1: 

2: 

3: \documentstyle[prl,aps,epsfig,multicol]{revtex}

4: %\documentclass[aps,prl,twocolumn]{revtex4}

5: %\textwidth=16truecm

6: %\textheight= 20truecm

7: %\hoffset=-1truecm

8: %\voffset=-2truecm

9: 

10: \begin{document}

11: %\preprint{IMSc-2003/06/16}

12: \def\be{\begin{equation}}

13: \def\ee{\end{equation}}

14: \def\bearr{\begin{eqnarray}}

15: \def\eearr{\end{eqnarray}}

16: \def\tc{$T_c~$}

17: \def\tcl{$T_c^{1*}~$}

18: \def\c2{ CuO$_2~$}

19: \def\ruo{ RuO$_2~$}

20: \def\lsco{LSCO~}

21: \def\bi{bI-2201~}

22: \def\tl{Tl-2201~}

23: \def\hg{Hg-1201~}

24: \def\sro{$Sr_2 Ru O_4$~}

25: \def\rc{$RuSr_2Gd Cu_2 O_8$~}

26: \def\mgb{$MgB_2$~}

27: \def\pz{$p_z$~}

28: \def\ppi{$p\pi$~}

29: \def\sqo{$S(q,\omega)$~}

30: \def\tperp{$t_{\perp}$~}

31: \def\cob{$CoO_2$~}

32: \def\nxcob{$Na_x CoO_2.yH_2O$~}

33: \def\ncob{$Na_{0.5} CoO_2$~}

34: \def\half{$\frac{1}{2}$~}

35: \def\nycob{$A_xCoO_{2+\delta}$~}

36: \def\naxcob{$Na_xCoO_2$~}

37: \def\wat{$H_2O$~}

38: \def\na{$Na$~}

39: \def\ewat{$\epsilon_{\rm H_2O}$~}

40: 

41: 

42: \title{How Ice enables Superconductivity in \nxcob by melting charge order:\\

43: Possibility of novel Electric Field Effects}

44: 

45: \author{ G. Baskaran\\

46: The Institute of Mathematical Sciences\\

47: C.I.T. Campus, Chennai 600 113, India }

48: 

49: %\date{today}

50: 

51: \maketitle

52: 

53: \begin{abstract}

54: 

55: Charge ordering in doped \cob planes near the commensurate fillings

56: $x = \frac{1}{4}$ and $\frac{1}{3}$ are considered for \nxcob and 

57: suggested to be competitors to superconductivity, leading to 

58: the experimentally seen narrow superconducting dome bounded by 

59: commensurate doping: $\frac{1}{4}<x< \frac{1}{3}$. 

60: Intercalated hydrogen bonded \wat 

61: network, by its enhanced dielectric constant, screen and frustrate 

62: local {\em charge order condensation energy} and replace a generic 

63: `charge glass order' by superconductivity in the dome.  An access 

64: to superconductivity and charge order, available through the new 

65: water channel, is used to predict novel effects such as 

66: `Electrical Modulation of Superconductivity' and `Electroresistance

67: Effect'.

68: \end{abstract}

69: 

70: %\pacs{PACS number: 13.20.He, 11.30.Er, 12.15Hh, 11.80Et}

71: 

72: %\maketitle

73: 

74: \begin{multicols}{2}[]

75: 

76: Discovery of superconductivity in \nxcob by Takada and 

77: collaborators\cite{takada} have opened the possibility of realizing

78: unconventional superconductivity and novel quantum states in 2D arising 

79: from strong electron correlations in doped \cob layers. Water of a 

80: right proportion ($y \approx \frac{4}{3}$) seems absolutely 

81: necessary\cite{cava1,cava2,chu,jin} for stabilizing 

82: superconductivity, suggesting $H_2O$'s critical role. While water does 

83: wonders in nature, its key role here is some what puzzling. Elucidating 

84: its role in this unusual superconductor is an important 

85: task from material science and physics point of view.  This is what 

86: the present paper attempts using phenomenological and theoretical

87: considerations.

88: 

89: Enthused by the remarkable discovery of superconductivity in \nxcob,

90: the present author\cite{gbcob} and others\cite{rvbcob}

91: have suggested a single band t-J model

92: as an appropriate model to understand superconductivity and low energy 

93: electronic phenomena. A phase diagram has been suggested using ideas 

94: of resonating valence bond (RVB) theory developed for 

95: cuprates\cite{pwagbrvb}. A recent 

96: experiment\cite{cava2} which shows superconductivity in a 

97: rather narrow range 

98: of doping $\frac{1}{4}<x< \frac{1}{3}$, than predicted by RVB theories, 

99: suggest that there are perhaps left out interactions and consequent 

100: competing phases which make a simple t-J model valid only for 

101: limited range of $x$. The situation is not unusual - even in cuprates 

102: a simple t-J modeling is strictly valid only in the neighborhood of 

103: optimal doping. Charge order phenomenon\cite{stripe,stripe1} is known 

104: in cuprate superconductors; and it has been suggested to  

105: compete\cite{gbstripe} with superconductivity.

106: 

107: Doped \cob, compared to $CuO_2$ layers of high \tc cuprates, has a 

108: narrower conduction band and a less polarizable valence band of oxygen. 

109: Consequently, short range coulomb repulsions among carriers  

110: are screened less. This is likely to stabilize a variety of frustrated 

111: charge ordering in the triangular lattice, as we discuss in this 

112: paper. NMR result of Ray et al.\cite{susc} indeed provides a first evidence 

113: for charge freezing in \naxcob family ($x = \frac{1}{2}$), \ncob below 

114: about $T \approx 300 K$, a large temperature scale. 

115: 

116: We estimate and include unscreened short range coulomb interactions 

117: in the t-J model for the study of unhydrated \naxcob. 

118: We show that in addition to superconductivity, charge ordering in the 

119: narrow conduction band of \cob layer is a major instability, for a 

120: range of higher doping than suspected (Wen et al. in ref 7). 

121: As the unscreened

122: near neighbor coulomb interactions are large and comparable to the

123: band width, the characteristic charge order temperatures are

124: $T_{\rm ch} \sim 400 K$. 

125: 

126: Fortunately, \wat, in hydrated \nxcob, makes t-J modeling valid 

127: for a range of doping. For reasons which we elaborate in the 

128: present paper, hydrogen bonded \wat dipoles of the ice layers screen and 

129: frustrate {\em charge order condensation energy}. That is,

130: they {\em effectively 

131: screen out short range repulsions}, and enable the physics of a simple 

132: t-J model to be realized in a narrow range of doping, 

133: $\frac{1}{4}<x< \frac{1}{3}$, as superconductivity. 

134: 

135: We discuss few important charge ordered states at commensurate 

136: fillings, $x = \frac{1}{4}$ and $ \frac{1}{3}$, which we believe are

137: competitors to the experimentally observed superconductivity, in the 

138: range $\frac{1}{4} < x < \frac{1}{3}$. These reference charge ordered 

139: states are strongly frustrated by the random potential from the 

140: neighboring $Na$ layers, resulting in a glassy phase in the region

141: $\frac{1}{4} < x < \frac{1}{3}$ and beyond. The charge glass phase

142: is likely to be an anomalous metal, very much like the spin gap phase

143: in cuprates, where there are local charge order activities

144: at low frequency scales. 

145: 

146: We estimate the enhancement of the background static dielectric 

147: constant at short distance due to hydrogen bonding in the \wat layers.

148: We find that this screening is sufficient to reduce the large charge

149: order transition temperature down to $\sim 1 K$ and allow 

150: superconductivity to emerge. Strong commensurability effects and

151: the associated short range charge order reduce superconducting \tc 

152: considerably as we approach the commensurate ends $ x = \frac{1}{4} $

153: and $\frac{1}{3}$.

154: 

155: As \wat stabilize superconductivity and discourages charge

156: order, we have a new access to the electronic phases of \cob layer

157: through water. This leads to the possibility of some novel effects: i) 

158: `Electrical Modulation of Superconductivity' 

159: by external electric field or microwave radiation and 

160: ii) `Electroresistance Effect' in the normal state.

161: We estimate that voltages $\sim 500 V$, applied capacitively 

162: to thin films of \nxcob of thickness $\sim~1$ micron will 

163: orient the water dipoles and reduce the short distance 

164: dielectric screening, resulting in stabilization of charge glass order

165: phase and destabilization of superconductivity. This interesting 

166: switching effect may have device potential.

167: 

168: Recently we modeled the low energy physics of doped \cob using 

169: a t-J model and discussed an RVB scenario for superconductivity including

170: a PT violating $d_1+id_2$ wave superconductivity and a $p_1+ip_2$ wave 

171: superconductivity at a higher doping. To study charge order we must 

172: include some leading short distance carrier-carrier and carrier-$Na$-ion 

173: screened coulomb interaction: 

174: \bearr

175: H_{{\rm tJV}} =

176:  -t  \sum_{\langle ij\rangle} C^{\dagger}_{i\sigma} C^{}_{j\sigma}

177: + H.c. +  J \sum_{\langle ij\rangle} 

178: ({\bf S}_{i }\cdot{\bf S}_{j} - \frac{1}{4}n_i n_j) \nonumber \\

179: \sum_{ij} V_{ij} (n_i-1)( n_j-1) + \sum_{i}\epsilon_i (n_i -1)

180: \eearr

181: Here $C$'s and ${\bf S}$'s are the electron and spin operators.

182: As we have an electron doped system we have the `zero occupancy' 

183: constraint $\sum_{\sigma} n^{}_{i\sigma} \neq 0 $ at every site i. 

184: 

185: Recall that doubly occupied $Co^{3+}$ sites carry a charge 

186: $-e$ with reference to the neutral \cob layer and $V_{ij}$ is the 

187: screened coulomb repulsion between them. We have ignored the small two body 

188: off-diagonal coulomb interaction terms. For practical purposes only 

189: the nearest and next nearest neighbor terms $ V_1 \approx  

190: \frac{e^2}{\varepsilon_{ab} R_{nn}} e^{-\frac{R_{nn}}{\lambda_{ab}}}$

191: and $ V_2 \approx  \frac{e^2}{\varepsilon_{ab} R_{nnn}}

192: e^{-\frac{R_{nnn}}{\lambda_{ab}}}$ are important.

193: Here $\varepsilon_{ab} \approx \varepsilon_O + \varepsilon_{\rm {H_2O}} $ 

194: represents the short distance dielectric screening arising from the filled 

195: oxygen bands of \cob layers and $H_2O$ layers in \nxcob. And 

196: $\lambda_{ab} \approx Co-Co$ distance is the Thomas Fermi screening length 

197: for our tight binding metallic layer. The random site energy $\epsilon_i$ 

198: of charge degree of freedom represents the screened 

199: coulomb attraction from neighboring $Na^+$ ions.

200: 

201: Electronic structure calculations\cite{singh} give a value of 

202: $t \approx -0.1~eV$ for the conduction band of the \cob layer. 

203: We estimate $V_1$ and $V_2$ for \naxcob, 

204: the non-hydrated case. The dielectric constants of oxides of $Fe$ and $Ni$ 

205: that flank $Co$ in the periodic table are $\sim 4$ to $12$. We assume a 

206: background (short distance) static dielectric constant of 

207: $\varepsilon_{ab} \approx 8$ for our \cob layer. Recall that in 

208: cuprates the background $\epsilon$ is large $\sim 20$, in view of 

209: a more polarizable octahedral oxygen network; in \cob the oxygen filled

210: band is less polarizable and relatively deep below the fermi level.

211: Using this 

212: dielectric constant and values of Co-Co distances in \ncob 

213: we get $V_1 \approx 0.8~eV$ and $V_2 \approx 0.4~eV$. The mean square

214: fluctuation of the carrier site energy due to disordered $Na$-ions

215: is ${\sqrt{\langle \delta \epsilon_i^2 \rangle}} \approx 0.2~eV $ 

216: 

217: In the absence of hopping, the dopant carriers $Co^{3+}$ will order 

218: classically and undergo order-disorder transition at a fairly high 

219: temperature $k_B T_{\rm ch}(\rm classical)

220: \approx 2{\bar V}\sim 10^3~K$; here 

221: ${\bar V} \equiv \frac{1}{2}(V_1 + V_2)$ is a mean short distance 

222: repulsion. However the electron dynamics reduce 

223: $T_{\rm ch}(\rm classical)$ considerably. To estimate this reduction we

224: perform a mean field analysis of the t-J-V model for a CDW order, 

225: pretending that an unfrustrated charge order arises from nesting 

226: instability for \ncob. This gives us a BCS like expression for \tc:

227: \be

228: k_BT_{\em ch}  

229: \approx \epsilon_F e^{-\frac{1}{{\bar V}\rho_o}} 

230: \ee

231: Here $\rho_o$ is

232: a fermi sea averaged particle-hole density of states corresponding to 

233: the ordering wave vector. Substituting $\epsilon_F \approx 0.5~eV$, 

234: $\rho_0 \approx \frac{1}{2\epsilon_F}$ and ${\bar V} \approx 0.4$, we get  

235:  $T_{\rm ch} \approx 480 K$. Frustration on the triangular lattice at 

236:  half filling

237: and disorder effect from $Na$ ions will further reduce this. Thus we get

238: a charge order temperature in the right range, 

239: $T_{\rm ch}(NMR) \approx 300~K$, seen in NMR, 

240: 

241: \begin{figure}[h]

242: \epsfxsize 5.5cm

243: \centerline {\epsfbox{watfig1.eps}}

244: \caption{`Classical' charge order for $x = \frac{1}{4}$.

245: Spin-0, charge $-e$ carriers ($Co^{3+}$) form a triangular lattice. 

246: Neutral sites with spin-\half moments ($Co^{4+}$) form a Kagome lattice. 

247: Quantum fluctuations will reduce 

248: the amplitude of charge order substantially. Accompanying charge order, 

249: we expect some interesting spin liquid phase or complex short range spin 

250: order at low temperatures}

251: \end{figure}

252: Before we consider influence of \wat we discuss some simple charge 

253: orders at $x = \frac{1}{4}$ and $\frac{1}{3}$ that are favored by 

254: electrostatics in the unhydrated \naxcob. 

255: We ignore the superexchange contribution, as $J <<  V_1, V_2$. 

256: For $x= \frac{1}{4}$, the $Co^{3+}$ sites are arranged on a triangular

257: lattice (figure 1) to minimize coulomb repulsions. Interestingly, 

258: the remaining sites carry spins and form spin-\half Heisenberg

259: antiferromagnet on a Kagome lattice.  In our convention, the classical 

260: energy of this state is $E_{\frac{1}{4}}(\rm Kagome) = 0$. 

261: In the real system, carrier delocalization will considerably 

262: reduce the amplitude of charge order. {\em In this sense the charge ordered

263: states shown in figure 1 and figure 2 are to be thought of as reference

264: classical states}.

265: 

266: Another ground state comparable in energy is an anisotropic metal. It has 

267: ordered stripes - alternating insulating and 0.5 electron doped chains. 

268: The electrostatic energy of this state per site is $ \frac{1}{4}(V_1 + V_2)$.

269: However, the carrier delocalization in the 0.5 electron doped chains leads 

270: to a gain in kinetic energy which is easily estimated when J is 

271: neglected in our t-J model. This case corresponds to a quarter filled 

272: infinite U Hubbard model, which can be converted into a half filled band of 

273: non-interacting spinless fermions giving us the delocalization energy

274: $ = -|t|\sum \cos k = -2{\frac{|t|}{\pi}}$. Thus we get a total energy

275: per site, 

276: $E_{\frac{1}{4}}({\rm stripe}) = \frac{1}{4}(V_1 + V_2) - 

277: 2{\frac{|t|}{\pi}}$.  

278: 

279: Figure 2 shows the case of $x = \frac{1}{3}$. This classical 

280: ground state minimizes electrostatic repulsion and $Co^{3+}$ sites fill

281: one of the three sublattices and the remaining hexagonal lattice is

282: the neutral spin-\half site. It is a hexagonal spin-\half quantum 

283: antiferromagnet. The energy of this state per site is $E_{\frac{1}{3}} = 

284: \frac{V_2}{2}$. We also find striped states which are local minima.

285: \begin{figure}[h]

286: \epsfxsize 5.5cm

287: \centerline {\epsfbox{watfig2.eps}}

288: \caption{`Classical' charge order for $x = \frac{1}{3}$.

289: Triangular lattice of localized charge $-e$ carriers and a 

290: Hexagonal lattice of neutral spin-\half moments. Quantum fluctuations

291: will reduce the amplitude of charge order substantially. Accompanying

292: charge order we expect short range AFM order at low temperatures}

293: \end{figure}

294: 

295: So far we studied charge order in the \cob plane at commensurate

296: fillings. As we move away from $ x = \frac{1}{3}$ and $ x = \frac{1}{4}$,

297: defects and discommensurations will be produced or we may go to an 

298: incommensurate charge ordered structure. There may be one or more 

299: first order phase boundary between $x = \frac{1}{3}$ and 

300: $ x = \frac{1}{4}$. However, all these nice charge ordered phases

301: will be challenged by a generically disordered arrangement of \na 

302: ions in a triangular lattice and the consequent random potential seen

303: by the mobile $Co^{3+}$ carriers as explained below . 

304: 

305: The energetically preferred sites of \na atoms\cite{cobstr}

306: in \naxcob form a triangular lattice that have 

307: the same lattice parameter as the triangular $Co$ layer. However,

308: the \na and $Co$ lattices are relatively shifted - if we project

309: the allowed positions of the \na atoms of the nearest top and 

310: bottom layer onto the $Co$ layer, these sites become the dual lattice

311: (hexagonal lattice) of the $Co$ triangular lattice. Because of this

312: {\em a sublattice order of the \na atoms does not couple to the

313: charge density wave order-parameter of the \cob lattice and in 

314: principle allows a finite temperature charge order-disorder 

315: phase transition}:

316: a 3-state Potts ($Z_3$ symmetry) model transition at $x = \frac{1}{3}$ 

317: and a 4-state Potts model ($Z_4$ symmetry) transition at 

318: $ x = \frac{1}{4}$. However, an inevitable  disorder in \na 

319: sublattice leads to, based on an Imry-Ma  type of argument, a glassy 

320: order at low temperatures rather than a genuine charge order phase 

321: transition. Thus we expect a phase diagram depicted in figure 3 for 

322: the range $\frac{1}{3} < x < \frac{1}{4}$. 

323: 

324: Let us move on to the hydrated case, \nxcob. As we mentioned 

325: earlier, the enhanced dielectric constant of the \wat layer

326: will screen short range coulomb repulsion and weaken and melt 

327: the high temperature charge ordering. It also screens and weakens

328: the random \na potential seen by the carriers.

329: For the appearance of low temperature superconductivity 

330: $T_{\rm ch}$ need not be reduced to nearly zero value. A 

331: sufficiently weakened charge ordered state may give up at low 

332: temperatures and superconductivity may emerge. Figure 3

333: shows sketches the change of phase diagram as we go to the

334: hydrated case. A strong resistance anomaly seen in a recent 

335: experiment\cite{jin}  at $T^* \approx 50~K$, may be 

336: the weakened charge order transition that we are discussing.

337: \begin{figure}[h]

338: \epsfxsize 8.0cm

339: \centerline {\epsfbox{watfig3.eps}}

340: \caption{Schematic phase diagram showing how hydration affects 

341: electronic phases: superconductivity is stabilized at low 

342: temperatures by a strong suppression of charge glass order.}

343: \end{figure}

344: Let us discuss nature of \wat ordering hydrogen bonding 

345: in \nxcob in some detail. In what follows we use a recent 

346: suggestion of Cava et al.\cite{cava3} that \wat may have a structure 

347: similar to the layers in hexagonal ice (1h ice). As mentioned earlier,

348: in \naxcob, energetically favorable interlayer sites of $Na$ form 

349: a triangular lattice. These sites are at the center of trigonal 

350: biprisms, capped by an oxygen atom at the top and one at the bottom. 

351: To understand \wat ordering, we consider

352: $Na_{\frac{1}{3}}.CoO_2.{\frac{4}{3}}H_2O$ as a reference compound.

353: Fill one of of the three sublattices by \na atoms to minimize 

354: electrostatic energy.  We are left with two empty sublattices that

355: form a hexagonal lattice. Two \wat molecules may be accommodated at the 

356: top and bottom of the capped trigonal biprism. By doing so we get two

357: hexagonal lattices of \wat sandwiching a triangular lattice of \na

358: ions. Thus we have a triangular lattice filled by \na and \wat in the

359: ratio $1:4$. We can view the above hexagonal sheets as the sheets in

360: hexagonal ice structure, as suggested by Cava and collaborators\cite{cava3}

361: in their

362: preliminary studies. {\em We find it very interesting that the nearest

363: neighbor $H_2O$-$H_2O$ distance in the above geometry, $ \approx 2.81~Au$,

364: is nearly the same\cite{icebook}  as that in real hexagonal ice 

365: $\approx 2.71$}.

366: No wonder water may freeze into ice in \nxcob ! Having nearly the same

367: \wat-\wat distance may also help ice sheets to have a good hydrogen 

368: bonding network like in hexagonal ice.

369: 

370: Thus it is likely that \wat molecules continue to have hydrogen bonding

371: in spite of the \cob and \na environment. As hydrogen bond energy is

372: substantial $\sim 0.5~ eV$, water tends to have hydrogen bonding 

373: activity in extreme environments. Examples are biological systems, 

374: clathrate hydrates and water containing charged ions, where \wat 

375: continue to maintain hydrogen bonding even though the 

376: local structure deviates considerably from the standard ice or water 

377: structure. Further the random \na environment in \nxcob may convert

378: the 2D ice layer into a 2D amorphous ice layer with a good short range 

379: hexagonal order.

380: 

381: Now we discuss how the dielectric property of the \wat layer may control

382: the low temperature electronic phases of the conducting \cob layer. We

383: are interested in finding how the electron-electron interactions 

384: at the charge order wave vector ${\bf q } = {\bf Q}$ 

385: get screened by the interacting water dipoles. The relevant static 

386: dielectric constant is $\epsilon_{\rm H_2O} (Q)$. In an ice system like 

387: ours with a large disorder in dipole orientations, we expect very small 

388: variation of $\epsilon_{\rm H_2O}(q)$ with $q$. Further random site 

389: disorder in the 

390: \na sublattice will produce Bjerrum defects; so we do not 

391: expect a 2D dipolar order-disorder transition, as in ideal 2D models 

392: of ice. 

393: 

394: Dielectric constant of ice has been studied extensively in the past 

395: and also recently\cite{frohlich,haymet,icebook}. A general expression 

396: for the dielectric 

397: constant of an interacting dipolar system is\cite{frohlich,haymet}:

398: \be

399: \epsilon = \epsilon_{\infty} + 

400: \frac{4\pi}{3Vk_BT} \langle ({\bf P} - \langle {\bf P}\rangle )^2 \rangle

401: \ee

402: Here $\epsilon_{\infty} \sim 1 - 2$ is the high frequency dielectric 

403: constant of the dipole, in our case \wat molecule. $\bf P$

404: is the total dipole moment of the system of volume V. And 

405: $\langle~...~\rangle$ denotes thermal average.  Static dielectric

406: constant of ice has not been measured at liquid He temperatures, as the

407: dielectric relaxation becomes too slow even around liquid air 

408: temperatures. Fortunately, extensive numerical study of \ewat

409: are available. For example, a recent calculation\cite{haymet} shows 

410: that for hexagonal ice, \ewat $\approx 220 $ at $T = 50 K$. 

411: 

412: We use this 3D result to get an approximate estimate for our weakly 

413: coupled hexagonal ice layers as follows.  We replace the volume 

414: $V$ by $\approx 6V$ to account for the c-axis expansion

415: in \nxcob. Missing hydrogen bonds along the c-axis reduces the number

416: of allowed proton configurations leading to a reduction of 

417: $\langle ({\bf P} - \langle {\bf P}\rangle )^2 \rangle$ to  

418: $\approx \frac{2}{3} \langle ({\bf P} - \langle {\bf P}\rangle )^2 \rangle$.

419: This gives us \ewat (hexagonal sheet) $\approx 20$ at $T = 50 K$. Our

420: system being strongly disordered, we do not expect \ewat (hexagonal sheet)

421: to change at lower temperatures. Thus the background dielectric constant

422: of \nxcob is $\epsilon = \epsilon_o + \epsilon_{\rm H_2O} \approx

423: 8 + 20$. This reduces the mean short range repulsion by nearly a factor

424: of 3, making the $T_{\rm ch} \approx 1 K$, in equation (2). Once the 

425: long range charge order

426: is disabled by a reduction of V, the simple t-J model and consequent

427: low temperature superconducting phase is realized, albeit with a reduced

428: \tc in the range $\frac{1}{3} < x < \frac{1}{4}$. The sharp reduction

429: in superconducting \tc at the commensurate boundaries of the dome should 

430: arise from the strong short range order and lesser discommensurations and

431: defects. The experimentally seen flat value of \tc $\approx 2 K$ 

432: for $x < \frac{1}{4}$ and for $x > \frac{1}{3}$is likely to be an effect 

433: of phase separation.

434: \begin{figure}[h]

435: \epsfxsize 6.0cm

436: \centerline {\epsfbox{watfig4.eps}}

437: \caption{Schematic experiment to observe the `Electrical  

438: Modulation of Superconducitivy'. The order voltage required

439: is $500~V$, when the film thickness is about a micron.}

440: \end{figure}

441: Our proposal of a critical and catalytic role of \wat layer suggests

442: ways to access and control low temperature electronic phases of the 

443: \cob layer. Based on this we suggest two effects: i) `Electrical 

444: Modulation of Superconductivity'. Here we control (figure 4) 

445: superconductivity 

446: and superconducting \tc by modifying the screening property of \wat

447: layer by external electric fields - DC, AC or pulsed fields. 

448: This has interesting consequences of 

449: being able to locally erase superconductivity by STM tips, dynamically

450: create Josephson networks or create 2D superconductivity of desired

451: shapes through appropriate capacitor shapes etc. 

452: Since microwaves are absorved by hydrogen bonded networks we can pump 

453: microwaves at appropriate frequencies ($\hbar \omega <

454: \Delta_{\rm sc}$, the superconducting gap) and dynamically polarize 

455: water dipoles and may influence its dielectric properties, and in turn 

456: control superconductivity.

457: 

458: ii) `Electroresistance Effect'. By modifying the amplitude of 

459: charge glass order as well as $T_{\rm ch}$ in the 

460: non-superconducting state, by influencing the \wat layer by

461: external DC or AC electric field,

462: we can change $\rho_{ab}$, the ab-plane resistivity. 

463: 

464: Below we estimate the electric field required to completely suppress 

465: superconductivity. Having established a hydrogen bonded

466: network it requires a finite energy to rotate a water molecule and

467: orient its dipole moment along an external electric field. In infrared 

468: absorption

469: and neutron scattering experiments\cite{icebook}

470: the absorption band corresponding to 

471: rotation of \wat molecules is in the range $60~{\rm to}~120~meV$. 

472: Assuming a random orientation, the average energy required to reorient 

473: a water molecule is $\approx 50~meV$, i.e., a potential of

474: $50~{\rm mV}$ applied over the length $\approx 1~Au$ of the

475: water dipole will orient the dipole moment of water along its field.

476: If we have c-axis oriented \nxcob film of thickness $1$ micron we need

477: to apply a voltage $\approx 500~Volts$ across the film, in order to 

478: orient the majority of dipoles. Strong polarization of water dipoles 

479: reduces the dielectric constant, as is evident from equation (3).

480: The resulting reduced screening of carriers in the \cob layer

481: allows charge order to grow and superconductivity gets suppressed.

482: 

483: A theoretical analysis, including some of the possible difficulties 

484: in observing the effects will be presented in a future publication. 

485: 

486: To get a clear understanding of this complex system, and to see

487: if our proposal is correct more experiments are necessary:

488: 

489: a) {\bf Charge Order:} It will be interesting to perform NMR,

490: NQR, STM, $\mu SR$ and other local probe measurements to look for 

491: charge order in the vicinity of the commensurate fillings 

492: $x = \frac{1}{4}, \frac{1}{3}, \frac{1}{2}, \frac{2}{3}$ and 

493: $\frac{3}{4}$ and see how they differ between the two systems

494: \nxcob and \naxcob. 

495: 

496: b) {\bf Spin order, singlets and gaps:} Accompanying local charge order 

497: we expect a spin order at 

498: low temperatures (the scale of J is small, $\sim 6$ to $7~meV$). 

499: In general the enhanced singlet stabilization by the superexchange

500: process will introduce some kind of spin gap phenomenon. If the charge 

501: order at $x = \frac{1}{4}$ leads to a Kagome lattice of spins it will be 

502: an interesting testing ground for some of the ideas of the spin

503: liquid phase of spin-\half Kagome antiferromagnet, including possible

504: novel excitations.

505: 

506: c) {\bf Lower Doping:} Experimentally, it has not been 

507: possible\cite{cava2}

508: to make \nxcob for $x < \frac{1}{4}$. It is likely\cite{cava2} that 

509: c-axis ionic bonding is weakened, by decreasing $x$ and 

510: presence of water layer, making a 3D structure unstable.

511: It will be important to synthesize, by 

512: non-equilibrium means, meta stable compounds for $x < \frac{1}{4}$

513: to test the validity of RVB theory and also test our hypothesis

514: of the role played by water.

515: 

516: d) {\bf Replacing \wat:} It will be desirable to have a stable solid

517: $Na_x CoO_2.y X$, where an intercalant `X' not only increases the 

518: dielectric constant but also provides additional bonding between 

519: \cob layers and make stable compounds for $x < \frac{1}{4}$.

520: 

521: e) {\bf Higher doping:} According to reference 6, the dopant 

522: induced dynamics, within the t-J model will favor ferromagnetic 

523: correlations and a consequent p-wave superconductivity at higher 

524: dopings slightly above $x = \frac{1}{3}$. It will be interesting 

525: to look for this.

526: 

527: f) {\bf Inhomogeneous Superconductivity:} If ice plays a central

528: role, as suggested in this paper, \wat density fluctuation in  

529: ice layer will directly influence superconductivity in nearby 

530: \cob layers resulting in a corresponding fluctuation in the 

531: superconducting order parameter and possible well grown local charge 

532: ordered phase. 

533: 

534: g) {\bf Slow Relaxation:} Since interacting water dipoles have very 

535: slow dielectric relaxation time scales\cite{icebook}, they may 

536: consequently affect superconductivity and impose some anomalous 

537: relaxation/aging effects. 

538: 

539: The present paper is phenomenological and qualitative in character.

540: Any detailed quantitative calculations of \tc and phase diagram

541: for this complex system needs further experimental guidance.

542: Issue of calculating local screening and dielectric constant in 

543: hydrogen bonded systems is known to have subtleties; added to 

544: this, we have conducting layers sandwiching water layers. We have 

545: made very crude estimates based on simple physical arguments and

546: very approximate considerations, as our primary aim is to focus 

547: and identify how water could play a crucial role in this complex 

548: system.

549: 

550: We thank A.K. Mishra, V.N. Muthukumar, Debanand Sa, Manas Sardar 

551: and R. Shankar for discussion and Latha Malar Baskaran for 

552: reference [18].

553: 

554: \begin{references}

555: \bibitem{takada} K. Takada et al., Nature, {\bf 422} 53 (03)

556: \bibitem{cava1} M. Foo et al., cond-mat/0304464

557: \bibitem{cava2} Y. Wang et al., cond-mat/0305455

558: \bibitem{chu} B. Lorenz et al., cond-mat/0304537;

559: F. Rivadulla et al., cond-mat/0304455

560: \bibitem{jin} R. Jin et al., cond-mat/0306066

561: \bibitem{gbcob} G. Baskaran, cond-mat/0303649

562: \bibitem{rvbcob}Brijesh Kumar, B.S. Shastry, cond-mat/0304210;

563: Qiang-Hua Wang, Dung-Hai Lee and Patrick A. Lee, cond-mat/0304377;

564: Masao Ogata, cond-mat/0304405

565: \bibitem{pwagbrvb} P.W. Anderson, Science, {\bf 235} 1196 (87)

566: G. Baskaran, Z. Zou and P.W. Anderson, 

567: Sol. St. Commn, {\bf 63} 973 (87); G. Baskaran and P.W. Anderson, 

568: Phys. Rev. {\bf B~37} 580 (88)

569: \bibitem{stripe} J. Zaanen and O. Gunnarsson, Phys. Rev.

570: {\bf B~46} 7391 (89); V.J. Emery, S. A. Kivelson and

571: O. Zachar, Phys. Rev. {\bf B~56} 6120 (97); S.R. White and

572: D.J. Scalapino, Phys. Rev. {\bf B~60} R753 (99);

573: \bibitem{stripe1} J.M. Tranquada et al., Nature {\bf 375} 561 (95)

574: \bibitem{gbstripe} G. Baskaran, Mod. Phys. Lett., {\bf B~14} 377 (00)

575: \bibitem{susc} R. Ray et al., Phys. Rev. {\bf B~59} 9454 (99);

576: Phys. stat. sol. {\bf B~215} 703 (99)

577: \bibitem{singh} D.J. Singh, Phys. Rev. {\bf B~61} 13397 (00)

578: \bibitem{cobstr} R.J. Balsys and R.L. Davis, Sol. St. Ionics,

579: {\bf 93} 279 (96)

580: \bibitem{cava3} R. Cava, M2S-Rio Meeting, 25-30 May 2003.

581: \bibitem{icebook} V. F. Petrenko and R. W. Whitworth, 

582: Physics of Ice, (Oxford University Press, 1999)

583: \bibitem{frohlich} H. Frohlich, Theory of Dielectrics and

584: Dielectric Loss (Oxford University Press, Oxford, 1958)

585: \bibitem{haymet} S.W. Rick and A.D.J. Haymet, 

586: J. Chem. Phys. {\bf 118} 9291 (2003)

587: \end{references}

588: \end{multicols}

589: \end{document}

590: 

591: