1: \documentstyle[aps,preprint,floats,psfig]{revtex}
2: %\documentstyle[aps,prl,twocolumn,psfig]{revtex}
3:
4: \draft
5: \begin{document}
6: \newcommand {\be}{\begin{equation}}
7: \newcommand {\ee}{\end{equation}}
8: \newcommand {\bea}{\begin{eqnarray}}
9: \newcommand {\eea}{\end{eqnarray}}
10: \newcommand {\nn}{\nonumber}
11:
12: %\twocolumn[\hsize\textwidth\columnwidth\hsize\csname@twocolumnfalse%
13: %\endcsname
14:
15: \title{Impurity Bound States and Symmetry of the Superconducting
16: Order Parameter in $\rm Sr_2RuO_4$}
17:
18: \author{Kazumi Maki and Stephan Haas}
19: \address{Department of Physics and Astronomy, University of Southern
20: California, Los Angeles, CA 90089-0484
21: }
22:
23: \date{\today}
24: \maketitle
25:
26: \begin{abstract}
27:
28: Recent experiments on $\rm Sr_2RuO_4$
29: have indicated the presence of linear nodes in the
30: superconducting order parameter. Among the possible spin
31: triplet states, 2D p-wave superconductivity with
32: $E_u$ symmetry appears to
33: be inconsistent with the experiments, whereas the 2D
34: f-wave order parameter with $\rm B_{1g} \times E_u$
35: symmetry turns out to be a more likely
36: candidate. Here it is shown how the quasi-particle
37: bound state wave functions around a single impurity
38: provide a clear signature of the symmetry of the underlying
39: superconducting order parameter.
40:
41: \end{abstract}
42: \pacs{}
43: %]
44:
45: \section{Introduction}
46:
47: In recent years, we have experienced
48: a rapid improvement in the quality of $\rm Sr_2RuO_4$
49: single crystals. This development has lead to a renewed discussion
50: regarding the symmetry of the underlying superconductiving order
51: parameter.
52: The data from specific heat measurements\cite{nishizaki},
53: NMR results on $\rm T_1^{-1}$\cite{ishida}, and magnetic
54: penetration depth experiments\cite{bonalde} in very pure
55: single crystals with $\rm T_c \simeq 1.5 K$ suggest a
56: linear nodal structure in the superconducting order
57: parameter, similar to the $\rm d_{x^2 - y^2}$-wave
58: superconductors\cite{won}.
59: Hence, the initially proposed fully gapped 2D p-wave
60: superconductivity with $E_u$ symmetry\cite{rice}
61: appears to be inconsistent with the experimentally observed nodal
62: structure. On the other hand, the spin triplet nature
63: of the superconductivity in $\rm Sr_2RuO_4$ has clearly been
64: established by muon rotation experiments\cite{luke} which
65: probe a spontaneous spin polarization, and by the flat
66: Knight shift seen in NMR\cite{ishida2}.
67:
68: Motivated by these recent experiments, several spin triplet
69: order parameters with linear nodes have been proposed.
70: \cite{hasegawa} In particular, from a study of the thermal conductivity
71: tensor in a planar magnetic field it has been suggested that
72: the 2D f-wave order parameter with $\rm B_{1g} \times E_u$
73: symmetry is a likely candidate.\cite{dahm} Furthermore, recent
74: measurements of the angular dependence of the upper
75: critical field have detected a
76: in-plane anisotropy consistent with a $\rm B_{1g} \times E_u$
77: component to the order parameter.\cite{mao}
78:
79: The objective of this paper is to calculate the quasi-particle
80: bound state wave function around non-magnetic impurities in
81: unconventional superconductors with the anisotropic order
82: parameters which have been
83: proposed for $\rm Sr_2RuO_4$. It has recently been
84: observed in a series of STM imaging experiments
85: that a Zn-impurity in the high-$\rm T_c$ compound
86: Bi2212 gives rise to such a bound state, leading to a distinct
87: four-fold symmetric pattern in the local tunneling conductance
88: around the impurity.\cite{pan}
89: From the theoretical side it was shown that this type of bound
90: state wave function can be interpreted in terms of the solutions
91: of the Bogoliubov-de Gennes equations for $\rm d_{x^2 - y^2}$-wave
92: superconductors.\cite{morita,haas}. Here we will perform an
93: analogous analysis for the proposed 2D p-wave and f-wave
94: order parameters.
95:
96: \section{Bound State Wave Functions}
97: \subsection{2D p-wave superconductors}
98:
99: For the initially proposed fully gapped odd-parity
100: 2D p-wave superconductivity with an order parameter
101: $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \exp{(\pm i \phi )}$
102: the Bogoliubov - de Gennes equations have already been worked
103: out.\cite{matsumoto,okuno} In particular, when $\rm \Delta({\bf r})
104: = \Delta = const. $\cite{footnote}, one obtains
105: \bea
106: E u({\bf r}) &=& \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)
107: u({\bf r}) + p_F^{-1} \Delta (i\partial_x - \partial_y )
108: v({\bf r}),\\
109: E v({\bf r}) &=& - \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)
110: v({\bf r}) + p_F^{-1} \Delta (i\partial_x + \partial_y )
111: u({\bf r}),
112: \eea
113: where $\mu$ is the chemical potential, and the
114: impurity potential, centered at the site ${\bf r} = 0$,
115: is approximated by $V({\bf r}) = a \delta^2({\bf r})$.
116: For simplicity, only the 2D system is considered here.
117:
118: A simple variational solution of Eqs. (1) and (2) is found to be
119: \bea
120: u({\bf r}) &=& A \exp{(- \gamma r)} J_0(p_F r), \\
121: v({\bf r}) &=& A \alpha \exp{(- \gamma r)} J_1(p_F r) \exp{(i \phi)},
122: \eea
123: where $p_F$ is the Fermi momentum, $J_l(z)$ are Bessel functions
124: of the first kind, and $A$ is a global normalization factor. By
125: inserting these variational wave functions $u({\bf r})$ and
126: $v({\bf r})$ into the Bogoliubov - de Gennes equations, it follows
127: that $E = -(V/2) \pm \sqrt{(K - V )^2 + \Delta^2}$ and
128: $\alpha = \Delta/(E + K) \simeq 1$,
129: where the kinetic $K$ and the potential
130: $V$ contributions to the energy are defined by
131: \bea
132: K &\equiv & \frac{\int_0^{\infty} dr r \left[
133: \left( \partial_r \exp{(-\gamma r)}
134: J_l(p_F r)\right)^2 + \left( l \exp{(-\gamma r)} J_l(p_F r)/r \right)^2
135: \right]}
136: {2m \int_0^{\infty} dr r \left( \exp{(-\gamma r)} J_l(p_F r)\right)^2}
137: -\mu
138: \simeq \frac{\gamma^3}{m p_F},\\
139: V &\equiv & \frac{\int_0^{\infty} dr r \exp{(-2 \gamma r)} J_0^2(p_F r)
140: V({\bf r})}
141: {\int_0^{\infty} dr r \exp{(-2 \gamma r)} J_0^2(p_F r)}
142: \simeq (2 \pi \gamma p_F) \int_0^{\infty} dr r \exp{(-2 \gamma r)}
143: J_0^2(p_F r) V ({\bf r}).
144: \eea
145:
146: The squares of the wave functions $u({\bf r})$ and $v({\bf r})$ can
147: be observed by scanning tunneling microscopy.\cite{pan}
148: The tunneling current,
149: $I({\bf r},{\it V}) \propto \int dE A_S({\bf r}, E)
150: A_N({\bf r}, E + e{\it V})$, is
151: a convolution of the local one-particle spectral function
152: of the normal-state tip, $A_N ({\bf r}, E) = \sum_k
153: \delta (E - E_k)$, and that of the
154: superconducting sample, $A_S ({\bf r}, E) = \sum_k
155: [ |u({\bf r})|^2 \delta (E - E_k) +
156: |v({\bf r})|^2 \delta (E + E_k) ]$.
157: The differential tunneling conductance is thus
158: obtained by taking the partial derivative of $I({\bf r},{\it V})$
159: with respect to the applied voltage ${\it V}$,
160: \bea
161: \frac{\partial I}{\partial {\it V}} ({\bf r}, {\it V}) \propto
162: {\rm sech}^2\left( \frac{e{\it V} - E_0}{2T} \right) |u({\bf r})|^2
163: + {\rm sech}^2\left( \frac{e{\it V} + E_0}{2T} \right) |v({\bf r})|^2.
164: \eea
165: At small temperatures, the local tunneling conductance around
166: the impurity site is dominated by $|u({\bf r})|^2$ for a fixed binding energy
167: $E_0$ and by $|v({\bf r})|^2$ for $-E_0$.
168:
169:
170: In Fig. 1(a) and (b), we show $|u({\bf r})|^2$ and
171: $|v({\bf r})|^2$ respectively. The Fermi wave vector
172: $p_F \simeq 2.7 /a$ was chosen to be consistent with
173: band structure calculations.\cite{mazin} The patterns
174: described by the squares of these wave functions are
175: concentric circles without a trace of four-fold symmetry.
176: If the underlying superconducting order parameter is
177: indeed
178: $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \exp{(\pm i \phi )}$
179: the patterns of the local tunneling current around non-magnetic
180: impurities should thus be featureless along the azimuthal direction
181: in the 2D plane.
182:
183: \subsection{2D f-wave superconductors}
184:
185: Let us now consider 2D f-wave superconductors with an
186: order parameter
187: $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \cos{(2 \phi)} \exp{(\pm i \phi )}$.
188: Following the above procedure, the corresponding Bogoliubov - de Gennes
189: equations are then given by
190: \bea
191: E u({\bf r}) &=& \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)
192: u({\bf r}) + p_F^{-3} \Delta (\partial_x^2 - \partial_y^2 )
193: (i\partial_x - \partial_y )v({\bf r}),\\
194: E v({\bf r}) &=& - \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)
195: v({\bf r}) + p_F^{-3} \Delta (\partial_x^2 - \partial_y^2 )
196: (i\partial_x + \partial_y )u({\bf r}).
197: \eea
198:
199: Bound state solutions for this case are found to be of the form
200: \bea
201: u({\bf r}) &=& A \exp{(- \gamma r)}\left( J_0(p_F r) + \beta
202: J_4(p_F r)\cos{(4 \phi )} \right), \\
203: v({\bf r}) &=& A \exp{(- \gamma r)}\left(\alpha
204: J_1(p_F r) \exp{(- i \phi)} + \delta J_3(p_F r) \exp{(3 i \phi)} \right).
205: \eea
206:
207: In analogy to Ref. \cite{haas}, we obtain
208: \bea
209: E &=& K - V -\frac{1}{\sqrt{2}}\Delta ( \alpha + \delta ), \\
210: E \alpha &=& - K \alpha - \frac{1}{\sqrt{2}}\Delta(1+\frac{\beta}{\sqrt{2}}),\\
211: E \beta &=& K \beta - \frac{1}{2} \Delta (\alpha + \delta ), \\
212: E \delta &=& -K \delta - \frac{1}{\sqrt{2}}\Delta(1+\frac{\beta}{\sqrt{2}}).
213: \eea
214: From Eqs. (12) - (15) we get
215: $\alpha = \delta = (V - K ) / ( \sqrt{2} \Delta ) $ and
216: $\beta = (1 - V/(K-E))/\sqrt{2}$.
217: For an impurity with a scattering strength in the unitary limit
218: the energy of the bound state is expected to be very small,
219: $E \simeq 0$, and $V \simeq \Delta$. This gives $\beta \simeq
220: -1/\sqrt{2}$. On the other hand, in the Born limit
221: $E\simeq \Delta$ and $V \simeq 0$, which gives
222: $\beta \simeq
223: 1/\sqrt{2}$.
224: In Fig. 2(a) and (b), we plot
225: $|u({\bf r})|^2$ and $|v({\bf r})|^2$ for $\beta = 1/\sqrt{2}$,
226: corresponding to weak scattering.
227: In the limit of strong impurity scattering, $\beta = - 1/\sqrt{2}$,
228: the patterns are
229: changed as shown in Fig. 3(a) and (b).
230: It is observed that for 2D f-wave superconductors with
231: $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \cos{(2 \phi)} \exp{(\pm i \phi )}$,
232: both $|u({\bf r})|^2$ and $|v({\bf r})|^2$ have a
233: four-fold symmetry. Depending on the impurity scattering strength,
234: they extend either
235: in the directions of the Ru-O bonds or are tilted by an angle of
236: 45 degrees.
237: The STM imaging of impurity bound states can thus provide a
238: clear signature of the order parameter symmetry for the
239: underlying superconductivity.
240:
241: \section{Concluding Remarks}
242:
243: Stimulated by the successful scanning tunneling microscopy imaging of
244: quasi-particle bound state wave functions around Zn-impurities
245: in Bi2212\cite{pan,hudson}, we have studied the analogous
246: patterns of impurity bound states
247: in $\rm Sr_2RuO_4$. By applying the appropriate Bogoliubov - de Gennes
248: equations, the characteristic patterns were distinguished
249: for two proposed order parameters: (i) gapped
250: 2D p-wave (or $E_u$) superconductors
251: with $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \exp{(\pm i \phi )}$,
252: and (ii) gapless 2D f-wave (or $\rm B_{1g} \times E_u$) superconductors
253: with $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \cos{(2\phi)}
254: \exp{(\pm i \phi )}$. While the tunneling conductance patterns of
255: the fully gapped odd-parity 2D p-wave superconductor are featureless
256: along the azimuthal direction, a clear four-fold symmetry in
257: $|u({\bf r})|^2$ and $|v({\bf r})|^2$ is predicted
258: for the
259: 2D f-wave superconductor.
260:
261: The experimentally observed in-plane anisotropy of the upper critical
262: field $H_{c2}$ in $\rm Sr_2RuO_4$
263: is quite small ($\approx 3 \%$). This may
264: indicate that there is
265: a combination of 2D f-wave and p-wave superconductivity in this
266: compound.\cite{mao}
267: Furthermore, a possible
268: alternative to the plain 2D p-wave superconductivity considered
269: above would be a 3D $\rm A_{1g} \times E_u$ f-wave order parameter of the form
270: $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \cos{(c k_z)}
271: \exp{(\pm i \phi )}$. The in-plane impurity bound state patterns for
272: 2D $\rm E_u$ and 3D $\rm A_{1g} \times E_u$
273: are the same, and a directional probe along the $\hat{k}_z$-direction would
274: be needed to distinguish between these two cases.
275:
276: Our study suggests that the Bogoliubov - de Gennes formalism in the
277: continuum limit is very useful in addressing the shape of
278: impurity induced bound states. The corresponding bound state
279: wave functions $u({\bf r})$ and $v({\bf r})$ clearly reflect the
280: symmetry of $\rm \Delta({\bf r}) $.
281: Therefore the imaging of these wave functions provides unique insight
282: into the underlying symmetry of the order parameter. A similar study
283: of the vortex state is in progress.
284:
285: We thank Kazushige Machida,
286: Manfred Sigrist, and Hyekyung Won for useful discussions.
287: K. M. acknowledges the hospitality and the support
288: of the Max-Planck Institut f\"ur Physik komplexer Systeme
289: at Dresden, and S. H. thanks the
290: Zumberge foundation for financial support.
291:
292:
293: \begin{references}
294:
295: \bibitem{nishizaki} S. Nishizaki, Y. Maeno, and Z. Mao,
296: J. Phys. Soc. Jpn. {\bf 60}, 572 (2000);
297: S. Nishizaki, Y. Maeno, and Z. Mao,
298: J. Low Temp. Phys. {\bf 117}, 1581 (1999).
299:
300: \bibitem{ishida} K. Ishida, H. Mukuda, Y. Kitaoka, Z. Q.
301: Mao, Y. Mori, and Y. Maeno, Phys. Rev. Lett. {\bf 84},
302: 5387 (2000).
303:
304: \bibitem{bonalde} I. Bonalde, B. D. Yanoff, M. B. Salamon,
305: D. J. van Harlingen, E. M. E. Chia, Z. Q. Mao, and Y. Maeno,
306: preprint.
307:
308: \bibitem{won} H. Won and K. Maki, Phys. Rev. B {\bf 49}, 1397
309: (1994).
310:
311: \bibitem{rice} T. M. Rice and M. Sigrist, J. Phys. Cond. Mat.
312: {\bf 7}, L643 (1995); M. Sigrist, D. Agterberg, A. Furusaki,
313: C. Honerkamp, K. K. Ng, T. M. Rice, and M. E. Zhitomirsky,
314: Physica C {\bf 317 - 318}, 134 (1999).
315:
316: \bibitem{luke} G. M. Luke, Y. Fudamoto, K. M. Kojima,
317: M. I. Larkin, J. Merrin, B. Nachumi, Y. J. Uemura,
318: Y. Maeno, Z. Q. Mao, Y. Mori, H. Nakamura, and M. Sigrist,
319: Nature {\bf 394}, 558 (1998).
320:
321: \bibitem{ishida2} K. Ishida, H. Mukuda, Y. Kitaoka,
322: K. Asayama, Z. Q. Mao, Y. Mori, and Y. Maeno,
323: Nature {\bf 396}, 653 (1998).
324:
325: \bibitem{hasegawa} Y. Hasegawa, K. Machida, and M. Ozaki,
326: J. Phys. Soc. Jpn. {\bf 69}, 336 (2000).
327:
328: \bibitem{dahm} T. Dahm, H. Won, and K. Maki,
329: cond-mat/0006301.
330:
331: \bibitem{mao} Z. Q. Mao, Y. Maeno, S. Nishizaki, T. Akima,
332: and T. Ishiguro, Phys. Rev. Lett. {\bf 24}, 991 (2000).
333:
334: \bibitem{pan} S. H. Pan, E. W. Hudson, K. M. Lang, H. Eisaki,
335: S. Uchida, and J. C. Davis, Nature {\bf 403}, 746 (2000).
336:
337: \bibitem{morita} Y. Morita, M. Kohmoto, and K. Maki,
338: Europhys. Lett. {\bf 40}, 207 (1997).
339:
340: \bibitem{haas} S. Haas and K. Maki, to be published
341: in Phys. Rev. Lett. (cond-mat/0004386).
342:
343: \bibitem{footnote}
344: For the high-$\rm T_c$ cuprates, it is known that
345: their superconducting coherence length is rather short,
346: $\xi \simeq 10 -20 \AA$. Therefore the impurity induced spatial
347: dependence of the gap function $\Delta({\bf r})$
348: was neglected in Ref. \cite{haas}.
349: In contrast, for $\rm Sr_2RuO_4$
350: $\xi \simeq 1 \mu m$. Hence a self-consistent treatment of
351: the spatial variation in $\Delta({\bf r})$
352: may be needed for a more
353: quantitative account of the impurity problem in this material.
354:
355: \bibitem{matsumoto} M. Matsumoto and M. Sigrist,
356: J. Phys. Soc. Jpn. {\bf 68}, 724 (1999);
357: M. Matsumoto and M. Sigrist,
358: Physica B {\bf 281 - 282}, 973 (2000).
359:
360: \bibitem{okuno} Y. Okuno, M. Matsumoto, and M. Sigrist,
361: J. Phys. Soc. Jpn. {\bf 68}, 3054 (1999). In this reference,
362: impurity induced circular currents were studied. The
363: corresponding wave functions are practically equivalent to our solutions.
364:
365: \bibitem{mazin} I. I. Mazin and D. J. Singh, Phys. Rev. Lett.
366: {\bf 79}, 733 (1997).
367:
368: \bibitem{hudson} E. W. Hudson, S. H. Pan, A. K. Gupta, K. W. Ng,
369: and J. C. Davis, Science {\bf 285}, 88 (1999).
370:
371: \end{references}
372:
373: \newpage
374:
375: \begin{figure}[h]
376: \centerline{\psfig{figure=11.eps,height=6cm}
377: \psfig{figure=12.eps,height=6cm}}
378: \caption{
379: Spatial variation of the local tunneling conductance, centered at a
380: non-magnetic impurity in a 2D p-wave superconductor
381: with $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \exp{(\pm i \phi )}$.
382: In the left figure, the dominant contribution
383: $|u({\bf r})|^2$ at the
384: positive bound state resonant frequency $E_0$ is shown.
385: On the right hand side, the dominant contribution
386: $|v({\bf r})|^2$ at the
387: negative bound state resonant frequency $-E_0$ is shown.
388: }
389: \end{figure}
390:
391: \begin{figure}[h]
392: \centerline{\psfig{figure=21.eps,height=6cm}
393: \psfig{figure=22.eps,height=6cm}}
394: \caption{
395: Spatial variation of the local tunneling conductance, centered at a
396: non-magnetic impurity in a 2D f-wave superconductor with
397: $\vec{\Delta} ({\bf k}) = \Delta \hat{z} \cos{(2 \phi)} \exp{(\pm i \phi )}$.
398: On the left hand side, the dominant contribution
399: $|u({\bf r})|^2$ at the
400: positive bound state resonant frequency $E_0$ is shown.
401: On the right hand side, the dominant contribution
402: $|v({\bf r})|^2$ at the
403: negative bound state resonant frequency $-E_0$ is shown.
404: The solution in this figure corresponds to the weak impurity
405: scattering limit.
406: }
407: \end{figure}
408:
409: \begin{figure}[h]
410: \centerline{\psfig{figure=31.eps,height=6cm}
411: \psfig{figure=32.eps,height=6cm}}
412: \caption{
413: Same as Fig. 2, but for strong impurity scattering.
414: }
415: \end{figure}
416:
417:
418:
419:
420:
421:
422: \end{document}
423:
424: