1: % RevTef (version 3.0) March 04, 1993
2: %
3: %\documentstyle[prb,aps,psfig]{revtex}
4: \documentstyle[aps,preprint,floats,psfig]{revtex}
5: %\documentstyle[prl,aps,psfig]{revtex}
6: %\input psfig
7: \begin{document}
8: \newcommand {\be}{\begin{equation}}
9: \newcommand {\ee}{\end{equation}}
10: \newcommand {\bea}{\begin{eqnarray}}
11: \newcommand {\eea}{\end{eqnarray}}
12: \newcommand {\nn}{\nonumber}
13:
14: \draft
15: %
16: %\twocolumn[\hsize\textwidth\columnwidth\hsize\csname @@twocolumnfalse\endcsname
17: %
18: %
19: % Title Page
20: %
21:
22: \title{Quasi-Particle Bound States around Impurities in
23: $\rm d_{x^2 - y^2}$-wave Superconductors
24: }
25:
26: \author{Stephan Haas and Kazumi Maki}
27: \address{Department of Physics and Astronomy, University of Southern
28: California, Los Angeles, CA 90089-0484
29: }
30:
31: \date{\today}
32: \maketitle
33:
34: \begin{abstract}
35: Zn and Ni impurities in the hole-doped high-temperature superconductors
36: are known to have strong effects on thermodynamic and transport properties.
37: A recent scanning tunneling microscope study of Zn-doped
38: Bi2212 (Pan {\it et al.}, Nature {\bf 403}, 746 (2000)) has provided
39: high-resolution images of the local density of states around
40: non-magnetic impurities in
41: $\rm d_{x^2 - y^2}$-wave superconductors. These pictures contain detailed
42: information about the spinor wave functions $u({\bf r})$ and
43: $v({\bf r})$ of bound states with energy $E_0 \sim \Delta/30$,
44: centered at
45: the Zn-sites. We show that this type of wave function follows from
46: the solutions of the Bogoliubov-de Gennes equations for
47: $\rm d_{x^2 - y^2}$-wave superconductors.
48:
49: \end{abstract}
50: \pacs{}
51: %\vskip2pc]
52:
53: {\it Introduction:}
54: Impurity-doping of anisotropic superconductors has turned out to
55: be a valuable probe, setting them apart from conventional superconductors
56: with an s-wave order parameter. For example, doping of
57: high-temperature superconductors, such as YBCO, LSCO, and BCCO,
58: with non-magnetic impurities has contributed greatly to establish the
59: underlying $\rm d_{x^2 - y^2}$-wave order parameter in these systems.
60: \cite{hirschfeld,sun,haas}
61: In particular, an analysis of the thermodynamic and transport properties
62: in these compounds suggests that
63: Zn-impurities can be modeled with a scattering potential in the
64: unitary limit.\cite{sun,maki} On the other hand,
65: little is known experimentally about the
66: local real-space structure of the impurity bound states around the
67: Zn-sites, in spite of numerous theoretical studies on this question.
68: \cite{byers,salkola,tsuchiura} Very recently, Pan {\it et al.} have
69: provided detailed scanning tunneling microscope (STM) images of the
70: local tunneling conductivity, fixed at $\pm E_0$, where
71: $E_0 \simeq \Delta/30$ is the binding energy of the impurity bound state
72: around a Zn site in Bi2212. The corresponding wave function exhibits
73: a fourfold symmetry, associated with the underlying $\rm CuO_2$ lattice.
74:
75: In the context of conventional s-wave superconductors, this type of
76: bound state, in this case associated with a magnetic impurity,
77: has been predicted by Shiba.\cite{shiba} Recently, this idea has found
78: experimental support by STM studies in the vicinity of Fe and Gd
79: impurity atoms on the surface of Nb superconductors.\cite{yazdani}
80:
81: The objective of this paper is to present a simple solution of the
82: Bogoliubov-de Gennes equations for $\rm d_{x^2 - y^2}$-wave
83: superconductors\cite{degennes,morita} in the presence of a single impurity.
84: This corresponds to an experimental situation where the impurity
85: concentration is dilute enough, such that interactions between the
86: impurities can be neglected. Within this approach,
87: we obtain a wave function which resembles very accurately the fourfold
88: symmetry which has recently been observed around Zn-impurities in
89: Bi2212\cite{pan},
90: where the energies of the associated bound states are very small
91: ($|E_0|/\Delta \ll 1$). We believe that this image of the local
92: tunneling conductivity around the Zn-site, as seen by STM, provides
93: strong evidence for $\rm d_{x^2 - y^2}$-wave symmetry. Therefore,
94: this method can also be used to test anisotropic order parameters
95: in organic superconductors, and may serve as a straightforward tool
96: to explore the underlying symmetry in other new superconductors where
97: the nature of the order parameter is still under debate.
98:
99: We also study the case of Ni impurities which may be modeled by a
100: weaker scattering potential, close to the Born limit.\cite{puchkaryov}
101: In this case, we find that although $|v({\bf r})|^2$ is not much different
102: from the Zn-case, $|u({\bf r})|^2$ is rotated by $\pi/4$ with respect
103: to the strong-scattering limit corresponding to Zn.
104: Finally, we find that for impurities with an impurity potential
105: of intermediate strength, the lowest harmonic in $|u({\bf r})|^2$
106: is suppressed, and therefore a dominant eight-fold symmetry may be
107: expected instead of the four-fold symmetry, observed in the weak and
108: strong scattering limits.
109:
110: {\it Bogoliubov-de Gennes equations and bound state wave functions:}
111: The Bogoliubov-de Gennes (BdG) equations for $\rm d_{x^2 - y^2}$-wave
112: superconductors in the continuum limit are given by\cite{degennes,morita}
113: \bea
114: E u({\bf r}) &=& \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)
115: u({\bf r}) + \frac{1}{p_F^2} \Delta (\partial_x^2 - \partial_y^2 )
116: v({\bf r}),\\
117: E v({\bf r}) &=& - \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)
118: v({\bf r}) + \frac{1}{p_F^2} \Delta (\partial_x^2 - \partial_y^2 )
119: u({\bf r}),
120: \eea
121: where $\mu$ is the chemical potential, and $V ({\bf r}) > 0$ is the
122: impurity potential, centered at the site ${\bf r} = 0$.
123:
124: Let us first consider the case of Zn impurities where the energy of the
125: bound states is known to be very small\cite{pan}, $E_0 \simeq 0$. In the
126: following, we use a variational ansatz for the solutions of the BdG equations:
127: \bea
128: u({\bf r}) &=& A \exp{(- \gamma r)} \left( J_0(p_F r) + \sqrt{2} \beta
129: J_4(p_F r) \cos{(4 \phi)} \right),\\
130: v({\bf r}) &=& \sqrt{2} A \alpha \exp{(- \gamma r)} J_2(p_F r) \cos{(2 \phi)},
131: \eea
132: where $J_l(z)$ are Bessel functions of the first kind, and $p_F$ is the
133: Fermi momentum.
134: ($A$ is the global normalization
135: factor for the wave functions which can be neglected.)
136: Note that for Bi2212 it is believed that $p_F \simeq 0.7 \AA^-1$. This
137: implies that $p_F a \simeq 2$.\cite{chiao} The coefficients
138: $\alpha$, $\beta$, and $\gamma$ are determined variationally.
139: Inserting Eqs. 3 and 4 into Eqs. 1 and 2, we find
140: \bea
141: E &=& K - V -\frac{1}{\sqrt{2}}\Delta \alpha, \nn \\
142: E \alpha &=& - K \alpha - \frac{1}{\sqrt{2}}\Delta(1+\frac{\beta}{\sqrt{2}}),\\
143: E \beta &=& K \beta - \frac{1}{2} \Delta \alpha \nn ,
144: \eea
145: where
146: \bea
147: K &\equiv & \frac{\int_0^{\infty} dr r \left[
148: \left( \partial_r \exp{(-\gamma r)}
149: J_l(p_F r)\right)^2 + \left( l \exp{(-\gamma r)} J_l(p_F r)/r \right)^2
150: \right]}
151: {2m \int_0^{\infty} dr r \left( \exp{(-\gamma r)} J_l(p_F r)\right)^2}
152: -\mu
153: \simeq \frac{\gamma^3}{m p_F},\\
154: V &\equiv & \frac{\int_0^{\infty} dr r \exp{(-2 \gamma r)} J_0^2(p_F r)
155: V({\bf r})}
156: {\int_0^{\infty} dr r \exp{(-2 \gamma r)} J_0^2(p_F r)}
157: \simeq (2 \pi \gamma p_F) \int_0^{\infty} dr r \exp{(-2 \gamma r)}
158: J_0^2(p_F r) V ({\bf r}).
159: \eea
160: In general, $K$ in Eq. 6 depends on which Bessel function $J_l(p_F r)$
161: is used. However, in the limit $\gamma/p_F \ll 1$, $K$ reduces to
162: $\gamma^3/(m p_F)$ for all $J_l(p_F r)$ with $l \ll p_F/\gamma$.
163: In the approximation for $V$,
164: only the dominant s-wave component ($l$ = 0) of the scattering potential
165: $V({\bf r})$ has been considered.
166: In the usual convention, which is used in this paper,
167: a positive sign of $V({\bf r})$ corresponds to the
168: attraction of electrons by the impurity. In impurity-doped
169: Bi2212\cite{pan} the charge carriers are holes,
170: and hence the opposite sign has to be chosen.
171:
172:
173: For the case of a Zn-impurity (strong scattering limit), we may assume
174: that $E \simeq 0$.\cite{pan} This gives $K \simeq V/2$,
175: $V \simeq 3\Delta/2 \approx \sqrt{2}\Delta$,
176: $\alpha \simeq \sqrt{2}(V - K )/\Delta$,
177: and $\beta \simeq -\sqrt{2}$.
178: These choices yield the approximate parameter set
179: $\alpha \simeq 1$, $\beta \simeq -1/\sqrt{2}$, and $\gamma \simeq
180: p_F (\sqrt{2} p_F \xi)^{-1/3}
181: \simeq 0.3 p_F$.
182: The tunneling conductance is given by\cite{gygi}
183: \bea
184: \frac{d I}{d V} ({\bf r}) \propto
185: {\rm sech}^2\left( \frac{eV - E_0}{2T} \right) |u({\bf r})|^2
186: + {\rm sech}^2\left( \frac{eV + E_0}{2T} \right) |v({\bf r})|^2.
187: \eea
188: Thus, at small temperatures, the local tunneling conductance around
189: the impurity site is dominated by $|u({\bf r})|^2$ for a fixed binding energy
190: $E_0$ and by $|v({\bf r})|^2$ for $-E_0$.
191: In Figs. 1 and 2 images of $|u({\bf r})|^2$ and
192: $|v({\bf r})|^2$ are shown. Both $|u({\bf r})|^2$ and $|v({\bf r})|^2$ have a
193: four-fold symmetry, and extend in the directions of the Cu-O bonds. This
194: strongly resembles the images seen by the STM experiments.\cite{pan}
195: Furthermore, weaker higher-harmonic
196: satellite peaks are observed in $|u({\bf r})|^2$,
197: along the ($\pm \pi/4, \pm \pi/4)$ directions, in accordance with
198: the experiments\cite{pan}. Note that with our sign convention for
199: the scattering potential $V({\bf r})$, the roles of
200: $|u({\bf r})|^2$ and $|v({\bf r})|^2$ are interchanged for the hole-doped
201: cuprates. Thus, in Bi2212 our $|u({\bf r})|^2$ corresponds to a bound
202: state at $-E_0$ and $|v({\bf r})|^2$ to a bound state at $E_0$.
203:
204: Now let us turn to the case of Ni impurities. Ni is considered to
205: be a weak scatterer. Therefore we may assume that $V({\bf r}) \rightarrow
206: 0$, and consequently $\alpha \simeq 2 \beta (K - E)/\Delta$ and
207: $\beta \simeq \sqrt{2}(1 - 4(E^2 - K^2)/\Delta^2)^-1$.
208: So, if $E^2 > K^2 + \Delta^2/4$
209: the coefficient
210: $\beta$ turns out to be positive, which is most likely the case for Ni.
211: Indeed, a sufficiently large magnitude of $\alpha$ implies that
212: $E^2 \gg K^2 \simeq 0$
213: and $E\simeq \sqrt{3}\Delta /2$.
214: This gives $\alpha \simeq -\sqrt{3/2}$, $\beta \simeq 1/\sqrt{2}$,
215: and $\gamma \simeq
216: 0.1 p_F$. The squares of the wave functions $u({\bf r})$ and $v({\bf r})$
217: are shown in
218: Figs. 3 and 4. Comparing with the corresponding images for the
219: strong scattering limit,
220: we note that $|v({\bf r})|^2$ is qualitatively similar in both
221: cases, whereas $|u({\bf r})|^2$ appears to be rotated by $\pi/4$ with
222: respect to the case of Zn impurities.
223:
224: For impurities with a scattering potential $V({\bf r})$ of intermediate
225: strength, the coefficient $\beta$ in the wave function $u({\bf r})$ is found
226: to be greatly enhanced. Consequently, the higher-harmonic term in
227: $u({\bf r})$ (Eq. 3), containing
228: a $\cos{(4\phi)}$-modulation dominates the tunneling response at positive
229: binding energies $+E_0$. In this case, an eight-fold instead of a
230: four-fold symmetry should be
231: observed in $|u({\bf r})|^2$, as shown in Fig. 5. On the other hand,
232: the shape of
233: $|v({\bf r})|^2$ is not affected. The weaker satellite peaks in Fig. 1
234: can be viewed as strong-coupling precursors of this phenomenon. It would be
235: interesting to test this particular prediction of an eight-fold
236: symmetry in $|u({\bf r})|^2$ experimentally, using appropriate candidate
237: impurity atoms with intermediate scattering strengths.
238:
239: {\it Conclusions:} In summary, we observe that (i) using the
240: Bogoliubov-de Gennes equations for single impurities in
241: $\rm d_{x^2 - y^2}$-wave superconductors in the continuum limit
242: is a very useful approach.
243: (ii) Both Zn and Ni impurities produce bound states
244: with a four-fold symmetry, localized around the impurity sites.
245: (iii) Impurities with intermediate scattering strength may have
246: a bound state with a dominant eight-fold symmetry. The simple analysis
247: of the BdG equations we have presented in this work provides a
248: semi-quantitative picture of the spatial structure of bound states
249: in anisotropic superconductors.
250:
251: Recent experiments suggest that the electron-doped high-temperature
252: superconductors NCCO and PCCO also have a d-wave order parameter.
253: \cite{tsuei,kokales,prozorov}
254: On the other hand, the effects of Zn and Ni doping in the electron-doped
255: systems appear to be opposite from the hole-doped case, i.e. Ni impurities
256: lead to a stronger suppression
257: of the superconducting ordering temperature than
258: Zn doping.\cite{dastuto}
259: It is therefore likely that the roles of Zn and Ni are
260: interchanged in the electron-doped high-$\rm T_c$ cuprates.
261: Furthermore, there are indications that the $\rm \kappa-(ET)_2$-salts
262: have a d-wave superconducting order parameter as well.
263: \cite{carrington,pinteric} Therefore, similar studies of impurity
264: bound states in these compounds would be of great interest.
265:
266: We thank Hyekyung Won for useful discussions.
267: S. H. acknowledges the Zumberge foundation for financial support.
268:
269:
270:
271:
272: \begin{references}
273:
274: \bibitem{hirschfeld} P. J. Hirschfeld and N. Goldenfeld, Phys. Rev. B
275: {\bf 48}, 4219 (1993).
276:
277: \bibitem{sun} Y. Sun and K. Maki, Phys. Rev. B {\bf 51}, 6059 (1995);
278: Europhys. Lett. {\bf 32}, 385 (1995).
279:
280: \bibitem{haas} S. Haas, A. V. Balatsky, M. Sigrist, and T. M. Rice,
281: Phys. Rev. B {\bf 56}, 5108 (1997).
282:
283: \bibitem{maki} K. Maki and H. Won, J. Phys. I (Paris), {\bf 6},
284: 2317 (1996).
285:
286: \bibitem{byers} I. M. Byers, M. E. Flatte, and D. J. Scalapino,
287: Phys. Rev. Lett. {\bf 712}, 3363 (1993);
288: M. E. Flatte and I. M. Byers, Solid State Phys. {\bf 52}, 137 (1999).
289:
290: \bibitem{salkola} M. Salkola, A. V. Balatsky, and J. R. Schrieffer,
291: Phys. Rev. B {\bf 55}, 12648 (1997);
292: I. Martin and A. V. Balatsky, cond-mat/0003142.
293:
294: \bibitem{tsuchiura} H. Tsuchiura, Y. Tanaka, M. Ogata, and
295: S. Kashimaya, Phys. Rev. Lett. {\bf 84}, 3165 (2000).
296:
297: \bibitem{pan} S. H. Pan, E. W. Hudson, K. M. Lang, H. Eisaki,
298: S. Uchida, and J. C. Davis, Nature {\bf 403}, 746 (2000).;
299: E. W. Hudson, S. H. Pan, A. K. Gupta, K. W. Ng, and J. C. Davis,
300: Science {\bf 285}, 88 (1999). In this second
301: reference, an asymptotic
302: exponential decay, $|u({\bf r})|^2 \sim \exp{(-2\gamma r)}$, is
303: observed with $\gamma = (1.97 nm)^{-1}$. With $\gamma \approx
304: \sqrt{\Delta^2 - E_0^2}/v_F$, we find that $v_F \approx
305: 3\times10^7 cm/s$ with $\Delta =45 meV$ and $E_0 = -1.3 meV$.
306: This number appears to be quite reasonable.
307:
308: \bibitem{shiba} H. Shiba, Prog. Theor. Phys. {\bf 40}, 435 (1968).
309:
310: \bibitem{yazdani} A. Yazdani, B. A. Jones, C. P. Lutz, M. F. Crommie,
311: and D. M. Engler, Science {\bf 275}, 1767 (1997).
312:
313: \bibitem{degennes} see for example P. G. de Gennes, ``Superconductivity of
314: Metals and Alloys" (W. A. Benjamin, New York, 1966), pp. 137 - 144.
315:
316: \bibitem{morita} Y. Morita, M. Kohmoto, and K. Maki,
317: Europhys. Lett. {\bf 40}, 207 (1997).
318:
319: \bibitem{puchkaryov} E. Puchkaryov and K. Maki, Europhys. J. B
320: {\bf 4}, 191 (1998).
321:
322: \bibitem{chiao} See for example M. Chiao, R. W. Hill,
323: C. Lupien, B. Popic, R. Gagnon, and L. Taillefer, Phys. Rev. Lett.
324: {\bf 82}, 2943 (1999); cond-mat/9910367.
325:
326: \bibitem{gygi} See for example F. Gygi and M. Schl\"uter,
327: Phys. Rev. B {\bf 41} 822 (1990).
328:
329: \bibitem{tsuei} C. C. Tsuei and J. R. Kirtley, cond-mat/0002341.
330:
331: \bibitem{kokales} J. D. Kokales, P. Fournier, L. V. Mercaldo,
332: V. V. Talanov, R. L. Greene, and S. M. Anlage, cond-mat/0002300.
333:
334: \bibitem{prozorov} R. Prozorov, R. W. Giannetta, P. Fournier,
335: and R. L. Greene, cond-mat/0002301.
336:
337: \bibitem{dastuto} For a comprehensive summary see M. D'Astuto and
338: M. Acquarone, ``Effects of the substitution of the copper sites in
339: $\rm Nd_{1-x}Ce_xCuO_4$", in {\it High-Temperature Superconductivity:
340: Models and Measurements}, edited by M. Acquarone
341: (World Scientific, Singapore, 1996),
342: pp. 343 - 350 . See also I. Felner,
343: D. Hechel, and V. Yaron, Physica C {\bf 165},
344: 247 (1990); J. M. Tarascon, E. Eang, S. Kivelson, B. G. Bagley, G. W. Hull,
345: and R. Ramesh, Phys. Rev. B {\bf 42}, 218 (1990).
346:
347: \bibitem{carrington} A. Carrington, I. J. Bonalde, R. Prozorov,
348: R. W. Giannetta,
349: A. M. Kini, J. Schlueter, H. H. Wang, U. Geiser, and J. M. Williams,
350: Phys. Rev. Lett. {\bf 83}, 4172 (1999).
351:
352: \bibitem{pinteric} M. Pinteric, S. Tomic, M. Prester, D. Drobac, O. Milat,
353: K. Maki, D. Schweitzer, I. Heinen, and W. Strunz, Phys. Rev. B {\bf 61},
354: 7033 (2000).
355:
356:
357:
358: \end{references}
359:
360: \newpage
361:
362: \begin{figure}[h]
363: \centerline{\psfig{figure=fig11.eps,width=6cm,angle=0}}
364: \vspace{0.3cm}
365: \caption{
366: Spatial variation of the local tunneling conductance, centered at a
367: strong-scattering impurity, such as Zn, in a $\rm d_{x^2 - y^2}$-wave
368: superconductor. In this figure, the dominant contribution
369: $|u({\bf r})|^2$ at the
370: positive bound state resonant frequency $+E_0$ is shown.
371: }
372: \end{figure}
373:
374: \begin{figure}[h]
375: \centerline{\psfig{figure=fig22.eps,width=6cm,angle=0}}
376: \vspace{0.3cm}
377: \caption{
378: Spatial variation of the local tunneling conductance, localized around a
379: strong-scattering impurity, such as Zn, in a $\rm d_{x^2 - y^2}$-wave
380: superconductor. Here the dominant contribution
381: $|v({\bf r})|^2$ at the
382: negative bound state resonant frequency $-E_0$ is shown.
383: }
384: \end{figure}
385:
386: \begin{figure}[h]
387: \centerline{\psfig{figure=fig33.eps,width=6cm,angle=0}}
388: \vspace{0.3cm}
389: \caption{
390: Spatial variation of the local tunneling conductance, centered at a
391: weak-scattering impurity, such as Ni, in a $\rm d_{x^2 - y^2}$-wave
392: superconductor. In this figure, the dominant contribution
393: $|u({\bf r})|^2$ at the
394: positive bound state resonant frequency $+E_0$ is shown.
395: }
396: \end{figure}
397:
398: \begin{figure}[h]
399: \centerline{\psfig{figure=fig44.eps,width=6cm,angle=0}}
400: \vspace{0.3cm}
401: \caption{
402: Spatial variation of the local tunneling conductance, localized around a
403: weak-scattering impurity, such as Ni, in a $\rm d_{x^2 - y^2}$-wave
404: superconductor. Here the dominant contribution
405: $|v({\bf r})|^2$ at the
406: negative bound state resonant frequency $-E_0$ is shown.
407: }
408: \end{figure}
409:
410: \begin{figure}[h]
411: \centerline{\psfig{figure=fig55.eps,width=6cm,angle=0}}
412: \vspace{0.3cm}
413: \caption{
414: Spatial variation of the local tunneling conductance, centered at an
415: impurity of intermediate scattering strength. Here
416: the dominant contribution
417: $|u({\bf r})|^2$ at the
418: positive bound state resonant frequency $+E_0$ is shown.
419: In contrast to the weak and strong scattering limits,
420: a dominant eight-fold symmetry is observed.
421: }
422: \end{figure}
423:
424: \end{document}
425:
426: