1: %
2: %&LaTeX
3: % ****** Start of file template.aps ****** %
4: %
5: % This file is part of the APS files in the REVTeX 3.1 distribution.
6: % Version 3.1 of REVTeX, July 1, 1996.
7: %
8: % Copyright (c) 1996 The American Physical Society.
9: %
10: % See the REVTeX 3.1 README file for restrictions and more information.
11: %
12: %
13: % This is a template for producing files for use with REVTEX 3.1.
14: % Copy this file to another name and then work on that file.
15: % That way, you always have this original template file to use.
16: %
17:
18: \documentstyle[prl, aps, amssymb]{revtex}
19: \begin{document}
20: % \draft command makes pacs numbers print
21: \draft
22: \wideabs{
23: \title{Systematic $^{63}$Cu NQR and $^{89}$Y NMR Study of
24: Spin Dynamics in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ Across
25: Superconductor-Insulator Boundary}
26: % repeat the \author\address pair as needed
27: \author{P.M. Singer and T. Imai}
28: \address{Department of Physics and Center for Materials Science and
29: Engineering, M.I.T., Cambridge, MA 02139}
30: \date{\today}
31: \maketitle
32: \begin{abstract}
33: We demonstrate that spin dynamics in underdoped
34: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ for $y \simeq$ 6.0 exhibit
35: qualitatively the same behavior to underdoped La$_{2-x}$Sr$_{x}$CuO$_{4}$
36: for an equal amount of hole concentration $p=z/2=x \leq 0.11$. However,
37: a {\it spin--gap} appears as more holes are doped
38: into the CuO$_{2}$ plane by increasing the oxygen concentration to
39: $y \simeq 6.5$
40: for a fixed value of Ca concentration $z$. Our results also suggest
41: that Ca doping causes disorder effects that enhance the
42: low frequency spin fluctuations.
43: \end{abstract}
44: % insert suggested PACS numbers in braces on next line
45: \pacs{76.60.-k, 74.72.Bk, 74.25.Dw}
46: }
47:
48: The mechanism of high temperature superconductivity remains a major
49: mystery in condensed matter physics. The fundamental
50: difficulty stems from the complexity of the electronic phase diagram,
51: particularly in the underdoped region. Earlier $^{63}$Cu NMR
52: (Nuclear Magnetic Resonance) and
53: NQR (Nuclear Quadrupole Resonance)
54: measurements of the nuclear spin-lattice relaxation rate $^{63}1/T_{1}$ led to the discovery of
55: the {\it pseudo--gap} phenomenon in the spin excitation spectrum of
56: bi-layer (Y,La)Ba$_{2}$Cu$_{3}$O$_{y}$
57: \cite{Horvatic,Warren,Yasuoka}. In the {\it spin pseudo--gap}, or {\it
58: spin--gap} regime, low energy spin excitations are
59: suppressed below the spin--gap temperature $T^{*}$ ($>T_{c}$) which
60: results in a decrease in $^{63}1/T_{1}T$
61: ($^{63}1/T_{1}$ divided by temperature $T$) below $T^{*}$.
62: Subsequent NMR
63: and optical charge transport measurements showed that a
64: pseudo--gap appears both in the spin and charge excitation spectrum
65: of a wide variety of high $T_{c}$ cuprates
66: \cite{Schlesinger,Homes,Itoh,Julien,Ishida}, with the most notable exception being the prototype
67: high $T_{c}$ cuprate La$_{2-x}$Sr$_{x}$CuO$_{4}$.
68: Moreover, the temperature scale $T^{*}$ of the spin--gap and charge--gap increases with decreasing
69: hole concentration towards the superconductor--insulator boundary
70: \cite{Yasuoka,Homes,Itoh,Ishida}.
71:
72: In various theoretical model analysis, the pseudo--gap is often considered
73: the key in understanding the mechanism of superconductivity. Unfortunately,
74: driving CuO$_{2}$ planes into the insulating regime in a controlled fashion
75: is technically difficult in many high $T_{c}$ cuprates. As such, the fate
76: of the pseudo-gap in the
77: heavily underdoped insulating regime has been highly controversial.
78: Attempts have been made to infer information on the
79: spin-gap based on uniform spin susceptibility $\chi'({\bf q}={\bf 0})$
80: \cite{Hwang,Oda}.
81: However, it is important to realize
82: that growth of short range spin order alone causes a reduction of $\chi'({\bf q}={\bf 0})$
83: {\it without} having any gaps.
84: For example, the undoped CuO$_{2}$ plane shows a roughly
85: linear decrease of $\chi'({\bf q}={\bf 0})$ with decreasing
86: temperature which is entirely consistent with the 2-d Heisenberg
87: model. By continuity, it is natural to associate the
88: decrease of $\chi'({\bf q}={\bf 0})$ in the heavily underdoped regime
89: to be mostly due to growth of short range order \cite{Johnston} and not to a spin-gap.
90:
91: %We recall that Johnston \cite{Johnston} used
92: %he 2-d Heisenberg model to describe
93: %the systematic variation of $\chi'({\bf q}={\bf 0})$ based on the
94: %effective energy scale $J(p)$ of short range order, which decreases
95: %with increasing $p$, and in the
96: %limit $p=0$, is equal to the spin-spin exchange interaction $J(0)=1500$ K.
97:
98: It has also become increasingly popular to infer $T^{*}$
99: for the charge sector based on scaling analysis of the Hall effect
100: \cite{Hwang} or resistivity data \cite{Cooper}. Some authors claim that there is a universal phase
101: diagram of $T^{*}$ with $p$ and $T$ being the only two parameters, even
102: in La$_{2-x}$Sr$_{x}$CuO$_{4}$. However, earlier $^{63}$Cu NQR
103: \cite{Ohsugi,Imai} and neutron scattering \cite{Yamada-PRL} experiments
104: revealed no hint of a spin--gap above $T_{c}$ in La$_{2-x}$Sr$_{x}$CuO$_{4}$.
105: Instead, La$_{2-x}$Sr$_{x}$CuO$_{4}$ exhibits an instability at low temperatures
106: towards the formation of the quasi--static {\it stripe} with incommensurate
107: spin and charge density waves \cite{Tranquada,Lee,Wakimoto}.
108: Careful NMR(NQR) experiments of spin--gap effects with controlled
109: doping near the superconductor-insulator boundary are necessary and would
110: allow for comparison with La$_{2-x}$Sr$_{x}$CuO$_{4}$.
111:
112: In this Letter, we report a systematic microscopic investigation of
113: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ utilizing $^{63}$Cu NQR
114: and $^{89}$Y NMR.
115: The advantage of the Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ system
116: is that one can control the hole concentration near the superconductor-insulator
117: boundary by fixing $y \simeq 6.0$ and varying $z$. In this case, the hole
118: concentration is given by $p=z/2$, because the chain
119: Cu sites (Cu(1)) with two-fold oxygen coordination remain insulating with a
120: 3d$^{10}$ configuration \cite{Vega}. In Fig. 1(a) we show the {\it absence} of a
121: spin-gap signature in
122: heavily underdoped Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$
123: based on measurements of $^{63}1/T_{1}T$ at the planar Cu site (Cu(2)).
124: Instead, we show that the low energy spin excitations exhibit similar behavior to underdoped
125: La$_{2-x}$Sr$_{x}$CuO$_{4}$, with equivalent
126: $p=z/2=x$, which monotonically grow with decreasing $p$ and $T$.
127: With further hole doping Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ by oxygen loading
128: to $y \simeq 6.5$ for fixed $z$,
129: however, we $do$ observe the spin-gap signature (Fig. 1(a)), even though it
130: appears somewhat suppressed compared to YBa$_{2}$Cu$_{3}$O$_{6.5}$
131: without Ca substitution (Fig. 1(b)).
132: This is the first time in the high $T_{c}$ cuprates where the appearance
133: of a spin--gap signature is experimentally tracked through the
134: insulator-superconductor boundary by increasing the hole doping. We
135: recall that in contrast with the present case, further hole doping
136: La$_{2-x}$Sr$_{x}$CuO$_{4}$ (or La$_{2}$CuO$_{4+\delta}$) does
137: $not$ result in a spin--gap signature,
138: therefore our finding challenges the popular argument that
139: charge disorder caused by the alloying effects of Sr$^{+2}$
140: substitution (i.e. "dirt effects") {\it alone}
141: suppresses the spin-gap and drives La$_{2-x}$Sr$_{x}$CuO$_{4}$ towards
142: the stripe instability.
143:
144: We synthesized our polycrystalline samples following \cite{Casalta}. The
145: oxygen concentration was controlled and determined following
146: \cite{Ueda} with precision $\Delta y \sim \pm$0.05.
147: The $^{63}$Cu NQR spectrum
148: of all our Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ samples
149: are very similar to those
150: reported earlier by Vega {\it et al.} \cite{Vega}, and the superconducting transition temperature
151: $T_{c}$, as determined by SQUID measurements, shows close agreement with \cite{Casalta}.
152: The temperature dependence of $^{63}1/T_{1}$ was measured with NQR near
153: the peak frequency of the $^{63}$Cu(2) site at $\omega_{n}/2\pi \simeq 25.5$ MHz \cite{Vega} by
154: applying an inversion pulse prior to the spin echo sequence. A
155: typical $\pi/2$-pulse width of 3 $\mu$s was used. NMR measurements
156: at 9 Tesla in uniaxially aligned powder gave
157: identical results to NQR within uncertainties.
158: $^{63}1/T_{1}T$ is given by
159: \begin{equation}
160: ^{63} \frac{1}{T_{1}T} = \frac{ 2 k_{B}}{g^{2} \mu_{B}^{2} \hbar^{2}}
161: \sum_{{\bf q}} |^{63}A({\bf q}) |^{2} \frac{\chi''({\bf q},\omega_{n})}{\omega_{n}}
162: \label{T1}
163: \end{equation}
164: where $^{63}A({\bf q})$ is the wave-vector dependent, geometrical form factor
165: of the electron-nucleus hyperfine
166: coupling \cite{Moriya,MMP}.
167: As shown in Fig. \ref{T1T}(a), $^{63}1/T_{1}T$ in underdoped
168: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ ($z\leq 0.22$) with nominal
169: hole concentration $p\leq 0.11$ does {\it not} exhibit a spin--gap.
170: Instead, $^{63}1/T_{1}T$
171: grows with decreasing temperature, exhibiting similar
172: values as La$_{2-x}$Sr$_{x}$CuO$_{4}$
173: for the equivalent hole concentration $p=z/2=x$ shown in Fig.
174: \ref{89Y}(a).
175: The fact that $^{63}1/T_{1}T$
176: grows with decreasing temperature indicates that
177: low energy spin excitations
178: continue to increase with decreasing temperature.
179: Moreover, the enhancement of low energy spin
180: excitations below 300 K is followed by the decrease of the
181: $^{63}$Cu NQR signal intensity below $T_{wipeout} (\gtrsim
182: 200$ K), i.e. {\it wipeout} effects \cite{Hunt,Singer,Hunt2}.
183: Wipeout effects can be caused by various
184: mechanisms \cite{Hunt} including the presence of nearly localized free spins
185: induced by hole localization (in
186: analogy with Cu NMR wipeout in Cu metal imbedded with
187: dilute Fe or Mn spins), as well as the onset of the
188: glassy slowing down of stripes. We note that as a consequence of
189: wipeout effects,
190: the value of $^{63}1/T_{1}T$ measured below $T_{wipeout}$ does {\it not}
191: represent that of the entire CuO$_{2}$ plane.
192:
193: The temperature dependence of $\chi'({\bf q}={\bf 0})$ was deduced
194: from the spin contribution
195: $^{89}K_{spin}=D\chi'({\bf q}={\bf 0})$ to the $^{89}$Y NMR Knight shift,
196: \begin{equation}
197: ^{89}K = ^{89}K_{orb}+^{89}K_{spin}
198: \label{T1}
199: \end{equation}
200: as shown in Fig. \ref{89Y}(b), where the powder averaged orbital contribution is
201: $^{89}K_{orb}=+150 \pm 5$ ppm \cite{Alloul} and $D$ ($< 0$) is the hyperfine coupling
202: constant. Our $^{89}K$ data, taken in a
203: magnetic field of 9 Tesla, for
204: Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{y}$ are consistent with
205: earlier results reported above $\sim$110 K for
206: Y$_{0.8}$Ca$_{0.2}$Ba$_{2}$Cu$_{3}$O$_{y}$
207: by Williams {\it et al.} \cite{Williams} also shown in Fig. \ref{89Y}(b).
208: Our new measurement conducted down to $T_{c}=75 $ K
209: in Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{6.9}$ shows clear
210: signature of saturation {\it below} 100 K,
211: similar to overdoped YBa$_{2}$Cu$_{3}$O$_{7}$ without Ca substitution
212: \cite{Auler}. The saturation of $\chi'({\bf q}={\bf 0})$ below 100 K is followed by a broad maximum at
213: $T_{max} = 90 \pm 10$ K, which according to the 2-d Heisenberg model
214: \cite{Singh}, implies an effective energy scale $J(p)=T_{max}/0.93 = 97\pm 11$ K.
215: We also deduce $J(p)$ in underdoped
216: Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ and
217: Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{6.5}$ by matching $\chi'({\bf
218: q}={\bf 0})$ to the low
219: temperature ($T \ll J(p)$) portion of the 2-d Heisenberg model
220: \cite{Johnston}, as shown in Fig. \ref{89Y}(b).
221: Our results of
222: $J(p)$ summarized in Fig. \ref{Phase} are consistent with those reported in
223: La$_{2-x}$Sr$_{x}$CuO$_{4}$ \cite{Johnston}.
224:
225: The $^{89}$Y NMR data also shows evidence for the glassy slowing of disordered magnetism.
226: Below the onset of glassy slowing at $T_{wipeout}
227: (\gtrsim 200$ K), we find a change in
228: curvature of $^{89}1/T_{1}T$ and an increase in $^{89}\Delta
229: f$, as shown in Fig. \ref{89Y}(c) and (d) respectively \cite{Carretta}. The change in
230: curvature of $^{89}1/T_{1}T$ is first followed by a minimum at
231: $T_{89}^{min}$ \cite{K}, and then a maximum at
232: $T_{89}^{max}$ where the glassy slowing has reached
233: the NMR time--scale.
234: At a similar temperature $T_{\mu SR}$, $\mu$SR measurements
235: observe local hyperfine fields \cite{Niedermayer} that are frozen on
236: the $\mu$SR time--scale. The enhanced values of $^{89}\Delta f$ at
237: 1.7 K also indicate that the frozen
238: hyperfine fields at the $^{89}$Y nuclear site have a substantial spatial distribution $\sim 70$ Oe.
239:
240: The sequence of anomalies starting at $T_{wipeout}$ followed by
241: $T_{89}^{min}$, $T_{89}^{max}$ and $T_{\mu SR}$ (all shown in Fig.
242: \ref{Phase}) are analogous to La$_{2-x}$Sr$_{x}$CuO$_{4}$
243: where the on-set of glassy slowing down of the stripe phase
244: at $T_{wipeout}$ \cite{Hunt,Singer,Hunt2} is
245: followed by a minimum then a maximum
246: in $^{139}1/T_{1}T$ \cite{Chou} and $\mu$SR observation of
247: frozen hyperfine fields \cite{Niedermayer}.
248: These set of results establish the
249: following three points: First, the paramagnetic Cu spin fluctuations in
250: underdoped CuO$_{2}$ planes
251: exhibit nearly universal $p$ and $T$ dependences in
252: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ and
253: La$_{2-x}$Sr$_{x}$CuO$_{4}$ for equivalent $p=z/2=x$, {\it
254: without} a spin--gap signature.
255: In the same temperature range, the $^{89}$Y NMR Knight shift decreases
256: monotonically and is most likely due to the
257: growth of short range spin order.
258: Second, the gradual slowing of Cu
259: spin fluctuations, as observed by the increase in
260: $^{63}1/T_{1}T$, is followed by glassy freezing of the Cu moments
261: starting at $T_{wipeout} \sim$ 200 K in
262: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ while similar behavior
263: is observed only below $\sim$100 K in
264: La$_{2-x}$Sr$_{x}$CuO$_{4}$ within a similar doping range.
265: The factor 2 higher temperature scale is consistent with the finding
266: based on $\mu$SR that the
267: spin freezing temperature $T_{\mu SR}$ in
268: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ is also a factor $\sim$2
269: higher than in La$_{2-x}$Sr$_{x}$CuO$_{4}$ \cite{Niedermayer}. Recalling that the
270: N\'{e}el temperature of $T_{N}$=420 K in undoped YBa$_{2}$Cu$_{3}$O$_{6.0}$ is
271: higher than $T_{N}$=320 K in La$_{2}$CuO$_{4}$ because of the bi-layer
272: coupling, the higher temperature scale of glassy spin freezing in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$
273: may also be due to the stronger 3-d coupling along the
274: $c$-axis. However, we cannot rule
275: out the possibility that Ca$^{+2}$ substitution causes stronger disorder in
276: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ than
277: Sr$^{+2}$ substitution in La$_{2-x}$Sr$_{x}$CuO$_{4}$, as suggested by
278: the factor $\sim 2$ broader $^{63}$Cu NQR spectrum \cite{Vega}, which may
279: enhance the tendency towards spin freezing. Third, the observed increase of
280: $^{89}1/T_{1}T$ implies that the Cu moments are not
281: slowing down towards the commensurate
282: antiferromagnetic spin structure with divergently large spin-spin correlation
283: length. In this context, it is important to recall that the critical
284: slowing down towards the N\'{e}el state does not cause a large
285: enhancement of $^{89}1/T_{1}T$ in undoped YBa$_{2}$Cu$_{3}$O$_{6.0}$ \cite{Ohno}
286: since the hyperfine form factor $^{89}A({\bf q})$ is zero for the commensurate wave vectors.
287: The strong increase of $^{89}1/T_{1}T$ shows that either
288: the spin structure is incommensurate, as expected for the stripe phase, or that the spin-spin
289: correlation length is limited to a relatively short length scale due
290: to disorder caused by the holes, or possibly both. Stripes are
291: dynamic at NMR time scales even at $\sim 350 $ mK as evidenced by
292: motional narrowing effects \cite{Hunt2}, therefore the exact spin
293: configuration cannot be distinguished using NMR.
294:
295: We have established that the slowing of the paramagnetic spin
296: dynamics in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ is
297: qualitatively similar to La$_{2-x}$Sr$_{x}$CuO$_{4}$. Most
298: importantly, we do not observe the signature of a spin gap. Instead, we
299: find signatures of glassy slowing of spin fluctuations similar to the
300: case of La$_{2-x}$Sr$_{x}$CuO$_{4}$.
301: We caution that the absence of a spin--gap signature in the form of a decrease
302: in $^{63}1/T_{1}T$ does not necessarily {\it prove} that there is no
303: global suppression of lower energy parts of the spin fluctuations.
304: $^{63}1/T_{1}T$ may grow monotonically with decreasing temperature as
305: long as very low frequency ($\sim \omega_{n}$) components of the spin fluctuations
306: grow, even if the global spin fluctuation spectrum is gapped below a
307: certain temperature $T^{*}$. On the other hand,
308: $T_{wipeout}$ sets an upperbound on
309: $T^{*}$ in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ because if $T^{*}$ is
310: significantly larger than $T_{wipeout}$, we should observe the
311: decrease of $^{63}1/T_{1}T$ prior to the influence of glassy slowing of
312: the spin dynamics which become visible below $T_{wipeout}$. Our finding
313: that the magnitude of $T^{*}$ is at most comparable to $T_{wipeout}$
314: in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$
315: is at odds with popularly held speculations, often based on
316: theoretical expectations or more indirect experimental information
317: such as the Hall effect, resistivity, and $\chi'({\bf q}={\bf 0})$, that $T^{*}$ blows up
318: towards $J(p=0)\sim 1500$ K.
319:
320: A potential common cause of the absence of the spin--gap signature in underdoped
321: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ and
322: La$_{2-x}$Sr$_{x}$CuO$_{4}$ is the random charge potential and/or
323: disorder induced by substitution of Ca$^{+2}$ or Sr$^{+2}$ ions into
324: Y$^{+3}$ or La$^{+3}$ sites respectively. It is worth recalling that
325: the absence of a spin--gap signature in La$_{2-x}$Sr$_{x}$CuO$_{4}$ has
326: often been attributed to ``dirt effects'' caused by Sr$^{+2}$.
327: However, our results in Fig. \ref{T1T}(a) also indicate
328: that disorder {\it alone} does not entirely suppress the spin--gap. Due to
329: the solubility limit of Ca$^{2+}$ into
330: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ with a maximum $T_{c} \sim
331: 30$ K \cite{Casalta},
332: we doped more holes into Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{6.0}$
333: by adding oxygen into the chain layers for the {\it same} sample
334: to obtain Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{6.50}$ with $T_{c}=59$ K. We found that
335: $^{63}1/T_{1}T$ in Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{6.50}$
336: decreases below $T^{*}\sim 130$ K, similar to the
337: spin--gap signature in
338: YBa$_{2}$Cu$_{3}$O$_{6.50}$ \cite{Yasuoka,ImaiData}.
339: The data therefore suggests that the spin--gap {\it does} develop when more holes are added
340: into the CuO$_{2}$ plane in
341: Y$_{0.78}$Ca$_{0.22}$Ba$_{2}$Cu$_{3}$O$_{y}$, even if Ca doping tends to suppress the spin--gap
342: signature.
343:
344: In order to test the effects of
345: Ca substitution in a
346: more systematic fashion, we compare $^{63}1/T_{1}T$
347: for Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.5\sim 6.55}$ with $z=0$,
348: 0.08, and 0.22 (Fig. \ref{T1T}(b)).
349: $^{63}1/T_{1}T$ is systematically enhanced with increasing $z$, especially at lower
350: temperatures. Our data suggest that Ca$^{+2}$ doping not
351: only introduces holes but also gives rise to disorder effects which tend to fill in the low frequency parts of
352: spin fluctuation spectrum, without affecting the
353: magnitude of $T^{*}$ significantly.
354: The Ca
355: substitution effects for $y \simeq 6.5$ is in remarkable
356: contrast with the lack of change in $^{63}1/T_{1}T$
357: observed for $y\simeq 6.9$ (Fig. 1). Our results for $y\simeq 6.9$ are
358: consistent with earlier reports \cite{Williams2}.
359: It is interesting to note the qualitative similarity with the Zn substitution effects
360: in YBa$_{2}$Cu$_{3}$O$_{y}$\cite{Mahajan}.
361: $^{89}$Y NMR by Mahajan {\it et al.} \cite{Mahajan} showed that Zn
362: substitution causes
363: $^{89}$Y line splitting in $T_{c} \simeq$60 K phase samples but
364: causes only $^{89}$Y NMR line broadening in the $T_{c} \simeq 90$ K phase
365: with $y\simeq 6.9$.
366: These results suggest that both random spinless
367: impurities in the CuO$_{2}$ plane (Zn$^{2+}$) and random Coulomb potentials
368: outside the CuO$_{2}$ plane (Ca$^{2+}$) are more effectively shielded
369: by a larger number of holes in the overdoped region. We mention
370: that a more detailed analysis of the $^{63}$Cu(2) spin-lattice
371: recovery, similar to that used for Zn doped
372: YBa$_{2}$Cu$_{4}$O$_{8}$ by Itoh {\it et al.} \cite{ItohZn}, is unfortunately not
373: possible in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$
374: due to the small overlap ($\sim 1 \%$) of the Cu(1) signal with
375: very long $^{63}T_{1}$ \cite{Vega}.
376:
377: To conclude, using both $^{63}$Cu NQR and $^{89}$Y NMR we
378: demonstrate the remarkable
379: similarity in the paramagnetic spin dynamics between
380: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{6.0}$ and La$_{2-x}$Sr$_{x}$CuO$_{4}$
381: for equivalent nominal hole concentration.
382: We do not observe any signatures of a spin--gap for $p=z/2=x\leq 0.11$.
383: Upon further hole doping by oxygen loading with fixed $z$, we demonstrate that a
384: spin--gap $does$ develop. Combining all our data, we deduce a phase diagram which crosses the
385: superconductor--insulator boundary and
386: includes the spin--gap temperature $T^{*}$, the effective energy scale
387: $J(p)$ ($\gg T^{*}$),
388: and the glassy freezing of the spin dynamics. Our systematic study
389: of Ca substitution suggest that charge disorder caused by
390: Ca$^{+2}$ ions tends to suppress the spin--gap signature while keeping
391: $T^{*}$ ($< T_{wipeout}$) roughly constant.
392:
393: T.I. thanks M. Greven, C. Nayak, S. Chakravarty, and X.-G. Wen for inspiring this project.
394: This work was supported by NSF DMR 99-71264 and NSF DMR 98-08941.
395:
396: \begin{references}
397: \bibitem{Horvatic}M. Horvati\'{c} {\it et al.}, Phys. Rev. B {\bf 39}, 7332 (1989).
398: \bibitem{Warren}W.W. Warren {\it et al.}, Phys. Rev. Lett. {\bf 62}, 1193 (1989).
399: \bibitem{Yasuoka}H. Yasuoka, T. Imai, and T. Shimizu, in {\it Strong
400: Correlation and Superconductivity}, Eds. H.
401: Fukuyama, S. Maekawa, and A. P. Malozemoff, Springer-Verlag (1989).
402: Also see A. Goto {\it et al.}, Phys. Rev. B {\bf 55}, 12736 (1997) and
403: A. Goto {\it et al.}, J. Phys. Soc. Jpn. {\bf 65}, 3043 (1996).
404: \bibitem{Schlesinger}Z. Schlesinger {\it et al.}, Phys. Rev. Lett. {\bf 65}, 801 (1990).
405: \bibitem{Homes}C.C. Homes {\it et al.}, Phys. Rev. Lett. {\bf 71}, 1645 (1993).
406: \bibitem{Itoh}Y. Itoh {\it et al.}, J. Phys. Soc. Jpn. {\bf 63}, 22 (1994)
407: and {\it ibid} J. Phys. Soc. Jpn. {\bf 65}, 3751 (1996).
408: \bibitem{Julien}M.-H. Julien {\it et al.}, Phys. Rev. Lett. {\bf 76}, 4238 (1996).
409: \bibitem{Ishida}K. Ishida {\it et al.} Phys. Rev. B {\bf 58}, R5960 (1998).
410: \bibitem{Hwang}H.Y. Hwang {\it et al.} Phys. Rev. Lett. {\bf 72}, 2636 (1994).
411: \bibitem{Oda}T. Nakano {\it et al.}, J. Phys. Soc. Jpn. {\bf 67}, 2622 (1998).
412: \bibitem{Johnston}D.C. Johnston, Phys. Rev. Lett. {\bf 62}, 957 (1989).
413: \bibitem{Cooper}J.R. Cooper {\it et al.}, Physica C {\bf 341}, 855 (2000).
414: \bibitem{Ohsugi}S. Ohsugi {\it et al.} J. Phys. Soc. Jpn. {\bf 60}, 2351 (1991).
415: \bibitem{Imai}T. Imai {\it et al.} Phys. Rev. Lett. {\bf 70}, 1002 (1993).
416: \bibitem{Yamada-PRL}K. Yamada {\it et al.} Phys. Rev. Lett. {\bf 75}, 1626 (1995).
417: \bibitem{Tranquada}J.M. Tranquada {\it et al.} Nature {\bf 375}, 561 (1995).
418: \bibitem{Lee}Y.S. Lee {\it et al.} Phys. Rev. B. {\bf 60}, 3643 (1999).
419: \bibitem{Wakimoto}S. Wakimoto {\it et al.} Phys. Rev. B {\bf 63}, 172501 (2001).
420: \bibitem{Vega}A.J. Vega {\it et al.}, Phys. Rev. B {\bf 39}, 2322 (1989)
421: and {\it ibid} Phys. Rev. B {\bf 40}, 8878 (1989).
422: \bibitem{ImaiData}T. Imai {\it et al.}, Physica (Amsterdam) {\bf 162C}, 169 (1989).
423: \bibitem{Casalta}H. Casalta {\it et al.}, Physica (Amsterdam) {\bf 204C}, 331 (1993).
424: \bibitem{Ueda}Y. Ueda {\it et al.}, Physica (Amsterdam) {\bf 156C}, 281 (1988).
425: \bibitem{Moriya}T. Moriya, J. Phys. Soc. Jpn. {\bf 18}, 516 (1963).
426: \bibitem{MMP}A.J. Millis {\it et al.}, Phys. Rev. B {\bf 42}, 167 (1990).
427: \bibitem{Hunt}A.W. Hunt {\it et al.}, Phys. Rev. Lett. {\bf 82}, 4300
428: (1999). Also see T. Imai {\it et al.}, J. Phys. Soc. Jpn. {\bf 59}, 3846 (1989).
429: \bibitem{Singer}P.M. Singer {\it et al.}, Phys. Rev. B {\bf 60}, 15345 (1999).
430: \bibitem{Hunt2}A.W. Hunt {\it et al.}, Phys. Rev. B {\bf 64}, 134525 (2001).
431: \bibitem{K}$T_{89}^{min}$ also signals the temperature below
432: which $^{89}1/T_{1}T \propto$ $^{89}K_{spin}$ ($\propto$ $^{89}\Delta
433: f$) scaling \cite{Alloul} no longer holds.
434: \bibitem{Williams}G.V.M. Williams {\it et al.}, Phys. Rev. B {\bf 54}, R6909 (1996)
435: and {\it ibid} Phys. Rev. B {\bf 57}, 8696 (1998).
436: \bibitem{Singh}R.R.P. Singh {\it et al.}, Phys. Rev. B. {\bf 42}, 996 (1990).
437: \bibitem{Alloul}H. Alloul {\it et al.}, Phys. Rev. Lett {\bf 70}, 1171 (1993).
438: \bibitem{Auler}T. Auler {\it et al.}, Phys. Rev. B {\bf 56}, 11294 (1997).
439: \bibitem{Niedermayer}Ch. Niedermayer {\it et al.}, Phys. Rev. Lett. {\bf 80}, 3843 (1998).
440: \bibitem{Carretta}Independent $^{89}$Y NMR data in
441: Y$_{0.85}$Ca$_{0.15}$Ba$_{2}$Cu$_{3}$O$_{6.1}$ by
442: A. Campana {\it et al.}, Int. J. Mod. Phys. B {\bf 14}, 2797 (2000) are
443: consistent with our data.
444: \bibitem{Chou}F.C. Chou {\it et al.}, Phys. Rev. Lett. {\bf 71}, 2323 (1993).
445: \bibitem{Ohno}T. Ohno {\it et al.}, J. Phys. Soc. Jpn. {\bf 60}, 2040 (1991).
446: \bibitem{Williams2}G.V.M. Williams {\it et al.}, Phys. Rev. B {\bf 63}, 104514 (2001).
447: \bibitem{Mahajan}A.V. Mahajan {\it et al.}, Euro. Phys. J. B {\bf 13}, 457 (2000).
448: \bibitem{ItohZn}Y. Itoh {\it et al.}, J. Phys. Soc. Jpn. {\bf 70}, 1881 (2001).
449: \end{references}
450:
451: \begin{figure}
452: \caption{(a) $^{63}1/T_{1}T$ above $T_{c}$ (vertical lines) at the
453: $^{63}$Cu(2) site for
454: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ (where $[z$ $y$ $T_{c}]$
455: is indicated in (b)) and for
456: La$_{1.885}$Sr$_{0.115}$CuO$_{4}$ (dotted curve).
457: ($\blacksquare$) and ($\square$) are taken from [20]. All lines are a guide
458: for the eye. Dashed line through ($\blacktriangle$) indicates region below
459: $T_{wipeout}$ where only partial
460: Cu(2) signal intensity exists.}
461: \label{T1T}
462: \end{figure}
463:
464: \begin{figure}
465: \caption{The same symbol assignment as Fig. 1 is used, and new
466: symbols are shown. (a) $^{63}1/T_{1}T$
467: in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$
468: and La$_{2-x}$Sr$_{x}$CuO$_{4}$.
469: (b) $\chi'({\bf q=0})$ in Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ as measured by the $^{89}$Y NMR Knight shift
470: ($^{89}K$)
471: taken above $T_{89}^{min}$ [28] with respect to a YCl$_{3}$ reference.
472: Data ($\times$) from [29] are a series of
473: $[0.20$ $y$ $T_{c}]$ samples with
474: $T_{c}$ = 47.5 K, 65.8 K, 83.2 K, 86 K, 72.1 K, 60 K, and 47.5 K (in order of
475: increasing $|^{89}K_{spin}|$).
476: Arrow indicates net spin contribution $^{89}K_{spin}$. Solid lines are fits to the
477: 2-d Heisenberg model [30], and all other lines in figure are guides for the eye.
478: (c) $^{89}1/T_{1}T$ (with same data plotted below 30 K in the inset) and
479: (d) full width at half maximum $^{89}\Delta f$ of the $^{89}$Y NMR
480: line--shape.}
481: \label{89Y}
482: \end{figure}
483:
484: \begin{figure}
485: \caption{Phase diagram of
486: Y$_{1-z}$Ca$_{z}$Ba$_{2}$Cu$_{3}$O$_{y}$ as a function of
487: Ca$_{z}$ substitution for fixed $y \simeq 6.0$ to the
488: left of dashed vertical line, and as a function of O$_{y}$ concentration
489: for fixed $z=0.22$ to the right.
490: The data includes $T_{N}$ ($\blacklozenge$) [21], $T_{\mu SR}$ ($\blacktriangle$)
491: [33], $T_{89}^{min}$
492: ($\blacktriangledown$), $T_{89}^{max}$ ($\triangledown$),
493: $T_{wipeout}$ (hatched region), $T^{*}$ ($\bullet$),
494: $T_{c}$ ($\lozenge$) and
495: $J(p)$ ($\ast$,$\times$) deduced according to $T_{max}$ and
496: fit to the 2-d Heisenberg model, respectively.
497: All data to the right, including $T_{c}$ ($+$) from [29], are
498: positioned linearly according to $|^{89}K_{spin}|$ at 300 K. All lines are a guide for the eye and
499: ($\circ$) is $T^{*}$ for [0.08 6.5 62K].}
500: \label{Phase}
501: \end{figure}
502:
503: \end{document}
504:
505:
506:
507:
508:
509:
510:
511:
512:
513:
514:
515: