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{Temperature dependence of impurity bound states 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: We study the evolution with temperature

36: of quasiparticle bound states around non-magnetic impurities 

37: in $\rm d_{x^2 - y^2}$-wave superconductors. The

38: associated local density of states has a fourfold symmetry 

39: which has recently been observed in Zn-doped Bi2212

40: using scanning tunneling microscopy (STM).  

41: From the corresponding Bogoliubov-de Gennes equation we find

42: that with increasing temperature the 

43: magnitude of the bound state energy increases and

44: the amplitude of the fourfold contribution to the spinor wave 

45: functions decreases. In the pseudogap regime above $\rm T_c$ the 

46: fourfold angular dependence of the local tunneling conductance 

47: persists as long as the superconducting fluctuations 

48: are sufficiently strong to support a finite local order 

49: parameter. Once the gap function vanishes completely, the

50: angular structure of the bound state

51: wave function becomes featureless. These effects should be observable

52: in STM studies of impurity doped high-temperature superconductors. 

53:   

54: \end{abstract}

55: \pacs{}

56: %\vskip2pc]

57: 

58: 

59: {\it Introduction:}

60: It is well known that impurity doping, such as Zn or Ni substitution

61: in the high-temperature superconductors LSCO, YBCO, and BSCCO, is

62: a useful tool in demonstrating the underlying nodal structure 

63: of their order parameter. It  also

64: provides a semi-quantitative test of the BCS theory for 

65: $\rm d_{x^2 - y^2}$-wave superconductors.\cite{hirschfeld,haas,sun,maki}

66: In particular, an examination of the thermodynamic and transport properties

67: in these compounds suggests that Zn impurities can be modeled with a 

68: scattering potential in the unitary limit.\cite{sun,maki}

69: On the other hand, little is known experimentally about the local 

70: structure of the impurity bound states around the Zn sites despite

71: numerous theoretical studies on this question.\cite{salkola,haas2}

72: Recently, Pan {\it et al.} have provided the first scanning tunneling

73: microscope (STM) images of the local tunneling conductance around

74: Zn impurities in Bi2212 at low temperatures, T $\approx$ 4.3 K, and fixed at 

75: $\pm E$, where $E \approx \Delta/30$ is the binding energy of

76: the impurity bound state.\cite{pan}

77: The corresponding wave function was shown to exhibit a fourfold angular 

78: symmetry associated with the underlying $\rm d_{x^2 - y^2}$-wave order

79: parameter.

80: We have found that 

81: this quasiparticle bound state wave

82: function around impurity sites can be described within the formalism

83: of the Bogoliubov

84: - de Gennes (BdG) equations for $\rm d_{x^2 - y^2}$-wave superconductors.

85: \cite{haas2} Within this framework, the main features of these observations,

86: in particular the characteristic angular patterns seen in

87: the STM images of Zn-doped

88: BSCCO, were reproduced. 

89: 

90: While this previous work was based on an analysis of the BdG equations

91: at zero temperature, the 

92: thermal evolution of the properties of the impurity bound state 

93: is of particular interest. 

94: Here we will study 

95: how the angular STM patterns around impurities may provide valuable

96: information regarding the nature of the pseudogap in the 

97: temperature regime above $T_c$. 

98: We will assume that

99: the origin of the pseudogap is dominated by standard 

100: superconducting fluctuations.\cite{won,varlamov,martin,morr} 

101: This assumption is motivated

102: by recent angle-resolved 

103: photoemission spectroscopy (ARPES) measurements

104: on underdoped BSCCO which have clearly indicated

105: an angular dependence of the pseudogap feature proportional to 

106: $\cos^2{(2 \phi)}$, analogous to $|\Delta({\bf k})|^2$ in the 

107: $\rm d_{x^2 - y^2}$-wave superconducting phase of this compound.

108: \cite{ding}  

109: In the following, we will address the consequences 

110: of the pseudogap phenomenon on the bound state wave function 

111: in a semi-phenomenological manner. In particular, we will investigate

112: how  the characteristic

113: fourfold symmetry pattern in the wave function gradually disappears

114: as the temperature is increased. 

115: 

116: {\it Bogoliubov - de Gennes equations:}

117: Let us start our analysis by 

118: examining the BdG equations for the spinor wave functions in

119: a $\rm d_{x^2 - y^2}$-wave superconductor, given by\cite{haas2}  

120: \bea

121: E u({\bf r}) &=& \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)

122: u({\bf r}) + \frac{1}{p_F^2} \Delta (\partial_x^2 - \partial_y^2 )

123: v({\bf r}),\\

124: E v({\bf r}) &=& - \left( -\frac{\nabla^2}{2 m} - \mu - V ({\bf r}) \right)

125: v({\bf r}) + \frac{1}{p_F^2} \Delta (\partial_x^2 - \partial_y^2 )

126: u({\bf r}),

127: \eea

128: where $\mu$ is the chemical potential, $p_F$ is the Fermi momentum, and 

129: $V ({\bf r}) = V_0 \delta^2 ({\bf r}) > 0$ is an isotropic

130: impurity scattering potential, centered at ${\bf r} = 0$.

131: In previous work it was shown that the spinor wave functions $u({\bf r})$

132: and $v({\bf r})$ can be expanded in terms of Bessel functions of the 

133: first kind, leading to the Ansatz\cite{haas2}

134: \bea

135: u({\bf r}) &=&  A \exp{(- \gamma r)} \left( J_0(p_F r) + \sqrt{2} \beta

136: J_4(p_F r) \cos{(4 \phi)} \right),\\

137:  v({\bf r}) &=&  \sqrt{2} A \alpha \exp{(- \gamma r)} J_2(p_F r) \cos{(2 \phi)},

138: \eea

139: where $J_l(p_F r)$ are Bessel functions of the first kind,

140: $\alpha$, $\beta$, and $\gamma$ are variational 

141: parameters, and $A$ is a global normalization factor.

142: In this expansion, only

143: the leading-order angle-dependent terms have 

144: been retained. 

145: 

146: Since the spinor wave function $v({\bf r})$ does not contain an 

147: s-wave component, it can be eliminated from Eq. (1), yielding  

148: \bea

149: E u({\bf r}) = \left( K - V

150: + \frac{\Delta^2 (1 + \cos{(4 \phi)})}{2 ( K + E)} \right)

151: u({\bf r}),

152: \eea

153: where the contributions of the kinetic energy and the impurity

154: scattering potential are given by

155: \bea

156: K &\equiv &  \frac{\int_0^{\infty} dr r \left[

157: \left( \partial_r \exp{(-\gamma r)}

158: J_l(p_F r)\right)^2 + \left( l \exp{(-\gamma r)} J_l(p_F r)/r \right)^2

159: \right]}

160: {2m \int_0^{\infty} dr r \left( \exp{(-\gamma r)} J_l(p_F r)\right)^2}

161: -\mu

162: \simeq \frac{\gamma^2}{2 m},\\

163: V &\equiv & \frac{\int_0^{\infty} dr r \exp{(-2 \gamma r)} J_0^2(p_F r)

164: V({\bf r})}

165: {\int_0^{\infty} dr r \exp{(-2 \gamma r)} J_0^2(p_F r)}

166: \simeq (2 \pi \gamma p_F) \int_0^{\infty} dr r \exp{(-2 \gamma r)}

167: J_0^2(p_F r) V ({\bf r}).

168: \eea

169: Note that the contribution of the impurity potential 

170: $V$ does not appear in the 

171: denominator of the last term in Eq. (5) because

172: the spinor wave function

173: $v({\bf r})$ does not have an s-wave component.

174: Furthermore, the kinetic energy contribution reduces to 

175: $\gamma^2/2 m$ for all $J_l(p_F r)$ with $l \ll p_F/\gamma$.\cite{haas2} 

176: The two parameters of interest which will be determined in the following

177: are the temperature-dependent binding energy $E(T)$ and the 

178: coefficient of the fourfold symmetric term $\beta (T)$. 

179: Because $u({\bf r})$ contains

180: two orthogonal angular components, and the kinetic energy $K$ is 

181: independent of angular momentum, it follows that 

182: \bea 

183: \left( E - K + V - \frac{\Delta^2}{2(E + K)} \right)

184: - \frac{\beta \Delta^2}{2\sqrt{2}(E + K)} & = & 0, \\

185: \left( E - K - \frac{\Delta^2}{2(E + K)} \right)\beta

186: - \frac{\Delta^2}{2\sqrt{2}(E + K)} & = & 0.

187: \eea

188: Solving these two equations for $E$ and $\beta$, one obtains

189: \bea

190: \beta &=& \frac{\Delta^2}{2\sqrt{2}(E^2 - K^2 - \Delta^2/2)},\\

191: \Delta^4/8 &=& \left( E^2 - K^2 + V(K + E) - \Delta^2/2 \right)

192: \left( E^2 - K^2 - \Delta^2/2 \right).

193: \eea

194: The quartic equation for $E$ contains 

195: one root which will determine the bound state energy.

196: 

197: To make further progress, 

198: we will assume that

199: (i) the impurity scattering is in the unitary limit, $V_0 \agt \Delta_0$,

200: leading to a 

201: small binding energy E especially at low temperatures, 

202: (ii) the energy contributions

203: $K$ and $V$ are independent of 

204: temperature in the regime of interest,

205: $T \in [0, 3 T_c]$,

206: and (iii) the local amplitude of the superconducting

207: gap function $\Delta (T)$ is small but finite in the 

208: pseudogap regime above $T_c$. 

209: 

210: {\it Superconducting regime:}

211: Let us first consider the zero-temperature limit, assuming that 

212: the impurity scattering potential is in the unitary scattering 

213: limit. In this case, $E \rightarrow 0$, leading to

214: \bea

215: V \rightarrow \frac{1}{K^2 + \Delta_0^2/2} \left[ K(K^2 + \Delta_0^2) 

216: + \Delta_0^4/8K \right].

217: \eea

218: Following the assumption of temperature-independent

219: energy contributions

220: $K$ and $V$, we can set $K = a \Delta_0$, where $\Delta_0$

221: is the gap amplitude at $T = 0$ , and $a$ is a proportionality

222: constant whose physical range can be inferred from experiments. 

223: It then follows that

224: \bea

225: V \simeq \frac{1 + 8 a^2(1 + a^2)}{4a (1 + 2 a^2)}\Delta_0.

226: \eea

227: Furthermore, it is known that at zero temperature

228: $\beta < 0$, leading to the characteristic fourfold symmetric

229: patterns in the local density of states around impurity

230: sites that have been observed in recent

231: STM measurements.\cite{pan} It is 

232: therefore natural to postulate that $\beta < 0$ for all temperatures,

233: imposing the constraint: $a >  2^{-3/4} = 0.5946$.

234: Therefore, we will only consider the

235: physical parameter regime $a >  2^{-3/4}$.

236: For example, we can set $a = 2/3$, yielding

237: $\beta = -9/(17\sqrt{2}) = -0.3743$ at T=0. We have verified

238: that the finite-temperature

239: properties of the bound state

240: that will be discussed in the following do not depend strongly on

241: the choice of $a$. 

242: 

243: {\it Normal state:}

244: In the opposite limit, $T \gg T_c$, the gap function vanishes

245: $\Delta^2 (T \rightarrow \infty)

246: \rightarrow 0$. Assuming that $K$ and $V$ are basically 

247: unaffected by temperature, this gives

248: \bea

249: E(T \rightarrow \infty )

250:  = K - V = - \frac{(1 + 4 a^2)}{4a (1 + 2 a^2)}\Delta_0.

251: \eea

252: Hence, the magnitude of the impurity bound state energy $E$

253: increases with increasing temperature.

254: The parameter $\beta$ vanishes in this limit, indicating that the 

255: bound state wave function loses its fourfold angular symmetry 

256: pattern at high temperatures when $\Delta^2(T) \rightarrow 0$.

257:  

258: {\it Pseudogap regime:}

259: In the pseudogap regime the amplitude of the local order 

260: parameter, $| \Delta(T)|$, is assumed to be small 

261: but finite, and by perturbing about the limit $T \rightarrow \infty$

262: one finds to leading order that

263: \bea

264: E(T > T_c) \approx E(T \rightarrow \infty ) + \frac{\Delta^2(T)}{2(K + E)}

265: \simeq  - \frac{(1 + 4 a^2)}{4a (1 + 2 a^2)}\Delta_0

266: + \frac{2a (1 + 2 a^2) \Delta^2(T) }{(8 a^4 - 1) \Delta_0},

267: \eea

268: and

269: \bea

270: \beta(T > T_c) \approx - \frac{16 a^2 (1 + 2 a^2)^2 \Delta^2(T) }

271: {2\sqrt{2} (64 a^6(1 + a^2) - (1 + 8a^2))\Delta_0^2 }.

272: \eea

273: Hence the magnitude of the binding energy $E(T)$ decreases as

274: $T \rightarrow 0$ due to the progressive opening of the energy gap. 

275: 

276: Provided that the pseudogap above $T_c$

277: arises mainly due to superconducting fluctuations

278: \cite{won,varlamov,martin,morr}, 

279: it is straightforward to incorporate its physical consequences

280: by assuming a small but finite amplitude of the local order parameter

281: above $T_c$.

282: This can be modeled by

283: \bea

284: \Delta^2(T)=\frac{1}{2}\left[\Delta^2_0(1-(T/T_c)^3)^2

285: + {\sqrt{\Delta^4_0(1 - (T/T_c)^3)^4 + C }} \right],

286: \eea

287: where $C = 0.027572$. The coefficient $C$ is

288: obtained from 

289: \bea

290: \frac{C}{4 \Delta^4_0((T/T_c)^3 - 1)^2} 

291: \simeq \frac{2\pi T}{m \xi^2_0 \ln{(T/T_c)}},

292: \eea

293: where the right hand side is

294: the spatial average of the fluctuation order parameter, and

295: $\xi_0 = 73 (3) v^2/2(2\pi T)^2$. The approximate

296: ratio $T_c/E_F \simeq 0.03$

297: in Bi2212 has been deduced from the low-temperature behavior

298: of the thermal conductivity.\cite{chiao} This interpolation 

299: formula for $\Delta^2(T)$ has already

300: proven successful in the analysis 

301: of the excess Dingle temperatures in the vortex state of the 

302: $\kappa - (ET)_2$ salts.\cite{maki2,ito}

303: 

304: In Fig. 1, the temperature-dependent gap amplitude is plotted 

305: along with the energy of the bound state and the coefficient 

306: of the fourfold symmetry term in the spinor wave function 

307: $u({\bf r})$. The gap amplitude at zero-temperature has been

308: set equal to unity, $\Delta_0 \equiv 1$. At very small temperatures,

309: we observe that

310: $\Delta(T \rightarrow 0 ) \rightarrow 1$, $E (T \rightarrow 0)

311: \rightarrow 0$, and $\beta (T \rightarrow 0) \rightarrow  

312: -(\sqrt{2}(a^2 +1))^{-1}$, as expected from the discussion 

313: of the $T \rightarrow 0$ limit.  In the limit $T \rightarrow \infty$, we

314: find that $\Delta(T \rightarrow 0 ) \rightarrow 0$,

315: $E (T \rightarrow -(1+4a^2)(4a + 8a^3)^{-1}$, and

316: $\beta (T \rightarrow 0) \rightarrow 0$, implying that the 

317: magnitude of the binding energy increases with temperature, and 

318: that the fourfold angular features in the spinor wave function 

319: disappear 

320: at high temperatures. The intermediate-temperature

321: regime of $E(T)$ and $\beta (T)$ in Fig. 1 has been

322: determined by a numerical solution of the coupled equations

323: (10) and (11). In the fluctuation regime

324: around $T_c$, the coefficient $\beta (T \approx T_c )$ is found

325: to be appreciable. Therefore, remnants of the characteristic

326: $\cos^2{(2 \phi)}$-dependence in the square of the bound state 

327: spinor wave function $u({\bf r})$ should be reflected 

328: in the local tunneling density of states, measured by STM

329: experiments in the pseudogap regime above $T_c$.

330: In Fig. 2, the corresponding temperature evolution of 

331: $|u({\bf r})|^2$ is shown. 

332: 

333: 

334: {\it Conclusions:}

335: In summary, we have studied the effect of temperature on the 

336: bound state wave function around impurities in 

337: $\rm d_{x^2 - y^2}$-wave superconductors. Within the framework 

338: of the BdG equations,

339: the magnitude of the binding energy is found to increase with 

340: temperature, but stays on the order of $\Delta_0$. On the other 

341: hand, the fourfold contribution to the local density of states

342: disappears gradually with increasing temperature. Therefore it 

343: appears to be possible to take a ``snapshot" of the pseudogap 

344: by STM imaging. The present analysis may also be applicable to other 

345: unconventional superconductors, such as the layered organic

346: superconductors $\rm \kappa-(ET)_2 Cu[N(CN)_2]Br$ and

347: $\rm \kappa-(ET)_2 Cu(NCS)_2$, and $\rm Sr_2RuO_4$.

348: 

349: 

350: We thank Alexander Balatsky, Ivar Martin,  Bruce Normand, and 

351: Hyekyung Won for useful discussions.

352: S. H. acknowledges the Zumberge foundation for financial support. 

353: 

354: 

355: 

356: 

357: \begin{references}

358: 

359: \bibitem{hirschfeld} P. J. Hirschfeld and N. Goldenfeld, 

360: Phys. Rev. B {\bf 48}, 4219 (1993).

361: 

362: \bibitem{haas} S. Haas, A. V. Balatsky, M. Sigrist, and

363: T. M. Rice, Phys. Rev. B {\bf 56}, 5108 (1997).

364: 

365: \bibitem{sun} Y. Sun and K. Maki, Phys. Rev. B {\bf 51},

366: 6059 (1995); Europhys. Lett. {\bf 32}, 385 (1995).

367: 

368: \bibitem{maki} K. Maki and H. Won, Ann. Phys. (Leipzig)

369: {\bf 5} 320 (1996); J. Phys. I (France), {\bf 6}, 2317

370: (1996).

371: 

372: \bibitem{salkola} A. V. Balatsky, M. Salkola, A. Rosengren,

373: Phys. Rev. B {\bf 51}, 15547; M. Salkola, A. V. Balatsky, and J. R. Schrieffer,

374: Phys. Rev. B {\bf 55}, 12648 (1997);

375: H. V. Kruis, I. Martin, A. V. Balatsky, cond-mat/0008349.

376:  

377: \bibitem{haas2} S. Haas and K. Maki, 

378: Phys. Rev. Lett. {\bf 85}, 2172 (2000); K. Maki and

379: S. Haas, to be published in Phys. Rev. B. 

380: 

381: \bibitem{pan} S. H. Pan, E. W. Hudson, K. M. Lang, H. Eisaki,

382: S. Uchida, and J. C. Davis, Nature {\bf 403}, 746 (2000).

383: 

384: \bibitem{won} H. Won and K. Maki, Physica C

385: {\bf 282 - 287}, 1839 (1997);

386: H. Won and K. Maki, J. Phys. IV France

387: {\bf 9}, Pr10-283 (1999).

388: 

389: \bibitem{varlamov} A. A. Varlamov, G. Balestino, E. Milani, and

390: D. V. Livanov, Adv. Phys. {\bf 48}, 655 (1999);

391: L. B. Ioffe, A. I. Larkin, A. A. Varlamov, and L. Yu,

392: Phys. Rev. B {\bf 47}, 8936 (1993);

393: V. V. Dorin, R. A. Klemm, A. A. Varlamov, A. I. Buzdin, and 

394: D. V. Livanov, Phys. Rev. B. {\bf 48}, 12951 (1993).

395: 

396: \bibitem{martin} I. Martin and A. V. Balatsky, Phys. Rev. B {\bf 62}, 

397: R6124 (2000).

398: 

399: \bibitem{morr} H. Westfahl Jr. and D.K. Morr, Phys. Rev. B {\bf 62}, 

400: 5891 (2000).

401: 

402: \bibitem{ding} H. Ding, M. R. Norman, T. Yokoya, T. Takeuchi, 

403: M. Randeria, J. C. Campuzano, T. Takahashi, T.

404: Mochiku, and K. Kadowaki, Phys. Rev. Lett. {\bf 78}, 2628 (1997);

405: T. Timusk and B. Statt, Rep. Prog. Phys. {\bf 62}, 61 (1999).

406: 

407: \bibitem{chiao} M. Chiao, R. W. Hill, C. Lupien, B. Popi\'c,

408: R. Gagnon, and L. Taillefer, Phys. Rev. Lett. {\bf 82}, 2943 (1999).

409: 

410: \bibitem{maki2} K. Maki, Phys. Rev. B {\bf 44} 2861 (1991).

411: 

412: \bibitem{ito} H. Ito, N. Clayton, P. J. Meeson, M. Springford,

413: and G. Saito, proceedings of ICSM 2000 (Badgastein, July 2000),

414: Synth. Metals (in press).

415: 

416: \end{references}

417: 

418: \newpage

419: 

420: \begin{figure}[h]

421: \centerline{\psfig{figure=fig1.eps,width=9cm,angle=0}}

422: \vspace{0.3cm}

423: \caption{

424: Temperature dependence of the gap amplitude $\Delta$ (solid line),

425: the bound state energy $E$ (dashed line), and the coefficient of

426: the fourfold symmetry term in the bound state wave function $\beta$

427: (dot-dashed line). The proportionality constant $a$ has been set

428: to $a = 2/3$.

429: }

430: \end{figure}

431: \begin{figure}[h]

432: \vspace{-0.9cm}

433: \centerline{\psfig{figure=0.4.eps,width=6cm,angle=0}

434: \psfig{figure=0.25.eps,width=6cm,angle=0}

435: \psfig{figure=0.0.eps,width=6cm,angle=0}}

436: \vspace{0.3cm}

437: \caption{

438: Temperature dependence of the bound state wave function,

439: $|u({\bf r})|^2$, localized around a

440: strong-scattering impurity in a $\rm d_{x^2 - y^2}$-wave

441: superconductor. left: superconducting regime ($T = 0$),

442: center: pseudogap regime ($T > T_c$), right:

443: normal state ($T = \infty $).

444: }

445: \end{figure}

446: 

447: 

448: 

449: 

450: 

451: \end{document}

452: 

453: