146: The remarkable edge--mode structure of the DW can already be seen

147: within a  non-interacting picture. The lowest Landau levels support

148: an electron-like and a hole-like mode for each spin, which cross

149: precisely at the Fermi level when Zeeman splitting is introduced

150: [see Fig. \ref{levels}] \cite{brey06,abanin06,mele05}. Accounting for

151: interactions within a Hartree-Fock description, the

152: counter--propagating states admix to form the DW \cite{fertig06}, with

153: an unusual collective mode. In particular,

154: %it is a broken symmetry state which (for sufficiently

155: %strong interactions) supports a gapless collective mode. Moreover,

156: due to the spin--charge coupling inherent in a state projected into

157: a single Landau level \cite{kane}, it supports topological {\em

158: charged} excitations which can propagate freely along the edge.

159: In addition, a spin gap may open in the collective excitation spectrum

160: when interactions are not too strong, with important consequences

161: for dissipative transport.

162: 

163: A prominent candidate for a dissipation mechanism in this system is

164: the scattering between counter--propagating edge states

165: \cite{kane_rev}, which one might expect to be enhanced

166: by the DW structure which strongly hybridizes forward and backward

167: propagating states. However, the requirement of spin conservation

168: forbids backscattering within a single DW by static impurities

169: \cite{abanin06}. In contrast, scattering from local {\it magnetic}

170: impurities {\it does} allow backscattering and generates

171: dissipation. As we show below, the coupling of a DW to a magnetic

172: impurity leads to an unusual ``chiral'' Kondo effect, since left and

173: right moving electrons each come in only a single spin flavor. The

174: resulting resistance at finite $T$ exhibits the variety of behaviors

175: summarized above, which we discuss below in more detail.

176: 

177: \textit{Model} -- As a first stage we derive an effective model for

178: a clean graphene sheet, with a straight edge along

179: the $y$ direction.

180: Our starting point for the analysis is to view

181: the system as a two-dimensional (2D) quantum Hall ferromagnet, in

182: which only the two Landau levels closest to zero energy are

183: included. These levels carry opposite spin, and

184: may be grouped together into a spinor whose

185: long-wavelength degrees of freedom are those of a Heisenberg

186: ferromagnet, with Hamiltonian \cite{kane,fertig06}

187: \begin{equation}

188: H_{2D}^{0}= \int d^2r \left\{{J \over 2} \sum_{\mu}|\partial_{\mu}

189: {\bf S}({\bf r})|^2+\Delta(x)S_z({\bf r})\right\}.

190: \label{2Dferr}

191: \end{equation}

192: Here $J$ represents an exchange stiffness due to interactions.

193: $\Delta(x)$ encodes the combined effects of the Zeeman interaction

194: and the edge potential (dictated by the boundary conditions and electrostatic

195: environment \cite{silvestrov}): it increases monotonically from the bulk

196: (negative) Zeeman energy for $x$ well inside the bulk to a positive

197: value at the edge, leading to the energy crossing  in the

198: non-interacting spectrum illustrated in Fig. \ref{levels}. Initially

199: treating the system classically, we parameterize the spin field by

200: ${\bf S}({\bf r}) = S {\bf \Omega}({\bf r})$, with $S=1/2$ and ${\bf

201: \Omega}= [\sin\psi({\bf r})\cos\chi({\bf

202: r}),\sin\psi({\bf r})\sin\chi({\bf r}),\cos\psi({\bf

203: r})]$. As shown in \cite{fertig06}, the form of $\Delta(x)$ in

204: the above energy functional dictates a minimum energy configuration

205: where $\chi$ is spatially constant, and $\psi({\bf

206: r})=\psi_0(x)$ exhibits a non-trivial topology: $\psi_0(x

207: \rightarrow -\infty) \rightarrow 0$, $\psi_0(x \rightarrow {\rm

208: edge}) \rightarrow \pi$. A domain wall (DW) is thus formed parallel

209: to the edge. The ground state energy is independent of $\chi$,

210: implying a broken symmetry and an associated collective mode

211: propagating with a wavevector $k_y$ along the DW \cite{falko}.

212: 

213: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

214: \begin{figure}[t]

215: %\begin{center}

216: \includegraphics[scale=.35]{levels.eps}

217: \caption{(Color online). Non--interacting energy levels near the

218: edge. Solid black lines are spin up states, dashed red lines are

219: spin down states. The dotted line denotes the chemical potential

220: at zero energy.} \label{levels}

221: %\end{center}

222: \end{figure}

223: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

224: 

225: %We next construct a low--energy Hamiltonian which accounts for

226: %distortions about the DW configuration, and project the resulting

227: %energy functional onto effectively one-dimensional (1D) spin--wave

228: %modes of well--defined $k_y$. The resulting 1D model can be cast in

229: %terms of an antiferromagnetic XXZ spin-1/2 chain Hamiltonian, in a

230: %discretized space where the effective lattice spacing is set by the

231: %magnetic length $\ell=\sqrt{\hbar c/eH}$:

232: 

233: 

234: The 2D semiclassical theory may be projected into its low-energy

235: subspace, containing the one-dimensional (1D) mode.  This theory is

236: identical in form to a semiclassical treatment of an XXZ antiferromagnetic

237: spin-1/2 chain;

238: an appropriate choice of coefficients for the latter allows the

239: models to coincide

240: at the semiclassical level.  As the spin chain

241: incorporates the spin-1/2 nature of the system,

242: we work with this model, with lattice spacing

243: set by the

244: magnetic length $\ell=\sqrt{\hbar c/eH}$:

245: \begin{equation}

246: H_{DW}=J_{xy}\sum_{j}\left[ S^x_{j} S^x_{j+1} + S^y_{j} S^y_{j+1} \right] +

247: J_z\sum_{j}S^z_{j} S^z_{j+1}\; . \label{XXZ}

248: \end{equation}

249: Here $J_{xy}$ is related to the exchange energy $J$ via

250: $J_{xy}=JS^2{\cal N}$, with ${\cal N}=\sum_x \sin^2

251: \psi_0(x)$

252: %\begin{equation}

253: %J_{xy}=JS^2{\cal N}\; ,\quad {\cal N}=\sum_x \sin^2 \varpsi_0(x)

254: % \label{Jxy}

255: %\end{equation}

256: ($\psi_0(x)$ the classical DW texture), and $J_z$ depends on

257: the edge potential: $J_{z}={1 \over {{\cal N}2S}} \sum_x \Delta(x)

258: \sin^2 \psi_0(x)[1-\cos \psi_0(x)]$.

259: % $\Delta(x)$:

260: %\begin{equation}

261: %J_{z}={1 \over {{\cal N}2S}} \sum_x \Delta(x) \sin^2

262: %\varpsi_0(x)[1-\cos \varpsi_0(x)]\; .

263: % \label{Jz}

264: %\end{equation}

265: Using a standard Bosonization scheme for the spin operators in the

266: continuum limit \cite{giamarchi_book},

267: \begin{eqnarray}

268: S_\pm(y)=\frac{e^{\mp

269: i\theta}}{\sqrt{2\pi\alpha}}[(-)^y+\cos(2\phi)]\nonumber\\

270: S_z(y)=-\frac{\partial_y

271: \phi}{\pi}+\frac{(-)^y}{\pi\alpha}\cos(2\phi)\; ,\label{boson}

272: \end{eqnarray}

273: we map $H_{DW}$ to a (single--flavored) Luttinger model with a

274: sine--Gordon correction. In units where $\hbar=1$,

275: \begin{equation}

276: H_{DW}=\int \frac{dy}{2 \pi} \left\{ uK (\pi\Pi) ^2+\frac{u}{K}

277: (\partial_y \phi)^2 -\frac{J_z} {\pi \alpha^2} \cos(4\phi)\right\}

278: \label{LL}

279: \end{equation}

280: where the Boson field $\phi(y)$ and $\Pi(y)={\partial_y\theta\over

281: \pi}$ are canonically conjugate.

282: % obey the commutation relations

283: %$[\phi(y),\Pi(y^\prime)]=i\delta(y^\prime-y)$.

284: The spin--wave velocity $u$ and Luttinger parameter $K$ are related

285: to the XXZ parameters by

286: \begin{equation}

287: uK=J_{xy}\; ,\quad \frac{u}{K}=J_{xy}+\frac{4J_z}{\pi}\; ,

288: \label{LLdef}

289: \end{equation}

290: and the short distance cutoff $\alpha\sim\ell$.

291: 

292: The effective model for the ideal DW system [Eq. (\ref{LL})]

293: exhibits a quantum phase transition upon tuning of

294: the parameter $K$ across a critical point $K_c$, from a Luttinger liquid (LL) phase at $K>K_c$

295: [where the $\cos(4\phi)$ term is irrelevant] to an ordered [gapped] phase below $K_c$.

296: In terms of the spin fields, the latter phase is characterized

297: by ordering of the field $S_z$, and the

298: opening of a gap $\Delta_s$ to spin-flip excitations. The

299: critical behavior near $K_c$ can be derived from a perturbative

300: renormalization group (RG) analysis \cite{giamarchi_book}, yielding

301: a $KT$--transition at $K_c\sim 1/2$. As $K$ approaches $K_c$ from below,

302: the gap tends to vanish according to the scaling law

303: \begin{equation}

304: \Delta_s(K) \simeq

305: \frac{u}{\alpha}\exp\left\{{-\frac{C}{\sqrt{K_c-K}}}\right\}

306: \label{Delta_s}

307: \end{equation}

308: with $C$ a constant of order unity [see Fig. \ref{phases}].

309: 

310: We now assume that a local moment at the origin (represented by a

311: spin-1/2 operator $\vec{\sigma}$) couples to the original 2D spin

312: system via an exchange interaction $\tilde{J} {\vec S} ({\bf

313: r}=0)\cdot {\vec \sigma}$. In order to project this into the

314: low--energy subspace, we express it in a rotated basis where the

315: local moment is coupled to fluctuations about the ground state DW

316: configuration. This yields an effective Kondo coupling to the 1D

317: spin chain operators:

318: \begin{equation}

319: H_K=\frac{J_K^{xy}}{2} (S_+(0)\sigma_- + S_-(0)\sigma_+) + J_K^z

320: S_z(0)\sigma_z\; . \label{kondo}

321: \end{equation}

322: The resulting Hamiltonian for the effective 1D system then takes the

323: form $H=H_0+H_{K}$, with

324: \begin{equation}

325: H_0=H_{DW}+\tilde{E}_z\sigma_z\; .\label{H0}

326: \end{equation}

327: Note that $\tilde{E}_z$ is an effective Zeeman energy of the

328: localized spin, associated with a {\it local} magnetic field which

329: depends on its position within the width of the DW.

330: 

331: \textit{Transport} -- To derive the transport properties of the

332: system, we first express the charge current operator in terms of the

333: DW fields. Using the standard definition $j_e=-{\delta

334: H_{DW}\over\delta A}$, with $A(y,t)$ a vector potential, we find

335: \begin{equation}

336: j_e(y,t)=2eJ_z S_z(y,t)\; .\label{je}

337: \end{equation}

338: This remarkable relation between spin and charge current encodes a

339: unique property of the DW: in the absence of spin-flip processes

340: (i.e. for $J_K^{xy}=0$), current is conserved and the DW behaves as

341: a perfect conducting channel. The distinction between its two phases

342: becomes apparent when spin impurities are added: in the gapped phase,

343: spin-flip and hence back-scattering excitations are suppressed for $T<\Delta_s$.

344: Its `perfect conduction' is hence more robust.

345: Indeed, $H_{DW}$ [Eq. (\ref{LL})] is

346: analogous to a low--energy model of a quasi 1D {\em superconducting}

347: system, where two parallel wires are Josephson coupled. The field

348: $\phi$ describes the relative superconducting phase, $u/K$ the

349: superfluid stiffness, $uK$ the charging energy \cite{1dSC_rev}, and

350: the last term in $H_{DW}$ is the inter-wire Josephson coupling.

351: There as well, a local perturbation analogous to

352: $H_K$ [Eq. (\ref{kondo})] is required to generate finite dissipation

353: via a phase--slip mechanism. This occurs when a Josephson vortex

354: (i.e., out-of-phase current configuration) penetrates into the bulk

355: of the double-wire system. In the gapped phase, where Josephson

356: coupling is relevant, such processes are suppressed beyond the

357: Josephson length ($\sim 1/\Delta_s$).

358: 

359: We next consider a finite but small Kondo interaction and

360: $T>\Delta_s$, and calculate the dissipative part of the electric

361: conductance to leading order in $H_K$:

362: \begin{equation}

363: \delta G=\lim_{\omega\rightarrow

364: 0}\frac{-1}{L\omega}\int_{-L/2}^{L/2}dy \Im

365: \left\{\chi(y,y^\prime;\omega)\right\}\; , \label{Kubo}

366: \end{equation}

367: where $\chi=\langle j_e(y);j_e(y^\prime)\rangle^0_{\omega}$ is the

368: retarded correlation function evaluated with respect to $H_{0}$

369: (see, e.g., \cite{giamarchi_book}), and $L$ the length of the

370: system, assumed finite. Note that $\delta G$ is a {\it negative}

371: correction to the ideal conductance, so that $|\delta G|$ is

372: proportional to the backscattering rate and consequently to the

373: longitudinal resistance $R_{xx}$ \cite{Gdef}. Using the bosonized

374: representation Eqs. (\ref{boson}),(\ref{LL}) and the expression for

375: the local spin correlator (for $t>0$)

376: \begin{equation}

377: \langle

378: \sigma_+(t)\sigma_-(0)\rangle^0=e^{i\tilde{E}_zt}f\left(\frac{\tilde{E}_z}{T}\right)\;

379: ,\quad f(z)\equiv \frac{1}{e^z+1}

380: \end{equation}

381: we obtain the $T$--dependent conductance

382: \begin{equation}

383: \delta G(T)\approx -\frac{e^2}{h}g_K^2 \left(\frac{\pi\alpha

384: T}{u}\right)^{\kappa-2}\mathcal{G}\left(\frac{\tilde{E}_z}{T},\frac{\alpha

385: T}{u}\right)\; .\label{deltaG}

386: \end{equation}

387: Here $g_K\propto J_K^{xy}$ is dimensionless, $\kappa=1/2K$ and

388: \begin{eqnarray}

389: \mathcal{G}(z,\epsilon)&\equiv &

390: \sin\left(\frac{\pi\kappa}{2}\right)\int_\epsilon^\infty

391: dt\frac{t\cos(zt)}{[\sinh(\pi t)]^\kappa}  \nonumber\\

392: &-& 2m(z)\cos\left(\frac{\pi\kappa}{2}\right)\int_\epsilon^\infty

393: dt\frac{t\sin(zt)}{[\sinh(\pi t)]^\kappa}\label{Gdim}

394: \end{eqnarray}

395: with $m(z)=[f(z)-f(-z)]$ the local spin magnetization.

396: 

397: In the finite $T$ regime where $T\gg \tilde{E}_z$, the leading

398: $T$--dependence of $\delta G(T)$ inferred from Eqs. (\ref{deltaG}),

399: (\ref{Gdim}) is a power--law, which changes sign at $\kappa=2$:

400: \begin{equation}

401: \delta G(T)\sim -T^{\kappa-2}\; .\label{GvsT}

402: \end{equation}

403: The conductance therefore exhibits a crossover from a metallic

404: behavior ($dG/dT<0$) for $\kappa>2$ to an insulating behavior

405: ($dG/dT>0$) for $\kappa<2$. This crossover

406: would be a true metal--insulator (MI) as $K$ is tuned

407: through $K_{MI} \approx 1/4$, as depicted

408: in Fig. (\ref{phases}), if not for the existence of the gap.

409: Note that $K_{MI}$ is perturbatively

410: renormalized to values below 1/4 by $J_K^{xy}$ and $J_K^{z}$, so

411: that one always finds $K_{MI}<K_c$.

412: The MI crossover then always occurs within the gapped

413: %SDW

414: phase. Thus the power--law behavior of $G(T)$ is

415: restricted to finite $T\gg \Delta_s$.

416: 

417: We now focus on the intermediate regime $K_{MI}<K<K_c$. Assuming

418: further that $\tilde{E}_z\sim 0$, we consider the low $T$ limit

419: $T\ll\Delta_s$ where the field $\phi$ is ordered, while $\theta$ is

420: strongly fluctuating. The power--law decay of the correlation

421: function $\langle e^{i\theta(t)}e^{-i\theta(0)}\rangle^0$

422: (characteristic of the LL) is modified by a rapidly oscillating

423: exponential factor:

424: \begin{equation}

425: \langle e^{i\theta(t)}e^{-i\theta(0)}\rangle^0\approx

426: \left(\frac{\alpha}{ut}\right)^\kappa

427: \exp\left\{-\frac{\kappa\Delta_s

428: t^2}{\tau_\phi}+i\kappa\left(\frac{\pi}{2}+\frac{t}{\tau_\phi}\right)\right\}

429: \label{lowTcorr}

430: \end{equation}

431: where the decoherence rate $\tau_\phi^{-1}\equiv\Delta_s^2L/2\pi u$

432: corresponds to a `condensation energy' in the superconductor

433: analogue. The resulting $\delta G$ saturates as $T\rightarrow 0$ and

434: becomes

435: \begin{equation}

436: \delta G\sim -\tau_\phi^{2-\kappa}\; .

437: \end{equation}

438: It then follows from the definition of $\tau_\phi$ and

439: Eq. (\ref{Delta_s}) that the resistance {\it diverges} when $K$

440: approaches $K_c$ as

441: \begin{equation}

442: R_{xx}\sim -\delta G\sim

443: \exp\left\{\frac{2C^\prime}{\sqrt{K_c-K}}\right\}\; .

444: \label{RvsK}

445: \end{equation}

446: Here $C^\prime=C(2-\kappa)$; for $\kappa\sim 1/2K_c\sim 1$,

447: $C^\prime\sim C$.

448: 

449: We finally comment on the possible relation of our results to the

450: experimentally observed $R_{xx}$ in the

451: zero--plateau state. Provided $K$ is tunable by a magnetic field $H$, the system

452: is shown to exhibit diverse transport properties, which could explain the

453: apparent discrepancy between the

454: data of Refs. \cite{abanin07} and  \cite{ong07}.

455: Furthermore, we find a quantum critical point associated with the

456: closing of a spin--gap, manifested by a divergence of

457: $R_{xx}(T\rightarrow 0)$ according to a scaling law characteristic of

458: a KT--transition (in $1+1$--dimensions). The latter is remarkably

459: reminiscent of the peculiar field-tuned transition to an insulator observed

460: in \cite{ong07}.

461: %Direct comparison of our theory to

462: %experiments involves an estimate of $K$ and its dependence on

463: %the field $H$.

464: A rough estimate of $K$ in terms of physical parameters yields $K\sim \sqrt{\frac{e^2}{\epsilon\ell \Delta_{av}}}$,

465: with $\Delta_{av}$ an average edge potential strongly dependent on sample details.

466: In particular, when $\Delta_{av}$ is dominated by electrostatic effects induced by the gate voltage

467: \cite{silvestrov}, it is typically larger than the exchange energy $\frac{e^2}{\epsilon\ell}$.

468: This would yield $K<1$ and monotonically increasing with $H$, as assumed by our theory.

469: %the analysis is complicated

470: %by competition between the different factors.

471: 

472: Another assumption of our theory which requires justification is the

473: neglect of $\tilde{E}_z$. A finite $\tilde{E}_z$ introduces a cutoff

474: which competes with the gap $\Delta_s$, and a finite magnetization

475: of the local moment which reduces its contribution to

476: backscattering. We note, however, that the realistic system

477: presumably contains a multitude of spin impurities with randomly

478: distributed local parameters. The dissipative resistance is then

479: dominated by the moments with minimal $\tilde{E}_z$, and our

480: analysis applies as long as it is smaller than all other energy

481: scales.

482: 

483: To summarize, we study a model for charge transport carried by

484: fluctuations of a DW propagating along the edge of a graphene sample

485: in the $\nu=0$ quantized Hall state. Backscattering induced by

486: localized magnetic impurities at the edge provides a dissipation

487: mechanism. Its competition with a spin--gap generated in the ideal DW

488: may possibly explain the rich variety of

489: conductance properties observed in recent experiments.

490: 

491: \bigskip

492: \begin{acknowledgments}

493: We gratefully acknowledge discussions with L. Brey, K.S. Novoselov

494: and N. P. Ong. This project was supported in part by the US NSF

495: under Grant No. DMR-0704033 (H.A.F.) and the German--Israeli

496: Foundation for Scientific Research and Development (E.S. and G.V.P.).

497: \end{acknowledgments}

498: 

499: \begin{thebibliography}{99}

500: 

501: \bibitem{review} A.H. Castro Neto

502: et al., arXive:0709.1163; V.P. Gusynin, S.G. Sharapov, and J.P.

503: Carbotte, Int. J. Mod. Phys. {\bf 27}, 4611 (2007).

504: 

505: \bibitem{zheng} Y. Zheng and T. Ando, Phys. Rev. B {\bf 65}, 245420 (2002);

506: V.P. Gusynin and S.G. Sharapov, Phys. Rev. Lett. {\bf 95}, 146801 (2005);

507: A.H.Castro Neto, F. Guinea and N.M.R. Peres, Phys. Rev. B {\bf 73}, 205408 (2006).

508: 

509: \bibitem{zhang}  K.S. Novoselov et al., Nature {\bf 438}, 197 (2005),

510: Y. Zhang et al., Nature {\bf 438}, 201 (2005).

511: 

512: \bibitem{zhang06} Y. Zhang et al., Phys. Rev. Lett. {\bf 96}, 136806 (2006).

513: 

514: \bibitem{abanin07} D.A. Abanin et al., Phys. Rev. Lett. {\bf 98}, 196806 (2007).

515: 

516: \bibitem{ong07} J.G. Checkelsky, L. Li, and N. P. Ong, Phys. Rev. Lett. {\bf 100}, 206801 (2008);

517: J.G. Checkelsky, L. Li, and N. P. Ong, Phys. Rev. B {\bf 79}, 115434 (2009).

518: 

519: \bibitem{novoselov_pr} K.S. Novoselov, private communication.

520: 

521: \bibitem{KT} J. M. Kosterlitz and D. J. Thouless, J. Phys. C {\bf 6}, 1181

522: (1973); J. M. Kosterlitz, J. Phys. C {\bf 7}, 1046 (1974).

523: 

524: \bibitem{lehtinen04} See, for example, B. Uchoa et al., Phys. Rev. Lett. {\bf 101}, 026805 (2008)

525: and references therein.

526: 

527: \bibitem{brey06} L. Brey, Bull. of the Am. Phys. Soc. {\bf 51}, 459 (2006).

528: 

529: \bibitem{abanin06} D.A. Abanin, P. A. Lee, and L. S. Levitov, Phys. Rev. Lett. {\bf 96}, 176803 (2006).

530: 

531: \bibitem{fertig06} H.A. Fertig and L. Brey,

532: Phys. Rev. Lett. {\bf 97}, 116805 (2006).

533: 

534: \bibitem{giamarchi_book} T. Giamarchi, {\it Quantum Physics in One Dimension}, (Oxford, New York, 2004).

535: 

536: %\bibitem{com} The spectrum illustrated in

537: %Fig. \ref{levels}, most easily obtained for armchair edges

538: %\cite{brey06,abanin06}, should arise generically when the bulk is

539: %partially spin-polarized and assuming particle-hole symmetry.

540: 

541: \bibitem{mele05}

542: Such helical edge states may also be induced by spin-orbit interaction; see,

543: e.g., C.L. Kane and E.J. Mele, Phys. Rev. Lett. {\bf 95}, 226801 (2005).

544: 

545: \bibitem{kane} D. H. Lee and C. L. Kane, Phys. Rev. Lett. {\bf 64},

546: 1313 (1990); see also S. M. Girvin and A. H. MacDonald in {\it

547: Perspectives in Quantum Hall Effects}, S. Das Sarma and A. Pinczuk,

548: Eds. (Wiley, New York, 1997).

549: 

550: \bibitem{kane_rev} For a review, see article by C. L. Kane and M.P.A. Fisher,

551: in {\it Perspectives in Quantum Hall Effects}, S. Das Sarma and A. Pinczuk, Eds.

552: (Wiley, New York, 1997).

553: 

554: \bibitem{silvestrov}

555: P. G. Silvestrov and K. B. Efetov, Phys. Rev. B {\bf 77}, 155436 (2008).

556: 

557: \bibitem{falko}

558: See also V. I. Fal'ko and S. V. Iordanskii, Phys. Rev. Lett. {\bf 82}, 402 (1999).

559: 

560: %\bibitem{furusaki94} A. Furusaki and N. Nagaosa, Phys. Rev. Lett. {\bf 72}, 892 (1994).

561: 

562: \bibitem{1dSC_rev} See, e.g., review by K.Yu. Arutyunov, D. S. Golubev and A. D.

563: Zaikin, Phys. Rep. {\bf 464}, 1 (2008).

564: 

565: \bibitem{Gdef} Note that a finite size perfect conductor may also have resistance

566: associated with its leads \cite{giamarchi_book,abanin06}.

567: 

568: \end{thebibliography}

569: \end{document}

570: 

571: 

572: