cond-mat0409650/cup.tex
1: \documentclass[epj]{svjour}
2: % Remove option referee for final version
3: %
4: % Remove any % below to load the required packages
5: %\usepackage{latexsym}
6: \usepackage{graphicx}
7: \usepackage{amssymb,amsbsy,amsmath}
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: \newcommand {\subhm} {$\mathcal{H}\left(M\right)$\ }
10: \newcommand {\subhs} {$\mathcal{H}\left(S\right)$\ }
11: \newcommand{\op}[1]{%
12:     \fontdimen12\textfont3=2pt\fontdimen12\scriptfont3=1.4pt%
13:     \!\null\mathop{\vphantom{#1}\smash{#1}}\limits_{\sim}\null\!}
14: \newcommand{\xref}[1]{\protect\ref{#1}}
15: \newcommand{\figref}[1]{Fig.~\protect\ref{#1}}
16: \newcommand{\fmref}[1]{(\protect\ref{#1})}
17: \def\bra#1{\langle \, {#1} \, | \,}
18: \def\ket#1{\, | \, {#1} \, \rangle}
19: \newcommand {\telcomp}{Sr$_{14}$Cu$_{24}$O$_{41}$}
20: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
21: %-----------------------------------------------------------
22: \journalname{Eur. Phys. J. B}
23: %-----------------------------------------------------------
24: %
25: \begin{document}
26: %
27: \title{Strong Coulomb effects in hole-doped Heisenberg chains}
28: \author{J\"urgen Schnack%
29: %\inst{1}
30: }                     % Do not remove
31: %
32: \offprints{J\"urgen Schnack}          % Insert a name or remove this line
33: %
34: \institute{Universit\"at Osnabr\"uck, Fachbereich Physik,
35: D-49069 Osnabr\"uck, Germany}
36: %
37: \date{Received: date / Revised version: date}
38: % The correct dates will be entered by Springer
39: %
40: \abstract{
41:   Substances such as the ``telephone number compound'' \telcomp\
42:   are intrinsically hole-doped. The involved interplay of spin
43:   and charge dynamics is a challenge for theory. In this article
44:   we propose to describe hole-doped Heisenberg spin rings by
45:   means of complete numerical diagonalization of a Heisenberg
46:   Hamiltonian that depends parametrically on hole positions and
47:   includes the screened Coulomb interaction among the holes.  It
48:   is demonstrated that key observables like magnetic
49:   susceptibility, specific heat, and inelastic neutron
50:   scattering cross section depend sensitively on the dielectric
51:   constant of the screened Coulomb potential.
52: %
53: \PACS{
54:       {75.10.Pq}{Spin chain models}   \and
55:       {75.40.Mg}{Numerical simulation studies}
56:      } % end of PACS codes
57: } %end of abstract
58: %
59: \maketitle
60: %
61: 
62: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
63: \section{Introduction and model}
64: \label{sec-1}
65: 
66: Substances hosting spin and charge degrees of freedom exhibit a
67: large variety of phenomena like magnetic and charge ordering,
68: metallic conductivity and superconductivity
69: \cite{Dag:RMP94,KCG:PRB01}. The ``telephone number compound'',
70: \linebreak \telcomp , contains two magnetic one-dimensional
71: structures, chains and ladders.  The stoichiometric formula
72: suggests 6 holes per formula unit. We will assume that for the
73: undoped compound all holes are located in the chain subsystem
74: (i.~e. 60~\% holes), although this is experimentally under
75: discussion since x-ray absorption (XAS) measurements suggest
76: that at room temperature some holes are located in the ladder
77: subsystem \cite{NMK:PRB00}, whereas it is necessary to assume
78: that all holes are in the chain subsystem in order to explain
79: neutron scattering data \cite{RBM:PRB99}.
80: 
81: At low temperatures the ladder subsystem is magnetically
82: inactive due to a large spin gap \cite{TME:PRB98}.  The
83: remaining dynamics of the hole-doped chain system is still
84: interesting as well as complicated enough to constitute a
85: challenge for theoretical investigations. Especially the
86: evaluation of thermodynamic quantities both as function of
87: temperature and magnetic field is prohibitively complicated even
88: for moderate system sizes. Therefore, mostly approximate
89: descriptions in terms of classical spin dynamics
90: \cite{HoS:PRE03,SPV:EPJB02}, spin-dimer models
91: \cite{CBC:PRL96,ABC:PRB00,Klingeler:2003} or spin-wave analysis
92: \cite{MYK:PRB99} have been applied.  Calculations based on the
93: Hubbard model aim at ground-state correlations at low hole
94: doping \cite{AKA:PRB98}.
95: 
96: A fundamental question in this context is how the charge order
97: in the CuO$_2$ chains of substances such as \telcomp\ is
98: established. These chains seem to be a sequence of rather
99: perfect antiferromagnetically coupled spin dimers separated by
100: holes, see \figref{F-1}.  Any proposed theoretical model should
101: also be able to describe excitations involving hole motion which
102: is crucial because interesting physical properties of these
103: compounds result from a competition of charge mobility and
104: magnetic interactions
105: \cite{KFE:Nature98,MYD:Science99,Zaanen:Science99,HEB:PRL04}.
106: One possible explanation is that the formation of dimers is
107: generated by structural modulations of the material via a strong
108: variation of the on-site orbital energies
109: \cite{IsT:JPSJ98,GeL:PRL04,GeL:EPJB05,GeL:05}.
110: 
111: In this article we investigate how a screened electrostatic
112: hole-hole repulsion along the chain would express itself in
113: thermodynamic quantities. It will turn out that a rather strong
114: Coulomb repulsion is needed in order to reproduce the
115: experimental magnetization. This is in accord with e.g.
116: Ref.~\cite{VFB:PRB03} or with
117: Refs.~\cite{MTM:PRB98A,MTM:PRB98B}, where a dielectric constant
118: of 3.3 is found to be realistic.
119: 
120: In order to be able to evaluate thermodynamic quantities we
121: propose to describe hole-doped spin rings with a Heisenberg
122: Hamiltonian that depends parametrically on hole positions.  This
123: ansatz is similar to a simple Born-Oppenheimer description where
124: the electronic Hamiltonian (here spin Hamiltonian) depends
125: parametrically on the positions of the classical nuclei (here
126: hole positions).  Each configuration $\vec{c}$ of holes and
127: spins defines a Hilbert space which is orthogonal to all Hilbert
128: spaces arising from different configurations. The Hamilton
129: operator $\op{H}(\vec{c})$ of a certain configuration $\vec{c}$
130: is of Heisenberg type and depends parametrically on the actual
131: configuration $\vec{c}$, i.~e.
132: %--------------------------------------------------------
133: \begin{eqnarray}
134: \label{E-1-1}
135: \op{H}
136: &=&
137: \sum_{\vec{c}}\;
138: \left(
139: \op{H}(\vec{c})
140: +
141: V(\vec{c})
142: \right)
143: \\
144: \label{E-1-2}
145: \op{H}(\vec{c})
146: &=&
147: -
148: \sum_{u,v}\;
149: J_{uv}(\vec{c})\;
150: \op{\vec{s}}(u) \cdot \op{\vec{s}}(v)
151: \ .
152: \end{eqnarray}
153: %--------------------------------------------------------
154: $J_{uv}(\vec{c})$ are the respective exchange parameters which
155: depend on the configuration of holes. $J<0$ describes
156: antiferromagnetic coupling, $J>0$ ferromagnetic coupling.
157: 
158: %===================    figure   =================================
159: \begin{figure}[ht!]
160: \centering
161: %\resizebox{0.75\textwidth}{!}{\includegraphics[clip,width=65mm]{fig-1.eps}}
162: \includegraphics[clip,width=65mm]{fig-1.eps}
163: \caption{L.h.s.: Ground-state hole configuration for 20 sites and 60\%
164:   holes. This configuration is also called dimer configuration
165:   since it consists of weekly interacting antiferromagnetic
166:   dimers. One dimer is highlighted (d). The hole-spin exchange
167:   processes \textbf{1} and \textbf{2} lead to energetically
168:   low-lying configurations.
169:   R.h.s.: Exchange parameters used in this article: $J=-67$~K,
170:   $J_\parallel=7$~K, and $J_{NN}=25$~K.
171: }
172: \label{F-1}
173: \end{figure}
174: %===================    figure   =================================
175: 
176: Figure~\xref{F-1} shows on the l.h.s. as an example the
177: ground-state hole configuration of \telcomp\
178: \cite{RBM:PRB99,MYK:PRB99,ABC:PRB00} which is a sequence of
179: spin-hole-spin dimers separated by two holes. Energetically
180: excited configurations arise if holes are moved to other sites
181: as depicted exemplarily by the exchange processes \textbf{1} and
182: \textbf{2}.  The r.h.s. of \figref{F-1} illustrates how the
183: exchange parameters depend on the actual hole configuration. In
184: this work three different exchange parameters are employed.
185: 
186: A key ingredient of the proposed model is the inclusion of the
187: electrostatic interaction between holes which is modeled by a
188: screened Coulomb potential
189: %--------------------------------------------------------
190: \begin{eqnarray}
191: \label{E-2}
192: V(\vec{c})
193: &=&
194: \frac{e^2}{4\pi\epsilon_0\,\epsilon_r\,r_0}
195: \frac{1}{2}\;
196: \sum_{u\ne v}\;
197: \frac{1}{|u-v|}
198: \ ,
199: \end{eqnarray}
200: %--------------------------------------------------------
201: where $r_0=2.75$~\AA\ is the distance between nearest neighbor
202: sites on the ring. The dielectric constant $\epsilon_r$ is
203: considered as the only free parameter in the present
204: investigation. Several attempts have been undertaken to estimate
205: the dielectric constant which yielded values for $\epsilon_r$ up
206: to 30 \cite{CPP:PRL89,BaS:PRB94,EKZ:PRB97}. In related projects
207: where the exchange interaction of chain systems in cuprates is
208: derived from hopping matrix elements between different orbitals
209: using a Madelung potential the dielectric constant is found to
210: be $\epsilon_r=3.3$ \cite{MTM:PRB98A,MTM:PRB98B}.
211: 
212: 
213: 
214: 
215: 
216: 
217: 
218: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
219: \section{Discussion of the model}
220: \label{sec-2}
221: 
222: The aim of the proposed model is to evaluate the complete
223: spectrum for reasonably large system sizes and thus to be able
224: to investigate thermodynamic quantities both as function of
225: temperature and field. The spectrum does not only consist of
226: those levels belonging to the ground-state hole distribution,
227: compare \figref{F-1}, but also of all levels arising from all
228: other hole configurations.  For small systems all such
229: configurations can be generated and the related spin
230: Hamiltonians \fmref{E-1-2} can be diagonalized completely. For 8
231: spins and 12 holes for instance this amounts to 6310 distinct
232: hole configurations and tiny Hilbert spaces of dimension 256.
233: For 16 spins and 24 holes the total number of hole
234: configurations is already too big to be considered completely.
235: Therefore, only the ground-state configuration and low-lying
236: excitations with their respective degeneracies are taken into
237: account. This is sufficient since hole configurations which
238: deviate considerably from the ground state configuration possess
239: very high excitation energies.
240: 
241: Correlated electrons are usually modeled with the Hubbard model
242: \cite{AKA:PRB98}, therefore looking at Eqs.~\fmref{E-1-1} and
243: \fmref{E-1-2} one might be tempted to ask: \emph{Where is the
244:   kinetic energy of the holes?} The Hamiltonian of the Hubbard
245: model \cite{Hubbard63,Hubbard64a,Hubbard64b},
246: %--------------------------------------------------------
247: \begin{eqnarray}
248: \label{E-3}
249: \op{H}
250: &=&
251: - \sum_{\langle ij \rangle,\sigma} t_{ij} 
252: \left(
253: \op{c}^\dagger_{i\sigma}\op{c}_{j\sigma} + \text{h.c.} \right) 
254: + U \sum_{i} \op{n}_{i\uparrow}\op{n}_{i\downarrow}
255: \ ,
256: \end{eqnarray}
257: %--------------------------------------------------------
258: transforms at large $U$ and half filling into a Heisenberg
259: Hamiltonian \fmref{E-1-2} with $J_{ij}=-4t_{ij}^2/U$. Therefore,
260: if working close to half filling, it is legitimate to say that
261: the kinetic energy is absorbed into the exchange coupling.
262: Nevertheless, the remaining hole motion is treated classically,
263: i.~e. superpositions of hole states are not taken into account.
264: For the actual compound which is far from superconductivity this
265: assumption of well-localized holes seems to be appropriate.
266: 
267: What is properly taken into account is the screened Coulomb
268: interaction between holes. But, if so: \emph{Wouldn't it be
269:   sufficient to consider nearest-neighbor Coulomb repulsion
270:   only?} Although this is a common method we find that a simple
271: nearest-neighbor repulsion results in unphysical ground states. 
272: It is experimentally verified by means of low-temperature
273: susceptibility \cite{Klingeler:2003}, neutron scattering
274: \cite{RBM:PRB99,MYK:PRB99} as well as thermal expansion
275: measurements \cite{ABC:PRB00}, that the highly symmetric dimer
276: configuration, compare \figref{F-1}, constitutes the ground
277: state of \telcomp. Using only nearest-neighbor Coulomb repulsion
278: yields an alternating sequence of spins and holes with the
279: remaining 10~\% holes assembling as a big cluster irrespective
280: how strong the repulsion is. The reason is that this strange
281: configuration has the same number of nearest hole-hole neighbors
282: as the dimer configuration.  Even the inclusion of a
283: next-nearest neighbor Coulomb repulsion does not improve the
284: situation, the Coulomb interaction is still proportional to the
285: number of sites and may be overcome by the antiferromagnetic
286: binding $J$. 
287: 
288: 
289: 
290: 
291: 
292: 
293: 
294: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
295: \section{Results}
296: \label{sec-3}
297: 
298: At temperatures and energies below 200~K the behavior of the
299: chain subsystem in \telcomp\ is usually discussed in terms of
300: weakly interacting dimers sometimes augmented by weak interchain
301: interactions, see e.~g. 
302: \cite{CBC:PRL96,MYK:PRB99,ABC:PRB00,Klingeler:2003}. Such a
303: picture, although rather successful, does not allow to discuss
304: the influence of mobile holes on thermodynamic observables. It
305: is clear that configurations like \textbf{1} and \textbf{2} in
306: \figref{F-1} will contribute to thermal averages, but how? 
307: 
308: %===================    figure   =================================
309: \begin{figure}[ht!] 
310: \centering
311: %\resizebox{0.75\textwidth}{!}{\includegraphics{fig-2.eps}}
312: \includegraphics[clip,width=60mm]{fig-2.eps}
313: \caption{Low-energy part of the spectrum of a chain of
314:   $N_{\text{tot}}=40$ sites with $N_{\text{s}}=16$ spins and
315:   $N_{\text{h}}=24$ holes for three choices of the dielectric
316:   constant $\epsilon_r$: bars -- $\epsilon_r=1$, bars together
317:   with x-symbols -- $\epsilon_r=5$, bars and crosses --
318:   $\epsilon_r=10$.  The arrow marks the singlet-triplet
319:   transition employed in dimer models.} 
320: \label{F-2}
321: \end{figure}
322: %===================    figure   =================================
323: 
324: Figure \xref{F-2} shows the low-energy part of the spectrum of a
325: chain of $N_{\text{tot}}=40$ sites with $N_{\text{s}}=16$ spins
326: and $N_{\text{h}}=24$ holes for three choices of the dielectric
327: constant $\epsilon_r$.  If $\epsilon_r=1$ the spectrum up to
328: several hundreds of Kelvin is solely given by the levels of the
329: ground-state dimer configuration (bars in \figref{F-2}).  With
330: increasing $\epsilon_r$ the Coulomb repulsion decreases and so
331: does the excitation energy of magnetic levels belonging to hole
332: configurations where one or two holes are moved. As an example
333: the levels resulting from such configurations are given as
334: x-symbols ($\epsilon_r=5$) and crosses ($\epsilon_r=10$) in
335: \figref{F-2}. It is clear that besides the singlet-triplet
336: transition (arrow in \figref{F-2}), which is the main ingredient
337: of the dimer model, transitions to states involving spin-holes
338: exchange processes will contribute to thermodynamic observables
339: like the inelastic neutron scattering cross section. In the
340: following we discuss the influence on three basic observables. 
341: 
342: %===================    figure   =================================
343: \begin{figure}[ht!] 
344: \centering
345: %\resizebox{0.75\textwidth}{!}{\includegraphics{fig-3.eps}}
346: \includegraphics[clip,width=60mm]{fig-3.eps}
347: \caption{Zero-field magnetic susceptibility $\chi_0(T)$: solid
348:   curve -- $\epsilon_r=1$, dashed curve -- $\epsilon_r=5$,
349:   dashed-dotted curve -- $\epsilon_r=10$. For a comparison
350:   experimental data, taken at $B=1$~T, are provided by x-symbols
351:   \cite{Klingeler:2003}. The corrected data, given by crosses,
352:   take also into account that due to impurities the number of
353:   dimers is less than theoretically possible
354:   \cite{Klingeler:2003}.}
355: \label{F-3}
356: \end{figure}
357: %===================    figure   =================================
358: 
359: Figure~\xref{F-3} presents the results for the magnetic
360: susceptibility $\chi(T,B=0)=\chi_0(T)$ at vanishing magnetic
361: field $B=0$. The solid curve shows the theoretical
362: susceptibility for $\epsilon_r=1$, the dashed curve for
363: $\epsilon_r=5$, and the dashed-dotted curve for $\epsilon_r=10$.
364: One realizes that with increasing $\epsilon_r$, i.~e. with
365: stronger screening of the Coulomb interaction, the
366: susceptibility increases at intermediate temperatures and that
367: the maximum shifts to lower temperatures. Although being rather
368: moderate it is astonishing that the effect is at all observable
369: since the responsible levels are at excitation energies well
370: above that temperature range, compare the spectrum in
371: \figref{F-2}.
372: 
373: It turns out that the high degeneracy of excited hole
374: configurations is the reason for the influence of the hole
375: dynamics even at low temperatures. Looking again at the hole
376: configuration shown in \figref{F-1} one notices that the
377: spin-hole exchange processes can happen at very different places
378: leading to a large geometric degeneracy. This degeneracy can
379: overcompensate a small Boltzmann factor and thus expresses
380: itself in a high thermal weight. 
381: 
382: Together with the theoretical results \figref{F-3} shows the
383: experimentally determined magnetization which was measured along
384: the $c$-axis of the material at a magnetic field of $B=1$~T
385: \cite{Klingeler:2003}. The x-symbols depict the uncorrected
386: values whereas the crosses represent the corrected values. The
387: correction includes a subtraction of impurities (free spins
388: $s=1/2$) as well as a rescaling because the number of dimers on
389: the chain is less than theoretically possible. The almost
390: perfect coincidence with the theoretical result for
391: $\epsilon_r=1$ suggests that the hole-hole repulsion is rather
392: strong. The uncertainties in the measurement and the correction
393: procedure leave some freedom for the actual value, but it is
394: clear that $\epsilon_r$ should not be bigger than about three.
395: This implies that energy levels which result from other than the
396: dimer configuration are well above the triplet excitation,
397: compare \figref{F-2}. Although this might seem to be unrealistic
398: one has to keep in mind that any theoretical model must explain
399: why the experimental susceptibility practically coincides with
400: that of free dimers. This is only possible if other excitations
401: are well separated from the triplet excitation.
402: 
403: %===================    figure   =================================
404: \begin{figure}[ht!] 
405: \centering
406: %\resizebox{0.75\textwidth}{!}{\includegraphics{fig-4.eps}}
407: \includegraphics[clip,width=60mm]{fig-4.eps}
408: \caption{Specific heat at $B=0$: $\epsilon_r=1$ -- solid curve,
409:   $\epsilon_r=2$ -- dotted curve, $\epsilon_r=3$ -- thin curve,
410:   $\epsilon_r=5$ -- dashed curve, and $\epsilon_r=10$ --
411:   dashed-dotted curve.} 
412: \label{F-4}
413: \end{figure}
414: %===================    figure   =================================
415: 
416: But even if higher-lying energy levels are well above the
417: triplet excitation, due to their vast degeneracy they can
418: substantially contribute to the specific heat, which is shown in
419: \figref{F-4}. In order to demonstrate how the thermal weight of
420: the excited hole configurations grows several cases are shown.
421: The solid curve, which is the lowest among all curves, again
422: depicts the result for $\epsilon_r=1$.  With increasing
423: dielectric constants thermal weight is shifted from higher
424: temperatures to lower ones. This can very clearly be seen for
425: the shown sequence: $\epsilon_r=2$ (dotted curve),
426: $\epsilon_r=3$ (thin curve), $\epsilon_r=5$ (dashed curve), and
427: $\epsilon_r=10$ (dashed-dotted curve). This result shows that
428: for dielectric constants of the order of $\epsilon_r\approx 5$
429: about one third of the specific heat at its maximum is due to
430: states involving hole motion.
431: 
432: Inelastic neutron scattering is a valuable tool to measure the
433: magnetic excitation spectrum of substances like \telcomp, see
434: e.~g. \cite{RBM:PRB99,BRL:PB04}. The transition from the singlet
435: ground state to the first excited triplet state, arrow in
436: \figref{F-2}, has been measured with high accuracy and used to
437: determine the exchange constants, especially $J$ (\figref{F-1},
438: r.h.s.). For recent results have a look at \cite{BRL:PB04}. 
439: 
440: %===================    figure   =================================
441: \begin{figure}[ht!] 
442: \centering
443: %\resizebox{0.75\textwidth}{!}{\includegraphics{fig-5.eps}}
444: \includegraphics[clip,width=60mm]{fig-5.eps}
445: \caption{Rough sketch of the lowest transitions observable with
446:   inelastic neutron scattering: $\epsilon_r=1$ -- solid curve,
447:   $\epsilon_r=5$ -- dashed curve, and $\epsilon_r=10$ --
448:   dashed-dotted curve. The arrow marks the singlet-triplet
449:   transition at about 135~K.} 
450: \label{F-5}
451: \end{figure}
452: %===================    figure   =================================
453: 
454: In addition to this fundamental transition of the dimer
455: configuration, transitions to states with spin-hole exchange
456: should be detectable, too. Figure~\xref{F-5} shows as a rough
457: sketch where such transitions could be expected. The
458: singlet-triplet transition at about 135~K -- arrow in
459: \figref{F-5} -- is clearly seen for each dielectric constant. 
460: But with increasing $\epsilon_r$ transitions to states with one
461: or two holes moved become accessible. The dashed curve shows
462: schematically the lowest transitions for $\epsilon_r=5$, the
463: dashed-dotted curve the lowest transitions if $\epsilon_r=10$. 
464: Although the singlet-triplet transition might have the largest
465: matrix element, again the huge geometric degeneracy could help
466: to make the other transitions visible. A possible drawback is,
467: nevertheless, due to the fact that states involving spin-hole
468: exchange break the translational symmetry, compare \figref{F-1}. 
469: The momentum dependency, therefore, might be fuzzy. 
470: 
471: 
472: 
473: 
474: 
475: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
476: \section{Summary and Outlook}
477: \label{sec-4}
478: 
479: For the compound under investigation the model
480: is capable to address much more questions. Since it is not clear
481: whether the hole content of the chain subsystem in \telcomp\ is
482: really 60~\%, one can study the influence of a reduced number of
483: holes on magnetic observables. 
484: 
485: %===================    figure   =================================
486: \begin{figure}[ht!] 
487: \centering
488: %\resizebox{0.75\textwidth}{!}{\includegraphics{fig-6.eps}}
489: \includegraphics[clip,width=60mm]{fig-6.eps}
490: \caption{Zero-field magnetic susceptibility $\chi_0(T)$ for
491:   $\epsilon_r=5$: solid curve -- $N_{\text{s}}=16$ and
492:   $N_{\text{h}}=24$, dashed curve  -- $N_{\text{s}}=16$ and
493:   $N_{\text{h}}=23$.} 
494: \label{F-6}
495: \end{figure}
496: %===================    figure   =================================
497: 
498: Figure~\xref{F-6} shows as an example how a reduction by just
499: one hole will influence the magnetic susceptibility. The perfect
500: symmetry of successive spin-hole-spin dimers separated by two
501: holes is destroyed and a low-lying triplet competes with the
502: former singlet ground state which leads to a low-temperature
503: divergence of the susceptibility. It might very well be that a
504: part of the experimentally observed low-temperature divergence
505: of the susceptibility \cite{Klingeler:2003} is due to the
506: reduced hole doping of the chain. 
507: 
508: The absence of perfect symmetry also leads to an increased
509: mobility of the holes. At the imperfections neighboring holes
510: can be moved without altering the Coulomb energy much. 
511: Therefore, the excitation energy for configurations where a hole
512: is moved at imperfections will be rather low. 
513: 
514: 
515: Summarizing, the main advantage of the proposed effective
516: spin-hole Hamiltonian is that it allows to evaluate
517: thermodynamic observables both as function of temperature and
518: magnetic field for reasonably large systems. Using this model it
519: could be shown how a hole-hole Coulomb repulsion along the chain
520: would express itself in thermodynamic observables. The
521: comparison with experimental magnetization data suggests that
522: the screening is weak. The actual choice of the exchange
523: parameters, compare \figref{F-1}, does not influence the general
524: conclusions about the effects of the Coulomb repulsion between
525: holes.  It is not yet clear whether a weekly screened Coulomb
526: repulsion is realistic \cite{VFB:PRB03} or whether a combination
527: of hole-hole Coulomb repulsion and modulation of on-site
528: energies \cite{IsT:JPSJ98,GeL:PRL04,GeL:EPJB05,GeL:05} has to be
529: used. An experimental determination of other excited states than
530: the triplet excitation of the dimers would be very helpful in
531: this respect.
532: 
533: 
534: 
535: 
536: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
537: \section*{Acknowledgement}
538: 
539: I would like to thank Bernd B\"uchner, R\"udiger Klingeler,
540: Fatiha Ouchni, and Louis-Pierre Regnault for fruitful
541: discussions, R\"udiger Klingeler for providing the experimental
542: magnetization data, and Heinz-J\"urgen Schmidt for carefully
543: reading the manuscript.
544: 
545: 
546: 
547: \begin{thebibliography}{10}
548: 
549: \bibitem{Dag:RMP94}
550: E. Dagotto, Rev. Mod. Phys. {\bf 66},  763  (1994).
551: 
552: \bibitem{KCG:PRB01}
553: V. Kataev {\it et~al.}, Phys. Rev. B {\bf 64},  104422  (2001).
554: 
555: \bibitem{NMK:PRB00}
556: N. N\"ucker {\it et~al.}, Phys. Rev. B {\bf 62},  14384  (2000).
557: 
558: \bibitem{RBM:PRB99}
559: L.~P. Regnault {\it et~al.}, Phys. Rev. B {\bf 59},  1055  (1999).
560: 
561: \bibitem{TME:PRB98}
562: M. Takigawa, N. Motoyama, H. Eisaki, and S. Uchida, Phys. Rev. B {\bf 57},
563:   1124  (1998).
564: 
565: \bibitem{HoS:PRE03}
566: M. Holtschneider and W. Selke, Phys. Rev. E {\bf 68},  026120  (2003).
567: 
568: \bibitem{SPV:EPJB02}
569: W. Selke, V.~L. Pokrovsky, B. B\"uchner, and T. Kroll, Eur. Phys. J. B {\bf
570:   30},  83  (2002).
571: 
572: \bibitem{CBC:PRL96}
573: S.~A. Carter {\it et~al.}, Phys. Rev. Lett. {\bf 77},  1378  (1996).
574: 
575: \bibitem{ABC:PRB00}
576: U. Ammerahl {\it et~al.}, Phys. Rev. B {\bf 62},  8630  (2000).
577: 
578: \bibitem{Klingeler:2003}
579: R. Klingeler, {\em Spin- und Ladungsordnung in \"Ubergangsmetalloxiden:
580:   Thermodynamische und magnetische Untersuchungen}, {\em Forschungsberichte aus
581:   den Naturwissenschaften / Physik} (Mensch \& Buch Verlag, Berlin, 2003).
582: 
583: \bibitem{MYK:PRB99}
584: M. Matsuda, T. Yosihama, K. Kakurai, and G. Shirane, Phys. Rev. B {\bf 59},
585:   1060  (1999).
586: 
587: \bibitem{AKA:PRB98}
588: R. Arita, K. Kuroki, H. Aoki, and M. Fabrizio, Phys. Rev. B {\bf 57},  10324
589:   (1998).
590: 
591: \bibitem{KFE:Nature98}
592: S.~A. Kivelson, E. Fradkin, and V.~J. Emery, Nature (London) {\bf 393},  550
593:   (1998).
594: 
595: \bibitem{MYD:Science99}
596: A. Moreo, S. Yunoki, and E. Dagotto, Science {\bf 283},  2034  (1999).
597: 
598: \bibitem{Zaanen:Science99}
599: J. Zaanen, Science {\bf 286},  251  (1999).
600: 
601: \bibitem{HEB:PRL04}
602: C. Hess {\it et~al.}, Phys. Rev. Lett. {\bf 93},  027005  (2004).
603: 
604: \bibitem{IsT:JPSJ98}
605: M. Isobe and E. Takayama-Muromachi, J. Phys. Soc. Jpn. {\bf 67},  3119  (1998).
606: 
607: \bibitem{GeL:PRL04}
608: A. Gelle and M.~B. Lepetit, Phys. Rev. Lett. {\bf 92},  236402  (2004).
609: 
610: \bibitem{GeL:EPJB05}
611: A. Gelle and M.~B. Lepetit, Eur. Phys. J. B {\bf 43},  29  (2005).
612: 
613: \bibitem{GeL:05}
614: A. Gelle and M.~B. Lepetit,   , unpublished, cond-mat/0410203.
615: 
616: \bibitem{VFB:PRB03}
617: B. Valenzuela, S. Fratini, and D. Baeriswyl, Phys. Rev. B {\bf 68},  045112
618:   (2003).
619: 
620: \bibitem{MTM:PRB98A}
621: Y. Mizuno {\it et~al.}, Phys. Rev. B {\bf 57},  5326  (1998).
622: 
623: \bibitem{MTM:PRB98B}
624: Y. Mizuno, T. Tohyama, and S. Maekawa, Phys. Rev. B {\bf 58},  R14713  (1998).
625: 
626: \bibitem{CPP:PRL89}
627: C.~Y. Chen {\it et~al.}, Phys. Rev. Lett. {\bf 63},  2307  (1989).
628: 
629: \bibitem{BaS:PRB94}
630: F. Barriquand and G.~A. Sawatzky, Phys. Rev. B {\bf 50},  16649  (1994).
631: 
632: \bibitem{EKZ:PRB97}
633: V.~J. Emery, S.~A. Kivelson, and O. Zachar, Phys. Rev. B {\bf 56},  6120
634:   (1997).
635: 
636: \bibitem{Hubbard63}
637: J. Hubbard, Proc. R. Soc. London Ser. A-Math. {\bf 276},  238  (1963).
638: 
639: \bibitem{Hubbard64a}
640: J. Hubbard, Proc. R. Soc. London Ser. A-Math. {\bf 277},  237  (1964).
641: 
642: \bibitem{Hubbard64b}
643: J. Hubbard, Proc. R. Soc. London Ser. A-Math. {\bf 281},  401  (1964).
644: 
645: \bibitem{BRL:PB04}
646: C. Boullier {\it et~al.}, Physica B {\bf 350},  40  (2004).
647: 
648: \end{thebibliography}
649: 
650: 
651: 
652: 
653: %\bibliographystyle{/usr/share/texmf/tex/latex/local/revtex/prsty}
654: %\bibliography{/home/schnack/tex/bibtex/js-own,/home/schnack/tex/bibtex/js-mag,/home/schnack/tex/bibtex/js-cup,/home/schnack/tex/bibtex/js-mis}
655: 
656: 
657: 
658: \end{document}
659: 
660: