cond-mat0011257/all.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \documentstyle[twocolumn,seceq,letter,epsf]{jpsj}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \def\runtitle{Impurity Site NMR Relaxation in Unconventional Superconductors}
5: \def\runauthor{Masashige {\sc Matsumoto}}
6: 
7: \title
8: {
9: Impurity Site NMR Relaxation in Unconventional Superconductors
10: }
11: 
12: \author
13: {
14: Masashige {\sc Matsumoto}
15: }
16: 
17: \inst
18: {
19: Department of Physics, Faculty of Science, Shizuoka University, 836 Oya,
20: Shizuoka 422-8529, Japan
21: }
22: 
23: \recdate
24: {May 8, 2001}
25: 
26: \abst
27: {
28: Impurity nuclear spin relaxation is studied theoretically.
29: A single impurity generates a bound state localized around the impurity atom
30: in unconventional superconductors.
31: With increasing impurity potential,
32: the relaxation rate $T_1^{-1}$ is reduced by the impurity potential.
33: However, it has a peak at low temperatures due to the impurity bound state.
34: The peak disappears at non-impurity sites.
35: The impurity site NMR measurement detecting a local electronic structure just on the impurity atom
36: is very useful for identifying  the unconventional pairing states.
37: }
38: 
39: \kword
40: {
41: NMR, NQR, unconventional superconductivity, impurity
42: }
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44: \begin{document}
45: \sloppy
46: \maketitle
47: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
48: \renewcommand{\theequation}{\arabic{equation}}
49: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
50: 
51: 
52: 
53: %%%%%%%%%%%%%%%%%%%%%%%%%%   Introduction   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
54: The study of unconventional superconductors
55: has become one of the most attractive issues in recent condensed matter physics.
56: These materials include heavy fermion compounds, high-$T_c$ cuprates,
57: organic superconductors and the recently discovered Sr$_2$RuO$_4$.
58: A great effort has been made to detect the unconventional superconductivity by several probes:
59: specific heat measurement, tunneling spectroscopy, nuclear magnetic resonance (NMR),
60: $\pi$-junction and so on.
61: NMR experiments were important in the early stage
62: of studying the unconventional superconductivity.
63: It is well known that $T_1^{-1}$ in $s$-wave superconductors
64: has a Hebel-Slichter peak just below $T_c$,
65: while it does not exhibit this feature in unconventional superconductors.
66: On the other hand, it was also reported that the peak can be suppressed
67: by the strong damping effect even for a $s$-wave state.
68: \cite{Kitaoka}
69: 
70: To clarify this point,
71: Ishida et al. performed a Cu site (bulk site) NMR experiment
72: with Zn-doped YBa$_2$Cu$_3$O$_{7-\delta}$ (YBCO).
73: \cite{Ishida93}
74: They found that $T_1^{-1}$ is proportional to $T$ at low temperatures.
75: Non-magnetic impurities do not break the $s$-wave superconductivity.
76: \cite{Anderson}
77: However, they break unconventional superconductivity and yield low-energy impurity states.
78: \cite{Hirschfeld,Schmitt-Rink,Hotta}
79: The observed $T$ linear dependence of $T_1^{-1}$ can therefore
80: be understood as evidence of extended low-energy impurity states
81: and supports an unconventional pairing state for YBCO.
82: In this case, the effect of many impurities was important.
83: 
84: A single impurity can also break the unconventional superconductivity locally
85: and generate a bound state around it.
86: \cite{Matsumoto1,Balatsky,Onishi}
87: Recently, Pan et al. observed such a low-energy state around a Zn atom
88: in Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$
89: using a low-temperature scanning tunneling microscope (STM),
90: showing a fourfold symmetrical structure
91: which is consistent with a $d_{x^2-y^2}$-wave pairing state.
92: \cite{Pan,Yazdani}
93: Nishida et al. also observed the fourfold structure
94: around a columnar defect by a STM.
95: \cite{Nishida}
96: At present, STM is the only instrument
97: which can reveal the local electronic structure around a single impurity atom.
98: 
99: Usually, NMR is used to probe bulk electronic structure.
100: For example, Cu site NMR measurements for high-$T_c$ cuprates
101: detect local electronic structure at the Cu atom.
102: Recently, impurity site NMR experiments were performed by several groups.
103: \cite{Ishida96,Bobroff}
104: However, there is, so far, no theory for the impurity site NMR.
105: The purposes of this paper are (1) to address the theory for the impurity site NMR
106: and (2) to propose a new NMR experiment
107: which probes the local quasiparticle states around a single impurity atom.
108: In this study we concentrate our attentions on a single impurity problem,
109: since we can solve it exactly.
110: For unconventional superconductivity,
111: we show that the impurity site NMR $T_1^{-1}$ has a peak,
112: while the peak disappears at the bulk site (non-impurity site).
113: This peak at the impurity site is a common feature of all types of unconventional superconductors.
114: The impurity site NMR can reveal the local quasiparticle structure as well,
115: and provides us with a new experimental method
116: for identifying unconventional superconductivity.
117: Recently, NMR $T_1^{-1}$ in a vortex state was studied theoretically.
118: \cite{Takigawa,Morr}
119: They proposed that the NMR can be used as a site-selective probe
120: by controlling the resonance frequency.
121: Thus, NMR is gradually attracting attention as a local probe.
122: %with the development in experiment.
123: 
124: %%%%%%%%%%%%%%%%%%%%%%%%%%   Formulation   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
125: In the unconventional pairing case,
126: the superconducting order parameter is suppressed around the impurity.
127: However, we present our theory assuming a uniform order parameter,
128: since we can capture the essential physics clearly without any complex analysis.
129: Actually, we have performed numerical self-consistent calculations
130: to include the spatial dependence of the order parameter.
131: We have confirmed that the NMR $T_1^{-1}$ has no qualitative difference
132: between in the uniform and non-uniform order parameter cases.
133: The most important point in the impurity site NMR is the existence of the impurity bound state,
134: which does not depend on details of unconventional pairing states.
135: 
136: For the unconventional pairings,
137: we focus on $p_x\pm\ {\rm i} p_y$-wave and $d_{x^2-y^2}$-wave pairing states.
138: The $p_x\pm\ {\rm i} p_y$-wave is the simplest and most essential symmetry
139: suggested for the Sr$_2$RuO$_4$ superconductor.
140: \cite{Maeno,Rice,Baskaran}
141: This state is a spin triplet state and breaks the time-reversal symmetry.
142: \cite{Luke,Ishida98}
143: For simplicity, we assume that the superconductor is basically
144: two-dimensional and has a cylindrical Fermi surface.
145: The Matsubara Green function in a $2\times 2$ matrix form is given by
146: \cite{Matsumoto1}
147: \begin{eqnarray}
148: &&{\hat G}(  {\rm i} \omega_m,\mbox{\boldmath$r$},\mbox{\boldmath$r$}')
149: = {\hat G}_0({\rm i} \omega_m,\mbox{\boldmath$r$},\mbox{\boldmath$r$}')
150:                        + {\hat G}_0({\rm i} \omega_m,\mbox{\boldmath$r$},0) U_0 {\hat \tau}_3 \cr
151: &&~~~~~~~~~~~\times{1 \over 1 - {\hat G}_0({\rm i} \omega_m,0,0) U_0 {\hat \tau}_3}
152: {\hat G}_0({\rm i} \omega_m,0,\mbox{\boldmath$r$}'),
153: \end{eqnarray}
154: where a single impurity is located at the origin of the coordinate.
155: $U_0$ represents the strength of the short-range impurity potential.
156: $\omega_m$ is the fermion Matsubara frequency
157: and ${\hat {\tau_i}}$ $(i=1,2,3)$ represents the Pauli matrix in a charge space.
158: For simplicity we use $\hbar=1$ and $k_{\rm B}=1$ units throughout this letter.
159: ${\hat G}_0$ is the non-perturbed ($U_0=0$) Green function given by
160: \begin{equation}
161: {\hat G}_0({\rm i} \omega_m,\mbox{\boldmath$r$},\mbox{\boldmath$r$}')
162: = -{ 1 \over \Omega} \sum_{\mbox{\boldmath$k$}}
163:    {\rm e}^{{\rm i} \mbox{\footnotesize \boldmath $k$}
164:     \cdot (\mbox{\boldmath$r$}-\mbox{\boldmath$r$}')}
165:      {{\rm i} \omega_m + \epsilon_{\mbox{\footnotesize \boldmath $k$}}{\hat \tau}_3
166:    + {\hat \Delta}_{\mbox{\footnotesize \boldmath $k$}}
167:    \over
168:      \omega_m^2 + \epsilon_{\mbox{\footnotesize \boldmath $k$}}^2
169:    + \Delta_{\mbox{\footnotesize \boldmath $k$}}^2}.
170: \end{equation}
171: Here, ${\hat \Delta}_{\mbox{\footnotesize \boldmath $k$}}$
172: is equal to $\Delta_{\mbox{\footnotesize \boldmath $k$}}{\hat \tau}_1$
173: and expresses the momentum-dependent order parameter.
174: For the $p_x\pm {\rm i} p_y$-wave ($d_{x^2-y^2}$-wave),
175: $\Delta_{\mbox{\footnotesize \boldmath $k$}}=\Delta_p(T){\rm e}^{{\rm i} \theta_k}$
176: [$\Delta_{\mbox{\footnotesize \boldmath $k$}}=\Delta_d(T)\cos(2\theta_k)$].
177: Here, $\theta_k$ is the angle of the Fermi wave vector measured from the $k_x$ axis.
178: We solve the gap equation in the bulk region
179: and use the temperature-dependent order parameter $\Delta_p(T)$ ($\Delta_d(T)$)
180: which is real and positive for the $p_x\pm {\rm i} p_y$-wave ($d_{x^2-y^2}$-wave).
181: $\epsilon_{\mbox{\footnotesize \boldmath $k$}}$ is the kinetic energy
182: and $\Omega$ represents the volume of the system.
183: NMR relaxation occurs via the hyperfine interaction
184: between the nuclear spin and the conduction electrons.
185: NMR $T_1^{-1}$ at the impurity site is given in an explicit form \cite{Leadon}
186: \begin{eqnarray}
187: &&T_1^{-1}=2\pi(\frac{4\pi}{3})^2(\gamma_e\gamma_n)^2 W, \cr
188: &&W=\int {\rm d} E { a_{11}(E) a_{22}(-E)-a_{12}(E) a_{21}(-E) \over 1+\cosh(E/T)},
189: \label{eqn:T1}
190: \end{eqnarray}
191: where $W$ is proportional to the nuclear spin flip transition probability.
192: $\gamma_e$ and $\gamma_n$ are the gyromagnetic ratios
193: for the electron and nucleon, respectively.
194: $a_{ij}(E) = -{\rm Im}\bigl[G_{ij}({\rm i} \omega_m \rightarrow E
195:              + {\rm i} \delta,0,0)\bigr]/\pi$,
196: where the subscript of $a_{ij}$ represents the matrix element of the Green function.
197: Here, we have omitted the $u_0$ dependence in $a_{ij}$ for simplicity,
198: where $u_0=\pi N_0 U_0$ ($N_0$ is the density of states per volume at the Fermi energy).
199: $\delta$ is a positive, small number which expresses a finite level broadening.
200: 
201: %%%%%%%%%%%%%%%%%%%%%%%%%%   s-wave   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
202: In the $s$-wave case,
203: NMR $T_1^{-1}$ has the Hebel-Slichter peak just below $T_c$ at the non-impurity site ($u_0=0$).
204: At the impurity site ($u_0\neq 0$),
205: $a_{ij}$ scales as $a_{ij}(E,u_0)=a_{ij}(E,0)/(1+u_0^2)$.
206: The local density of states at the impurity is then simply reduced,
207: keeping the same energy dependence.
208: This is consistent with Anderson's theorem.
209: \cite{Anderson}
210: In the same manner, $W$ scales as
211: \begin{equation}
212: W(u_0,T)=W(0,T)/(1+u_0^2)^2.
213: \label{eqn:Ws}
214: \end{equation}
215: Therefore, $T_1^{-1}$ is simply reduced at the impurity site,
216: while the temperature dependence does not change
217: between the impurity site and the bulk site for the $s$-wave, as shown in Fig. \ref{fig:1}(a).
218: %%%%%%%%%%%%%%%%%%%%%%%%%%   Fig. 1   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
219: \begin{figure}[t]
220: \begin{center}
221: %
222: \begin{minipage}{6.2cm}
223: \epsfxsize=6.2cm
224: \epsfbox{fig1a.eps}
225: \end{minipage}
226: %
227: \begin{minipage}{6.2cm}
228: \epsfxsize=6.2cm
229: \epsfbox{fig1b.eps}
230: \end{minipage}
231: %
232: \end{center}
233: \caption{
234: Temperature dependence of $W$ with various $u_0$ with a fixed damping rate $\delta/T_c=0.1$.
235: The number represents a value of $u_0$.
236: (a) and (b) are for the $s$-wave and $p_x \pm {\rm i} p_y$-wave, respectively.
237: The peak of $T_1^{-1}$ in (b) becomes higher with the decrease of $\delta/T_c$
238: [see eq. (\ref{eqn:Wb})].
239: The value of $\delta/T_c$ does not change the temperature dependence of $T_1^{-1}$.
240: }
241: \label{fig:1}
242: \end{figure}
243: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
244: 
245: %%%%%%%%%%%%%%%%%%%%%%%%%%   p-wave   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
246: For the $p_x \pm {\rm i} p_y$-wave,
247: we show the results in Fig. \ref{fig:1}(b).
248: In this case,
249: the temperature dependence at the impurity site
250: is quite different from that of the bulk site.
251: In the unconventional pairing case, both $a_{12}$ and $a_{21}$ are zero,
252: since pair electrons cannot possess the same position.
253: Therefore the coherence factor vanishes for unconventional pairings,
254: resulting in the absence of the Hebel-Slichter peak.
255: $W$ is then expressed by the following simple form:
256: \begin{equation}
257: W=N_0^2\int {\rm d}E \frac{N_{\rm imp}^2(E)}{1+\cosh(E/T)},
258: \label{eqn:W}
259: \end{equation}
260: where $N_{\rm imp}(E)=a_{11}(E)/N_0=a_{22}(-E)/N_0$
261: is the dimensionless local density of states at the impurity atom.
262: For the $p_x\pm {\rm i} p_y$-wave, a bound state appears around the single impurity.
263: The energy position of the bound state is given by
264: $E_{\rm B}=-{\rm sgn}(u_0) \Delta_p(T) /\sqrt{1+u_0^2}$.
265: \cite{Okuno}
266: In fact, an impurity effect is observed in Sr$_2$RuO$_4$
267: as a reduction of its $T_c$.
268: \cite{Mackenzie}
269: $N_{\rm imp}$ in eq. (\ref{eqn:W}) for the $p_x\pm {\rm i} p_y$-wave
270: has the following analytic form:
271: \begin{eqnarray}
272: N_{\rm imp}(E)
273: &=&\frac{N_p(E)}{1+[u_0 N_p(E)]^2}\theta(|E|-\Delta_p(T)) \cr
274: &+&\frac{\pi|u_0|\Delta_p(T)}{(1+u_0^2)^{\frac{3}{2}}}\delta(E-E_{\rm B}),
275: \label{eqn:Np}
276: \end{eqnarray}
277: where $N_p(E)=|E|/\sqrt{E^2-\Delta_p^2(T)}$
278: is the dimensionless $p_x\pm {\rm i} p_y$-wave bulk density of states.
279: $\theta$ and $\delta$ in eq. (\ref{eqn:Np})
280: are the Heaviside and $\delta$-functions, respectively.
281: The first and second terms in eq. (\ref{eqn:Np})
282: are the continuum and bound states, respectively.
283: Introducing $u_0$, we can see in Fig. \ref{fig:2}(a)
284: that the density of states from the quasiparticle continuum is decreased,
285: while that of the impurity bound state increases.
286: %%%%%%%%%%%%%%%%%%%%%%%%%%   Fig. 2   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
287: \begin{figure}[t]
288: \begin{center}
289: %
290: \begin{minipage}{6.2cm}
291: \epsfxsize=6.2cm
292: \epsfbox{fig2a.eps}
293: \end{minipage}
294: %
295: \begin{minipage}{6.2cm}
296: \epsfxsize=6.2cm
297: \epsfbox{fig2b.eps}
298: \end{minipage}
299: %
300: \end{center}
301: \caption{
302: (a) Dimensionless local density of states for the $p_x\pm i p_y$-wave state.
303: $N_p$ is multiplied by $(1+u_0^2)^{-1}=N_{\rm imp}(\infty)$ for convenience.
304: Parameters are chosen as $u_0=1$ and $\delta/\Delta_p=0.01$.
305: (b) Temperature dependence of $W_{\rm B}$.
306: Parameters are chosen as $u_0=1$ and $\delta/T_c=0.1$.
307: }
308: \label{fig:2}
309: \end{figure}
310: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
311: As in the $s$-wave case,
312: $W$ for the $p_x\pm {\rm i} p_y$-wave state is reduced with the increase of $u_0$.
313: However, as shown in Fig. \ref{fig:1}(b),
314: it drastically changes at low temperatures due to the impurity bound state.
315: The bound state contribution to $W$ behaves as
316: \begin{equation}
317: W_{\rm B}
318: \simeq
319: N_0^2 \Bigl[\frac{\Delta_p(T) u_0}{(1+u_0^2)^{\frac{3}{2}}}\Bigr]^2
320: \frac{\pi}{2\delta} [1+\cosh(E_{\rm B}/T)]^{-1},
321: \label{eqn:Wb}
322: \end{equation}
323: where $\delta(E-E_{\rm B})=(\delta/\pi)[(E-E_{\rm B})^2+\delta^2]^{-1}$ has been used.
324: Figure \ref{fig:2}(b) clearly shows that $W_{\rm B}$ has a peak below $T_c$.
325: Notice that the peak in $W$ in Fig. \ref{fig:1}(b)
326: is closely related to the local impurity bound state.
327: It is different from the Hebel-Slichter peak and can be distinguished,
328: since $T_1^{-1}$ does not exhibit such a peak at the bulk site
329: for unconventional superconductivity.
330: The peak appears only at the impurity site.
331: At low temperatures, impurity nuclear spin relaxation occurs
332: via the impurity bound state at the impurity site.
333: As $u_0$ is increased,
334: the peak position of $W$ shifts to a lower temperature region [see Fig. \ref{fig:1}(b)],
335: which reflects the energy shift of the bound state.
336: For a large $u_0$, $W_{\rm B}$ can survive if $\delta$ is sufficiently small.
337: 
338: %%%%%%%%%%%%%%%%%%%%%%%%%%   d-wave   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
339: Next we study the $d_{x^2-y^2}$-wave case,
340: which is the most favorable pairing symmetry for high-$T_c$ cuprates.
341: In this case, $N_{\rm imp}$ in eq. (\ref{eqn:W}) takes the following form:
342: \begin{equation}
343: N_{\rm imp}(E)=\frac{1}{u_0^2}\frac{N_d(E)}{[A(E)+1/u_0]^2+N_d^2(E)},
344: \end{equation}
345: where $A(E)$ and $N_d(E)$ are the real and imaginary parts of
346: $\bigl\langle{(E+ {\rm i} \delta) / \sqrt{(-{\rm i} E+\delta)^2
347: +|\Delta_{\mbox{\footnotesize \boldmath $k$}}|^2}}\bigr\rangle_{\rm FS}$,
348: respectively.
349: Here, $\langle \cdots \rangle_{\rm FS}$ represents an average over the Fermi surface.
350: The energy position of the bound state is determined by the zero of $A(E)+1/u_0$
351: and its effective broadening is $N_d(E)$
352: which is the dimensionless $d_{x^2-y^2}$-wave bulk density of states.
353: Similar to the $p_x \pm {\rm i} p_y$-wave case,
354: the impurity site ($u_0 \neq 0$) NMR $T_1^{-1}$ for the $d_{x^2-y^2}$-wave has a peak
355: due to the impurity bound state, as shown in Fig. \ref{fig:3}.
356: %%%%%%%%%%%%%%%%%%%%%%%%%%   Fig. 3   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
357: \begin{figure}[t]
358: \begin{center}
359: %
360: \begin{minipage}{6.2cm}
361: \epsfxsize=6.2cm
362: \epsfbox{fig3.eps}
363: \end{minipage}
364: %
365: \end{center}
366: \caption{
367: Temperature dependence of $W$ with various $u_0$ for the $d_{x^2-y^2}$-wave
368: with a fixed damping rate $\delta/T_c=0.1$.
369: The number represents a value of $u_0$.
370: }
371: \label{fig:3}
372: \end{figure}
373: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
374: 
375: %%%%%%%%%%%%%%%%%%%%%%%%%%   Discussion   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
376: The important point in this letter is that
377: a single impurity generates a bound state in unconventional superconductors,
378: and that the NMR experiment can detect the local impurity bound state
379: by the impurity site $T_1^{-1}$ measurement.
380: We emphasize that the peak in $T_1^{-1}$ at low temperatures
381: in the impurity site NMR is a common feature
382: exhibited by all types of unconventional superconductors.
383: It does not depend on the details of unconventional pairing states.
384: We have demonstrated this
385: by examining the $p_x \pm {\rm i} p_y$-wave and $d_{x^2-y^2}$-wave pairing states.
386: The former is a fully gapped unconventional state,
387: while the latter is a gapless one.
388: 
389: Recently Ishida et al. carried out NMR and NQR measurements on CeRu$_2$
390: with 1\% Al (impurity) substitution for Ru.
391: \cite{Ishida96}
392: They reported that the $T_1^{-1}$ at the Al site
393: was reduced to 10\% of that of the Ru (bulk) site.
394: The strength of the impurity scattering can be estimated as $|u_0|=1.5$
395: by using eq. (\ref{eqn:Ws}).
396: If the pairing is unconventional,
397: we can expect the appearance of a peak below $T_c$ for $|u_0|=1.5$,
398: as shown in Figs. \ref{fig:1}(b) and \ref{fig:3}.
399: However, the Al site $T_1^{-1}$ does not exhibit such a peak.
400: It is simply reduced, keeping the same exponential temperature dependence as that of the Ru site.
401: This feature is consistent with our theory of impurity site NMR for the $s$-wave.
402: 
403: Very recently, Bobroff et al. succeeded in Li (impurity) site NMR measurement
404: in YBCO with Li substituted for Cu.
405: \cite{Bobroff}
406: The Li concentrations of the samples were 0.85\% and 1.86\% per CuO$_2$ layer.
407: For YBCO, Li induces a magnetic moment around it.
408: Unfortunately, it is difficult to extract the contribution from the conduction electrons in $T_1$,
409: since the induced moment also generates the relaxation.
410: However, this Li site experiment shows that
411: the recent NMR experiment has enough sensitivity
412: to detect $T_1$ at the impurity site in unconventional superconductors.
413: The impurity concentration is about 1\% for both CeRu$_2$ and YBCO.
414: These experiments indicate
415: that the impurity site NMR is possible at least for such impurity doping.
416: 
417: For the bulk site NMR $T_1$ measurement,
418: it is important to observe whether the Hebel-Slichter peak is present or not.
419: Examining the temperature dependence (exponential or power) at low temperatures is also important.
420: (I) In the absence of the Hebel-Slichter peak,
421: it is difficult to distinguish unconventional pairings from the $s$-wave
422: for the bulk site measurement,
423: since strong damping can suppress the coherence peak for the $s$-wave.
424: (II) The temperature dependence of the bulk site $T_1^{-1}$ is not useful
425: for distinguishing fully gapped unconventional states from the $s$-wave.
426: We cannot distinguish gapless unconventional states from the anisotropic $s$-wave, either.
427: In contrast to the bulk site measurement,
428: the impurity site measurement can distinguish for both (I) and (II) cases.
429: The feature of unconventional pairings appears as a peak in $T_1^{-1}$ only at the impurity site,
430: while the peak disappears at the bulk site.
431: On the other hand, for the $s$-wave,
432: the temperature dependence of $T_1^{-1}$ does not change
433: for either impurity or bulk sites [see eq. (\ref{eqn:Ws}) and Fig. \ref{fig:1}(a)].
434: If we combine the two (bulk and impurity sites) NMR measurements,
435: we can identify the unconventional pairing without ambiguity.
436: Thus, the impurity site NMR is a powerful experimental method
437: for probing unconventional superconductivity.
438: 
439: Our theory for a single impurity is applicable to a case with low impurity concentration.
440: With the increase of the impurity concentration,
441: extended impurity states are formed by the local impurity bound states,
442: resulting in additional continuum energy levels below the energy gap.
443: In this case,
444: the impurity effect appears as a constant $T_1 T$ at low temperatures
445: for both the impurity and bulk sites.
446: However, the impurity site measurement is expected to exhibit the impurity effect
447: more obviously than the bulk site,
448: since the electronic states at the impurity site are strongly affected by the impurity atom.
449: We note that the peak in $T_1^{-1}$ of our theory
450: for the single impurity (low impurity concentration)
451: is much more drastic than the constant $T_1 T$,
452: both of which are evidence of unconventional superconductivity.
453: 
454: The local electronic structure around the impurity atom
455: reflects the superconducting pairing symmetry.
456: So far, STM is the only probe of the local impurity state.
457: However, STM experiments are strongly affected by the surface roughness.
458: We note that the impurity site NMR can be used as a local probe
459: regardless of annoying surface conditions.
460: 
461: In conclusion, we have addressed the theory for the impurity site NMR $T_1^{-1}$
462: in unconventional superconductors.
463: We have concentrated our attention on a single impurity effect.
464: There are two characteristic points of the impurity site NMR.
465: The first one is that the relaxation rate $T_1^{-1}$ is reduced by the impurity potential.
466: The second is the peak in $T_1^{-1}$.
467: It appears as direct result of the relaxation process via the impurity bound state.
468: On the other hand, the peak vanishes at a non-impurity site.
469: This is the remarkable difference in $T_1^{-1}$
470: between the impurity site and non-impurity site NMR in unconventional superconductors.
471: In contrast,
472: the temperature dependence of $T_1^{-1}$ does not change
473: between the two sites for the conventional $s$-wave.
474: Therefore, observation of the peak by the impurity site NMR measurement
475: can provide strong evidence of an unconventional superconducting state.
476: 
477: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
478: The author expresses his sincere thanks to Y. Okuno and M. Sigrist
479: for many discussions on the impurity effect in unconventional superconductors.
480: He also thanks N. Hayashi and M. Takigawa
481: for interesting discussions on NMR $T_1^{-1}$.
482: He is grateful for the helpful discussions with H. Alloul and K. Ishida regarding NMR experiments.
483: He would like to thank M. Koga for his helpful comments
484: and critical reading of the manuscript.
485: This work was supported by JSPS for Encouragement of Young Scientists (No. 10740169).
486: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
487: 
488: \begin{thebibliography}{99}
489: 
490: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
491: 
492: \bibitem{Kitaoka}
493:   Y. Kitaoka et al.:
494:   Physica C {\bf 192} (1992) 272.
495: 
496: \bibitem{Ishida93}
497:   K. Ishida et al.:
498:   J. Phys. Soc. Jpn. {\bf 62} (1993) 2803.
499: 
500: \bibitem{Anderson}
501:   P. W. Anderson:
502:   J. Phys. Chem. Solids {\bf 11} (1959) 26.
503: 
504: \bibitem{Hirschfeld}
505:   P. J. Hirschfeld, D. Vollhardt and P. W\"{o}lfle:
506:   Solid State Commun. {\bf 59} (1986) 111;
507:   P. J. Hirschfeld, P. W\"{o}lfle and D. Einzel:
508:   Phys. Rev. B {\bf 37} (1988) 83;
509:   P. J. Hirschfeld and N. Goldenfeld:
510:   Phys. Rev. B {\bf 48} (1993) 4219.
511: 
512: \bibitem{Schmitt-Rink}
513:   S. Schmitt-Rink, K. Miyake and C. M. Varma:
514:   Phys. Rev. Lett. {\bf 57} (1986) 2575.
515: 
516: \bibitem{Hotta}
517:   T. Hotta:
518:   J. Phys. Soc. Jpn. {\bf 62} (1993) 274.
519: 
520: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
521: 
522: \bibitem{Matsumoto1}
523:   M. Matsumoto and H. Shiba:
524:   J. Phys. Soc. Jpn. {\bf 64} (1995) 1703.
525: 
526: \bibitem{Balatsky}
527:   A. V. Balatsky, M. I. Salkola and A. Rosengren:
528:   Phys. Rev. B {\bf 51} (1995) 15547;
529:   M. I. Salkola, A. V. Balatsky and D. J. Scalapino:
530:   Phys. Rev. Lett. {\bf 77} (1996) 1841.
531: 
532: \bibitem{Onishi}
533:   Y. Onishi et al.:
534:   J. Phys. Soc. Jpn. {\bf 65} (1996) 675.
535: 
536: \bibitem{Pan}
537:   S. H. Pan et al.:
538:   Nature (London) {\bf 403} (2000) 746.
539: 
540: \bibitem{Yazdani}
541:   A. Yazdani et al.:
542:   Phys. Rev. Lett. {\bf 83} (1999) 176.
543: 
544: \bibitem{Nishida}
545:   N. Nishida et al.:
546:   Physica B {\bf 284-288} (2000) 967;
547:   M. Matsumoto, S. Kaneko and N. Nishida:
548:   J. Phys. Soc. Jpn. {\bf 66} (1998) 105.
549: 
550: \bibitem{Ishida96}
551:   K. Ishida et al.:
552:   Z. Naturforsch {\bf 51 a} (1996) 793;
553:   H. Mukuda et al.:
554:   J. Phys. Soc. Jpn. {\bf 67} (1998) 2101.
555: 
556: \bibitem{Bobroff}
557:   J. Bobroff et al.:
558:   Phys. Rev. Lett. {\bf 86} (2001) 4116.
559: 
560: \bibitem{Takigawa}
561:   M. Takigawa, M. Ichioka and K. Machida:
562:   Phys. Rev. Lett. {\bf 83} (1999) 3057;
563:   M. Takigawa, M. Ichioka and K. Machida:
564:   J. Phys. Soc. Jpn. {\bf 69} 3943 (2000).
565: 
566: \bibitem{Morr}
567:   D. K.  Morr and R. Wortis:
568:   Phys. Rev. B {\bf 61} (2000) R882;
569:   R. Wortis, A. J. Berlinsky and C. Kallin:
570:   Phys. Rev. B {\bf 61} (2000) 12342;
571:   D. K. Morr:
572:   Phys. Rev. B {\bf 63} (2001) 214509.
573: 
574: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
575: 
576: \bibitem{Maeno}
577:   Y. Maeno et al.:
578:   Nature (London) {\bf 372} (1994) 532.
579: 
580: \bibitem{Rice}
581:   T. M. Rice and M. Sigrist:
582:   J. Phys.: Condens. Matter {\bf 7} (1995) L643;
583:   T. M. Rice:
584:   Nature (London) {\bf 396} (1998) 627;
585:   M. Sigrist et al.:
586:   Physica C {\bf 317-318} (1999) 134.
587: 
588: \bibitem{Baskaran}
589:   G. Baskaran:
590:   Physica B {\bf 223-224} (1996) 490.
591: 
592: \bibitem{Luke}
593:   G. M. Luke et al.:
594:   Nature (London) {\bf 394} (1998) 558.
595: 
596: \bibitem{Ishida98}
597:   K. Ishida et al.:
598:   Nature (London) {\bf 396} (1998) 658.
599: 
600: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
601: 
602: \bibitem{Leadon}
603:   R. Leadon and H. Suhl:
604:   Phys. Rev. {\bf 165} (1968) 596.
605: 
606: \bibitem{Okuno}
607:   Y. Okuno, M. Matsumoto and M. Sigrist:
608:   J. Phys. Soc. Jpn. {\bf 68} (1999) 3054.
609: 
610: \bibitem{Mackenzie}
611:   A. P. Mackenzie et al.:
612:   Phys. Rev. Lett. {\bf 80} (1998) 161.
613: 
614: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
615: 
616: \end{thebibliography}
617: 
618: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
619: 
620: \end{document}
621: