1: %% LyX 1.1 created this file. For more info, see http://www.lyx.org/.
2: %% Do not edit unless you really know what you are doing.
3: \documentclass[twocolumn,english,aps]{revtex4}
4: \usepackage[T1]{fontenc}
5: \usepackage[latin1]{inputenc}
6:
7:
8: \makeatletter
9:
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% LyX specific LaTeX commands.
11: \providecommand{\LyX}{L\kern-.1667em\lower.25em\hbox{Y}\kern-.125emX\@}
12:
13: \makeatother
14: \begin{document}
15:
16: \title{Magnetic effect from non-magnetic impurity in superconducting CuO\( _{2} \)
17: plane}
18:
19:
20: \author{V.M. Loktev}
21:
22:
23: \email{vloktev@gluk.org}
24:
25:
26: \affiliation{Bogolyubov Institute for Theoretical Physics, 14b Metrologichna str.,03143
27: Kiev-143, Ukraine}
28:
29:
30: \author{Yu.G. Pogorelov}
31:
32:
33: \email{ypogorel@fc.up.pt}
34:
35:
36: \affiliation{CFP/Departamento de Fisica, Unversidade do Porto, Rua do Campo Alegre
37: 687, 4169-007 Porto, Portugal}
38:
39: \begin{abstract}
40: We propose a new model for impurity center formed by a cathion substitute
41: for Cu, like Zn, in CuO\( _{2} \) planes. Its main effect on superconducting
42: electrons is due to the non-zero exchange field on O sites, neighbors
43: to the (non-magnetic) impurity. We discuss a strong suppression of
44: \( d \)-wave order parameter, a zero-energy resonance in local density
45: of states, and spin polarization of charge carriers, which can be
46: related to the experimentally observed effects in Zn-doped copper
47: oxides. These results are obtained \emph{without} using the unitary
48: scattering limit.
49: \end{abstract}
50: \maketitle
51:
52:
53: It is recognized that there are two types of impurities in high-Tc
54: superconductors (HTSC): i) intrinsic or {}``own{}'', and ii) extrinsic
55: or {}``foreign{}''. The first type are the heterovalent impurities
56: or oxygen vacancies, that is the dopants. They supply charge carriers
57: into insulating antiferromagnetic (AFM) cuprate planes enabling their
58: metallization \cite{And1}, but also they act as scattering centers
59: for carriers. Previously, we showed \cite{LP1, LP2, LP3} that formation
60: of superconducting (SC) condensate, either of \( s \)- and \( d \)-
61: symmetry, is possible at low enough concentration \( c \) of such
62: impurities but is prevented by growing fluctuations of SC order parameter
63: at higher \( c \).
64:
65: The second type are the homovalent impurities, which only produce
66: scattering of existing charge carriers and so can depress the SC properties
67: of HTSC systems. They are well studied in common SC metals, where
68: it was stated by Anderson \cite{And} that non-magnetic impurities
69: have practically no effect on SC characteristics. At the same time,
70: even low concentration of paramagnetic ions can completely destroy
71: SC order, as initially shown in the Born approximation by Abrikosov
72: and Gor'kov \cite{AG} and then confirmed by many authors under more
73: general scope \cite{Maki, Shiba, Z-MH}.
74:
75: But in the case of SC copper oxides, an apparent violation of this
76: so well theoretically based phenomenological principle was detected.
77: Thus, introducing of non-magnetic Zn\( ^{2+} \) ions instead of Cu\( ^{2+} \)
78: into the cuprate planes has a suppression effect on HTSC not weaker
79: but rather stronger than that by magnetic Ni\( ^{2+} \) ions \cite{Zhang, Bonn}.
80: This triggered an idea of viewing the non-magnetic impurity ions in
81: HTSC as extremely strong scatterers \cite{Chien} so that their perturbation
82: potential \( V_{imp} \) is the biggest energy parameter, treated
83: in the unitary limit: \( V_{imp}/W\gg 1 \) (\( W \) the bandwidth).
84: This concept was extensively elaborated \cite{Lee, Hirsch, Fehr},
85: the principal conclusions being the finite density of quasiparticle
86: states (DOS) at the Fermi level \( \varepsilon _{{\rm F}} \): \( \rho (\varepsilon \to \varepsilon _{{\rm F}})\to \rho _{u}\neq 0 \),
87: and the universal value of quasiparticle conductivity \( \sigma (\omega \to 0)\to \sigma _{u}\neq 0 \).
88: However, apart from the still existing controversies about those predictions
89: \cite{Atk}, it should be noted that, unlike the dopants, the foreign
90: impurity centers are formed in the CuO\( _{2} \) plane by a \emph{homovalent}
91: substitution (as Zn\( ^{2+} \) or Ni\( ^{2+} \) for Cu\( ^{2+} \)),
92: and it is problematic how they could produce such a strong perturbation
93: potential. Also we notice that the heterovalent non-magnetic scatterers
94: by dopants can not produce such effects \cite{LP2}.
95:
96: This Letter is aimed on an alternative approach to the problem of
97: foreign impurities. It will be shown that irrespectively of the type
98: (magnetic or non-magnetic) of the cathion substitute in CuO\( _{2} \)
99: plane, the resulting center acts on charge carriers as \emph{magnetic}.
100: In accordance with the generally accepted notion, such center should
101: in fact strongly suppress SC order either of \( s \)- or \( d \)-type,
102: as was first qualitatively stated yet by Mahajan \emph{et al} \cite{Mah}.
103: We note that similar views on the effect of Zn impurities in HTSC
104: cuprates were expressed in some recent publications \cite{McFarl, Schr, Park},
105: though still focused on unitary scattering. Below we consider the
106: problem of isolated non-magnetic impurity ion in a CuO\( _{2} \)
107: plane and its local effects on the \( d \)-wave SC order parameter,
108: local density of states (LDOS), and the itinerant spin polarization.
109: Our treatment does not need using the unitary limit, nevertheless
110: the effects can be quite strong.
111:
112: Fig. 1 shows a cathion impurity substitute for Cu in a CuO\( _{2} \)
113: plane, like real Zn, Fe, or Ni impurities in high-\( T_{c} \) compounds.
114: Associating the charge carriers mostly to O\( ^{-} \) holes, we conclude
115: that the main perturbation by such impurity (both magnetic and non-magnetic)
116: is due to the fact that its neighbor O sites occur in a non-zero exchange
117: field by Cu\( ^{2+} \) ions \cite{Wal, Jul}, which is equivalent
118: to the effect of magnetic impurity in a common superconductor. On
119: the other hand, there are no reasons to consider any sizeable spin-independent
120: perturbation from such isovalent impurity. The respective model Hamiltonian
121: is \( H=H_{sc}+H_{c}+H_{int} \). The unperturbed SC term \( H_{sc}=\sum _{\mathbf{k}}\psi _{\mathbf{k}}^{\dagger }(\xi _{\mathbf{k}}\hat{\tau }_{3}+\Delta _{\mathbf{k}}\hat{\tau }_{1})\psi _{\mathbf{k}} \)
122: couples the Nambu spinors \( \psi _{\mathbf{k}}^{\dagger }=(a_{\mathbf{k},\uparrow }^{\dagger },a_{-\mathbf{k},\downarrow }) \)
123: with Pauli matrices \( \hat{\tau }_{i} \). The normal dispersion
124: \( \xi _{\mathbf{k}}=W(2-\cos ak_{x}-\cos ak_{y})/4-\varepsilon _{{\rm F}} \)
125: leads to the density of states (DOS) \( \rho _{0}=4/(\pi W) \). The
126: \( d \)-wave gap function \( \Delta _{\mathbf{k}}=\Delta \theta (\varepsilon _{{\rm D}}^{2}-\xi _{\mathbf{k}}^{2})\gamma _{\mathbf{k}}/\gamma _{m} \)
127: includes the BCS-shell factor \( \theta (\varepsilon _{{\rm D}}^{2}-\xi _{\mathbf{k}}^{2}) \)
128: with the {}``Debye energy{}'' \( \varepsilon _{{\rm D}} \), the
129: symmetry factor \( \gamma _{\mathbf{k}}=\cos ak_{x}-\cos ak_{y} \)
130: with maximum absolute value \( \gamma _{m}=\pi \varepsilon _{{\rm F}}\rho _{0} \),
131: and the gap parameter\begin{equation}
132: \label{delta}
133: \Delta =VN^{-1}\sum _{\mathbf{k}}\gamma _{\mathbf{k}}\langle a_{-\mathbf{k},\downarrow }a_{\mathbf{k},\uparrow }\rangle
134: \end{equation}
135: where \( V \) is the attraction between two carriers with opposite
136: spins on neighbor O sites. The term \( H_{c}=-hS_{z} \) models the
137: (AFM) correlation between the impurity center and its environment,
138: where \( h\sim J_{dd} \), the Cu-Cu exchange constant, and \( \mathbf{S} \)
139: is the spin of a fictitious {}``magnetic impurity{}''. It can be
140: seen as a cluster of four 1/2 spins of Cu nearest neighbors to real
141: non-magnetic impurity (Fig. 1). In reality, its quantization axis
142: \( z \) is only defined over time periods no longer than \( \tau _{s}\sim \hbar \xi _{s}/(aJ_{dd})\sim 10^{-13} \)
143: s for spin correlation length \( \xi _{s}\sim a/\sqrt{x} \) \cite{Birg}
144: and doping levels \( x\sim 0.1 \) (this also agrees with the NMR
145: data \cite{McFarl}). However this \( \tau _{s} \) is much longer
146: than typical electronic times \( \sim \hbar /\varepsilon _{{\rm F}}\sim 10^{-15} \)
147: s for HTSC compounds. For \( h>0 \) we have \( \langle S_{z}\rangle \equiv s \)
148: and \( 0<s<S \), which accounts for the short-range AFM order, whereas
149: \( s\to 0 \) in the paramagnetic limit \( \beta h\ll 1 \).
150:
151: We separate the spin-dependent interaction between charge carriers
152: and impurity into three parts:\begin{equation}
153: \label{int}
154: H_{int}=H_{int}^{{\rm MF}}+H_{int}^{\parallel }+H_{int}^{\perp },
155: \end{equation}
156: where \[
157: H_{int}^{{\rm MF}}=JsN^{-1}\sum _{\mathbf{k},\mathbf{k}'}\sum _{\sigma =\pm }\alpha _{j,\mathbf{k}}\alpha _{j,\mathbf{k}'}\sigma a_{\mathbf{k}',\sigma }^{\dagger }a_{-\mathbf{k},\sigma }\]
158: is the {}``mean-field{}'' (MF) polarization of carrier spins by
159: the impurity center, and \[
160: H_{int}^{\parallel }=JN^{-1}\sum _{\mathbf{k},\mathbf{k}'}\sum _{\sigma =\pm }\alpha _{j,\mathbf{k}}\alpha _{j,\mathbf{k}'}\sigma (S_{z}-s)a_{\mathbf{k}',\sigma }^{\dagger }a_{\mathbf{k},\sigma },\]
161: \[
162: H_{int}^{{\rm MF}}=JN^{-1}\sum _{\mathbf{k},\mathbf{k}'}\sum _{\sigma =\pm }\alpha _{j,\mathbf{k}}\alpha _{j,\mathbf{k}'}S_{\sigma }a_{\mathbf{k}',-\sigma }^{\dagger }a_{\mathbf{k},\sigma },\]
163: are their interactions with longitudinal and transversal fluctuations
164: of \( \mathbf{S} \). In the paramagnetic limit: \( s\to 0 \), Eq.
165: \ref{int} is reduced to the common Kondo interaction \cite{Kon, Nag}.
166: For definiteness, the Cu-O \( p \)-\( d \) exchange parameter \( J \)
167: is considered positive. The functions\begin{eqnarray*}
168: \alpha _{1,\mathbf{k}}=2\cos \frac{ak_{x}}{2}\cos \frac{ak_{y}}{2},\quad \alpha _{2,\mathbf{k}}=2\cos \frac{ak_{x}}{2}\sin \frac{ak_{y}}{2}, & & \\
169: \alpha _{3,\mathbf{k}}=2\sin \frac{ak_{x}}{2}\cos \frac{ak_{y}}{2},\quad \alpha _{4,\mathbf{k}}=2\sin \frac{ak_{x}}{2}\sin \frac{ak_{y}}{2}, & &
170: \end{eqnarray*}
171: realize 1D irreducible representations of the \( C_{4v} \) point
172: group (of the plaquette surrounding the impurity), so that \( N^{-1}\sum _{\mathbf{k}}\alpha _{j,\mathbf{k}}\alpha _{j',\mathbf{k}}=\delta _{jj'} \).
173: Distinctive features of the perturbation, Eq. \ref{int}, compared
174: to the commonly used impurity models, are: i) its spatial extension
175: expressed by the factors \( \alpha _{j,\mathbf{k}} \), ii) additional
176: degrees of freedom by spin \( \mathbf{S} \), and iii) coupling of
177: \( \mathbf{S} \) to the local AFM correlations.
178:
179: In principle, this impurity center can produce yet another perturbation,
180: due to a possible role of AFM correlated Cu\( ^{2+} \) spins in the
181: SC coupling between charge carriers. Lacking one such spin would locally
182: perturb the \( \Delta _{\mathbf{k}}\hat{\tau }_{1} \) term in \( H_{sc} \)
183: by some expansions in \( \alpha _{j,\mathbf{k}}\alpha _{j,\mathbf{k}'} \).
184: This can influence the SC order, alike the simpler case of point-like
185: perturbation of \( s \)-wave SC coupling \cite{Pog}. However, for
186: simplicity, we leave this kind of perturbation for a separate study.
187:
188: We calculate the averages, like Eq. \ref{delta}, by simple spectral
189: formula at \( T=0 \):\begin{equation}
190: \label{av}
191: \langle ab\rangle =\pi ^{-1}\int _{0}^{\varepsilon _{{\rm F}}}\Im \langle \langle b|a\rangle \rangle _{\varepsilon }d\varepsilon ,
192: \end{equation}
193: where \( \Im f(\varepsilon )=\lim _{\delta \to 0}[f(\varepsilon -i\delta )-f(\varepsilon +i\delta )]/2 \)
194: and \( \langle \langle b|a\rangle \rangle _{\varepsilon \pm i\delta } \)
195: are the retarded and advanced two-time Green functions (GF's). The
196: relevant GF matrix \( \hat{G}_{\mathbf{k},\mathbf{k}'}=\langle \langle \psi _{\mathbf{k}}|\psi _{\mathbf{k}'}^{\dagger }\rangle \rangle \)
197: in absence of impurity perturbation (\( J=0 \)) is momentum-diagonal:
198: \( \hat{G}_{\mathbf{k},\mathbf{k}'}=\delta _{\mathbf{k},\mathbf{k}'}\hat{G}_{\mathbf{k}} \),
199: with \( \hat{G}_{\mathbf{k}}=(\varepsilon -\xi _{\mathbf{k}}\hat{\tau }_{3}-\Delta _{\mathbf{k}}\hat{\tau }_{1})^{-1} \).
200: The same expression holds for the momentum-diagonal GF \( \hat{G}_{\mathbf{k},\mathbf{k}} \)
201: in presence of single impurity, whose effect \( \sim 1/N \) is negligible
202: for this quantity. However it is only this small impurity effect that
203: gives rise to a momentum-non-diagonal GF's \( \hat{G}_{\mathbf{k},\mathbf{k}'} \).
204: They are found from the equation of motion \( \hat{G}_{\mathbf{k},\mathbf{k}'}=JN^{-1}\sum _{\mathbf{k}'',j}\alpha _{j,\mathbf{k}}\hat{G}_{\mathbf{k}}(s\hat{G}_{\mathbf{k}'',\mathbf{k}'}+\hat{G}_{\mathbf{k},\mathbf{k}'}^{(z)}+\hat{G}_{\mathbf{k},\mathbf{k}'}^{(-)})\alpha _{j,\mathbf{k}''} \)
205: including three scattered GF's: the MF one \( \hat{G}_{\mathbf{k}'',\mathbf{k}'} \),
206: the longitudinal \( \hat{G}_{\mathbf{k}'',\mathbf{k}'}^{(z)}=\langle \langle \psi _{\mathbf{k}''}(S_{z}-s)|\psi _{\mathbf{k}'}^{\dagger }\rangle \rangle \)
207: and the transversal \( \hat{G}_{\mathbf{k}'',\mathbf{k}'}^{(-)}=\langle \langle \bar{\psi }_{\mathbf{k}''}S_{-})|\psi _{\mathbf{k}'}^{\dagger }\rangle \rangle \)
208: with \( \bar{\psi }_{\mathbf{k}}^{\dagger }=(a_{\mathbf{k},\downarrow }^{\dagger },a_{-\mathbf{k},\uparrow }) \).
209: The two last terms are analogous to the well known Nagaoka's \( \Gamma \)-term
210: \cite{Nag, Z-MH} and treating them with a similar decoupling procedure
211: gives:\begin{eqnarray*}
212: \hat{G}_{\mathbf{k},\mathbf{k}'}^{(z)}=\frac{J\Sigma ^{2}}{N}\sum _{\mathbf{k}'',j}\alpha _{j,\mathbf{k}}\hat{G}_{\mathbf{k}}\hat{G}_{\mathbf{k}'',\mathbf{k}'}\alpha _{j,\mathbf{k}''},\quad \quad \quad \quad \quad & & \\
213: \hat{G}_{\mathbf{k},\mathbf{k}'}^{(-)}=\frac{J}{N}\sum _{\mathbf{k}'',j}\alpha _{j,\mathbf{k}}\hat{G}_{\mathbf{k}}(\varepsilon +h)\hat{X}_{\mathbf{k}''}\hat{G}_{\mathbf{k}'',\mathbf{k}'}\alpha _{j,\mathbf{k}''}, & &
214: \end{eqnarray*}
215: where \( \Sigma ^{2}=\langle S_{z}^{2}\rangle -s^{2} \), \( \hat{X}_{\mathbf{k}}=S(S+1)-s(s+1)-\Sigma ^{2}+(1+2\xi _{\mathbf{k}}/E_{\mathbf{k}})\hat{\tau }_{3} \),
216: \( E_{\mathbf{k}}=\sqrt{\xi _{\mathbf{k}}^{2}+\Delta _{\mathbf{k}}^{2}} \),
217: and one energy argument is shifted: \( \varepsilon \to \varepsilon +h \),
218: due to the AFM stiffness. Finally, we obtain the decoupled equation
219: of motion:\[
220: \hat{G}_{\mathbf{k},\mathbf{k}'}=N^{-1}\sum _{\mathbf{k}'',j}\alpha _{j,\mathbf{k}}\hat{G}_{\mathbf{k}}[Js+J^{2}(\Sigma ^{2}\hat{G}_{j}+\hat{X}_{j})]\hat{G}_{\mathbf{k}'',\mathbf{k}'}\alpha _{j,\mathbf{k}''}\]
221: where \( \hat{G}_{j}=N^{-1}\sum _{\mathbf{k}}\alpha _{j,\mathbf{k}}^{2}\hat{G}_{\mathbf{k}} \),
222: \( \hat{X}_{j}=N^{-1}\sum _{\mathbf{k}}\alpha _{j,\mathbf{k}}^{2}\hat{G}_{\mathbf{k}}(\varepsilon +h)\hat{X}_{\mathbf{k}} \),
223: and its standard iteration yields in the result:\begin{equation}
224: \label{Tmat}
225: \hat{G}_{\mathbf{k},\mathbf{k}'}=N^{-1}\sum _{j}\alpha _{j,\mathbf{k}}\hat{G}_{\mathbf{k}}\hat{T}_{j}\hat{G}_{\mathbf{k}'}\alpha _{j,\mathbf{k}'},
226: \end{equation}
227: with the partial T-matrices \( \hat{T}_{j}=[Js+J^{2}(\Sigma ^{2}\hat{G}_{j}+\hat{X}_{j})][1-Js-J^{2}(\Sigma ^{2}\hat{G}_{j}+\hat{X}_{j})]^{-1} \).
228: By the definition of our model, the parameter \( Js \) is positive.
229: It is interesting to trace the behavior of \( \hat{T}_{j} \) in the
230: two characteristic limits for AFM correlations between Cu\( ^{2+} \)spins.
231:
232: In the paramagnetic limit: \( h\to 0 \), \( s\to 0 \), we have \( \Sigma ^{2}\to S(S+1)/3 \)
233: and \( \hat{X}_{j}\to 2S(S+1)/3-N^{-1}\sum _{\mathbf{k}}\alpha _{j,\mathbf{k}}^{2}(1+2\xi _{\mathbf{k}}/E_{\mathbf{k}})\hat{G}_{\mathbf{k}}\hat{\tau }_{3} \).
234: In neglect of the small last term we arrive at:\[
235: \hat{T}_{j}\to J^{2}S(S+1)\hat{G}_{j}[1-J^{2}S(S+1)\hat{G}_{j}]^{-1}\]
236: generalizing the known results \cite{AG, Z-MH} for the case of extended
237: impurity center.
238:
239: Another limit, fully polarized, \( h\to \infty \), \( s\to S \),
240: corresponds to \( \Sigma ^{2}\to 0 \), \( \hat{X}_{j}\to 0 \) and
241: results in\begin{equation}
242: \label{pol}
243: \hat{T}_{j}\to JS(1-JS\hat{G}_{j})^{-1}
244: \end{equation}
245: which is only due to the effect of MF magnetic scattering. The obvious
246: validity condition for this limit, \( JS\gg k_{{\rm B}}T \), well
247: applies in the SC phase at \( T<T_{c}\sim \Delta /k_{{\rm B}} \),
248: so we use Eq. \ref{pol} for the T-matrices in what follows.
249:
250: The local SC correlation is characterized by the average \( \Delta _{12}=2V\langle a_{\delta _{1},\downarrow }a_{\delta _{2},\uparrow }\rangle \)
251: (see Fig. 1) where a site operator \( a_{\mathbf{n},\sigma } \) is
252: expressed through band operators: \( a_{\mathbf{n},\sigma }=N^{-1/2}\sum _{\mathbf{k}}{\rm e}^{i\mathbf{k}\cdot \mathbf{n}}a_{\mathbf{k},\sigma } \).
253: Since the phase of \( \Delta _{\mathbf{k}} \) is chosen zero, \( \Delta _{12} \)
254: is real. For \( J=0 \), this average does not differ from the uniform
255: gap parameter:\[
256: \Delta _{12}\to \frac{2V}{N}\sum _{\mathbf{k}}\langle a_{-\mathbf{k},\downarrow }a_{\mathbf{k},\uparrow }\rangle {\rm e}^{i\mathbf{k}\cdot (\delta _{2}-\delta _{1})}=\Delta ,\]
257: whereas for \( J\neq 0 \) the maximum perturbation of SC order near
258: the impurity is given by the suppression parameter \( \eta _{sup}=1-\Delta _{12}/\Delta \).
259: Its value is confined between 0 (for pure SC) and 1 (for complete
260: local suppression of SC order), and it only results from non-diagonal
261: GF's:\begin{eqnarray}
262: \eta _{sup}=\frac{2V}{N\Delta }\sum _{\mathbf{k},\mathbf{k}'\neq \mathbf{k}}\langle a_{-\mathbf{k},\downarrow }a_{\mathbf{k}',\uparrow }\rangle {\rm e}^{i(\mathbf{k}\cdot \delta _{2}-\mathbf{k}'\cdot \delta _{1})}= & & \label{hsupT} \\
263: =\frac{V}{2\pi \Delta }\sum _{j}(-1)^{j}\int _{-\infty }^{0}d\varepsilon {\rm Im\, Tr}\hat{G}_{j}\hat{T}_{j}\hat{G}_{j}\hat{\tau }_{1}, & & \nonumber
264: \end{eqnarray}
265: where the trace is in Nambu indices and Eqs. \ref{av},\ref{Tmat}
266: were used. Expansion of \( \hat{G}_{j} \) in Pauli matrices only
267: contains \( \hat{\tau }_{1} \)terms at \( j=2,3 \) (by the parity
268: of \( \alpha _{j,\mathbf{k}} \) with respect to the permutation \( k_{x}\leftrightarrow k_{y} \)):
269: \( JS\hat{G}_{2,3}=A+B\hat{\tau }_{3}\pm C\hat{\tau }_{1} \). Hence
270: only \( j=2,3 \) actually contribute in Eq. \ref{hsupT} by\begin{equation}
271: \label{hsup}
272: \eta _{sup}=\frac{V\varepsilon _{{\rm F}}\rho _{0}^{2}}{4}\int _{0}^{\varepsilon _{{\rm F}}}\frac{F(\varepsilon )}{\varepsilon }d\varepsilon ,
273: \end{equation}
274: where
275:
276: \[
277: F(\varepsilon )=\frac{16\varepsilon }{\pi JS\varepsilon _{{\rm F}}\Delta \rho _{0}^{2}}{\rm Im}[1-\frac{1}{(1-A)^{2}-B^{2}-C^{2}}]C,\]
278: and the complex coefficients \( A,B,C \) as functions of energy are
279: estimated in the relevant range \( \Delta <\varepsilon <\varepsilon _{{\rm F}} \),
280: setting \( E_{\mathbf{k}}\approx \xi _{\mathbf{k}}\approx Wa(k^{2}-k_{{\rm F}}^{2})/8 \):\begin{eqnarray*}
281: A\approx \frac{JS\rho _{0}\varepsilon }{8}(\ln |\frac{\varepsilon _{{\rm F}}+\varepsilon }{\varepsilon _{{\rm F}}-\varepsilon }|+2i\pi ),\quad \quad \quad \quad & & \\
282: B\approx \frac{JS\rho _{0}}{8}\ln \frac{\varepsilon _{{\rm F}}^{2}-\varepsilon ^{2}}{(W-\varepsilon _{{\rm F}})^{2}},\quad \quad \quad & & \\
283: C\approx \frac{\pi JS\rho _{0}\varepsilon _{{\rm F}}\Delta }{E\varepsilon }[\ln |\frac{\varepsilon _{{\rm D}}+\varepsilon }{\varepsilon _{{\rm D}}-\varepsilon }|+i\pi \theta (\varepsilon _{{\rm D}}-\varepsilon )]. & &
284: \end{eqnarray*}
285: Numeric analysis of these expressions with realistic parameter values:
286: \( W\sim 2 \) eV, \( JS\rho _{0}\sim 1 \), \( \varepsilon _{{\rm F}}\sim 0.3 \)
287: eV, \( \varepsilon _{{\rm D}}\sim 0.15 \) eV, shows that the main
288: contribution into the integral, Eq. \ref{hsup}, comes from the BCS
289: shell \( \Delta <\varepsilon <\varepsilon _{{\rm D}} \), where we
290: have: \( F(\varepsilon )\approx 1,\: 0<1-F(\varepsilon )\ll 1 \),
291: while \( 0<-F(\varepsilon )\ll 1 \) out of this shell, \( \varepsilon _{{\rm D}}<\varepsilon <\varepsilon _{{\rm F}} \)
292: (Fig. 2). Taking in mind the relation \( \ln (\varepsilon _{{\rm D}}/\Delta )\approx W/(V\gamma _{m})=W^{2}/(4V\varepsilon _{{\rm F}}) \)
293: which follows from Eq. \ref{delta}, we obtain from Eq. \ref{hsup}
294: \emph{almost complete} suppression: \( \eta _{sup}\approx 1 \). A
295: small residual part \( 0<1-\eta _{sup}\ll 1 \) is only due to a small
296: negative deviation of \( F(\varepsilon ) \) from unity within the
297: shell and to a small negative out-of-shell contribution. Thus, for
298: the above indicated choice of parameters we have \( \eta _{sup}\approx 96\% \).
299: The value of \( 1-\eta _{sup} \) yet diminishes with growing \( JS \),
300: but it should be stressed that no unitary limit \( JS\rho _{0}\gg 1 \)
301: is needed to get such a strong effect.
302:
303: The decay of this maximum effect with separation \( \mathbf{R} \)
304: from the impurity is given, in similarity with Eq. \ref{hsupT}, by\begin{equation}
305: \label{etaR}
306: \eta _{sup}(\mathbf{R})=\frac{V}{\pi \Delta }\int _{-\infty }^{0}d\varepsilon {\rm Im\, Tr}\hat{G}_{2}(\mathbf{R})\hat{T}_{2}\hat{G}_{2}(\mathbf{R})\hat{\tau }_{1},
307: \end{equation}
308: Here the matrix \( \hat{G}_{j}(\mathbf{R})=N^{-1}\sum _{\mathbf{k}}{\rm e}^{i\mathbf{k}\cdot \mathbf{R}}\alpha _{j,\mathbf{k}}\hat{G}_{\mathbf{k}} \),
309: and for \( R\gg \xi _{\pm }=a\sqrt{W/[8(\varepsilon _{{\rm F}}\pm \varepsilon )]} \)
310: it is mainly contributed by two saddle points in the complex \( k \)-plane:
311: \( \pm \xi _{\pm }^{-1}-iR^{-1} \), hence all its matrix elements
312: decay asymptotically like \( \cos (R/\xi _{\pm })/\sqrt{R/\xi _{\pm }} \),
313: and the non-diagonal elements contain yet the anisotropic factor \( (\Delta /\varepsilon )\cos 2\psi \)
314: where \( \psi =\arctan R_{y}/R_{x} \). However, for the energies
315: \( \varepsilon \sim \varepsilon _{{\rm D}} \) relevant here, such
316: anisotropy is less pronounced than that in the limit \( \varepsilon \to 0 \)
317: considered by Balatsky \emph{et al} \cite{Bal}. Integration in Eq.
318: \ref{etaR} results in asymptotic \( \eta _{sup}(\mathbf{R})\approx \eta _{sup}\sqrt{W/(8\varepsilon _{{\rm F}})}\sum _{i=\pm }(u_{i}+f_{i}\cos 2\psi +h_{i}\cos 4\psi )(a/R) \)
319: where \( h_{i}\ll f_{i}\sim u_{i}\sim 1 \). This angular dependence
320: resembles that for LDOS around Zn impurity, suggested by Haas and
321: Maki from continuous Bogolyubov-de Gennes equations \cite{Haas},
322: while the anomalously slow radial decay should enhance the overall
323: suppression of SC order.
324:
325: Besides the considered local suppression of the order parameter, related
326: to the non-diagonal (in Nambu indices) elements of GF's \( \hat{G}_{\mathbf{k},\mathbf{k}'} \),
327: there are also local effects related to their diagonal elements. Thus,
328: the variation of local DOS (LDOS): \( \rho (\mathbf{n})=(\pi N)^{-1}\sum _{\mathbf{k},\mathbf{k}'\neq \mathbf{k}}{\rm Im\, Tr\, e}^{i(\mathbf{k}-\mathbf{k}')\cdot \mathbf{n}}\hat{G}_{\mathbf{k},\mathbf{k}'}\hat{\tau }_{3} \),
329: attains its maximum at \( \mathbf{n}={\bf \delta } \), nearest neighbor
330: sites to the impurity, and Eq. \ref{Tmat} mainly contributes there
331: by \( j=1 \): \( \rho ({\bf \delta })\approx {\rm Im\, Tr}\hat{G}_{1}({\bf \delta })\hat{T}_{1}\hat{G}_{1}({\bf \delta })\hat{\tau }_{3} \).
332: The relevant GF's are \( \hat{G}_{1}({\bf \delta })\approx 2N^{-1}\sum _{\mathbf{k}}\hat{G}_{\mathbf{k}}=2(g_{0}+g_{3}\hat{\tau }_{3}) \),
333: where \( g_{0}\approx \rho _{0}\varepsilon [\varepsilon _{{\rm F}}^{-1}+i(\pi /\Delta )\arcsin (\Delta /\varepsilon )] \)
334: and \( g_{3}\approx \rho _{0}\ln (W/\varepsilon _{{\rm F}}) \) \cite{LP2}.
335: Considered as a function of energy, \( \rho ({\bf \delta }) \) displays
336: a very sharp resonance in the denominator of \( \hat{T}_{1} \): \( {\rm Re}[1-2JSg_{0}(\varepsilon )]^{2}-(2JSg_{3})^{2}\to 0 \)
337: at \( \varepsilon \to 0 \) (cf. Fig. 3) with the observed peak in
338: the related tunnel conductivity \cite{Pan}) if the impurity perturbation
339: parameter \( J \) is close to \( J_{cr}=1/(2Sg_{3}) \). This refers
340: to a \emph{fine tuned} rather than unitary perturbation \( J \) and
341: agrees with its choice made to estimate \( \eta _{sup} \). The \( j=1 \)
342: contribution also dominates in the Kondo-like local polarization of
343: itinerant spins: \( m({\bf \delta })\approx \int _{0}^{\varepsilon _{{\rm F}}}d\varepsilon {\rm Im\, Tr\, e}^{i(\mathbf{k}-\mathbf{k}')\cdot \mathbf{n}}\hat{G}_{1}({\bf \delta })\hat{T}_{1}\hat{G}_{1}({\bf \delta }) \),
344: which should explain the observed enhancement of exchange fields on
345: \( ^{63} \)Cu \cite{Wal} and \( ^{89} \)Y \cite{Mah} nuclei close
346: to Zn impurities. A more detailed treatment of these phenomena will
347: be presented elsewhere.
348:
349: The proposed model can be equally applied to isovalent substitutes
350: for Cu, which are magnetic themselves, as Ni\( ^{2+} \) or Fe\( ^{2+} \).
351: But, since the net MF on neighbor O sites in this case is due to incomplete
352: AFM compensation of exchange fields by different magnetic ions, the
353: perturbation parameter \( JS \) may be \emph{weaker} than that for
354: non-magnetic Zn\( ^{2+} \) and so the resulting suppression of SC
355: order, as is observed in the experiment \cite{Bonn}.
356:
357: In conclusion, we developed a microscopic model of spin dependent
358: perturbation on charge carriers in CuO\( _{2} \) planes, produced
359: by a non-magnetic substitute for Cu. An almost complete suppression
360: of \( d \)-wave order parameter at nearest neighbor sites to the
361: impurity atom is obtained, as a result of parallel alignment of carrier
362: spins in the exchange field \( JS \) by non-compensated Cu\( ^{2+} \)
363: spins, and this strong effect is achieved with moderate \( JS \)
364: values. It decays with distance from impurity rather slowly, which
365: can explain the fast destroying of SC order in cuprates already at
366: low Zn concentration. The model also provides explanation for other
367: local effects, such as a sharp resonance of LDOS and local polarization
368: of charge carrier spins close to impurity.
369:
370: \begin{acknowledgments}
371: We acknowledge the partial support for this work from Portuguese Fundação
372: de Ciência e Tecnologia through the project 2/2.1/FIS/302/94 and from
373: Swiss Science Foundation under SCOPES-project 7UKPJ062150.00/1.
374: \end{acknowledgments}
375: \begin{thebibliography}{10}
376: \bibitem{And1}P.W. Anderson, The Theory of Superconductivity in High-\( T_{c} \)
377: Cuprates. Princeton Univ. Press, Princeton (1997) p. 34.
378: \bibitem{LP1}V.M. Loktev and Yu.G. Pogorelov, Physica \textbf{C 272}, 151 (1996).
379: \bibitem{LP2}V.M. Loktev and Yu.G. Pogorelov, Low Temp. Phys. \textbf{27}, 1039
380: (2001).
381: \bibitem{LP3}V.M. Loktev and Yu.G. Pogorelov, Europhys. Lett. (in press, 2002).
382: \bibitem{And}P.W. Anderson, J. Phys. Chem. Solids, \textbf{11}, 26 (1959).
383: \bibitem{AG}A.A. Abrikosov and L.P. Gor'kov, Sov. Phys. JETP \textbf{12}, 1243
384: (1961).
385: \bibitem{Maki}K. Maki, Phys. Rev. \textbf{153}, 428 (1967).
386: \bibitem{Shiba}H. Shiba, Progr. Theoret. Phys. (Kyoto) \textbf{40}, 435 (1968).
387: \bibitem{Z-MH}J. Zittartz and E. Mueller-Hartmann, Z. Phys. \textbf{232}, 11 (1970).
388: \bibitem{Zhang}Kuan Zhang and D.A. Bonn, Phys. Rev. Lett. \textbf{73}, 2484 (1994).
389: \bibitem{Bonn}D.A. Bonn, S. Kamal, Kuan Zhang, Ruixing Liang, D.J. Baar, E. Klein,
390: and W.N. Hardy, Phys. Rev. \textbf{B 50}, 4051 (1994).
391: \bibitem{Chien}T.R. Chien, Z.Z. Chang, and N.P. Ong, Phys. Rev. Lett. \textbf{67},
392: 2088 (1991).
393: \bibitem{Hirsch}P.J. Hirschfeld and N. Goldenfeld, Phys. Rev. \textbf{B 48}, 4219
394: (1993).
395: \bibitem{Lee}P.A. Lee, Phys. Rev. Lett. \textbf{71}, 1887 (1993).
396: \bibitem{Fehr}R. Fehrenbacher and M.R. Norman, Phys. Rev. \textbf{B 50}, 3495 (1994).
397: \bibitem{Atk}W.A. Atkinson, P.J. Hirschfeld, and A.H. McDonald, Phys. Rev. Lett.
398: \textbf{85}, 3922 (2000).
399: \bibitem{Mah}A.V. Mahajan, H. Alloul, G. Collin, and J.F. Marucco, Phys. Rev. Lett.
400: \textbf{72}, 3100 (1994).
401: \bibitem{McFarl}W.A. McFarlane, J. Bobroff, H. Alloul, P. Mendels, N. Blanchard, G.
402: Collin, and J.F. Marucco, Phys. Rev. Lett. \textbf{85}, 1108 (2000).
403: \bibitem{Schr}Quijin Chen, R.J. Schrieffer, cond-mat/0202541.
404: \bibitem{Park}Kwon Park, cond-mat/0203142.
405: \bibitem{Wal}R.E. Walstedt, R.F. Bell, L.F. Schneemeyer, J.V. Waszczak, W.W. Warren
406: Jr., R. Dupree, and A. Gencten, Phys. Rev. \textbf{B 48}, 10646 (1993).
407: \bibitem{Jul}M.-H. Julien, T. Fehér, M. Horvatic, C. Berthier, O.N. Bakharev, P.
408: Sérgansan, G. Collin, and J.F. Marucco, Phys. Rev. Lett. \textbf{84},
409: 3422 (2000).
410: \bibitem{Birg}R.J. Birgeneau and G. Shirane, in: \emph{Physical Properties of High
411: Temperature Superconductors}, Ed. D.M. Ginsberg, World Scientific,
412: Singapore (1989) p. 151.
413: \bibitem{Kon}J. Kondo, Progr. Theoret. Phys. (Kyoto) \textbf{32}, 37 (1964).
414: \bibitem{Nag}Y. Nagaoka, Phys. Rev. \textbf{138}, A1209 (1965).
415: \bibitem{Pog}Yu.G. Pogorelov, Solid State Commun. \textbf{95}, 245 (1995).
416: \bibitem{Bal}A.V. Balatsky, M.I. Salkola, and A. Rosengren, Phys. Rev. \textbf{B}
417: \textbf{51}, 15547 (1995).
418: \bibitem{Haas}S. Haas and K. Maki, Phys. Rev. Lett. \textbf{85}, 2172 (2000).
419: \bibitem{Pan}S.H. Pan, E.W. Hudson, K.M. Lang, H. Eisaki, S. Uchida and J.C. Davis,
420: Nature \textbf{403}, 746 (2000).\end{thebibliography}
421:
422: \end{document}
423: