1: \documentclass[aps,prl,twocolumn,superscriptaddress,showpacs]{revtex4}

2: %\documentclass[aps,prl,preprint,superscriptaddress,showpacs]{revtex4}

3: \usepackage{graphicx,bm}

4: 

5: \begin{document}

6: \title{Momentum Dependence of Charge Excitations in 

7: the Electron-Doped Superconductor Nd$_{1.85}$Ce$_{0.15}$CuO$_4$: a RIXS Study}

8: \author{K. Ishii}

9: \email{kenji@spring8.or.jp}

10: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

11: Energy Research Institute, Hyogo 679-5148, Japan}

12: \author{K. Tsutsui}

13: \affiliation{Institute for Materials Research, Tohoku University,

14: Sendai 980-8577, Japan}

15: \author{Y. Endoh}

16: \affiliation{International Institute for Advanced Studies,

17: Kizugawadai, Kizu, Kyoto 619-0025, Japan}

18: \author{T. Tohyama}

19: \affiliation{Institute for Materials Research, Tohoku University,

20: Sendai 980-8577, Japan}

21: \author{S. Maekawa}

22: \affiliation{Institute for Materials Research, Tohoku University,

23: Sendai 980-8577, Japan}

24: \author{M. Hoesch}

25: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

26: Energy Research Institute, Hyogo 679-5148, Japan}

27: \author{K. Kuzushita}

28: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

29: Energy Research Institute, Hyogo 679-5148, Japan}

30: \author{M. Tsubota}

31: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

32: Energy Research Institute, Hyogo 679-5148, Japan}

33: \author{T. Inami}

34: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

35: Energy Research Institute, Hyogo 679-5148, Japan}

36: \author{J. Mizuki}

37: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

38: Energy Research Institute, Hyogo 679-5148, Japan}

39: \author{Y. Murakami}

40: \affiliation{Synchrotron Radiation Research Center, Japan Atomic

41: Energy Research Institute, Hyogo 679-5148, Japan}

42: \affiliation{Department of Physics, Tohoku University, Sendai

43: 980-8578, Japan}

44: \author{K. Yamada}

45: \affiliation{Institute for Materials Research, Tohoku University,

46: Sendai 980-8577, Japan}

47: \date{\today}

48: 

49: \begin{abstract}

50: We report a resonant inelastic x-ray scattering (RIXS) study of charge

51: excitations in the electron-doped high-$T_{\mathrm c}$ superconductor

52: Nd$_{1.85}$Ce$_{0.15}$CuO$_4$.  The intraband and interband excitations

53: across the Fermi energy are separated for the first time by tuning the

54: experimental conditions properly to measure charge excitations at low

55: energy.  A dispersion relation with ${\mathbf q}$-dependent width

56: emerges clearly in the intraband excitation, while the intensity of the

57: interband excitation is concentrated around 2~eV near the zone center.

58: The experimental results are consistent with theoretical calculation of

59: the RIXS spectra based on the Hubbard model.

60: \end{abstract}

61: 

62: \pacs{78.70.Ck, 74.25.Jb, 74.72.Jt}

63: 

64: \maketitle

65: 

66: The asymmetric features of the electronic phase diagram of the doping

67: dependence of the Mott insulating Cu oxides between hole- and

68: electron-doping have been an issue of debate for a long time. Their

69: exploration is very important for the understanding not only of the

70: mechanism of high-$T_{\mathrm c}$ superconductivity but also of the

71: effects of doping on a Mott insulator.  Experimental studies

72: investigating the reconstruction of the electronic bands by the carrier

73: doping have extensively been pursued~\cite{Uchida1,Onose2}, and now the

74: comprehension was reached that the Mott gap feature remains up to

75: considerable doping levels.  Recently angle-resolved photoemission

76: spectroscopy (ARPES) has provided plenty of information on the momentum

77: dependence of the occupied states~\cite{Damascelli1}. On the other hand,

78: the electronic band structure above the Fermi energy is still unclear

79: mainly due to the lack of convincing experimental data.  Electron energy

80: loss spectroscopy (EELS) suffers from multiple scattering at large

81: momentum transfers, while conventional optical methods, such as

82: photo-absorption, electronic Raman scattering and so forth, can observe

83: only the excitation at zero momentum transfer.  In this respect,

84: resonant inelastic x-ray scattering stands out as a unique and ideal

85: probe to be able to measure the momentum dependence of the electronic

86: excitations, in which the band structure of the unoccupied state is

87: elucidated through the two-particle excitation spectra. Current

88: experimental developments in the measurement of such charge dynamics or

89: electronic excitation has been directed to the doped Mott insulators, in

90: particular to a number of cuprates~\cite{Kim4,Hasan3}.

91: 

92: The Cu $K$-edge RIXS experiments of the parent compounds of the

93: high-$T_{\mathrm c}$ cuprates, such as Ca$_2$CuO$_2$Cl$_2$~\cite{Hasan1}

94: and La$_2$CuO$_4$~\cite{Kim1} showed a clear energy gap between the

95: occupied lower Hubbard band (LHB), more precisely the Zhang-Rice singlet

96: band, to the unoccupied upper Hubbard band (UHB).  Recent RIXS

97: measurements for the hole doped cuprates showed that the Mott gap

98: feature is robust even at considerable doping levels, besides

99: appreciable low energy excitations corresponding to the hole

100: doping~\cite{Kim4,Hasan3}.  In this letter, we extended the effort to

101: investigate the electron doped cuprates using RIXS and observe how the

102: Mott gap feature changes with the electron doping. Combined with an

103: extensive theoretical analysis, we can observe not only the interband

104: excitation across the Mott gap but also the intraband excitation within

105: the UHB throughout the whole Brillouin zone.

106: 

107: The RIXS experiments were carried out on the IXS spectrometer installed

108: at the beam line 11XU of SPring-8~\cite{Inami1}.  A Si (111)

109: double-crystal monochromator and a Si (400) channel-cut secondary

110: monochromator were utilized.  Horizontally scattered x-rays were

111: analyzed in energy by a bent Ge (733) crystal.  The overall energy

112: resolution is about 400 meV estimated from the full width at half

113: maximum (FWHM) of the quasielastic scattering.  Single crystals of

114: Nd$_{2-x}$Ce$_x$CuO$_4$ of $x$ = 0.15 and 0.075 were prepared, which

115: show superconductivity below $T_{\mathrm c}$ = 25 K, and

116: antiferromagnetic order below $T_{\mathrm N} \sim$ 120 K, respectively.

117: The surface of the crystal is normal to the {\it c}-axis, which was kept

118: in the scattering plane so as to be scanned in the reciprocal lattice

119: space spanned by either the [100]-[001] or the [110]-[001] axes.  All

120: spectra were collected at room temperature.

121: 

122: \begin{figure}

123: %\includegraphics[scale=0.4]{fig/eidep.eps}

124: %\includegraphics[scale=0.4]{fig/fluo.eps}

125: \includegraphics[width=7.cm]{1a.eps}

126: \includegraphics[width=6.cm]{1b.eps}

127: \caption{(a) Resonant inelastic x-ray scattering spectra of

128: Nd$_{1.85}$Ce$_{0.15}$CuO$_4$ as a function of energy loss at some

129: representative incident x-ray energies $E_\mathrm{i}$.  The scattering

130: vector is fixed at $\mathbf{Q}=(0.5,0,12)$.  The strong intensity in the

131: spectra of $E_\mathrm{i}=8986$ eV at high excitation energy comes from

132: the Cu $K_{\beta 5}$ fluorescence. (b) X-ray absorption spectrum

133: measured by the fluorescence method.  The polarization of the x-rays

134: ($\epsilon$) is parallel to the ${\mathbf a}$- or ${\mathbf c}$-axes of

135: the crystal.  The thick arrow indicates the energy used for

136: momentum-dependent RIXS measurements.}

137: \label{fig:eidep}

138: \end{figure}

139: 

140: Figures~\ref{fig:eidep} show the incident energy ($E_\mathrm{i}$)

141: dependence of RIXS, together with the fluorescence spectra.  The

142: absolute momentum transfer is fixed at $\mathbf{Q}=(0.5,0,12)$.

143: Resonantly enhanced peaks at around 2~eV and 6~eV can be seen.  The

144: latter peak was also observed in the undoped Nd$_2$CuO$_4$, and it was

145: identified as a charge transfer excitation to the antibonding

146: state~\cite{Hill1}.  We find two resonances at around 8990~eV and

147: 9000~eV for the 6~eV peak.  These energies correspond to the resonance

148: energies for $\epsilon_\mathrm{i} \parallel c$ and $\epsilon_\mathrm{i}

149: \parallel ab$~\cite{Hamalainen1}, $\epsilon_\mathrm{i}$ being the

150: polarization of the incident x-ray.  Because the $\epsilon_\mathrm{i}$

151: is close to $\frac{{\hat \mathbf{a}}+{\hat \mathbf{c}}}{2}$ in our

152: experimental configuration of RIXS, where ${\hat \mathbf{a}}$ and ${\hat

153: \mathbf{c}}$ are the unit vectors along the $a$- and $c$-axes,

154: respectively, it is reasonable that both resonances are observed.  The

155: first two peaks in the fluorescence spectra are assigned to the

156: $1s-4p_{\pi}$ transition, while the next two are the $1s-4p_{\sigma}$

157: transition. The intensity of the 2~eV peak shows an enhancement at

158: around 8991~eV which corresponds to the absorption edge for

159: $4p_{\sigma}$.  Hereafter $E_\mathrm{i}$ is fixed at 8991~eV to focus on

160: the low energy excitations.

161: 

162: \begin{figure*}

163: \includegraphics[scale=0.3, trim=0 -5 0 0, clip]{2a.eps}

164: \includegraphics[scale=0.3, trim=50 -5 0 0, clip]{2b.eps}

165: \includegraphics[scale=0.3, trim=50 -5 0 0, clip]{2c.eps}

166: \begin{minipage}[b]{0.25\linewidth}

167: \begin{center}

168: \includegraphics[width=5cm]{2d.eps}

169: \includegraphics[scale=0.3]{2e.eps}

170: \end{center}

171: \end{minipage}

172: \caption{Momentum dependence in Nd$_{1.85}$Ce$_{0.15}$CuO$_4$. (a) Along

173: the $c$-axis, (b) $(\pi,0)$, and (c) $(\pi,\pi)$ directions.

174: $E_\mathrm{i}=8991$~eV.  The filled symbols are raw data, and the open

175: ones in (b) and (c) are data from which the elastic scattering and the

176: scattering at higher energy (dotted lines) are subtracted. (d) Contour

177: plot of the RIXS intensity. After subtraction of the elastic and

178: high-energy contributions (open symbols in (b) and (c)) the data are

179: normalized for the maximum intensity in each momentum and interpolated

180: smoothly. (e) Comparison of the RIXS spectra to

181: Nd$_{1.925}$Ce$_{0.075}$CuO$_4$.}  \label{fig:qdep}

182: \end{figure*}

183: 

184: As shown in Fig.~\ref{fig:qdep}(a), the momentum dependence of RIXS

185: along the $c$-axis is weak, as expected from the strong

186: two-dimensionality of the CuO$_2$ plane.  However it should be noted

187: that the quasielastic tail is appreciably suppressed for the scan at

188: $l=12.5$.  In our experimental condition ($\pi$-polarization of the

189: incident x-ray), the intensity of the elastic scattering, whose major

190: component is Thomson scattering, is mostly proportional to $\cos^2

191: 2\theta$, where $2\theta$ is the scattering angle and thus decreases

192: when $2\theta$ is close to 90~degree. It is crucially important to

193: reduce the elastic scattering to measure the low energy excitations in

194: RIXS.  For this reason, we selected $l=12.5$ to measure the momentum

195: dependence in the CuO$_2$ plane, though it is not a high symmetry plane.

196: 

197: Figures~\ref{fig:qdep} (b) and (c) show the momentum dependence of the

198: RIXS spectra along a line in the $\mathbf{q}=(\pi,0)$ and $(\pi,\pi)$

199: directions, respectively, where $\mathbf{q}$ represents the reduced

200: momentum transfer in the $ab$-plane.  Except for the spectrum at

201: $\mathbf{Q}=(0.75,0,12.5)$, all the spectra were measured at $h<0.5$ of

202: $\mathbf{Q}=(h,0,12.5)$ or $\mathbf{Q}=(h,h,12.5)$.  The spectrum at

203: $\mathbf{Q}=(0.75,0,12.5)$ lies between those of

204: $\mathbf{Q}=(0.2,0,12.5)$ and $\mathbf{Q}=(0.3,0,12.5)$, which indicates

205: that the electronic structure is symmetric with respect to $h=0.5$.  In

206: order to examine the momentum dependence more clearly, we subtract the

207: elastic contribution near 0~eV and the high-energy contribution above

208: 4~eV from the raw data (dashed lines in Figs.~\ref{fig:qdep}(b) and

209: (c)).  The former is estimated from the anti-Stokes region, and the

210: latter is treated as a tail of the excitation at 6~eV by extrapolating

211: smoothly to the lower-energy region.  The open symbols in

212: Figs.~\ref{fig:qdep}(b) and (c) are the resulting spectra, where the

213: data below 0.6~eV are not shown due to the uncertainty in the assignment

214: of the quasielastic contributions. The spectra are replotted in

215: Fig.~\ref{fig:qdep}(d) as a contour map, where the maximum intensity at

216: each momentum point is normalized to unity and a smoothing procedure is

217: applied.  We can clearly see two characteristic excitations. One is the

218: excitation at 2~eV observed at the zone center. Its intensity rapidly

219: decreases with increasing $\mathbf{q}$. The other one is a broad but

220: dispersive excitation along the $(\pi,0)$ and $(\pi,\pi)$ directions.

221: As a function of $\vert \mathbf{q} \vert$, the latter excitation shifts

222: to higher energy up to 2-2.5~eV at the zone boundary, accompanied by an

223: increase of the spectral width. The upper edges of the excitations are

224: dispersive with a width of more than 2~eV from the zone center to the

225: zone boundary.

226: 

227: Two characteristic excitations just described above have been elucidated

228: further by duplicating the similar scans for $x=0.075$ crystal.  Typical

229: data of scans are shown in Fig.~\ref{fig:qdep}(e). The excitation

230: spectra at the zone center superpose each other and it is essentially

231: independent of $x$.  On the other hand, the spectra at finite

232: $\mathbf{q}$s show the weaker intensities for $x=0.075$ in lower-energy

233: region and the intensity seems to be proportional to $x$.  Such

234: dependence on doping indicates different nature between two excitations

235: which is identified by the following theoretical analysis of RIXS from

236: the electron-doped CuO$_2$ plane.

237: 

238: Keeping in mind the nearly monotonic dispersion like excitation mode, we

239: performed calculations of the RIXS spectrum using the numerically exact

240: diagonalization technique on a $4\times 4$ cluster of a Hubbard model

241: with the electron density $18/16=1.125$.  The model includes the hopping

242: of the electrons between first, second, and third nearest neighbor sites

243: ($t$, $t'$, and $t''$, respectively) and the on-site Coulomb interaction

244: $U$.  The RIXS spectrum is expressed as a second-order process of the

245: dipole-transition between Cu $1s$ and $4p$ orbitals, where a Coulomb

246: interaction between a $1s$ core-hole and a $3d$ electron, $U_c$, is

247: explicitly included~\cite{Tsutsui1}.  We use $t'/t=-0.25$ and

248: $t''/t=0.12$, which are obtained from shape of the Fermi

249: surface~\cite{Armitage1}.  For other parameters, we take $U/t=8$,

250: $U_c/t=10$, and $t=0.3$~eV.  The inverse of the life time of the

251: intermediate state is assumed to be $\Gamma=3t$.

252: 

253: Figure~\ref{fig:theory} shows the calculated RIXS spectrum, where

254: $E_\mathrm{i}$ is set to a value denoted by the arrow in the absorption

255: spectrum shown in the inset.  The RIXS spectrum shows two characteristic

256: excitations similar to the observed ones: One is a 2~eV excitation at

257: $\mathbf{q}=(0,0)$, and the other is a broad band of excitations up to

258: $\sim3$~eV for all momenta except $(0,0)$.  The 2eV excitation is the

259: Mott gap excitation from LHB to UHB, as discussed in

260: Ref.~\cite{Tsutsui2}.  The broad excitations can be assigned to the

261: intraband excitation.  To confirm this, we calculated the dynamical

262: density response function $N(\mathbf{q},\omega)$, which can describe the

263: momentum-dependent intraband and interband density fluctuations

264: separately when $U$ is

265: large~\cite{Tohyama1,Eder1,Khaliullin1,DHKim1,Grober1}.  Comparing RIXS

266: and $N(\mathbf{q},\omega)$ (broken lines), we find qualitatively similar

267: behavior for all momentum and energy regions except for the 2~eV

268: excitation at $\mathbf{q}=(0,0)$.  This means that the broad and

269: dispersive excitations observed in Fig.~\ref{fig:qdep} come from the

270: charge fluctuations in the metallic phase.

271: 

272: \begin{figure}

273: \includegraphics[width=8.5cm]{3.eps}

274: \caption{The RIXS spectrum of an electron-doped $4\times 4$ Hubbard

275: cluster with long-range hopping terms.  The electron density is

276: $18/16=1.125$, and the model parameters are $t'/t=-0.25$, $t''/t=0.12$,

277: $U/t=8$, $U_c/t=10$, and $\Gamma/t=3$ with $t=0.3$~eV.  The

278: $\delta$-functions are convoluted with a Lorentzian broadening of

279: $0.8t$.  Broken lines: $N(\mathbf{q},\omega)$ for the same cluster.  The

280: inset of the left panel: the calculated absorption spectrum with a

281: Lorentzian broadening of $3t$. The vertical arrow denotes $E_\mathrm{i}$

282: for RIXS.  The inset of the right panel: momentum dependence of the peak

283: positions of RIXS (the squares), together with those of

284: $N(\mathbf{q},\omega)$ for $4\times 4$ (the circles) and

285: $\sqrt{18}\times\sqrt{18}$ (the crosses) clusters.}

286: \label{fig:theory}

287: \end{figure}

288: 

289: The energy position of the highest peak in RIXS and

290: $N(\mathbf{q},\omega)$ is plotted in the inset of the right panel of

291: Fig.~\ref{fig:theory}.  In addition to the positions taken from the main

292: panels, we plot peak positions obtained from a

293: $\sqrt{18}\times\sqrt{18}$ cluster with an electron density $20/18$.

294: The resulting momentum dependence is found to trace out the center the

295: of broad spectra in the contour plot of Fig.~\ref{fig:qdep}(d).  This

296: agreement justifies the assignment of the observed two structures to the

297: Mott-gap and intraband excitations.

298: 

299: As emphasized above, the intraband excitations show broad features. Such

300: a broadness comes from strong correlation common to Hubbard-type

301: models~\cite{Tohyama1,Eder1,Khaliullin1,DHKim1,Grober1}, and thus it is

302: independent of the presence of the long-range hoppings.  Instead, the

303: effect of $t'$ and $t''$ may appear in the very low-energy region, where

304: excitations are predominately controlled by the Fermi surface topology

305: inducing $2k_\mathrm{F}$ and $4k_\mathrm{F}$ branches~\cite{Tohyama1},

306: $k_\mathrm{F}$ being the Fermi momentum.  Unfortunately, at present the

307: energy resolution of RIXS is limited so that it is difficult to resolve

308: the branches.  We continue our efforts to improve the energy

309: resolution and this will open a new view of the intraband charge

310: excitations in the high-$T_c$ cuprates in the future.

311: 

312: Although we made use of electron-doped Nd$_{1.85}$Ce$_{0.15}$CuO$_4$, it

313: may also be possible to use hole-doped materials such as

314: La$_{2-x}$Sr$_{x}$CuO$_4$ to detect the intraband excitation. However,

315: it is important to notice that there is an advantage of electron doping

316: over hole doping.  As seen in the inset of the left panel of

317: Fig.~\ref{fig:theory}, the absorption spectrum shows three components:

318: $\omega-\epsilon_{\mathrm{1s-4p}}=-17t$, $-12t$, and $-8t$.  The latter

319: two components are also seen in the undoped system.  On the other hand,

320: the former appears upon electron doping only, corresponding to a final

321: state where the core hole attracts a doped electron on the same site.

322: Therefore, the final state hardly contains any pair of empty and doubly

323: occupied sites, i.e., no excitations across the Mott gap.  This means

324: that, the intraband charge excitations dominate the RIXS spectrum if

325: $E_\mathrm{i}$ is tuned to the lowest-energy peak.  As long as we fix

326: $E_\mathrm{i}$ to the absorption-edge region, the contribution from the

327: lowest-energy peak induces large intraband charge excitations.  In our

328: case we select the incident energy to the absorption edge for

329: $4p_{\sigma}$ as shown in Fig.~\ref{fig:eidep}(b), and this condition is

330: satisfied.  In contrast, a corresponding final state in the hole-doped

331: system, which emerges after hole doping, exists in a much higher-energy

332: region at around $\omega-\epsilon_{\mathrm{1s-4p}}=0$~\cite{Tsutsui2}.

333: Experimentally tuning the incident photon energy to this region seems to be

334: difficult because of the overlap of other absorption processes.

335: 

336: Finally, we compare the interband excitations across the Mott gap

337: between hole- and electron-doping. Recently RIXS experiments for

338: La$_{2-x}$Sr$_{x}$CuO$_4$ have been reported~\cite{Kim4,Hasan3}. Their

339: results showed that the spectral shape is almost independent of the

340: momentum transfer except for small shifts in energy. On the other hand,

341: the interband excitation of Nd$_{1.85}$Ce$_{0.15}$CuO$_4$ concentrates

342: on a energy ($\sim 2$~eV) at the zone center and becomes broad in energy

343: with increasing momentum transfer.  Such a difference in momentum

344: dependence is consistent with a previous theoretical

345: result~\cite{Tsutsui2}, where the difference in the strength of

346: antiferromagnetic correlation plays a crucial role.

347: 

348: In summary, we have performed a RIXS study for the electron-doped

349: superconductor Nd$_{1.85}$Ce$_{0.15}$CuO$_4$, and found

350: characteristics of the intraband and interband excitations. The

351: intraband excitation shifts to higher energy with the increase of the

352: peak width as a function of momentum transfer.  The spectral shape of

353: the intraband excitation has a similarity to $N(\mathbf{q},\omega)$ of

354: the two-dimensional Hubbard model. This demonstrates that RIXS is a

355: good tool to measure momentum-dependent density fluctuations in

356: strongly correlated metallic systems.  On the other hand, the

357: interband excitation across the Mott gap is enhanced in intensity at

358: the zone center, which is in contrast to the momentum-independent

359: spectral shape in hole-doped La$_{2-x}$Sr$_x$CuO$_4$.

360: 

361: The authors thank T. Uefuji for supplying a crystal of

362: Nd$_{1.925}$Ce$_{0.075}$CuO$_4$.  K. T., T. T., S. M., and K. Y. were

363: supported by the Japanese Ministry of Education, Culture, Sports, Science

364: and Technology, Grant-in-Aid for Scientific Research.  K. T., T. T., and

365: S. M. were also supported by CREST, NAREGI Nanoscience Project.

366: M. H. acknowledges support from the Japanese Society for the Promotion

367: of Science.  The numerical calculations were partially performed in the

368: supercomputing facilities of ISSP, University of Tokyo and IMR, Tohoku

369: University.

370: 

371: \bibliography{ixs,cuprate}

372: \end{document}

373: