1: %% ****** Start of file authguide.tex ****** %
2: %%
3: %% This file is part of the APS files in the REVTeX 4 distribution.
4: %% Version 4.0 of REVTeX, August, 2001
5: %%
6: %% Copyright (c) 2000,2001 The American Physical Society.
7: %%
8: %% See the REVTeX 4 README file for restrictions and more information.
9: %%
10: \listfiles
11: \documentclass[%
12: %prl%
13: %,preprint%
14: ,twocolumn%
15: ,secnumarabic%
16: %,tightenlines%
17: ,amssymb,nobibnotes, prl, aps]{revtex4}
18: %\usepackage{acrofont}%NOTE: Comment out this line for the release version!
19: %\usepackage{docs}%
20: \usepackage{bm,epsf}%
21: %
22: \newcommand{\bx}{{\bf x}}
23: \newcommand{\bv}{{\bf v}}
24: \newcommand{\ba}{{\bf a}}
25: \newcommand{\bk}{{\bf k}}
26: \newcommand{\bq}{{\bf q}}
27: \newcommand{\bp}{{\bf p}}
28: \newcommand{\bA}{{\bf A}}
29: \newcommand{\bB}{{\bf B}}
30: \newcommand{\eps}{{\epsilon}}
31: \newcommand{\kxy}{{\kappa_{xy}}}
32: \newcommand{\sxy}{{\sigma^{spin}_{xy}}}
33: \newcommand{\sone}{{\sigma_1}}
34: \newcommand{\stwo}{{\sigma_2}}
35: \newcommand{\sthree}{{\sigma_3}}
36: %
37: %\usepackage[colorlinks=true,linkcolor=blue]{hyperref}%
38: %\nofiles
39: \expandafter\ifx\csname package@font\endcsname\relax\else
40: \expandafter\expandafter
41: \expandafter\usepackage
42: \expandafter\expandafter
43: \expandafter{\csname package@font\endcsname}%
44: \fi
45:
46: \begin{document} \draft
47:
48: \title{Thermal Hall Conductivity of High Temperature Superconductors: \\
49: Quantization and Scaling}%
50:
51: \author{Oskar Vafek and Zlatko Te\v sanovi\'c}
52: \email{ovafek@pha.jhu.edu, zbt@pha.jhu.edu}
53: \affiliation{Department of Physics and Astronomy, Johns Hopkins University,
54: Baltimore, MD 21218
55: \\ {\rm(\today)}
56: }
57: \begin{abstract}
58: \medskip
59: Presented is the theory of thermal Hall conductivity $\kappa_{xy}$ in the vortex state
60: of high temperature cuprate superconductors in the clean limit.
61: We show that $\lim_{T \to 0} \kappa_{xy}/T$
62: is a {\em staircase} function of $1/B$ with an envelope that scales as
63: $1/B$. The relation to the experiments is discussed.
64: \end{abstract}
65:
66: \maketitle
67:
68: Fifteen years after the discovery of high temperature superconductivity
69: in cuprates \cite{BednorzMuller}, the search for the ``final''
70: theory continues unabated.
71: Nevertheless, certain important clues about the elements that such a
72: theory must contain appear to be established. One of the most significant
73: results of the past few years is that, in the superconducting state, cuprates
74: seem to be well described by the familiar BCS-type formalism with
75: Cooper pairs binding in the d-wave channel \cite{harlingen}.
76: This is a far from trivial observation in the materials known to exhibit
77: strong correlations.
78: Soon after the original discovery, Anderson \cite{anderson1} identified
79: one of the key features of the new superconductivity: the materials
80: are basically two-dimensional with most of the action relevant
81: for superconductivity taking place
82: in the $CuO_2$ planes. The reduced dimensionality and
83: the presence of nodal points on an otherwise gapped Fermi surface
84: governs the low temperature properties of
85: the high-$T_c$ superconductors (HTS).
86: Corroborating evidence
87: for the BCS character of the fermionic excitations comes from diverse
88: spectroscopic and transport experiments \cite{walker, ding, taillefert1, millis}.
89:
90: Especially informative probe of the quasiparticle dynamics is the thermal Hall
91: conductivity, $\kxy$ \cite{ong}, particularly since phonons,
92: which usually contribute to the longitudinal
93: thermal conductivity, $\kappa_{xx}$, do not contribute to $\kxy$ by the virtue
94: of their being electrically neutral.
95: In addition, from the theoretical point of view, in the leading ``nodal'' approximation,
96: $\kxy$ vanishes \cite{simonlee,jinwu,ashvin} and one must consider curvature terms.
97: For these reasons a special opportunity emerges to scrutinize,
98: both theoretically and
99: experimentally, the extent to which the fermionic quasiparticles
100: in HTS exhibit the BCS-type behavior.
101:
102: Our starting point is a 2-dimensional
103: Bogoliubov-de~Gennes Hamiltonian in the vortex state
104: \begin{equation}\label{bdg}
105: H_{BdG}=\sthree\frac{\left(\bp-\sthree\frac{e}{c}\bA\right)^2}{2m^*}-\sthree\mu+
106: \sone e^{-i\sthree\phi/2}\Delta(\bp) e^{-i\sthree\phi/2}
107: \end{equation}
108: where ${\bf A}({\bf x})$ is the vector
109: potential associated with the uniform external magnetic field $\bB$,
110: $\bp$ is a momentum operator, and $\sigma$'s are Pauli matrices
111: operating in the Nambu space.
112: The vortex phase fields satisfy
113: $\nabla\times\nabla\phi ({\bx}) = 2\pi\hat{z}\sum_i
114: \delta ({\bx} -{\bx}_i)$ with $\bx_i$ denoting the vortex positions
115: and $\delta ({\bx} -{\bx}_i) $ being a 2D Dirac delta function. For a $d_{x^2-y^2}$-wave
116: superconductor $\Delta(\bp)=\frac{\Delta_0}{p_F^2}(p_x^2-p_y^2)$.
117:
118: Following Franz and Tesanovic \cite{ft} (see also \cite{luca,ashvin,vmt}),
119: we divide the vortices into two sets $A$ and $B$, positioned at $\{ \bx^A_i\}$ and
120: $\{ \bx^B_i\}$ respectively (see Fig. \ref{transf}) and define two phase fields
121: $\phi^A(\bx)$ and $\phi^B(\bx)$ such that
122: $\nabla\times\nabla\phi^{\alpha} ({\bx}) = 2\pi\hat{z}\sum_i\delta ({\bx} -{\bx}^{\alpha}_i),$
123: $\alpha=A,B$.
124: The Hamiltonian (\ref{bdg}) is then transformed into \cite{ft}
125: \begin{equation}\label{bdgsg}
126: {\cal H}=U^{\dagger}H_{BdG} U=
127: \sthree\frac{\left(\bp-\sthree\bv+\ba\right)^2}{2m^*}-\sthree\mu+
128: \sone \Delta(\bp+\ba)
129: \end{equation}
130: using the unitary operator
131: \begin{equation}
132: U=e^{i\chi+i\sthree\Phi}
133: \end{equation}
134: where $\chi=\frac{1}{2}(\phi_A-\phi_B)$ and $\Phi=\frac{1}{2}(\phi_A+\phi_B)=\frac{1}{2}\phi$.
135: In Eq. (\ref{bdgsg}), $\bv=\frac{1}{2}\nabla\phi-\frac{e}{c}\bA$
136: is the superfluid velocity and $\ba=\frac{1}{2}(\nabla\phi_A-\nabla\phi_B)$
137: is the Berry gauge field
138: which imposes the condition that a quasiparticle wavepacket
139: encircling an $\frac{hc}{2e}$ vortex must pick up an overall minus sign.
140:
141: In addition to keeping the wavefunctions single-valued, the advantage of
142: this ``bipartite'' singular gauge transformation is that for a periodic
143: arrangement of vortices (Abrikosov vortex lattice) the resulting
144: Hamiltonian (\ref{bdgsg}) is itself periodic.
145: This is not true of Hamiltonian (\ref{bdg}).
146: Therefore, the eigenstates of (\ref{bdgsg}) can be expressed through
147: the familiar Bloch waves $\psi_{n\bk}(\bx)=e^{i\bk\cdot\bx}u_{n\bk}(\bx)$ labeled by one discrete quantum number,
148: the band index $n$, and two continuous quantum numbers, the vortex crystal
149: momenta $k_x$, $k_y$ that range in the first Brillouin zone.
150: Note that $u_{n\bk}(\bx)$ is not periodic with the periodicity of the
151: vortex lattice, but with the periodicity of the vortex sublattice containing a unit of
152: {\em electronic} flux $\frac{hc}{e}$ (see Fig. \ref{transf}).
153: %
154: \begin{figure}[t]
155: \epsfxsize=4.5cm
156: \hfil\epsfbox{fig1.eps}\hfill
157: \caption{Example of $A$ and $B$ sublattices for the square vortex arrangement.}
158: \label{transf}
159: \end{figure}
160: %
161:
162: By extracting the plane wave part of the Bloch states we can define a
163: Hamiltonian
164: \begin{equation}\label{bdgk}
165: {\cal H}(\bk)=\sthree\frac{\left(\bp+\bk-\sthree\bv+\ba\right)^2}{2m^*}-\sthree\mu+
166: \sone \Delta(\bp+\bk+\ba)
167: \end{equation}
168: where ${\cal H}(\bk)=e^{-i\bk\cdot\bx}{\cal H}e^{i\bk\cdot\bx}$.
169: We now wish to point out an important feature of ${\cal H}(\bk)$ in the case when the
170: vortex lattice has an inversion symmetry.
171: As shown in Ref.\cite{vmt} an operator
172: ${\cal P}=i\stwo {\cal C}{\cal I}$, where ${\cal C}$ represents charge
173: conjugation and ${\cal I}$ takes $\bx$ into $-\bx$, with the following property
174: \begin{equation}
175: {\cal P}^{-1} {\cal H}(\bk) {\cal P}= -{\cal H}(\bk);\;\;
176: \langle u_{n\bk} |{\cal P}| u_{n\bk} \rangle=0.
177: \end{equation}
178: This means that for every energy
179: band $\eps_{n\bk}$ there is another $\eps_{m\bk}$ such that $\eps_{n\bk}=-\eps_{m\bk}$, i.e.
180: the band spectrum has reflection symmetry about the Fermi level and
181: if a band were to cross the Fermi level ($\eps_{n\bq}=0$ for some $\bq$)
182: then there would have to be a twofold degeneracy at $\bq$. However, by the von~Neumann-Wigner
183: ``non-crossing'' theorem \cite{noncrossing}, the degeneracy unrelated to symmetry
184: cannot happen for a time reversal breaking Hamiltonian which depends
185: only on two continuous parameters. Thus, the quasiparticle spectrum of the $d_{x^2-y^2}$
186: superconductor in the vortex state is gapped by a (direct) gap $\Delta_m \ll \Delta_0$
187: unless a third parameter is fine tuned,
188: e.g. the band crossing at the Fermi
189: level can be achieved by fine-tuning the chemical potential $\mu$.
190:
191: If we rescale all coordinates by the magnetic length
192: $l=\sqrt{c/eB}$ (see Fig. 1), $\bx \rightarrow l\bx$, then
193: \begin{equation}\label{operatorscaling}
194: {\cal H}(\bk,l,\mu,\Delta_0) \rightarrow \frac{1}{l^2}f(l\bk,l^2\mu,\frac{\Delta_0}{p_F^2}),
195: \end{equation}
196: where $f$ is a universal (operator) function.
197: Consequently, the small band gap $\Delta_m$ at the Fermi level,
198: which is equal to the minimum of the lowest positive
199: eigen-energy of ${\cal H}(\bk,l,\mu)$ with respect to $\bk$,
200: is also a universal function that can be written as
201: \begin{equation}\label{scaling}
202: \Delta_m(B,\mu,\Delta_0)=B \times F\left(\frac{\mu}{B},\frac{\Delta_0}{p_F^2}\right).
203: \end{equation}
204: Thus, in order to determine the dependence of $\Delta_m$ on the magnetic field $B$,
205: all we need to know is
206: the dependence of $\Delta_m$ on $\mu$ for some fixed B and $\frac{\Delta_0}{p_F^2}$,
207: and then use the scaling relation (\ref{scaling}).
208: This dependence is determined numerically by diagonalizing
209: a tight-binding version \cite{vmt} of the Hamiltonian (\ref{bdgk})
210: close to the bottom of the tight-binding band where we expect to
211: recover the continuum theory (\ref{bdgk}).
212: The numerical calculations confirm the scaling (\ref{scaling}) which gives us
213: confidence in relating the continuum results to the tight-binding approximation.
214: As shown in Fig. \ref{scalingplot}, for a fixed value of
215: $\frac{t}{\Delta_0}=\alpha_D=10$, where $t$
216: is the tight-binding hopping constant,
217: $\frac{1}{B}\Delta_m(\frac{\mu}{B})$ is a saw-tooth function with a period of approximately $4\pi$.
218: %
219: \begin{figure}[t]
220: \epsfxsize=8.5cm
221: \hfil\epsfbox{fig2.eps}\hfill
222: \caption{Scaling of the field induced quasiparticle gap $\Delta_m$ (open circles) with
223: the magnetic field and the chemical potential $\mu$ (measured from the bottom of the
224: tight-binding band). The quantized values of
225: the $\sxy$ are also shown (solid line).
226: Magnetic length, $l$, is in units of the tight-binding lattice
227: constant $a$ and energies are in units of the hoping energy $t$; the results are
228: for a square vortex lattice with $\Delta_0=0.1t$. The dashed line is the envelope
229: scaling function.}
230: %
231: \label{scalingplot}
232: \end{figure}
233:
234: In Ref. \cite{vmt} it was demonstrated that the gapped quasiparticle
235: spectrum leads to the quantization of the spin Hall conductivity
236: $\sxy$ (see also Ref. \cite{ashvin}),
237: which is in turn related to the thermal Hall conductivity, $\kxy$, via
238: the Wiedemann-Franz law
239: \begin{equation}\label{wf}
240: \lim_{T \to 0}\frac{\kxy}{T}=\frac{4\pi^2}{3}\left(\frac{k_B}{\hbar}\right)^2\sxy(B).
241: \end{equation}
242: Because the quantization of $\sxy$ is topological in origin \cite{vmt},
243: any continuous change in a parameter of the Hamiltonian results in
244: the same quantized value of $\sxy$ unless the quasiparticle gap collapses.
245: Therefore, for a fixed Dirac anisotropy $\alpha_D$, by virtue of Eq. (\ref{operatorscaling}),
246: the dependence of $\sxy$ on $B$ and $\mu$ can be expressed entirely as a
247: dependence of $\sxy$ on $\frac{1}{B}\Delta_m(B,\mu)$, i.e.
248: \begin{equation}
249: \sxy(B,\mu)=G\left(\frac{\mu}{B} \right),
250: \end{equation}
251: where $G$ is a universal function of its argument that in addition
252: depends on the geometry of the vortex lattice.
253:
254: For a fixed (large) value of B the numerical computation of $\sxy$ can be readily
255: accomplished. The results for a square vortex lattice and $\alpha_D=10$
256: are displayed in Fig. \ref{scalingplot} where it is seen that
257: $\sxy$ is a stair-case function of $\frac{\mu}{B}$ which
258: starts at zero at empty filling and then jumps by a multiple of $2\frac{\hbar}{8\pi}$.
259: The band crossings, which are responsible for the changes in $\sxy$,
260: occur at symmetry points in the Brillouin zone,
261: and by the symmetry of the square vortex lattice, the steps
262: are bound to come in even integers \cite{notsu2}.
263:
264: Thus we arrive at the main result of this Letter shown in
265: Fig. \ref{scalingplot}: the envelope of the $\sxy(B)$ scales as $B^{-1}$ and
266: by Eq. (\ref{wf})
267: \begin{equation}
268: \lim_{T \to 0} \left(\frac{\kxy}{T}\right) \propto \frac{1}{B}
269: \end{equation}
270: This expression is reminiscent of the electrical Hall conductivity in the 2D electron gas.
271:
272: Recently, measurements of $\kappa_{xy}$ were conducted by Ong's group \cite{ong}
273: on YBCO samples with a very long mean free path.
274: These experiments were carried out
275: over a wide range of magnetic fields
276: (up to $\sim 14T$) and at temperatures from $T \sim 12.5K$ to above the
277: superconducting transition
278: $T_c \sim 90K$. Unfortunately, the experiments
279: are resolution limited below
280: $12.5K$ as signal becomes too weak.
281: In Fig. \ref{ongsdata} we re-analyze the data of Ref. \cite{ong}.
282: It is seen that for the magnetic fields above $4T$, $\kxy\propto C+B^{-1}$ which is, apart
283: from the finite intercept, in agreement with our theory.
284: %
285: \begin{figure}[t]
286: \epsfxsize=7.5cm
287: \hfil\epsfbox{fig3.eps}\hfill
288: \caption{Thermal Hall conductivity data of Zhang et. al. \cite{ong} for $B>4T$ plotted
289: vs. $B^{-1}$.}
290: \label{ongsdata}
291: \end{figure}
292: %
293: At fields $B<4T$, $\kxy$ decreases with decreasing $B$ and vanishes at $B=0$ \cite{ong}.
294: At low magnetic fields $\frac{1}{B}$ divergence of $\kxy$ is cut-off by disorder i.e.
295: $\kxy=\frac{\rho_{xy}}{\rho_{xx}^2+\rho_{xy}^2}$ with non-vanishing
296: longitudinal quasiparticle thermal resistivity $\rho_{xx}$. Even in the presence of weak disorder,
297: we still expect $\rho_{xy}\propto B$ and $\rho_{xx}$ to be approximately field
298: independent. Therefore, the transverse thermal conductivity in the presence of disorder should vanish
299: in the limit of vanishing $B$.
300: We must admit that the origin of the residual thermal conductivity
301: at $\frac{1}{B}=0$, evident in Fig. \ref{ongsdata},
302: remains mysterious to us and we can only
303: speculate that it is not of quasiparticle origin.
304:
305: In summary, based on a general analysis of the full {\em non-linearized}
306: Bogoliubov-de~Gennes equations,
307: we argued that magnetic field $B$ will induce a small gap $\Delta_m$
308: in the quasiparticle spectrum which is a non-monotonic function of the
309: magnetic field. Rather, $\Delta_m$ vanishes at some special values of
310: $\mu/B$ which depend on the details of the vortex lattice and the Dirac cone
311: anisotropy (Fig. \ref{scalingplot}).
312: The dependence of $\kappa_{xy}$ on $B$ is entirely
313: due to this non-monotonic behavior of $\Delta_m$, since $\kappa_{xy}/T$ is
314: a topological quantity and therefore
315: can change only at the band crossings \cite{vmt}.
316: The topological quantization of $\kappa_{xy}$ was also
317: discussed in the Ref. \cite{ashvin}.
318: However, we wish to point out a significant
319: qualitative difference between the
320: results presented here and those of Ref. \cite{ashvin}:
321: as we argued before $\kappa_{xy}$ is a quantized staircase function of $1/B$
322: with an {\em envelope} that itself
323: scales as $1/B$, while Ref. \cite{ashvin} states that $\kappa_{xy}/T$
324: can take on only one of the three allowed values $0, \pm 2$
325: and is otherwise field-independent. Consequently, in a
326: typical experimental situation,
327: the explicit values for $\kappa_{xy}$ obtained from
328: these two works will differ significantly.
329: Still, there is an agreement on the fact that
330: since the quasiparticle spectrum is gapped $\kappa_{xy}$ does
331: not obey simple Simon-Lee scaling, at least when the temperature
332: is comparable to the quasiparticle gap.
333:
334: We thank Marcel Franz, Ashot Melikyan and Ady Stern
335: for comments and discussions. This work was supported in part by NSF
336: grant DMR00-94981.
337:
338: \vspace{-0.5cm}
339: %%%%%%%%%%%%%%
340: \begin{references}
341: \bibitem{BednorzMuller} J. G. Bednorz and K. A. Muller, Z. Phys. {\bf 64} 188
342: (1986).
343: \bibitem{harlingen} D. J. Van Harlingen, Rev. Mod. Phys. {\bf 67}, 515 (1995).
344: \bibitem{anderson1} P. W. Anderson, Science {\bf 235}, 1196 (1987).
345: \bibitem{walker} M. B. Walker and M. F. Smith, \prb {\bf 61}, 11285 (2000).
346: \bibitem{ding} H. Ding {\it et. al.}, \prb {\bf 54}, R9678 (1996).
347: \bibitem{taillefert1} M. Chiao {\it et. al.}, cond-mat/9910367.
348: \bibitem{millis} J. Orenstein and A. J. Millis, Science {\bf 288}, 468 (2000).
349: \bibitem{ong} Y. Zhang {\it et. al.} \prl {\bf 86}, 890 (2001).
350: \bibitem{simonlee} S. H. Simon and P. A. Lee, \prl {\bf 78}, 1548 (1997).
351: \bibitem{jinwu} J. Ye, \prl {\bf 86}, 316 (2001).
352: \bibitem{ashvin} A. Vishwanath, \prl {\bf 87}, 217004 (2001);
353: A. Vishwanath cond-mat/0104213.
354: \bibitem{ft} M. Franz and Z. Te\v{s}anovi\'c, \prl {\bf 84}, 554 (2000);
355: O. Vafek, A. Melikyan, M. Franz, Z. Te\v{s}anovi\'c, \prb {\bf 63},
356: 134509 (2001).
357: \bibitem{luca} L. Marinelli, B. I. Halperin and S. H. Simon, \prb {\bf 62}
358: 3488 (2000).
359: \bibitem{vmt} O. Vafek, A. Melikyan, and Z. Te\v{s}anovi\'c, \prb {\bf 64},
360: 224508 (2001) and references therein.
361: \bibitem{noncrossing} J. von Neumann and
362: E. Wigner, {\it Physik. Z.}, {\bf 30} 467 (1927).
363: \bibitem{notsu2} The trivial double degeneracy of each band due to spin
364: is ignored in this calculation.
365: \end{references}
366: \end{document}
367:
368: