0805.3852/fkm.tex
1: \documentstyle[12pt]{article}
2: \textwidth=430pt
3: \textheight=640pt
4: \topmargin=-1pt
5: 
6: \title{Hartree-Fock study of electronic ferroelectricity in the
7: Falicov-Kimball model with $f$-$f$ hopping}
8: \author{Pavol Farka\v sovsk\'y\\
9: Institute  of  Experimental  Physics,  Slovak   Academy   of
10: Sciences\\
11: Watsonova 47, 040 01 Ko\v {s}ice, Slovakia}
12: \date{}
13: \begin{document}
14: \baselineskip=24pt
15: \maketitle
16: 
17: \begin{abstract}
18: 
19: The Hartree-Fock (HF) approximation with the charge-density-wave (CDW) instability
20: is used to study the ground-state phase diagram of the spinless Falicov-Kimball
21: model (FKM) extended by $f$-$f$ hopping in two and three dimensions. It is 
22: shown that the HF solutions with the CDW instability reproduce perfectly the
23: two-dimensional intermediate coupling phase diagram of the FKM model with 
24: $f$-$f$ hopping calculated recently by constrained path Monte Carlo
25: (CPMC) method. Using this fact we have extended our HF study on cases
26: that have been not described by CPMC, and namely, (i) the case of small
27: values of $f$-electron hopping integrals, (ii) the case of weak Coulomb 
28: interactions and (iii) the three-dimensional case. We have found that
29: ferroelectricity remains robust with respect to the reducing strength of
30: coupling ($f$-electron hopping) as well as with respect to the increasing
31: dimension of the system.
32: \end{abstract}
33: \thanks{PACS nrs.: 71.27.+a, 71.28.+d, 71.30.+h}
34: 
35: \newpage
36: 
37: %\section{Introduction}
38: The Falicov-Kimball model (FKM) is a paradigmatic example of simple
39: model to study correlation effects in interacting fermion systems
40: on a lattice~\cite{Falicov}. The model was originally proposed to 
41: describe metal-insulator transitions and has since been investigated 
42: in connection with a variety of problems such as binary 
43: alloys~\cite{Freericks}, the formation of ionic crystals~\cite{Gruber}, 
44: and ordering in mixed-valence systems~\cite{Farkas}. 
45: In the last few years the FKM was extensively studied in connection with 
46: the exciting idea of electronic ferroelectricity~\cite{Cz,F1,Zl,B1,B2,Yin}. 
47: The motivation for these studies comes from the pioneering work of 
48: Portengen at al.~\cite{P1,P2} who studied the FKM with a k-dependent 
49: hybridization in the Hartree-Fock (HF) approximation and found 
50: that the Coulomb interaction $U$ between the itinerant $d$-electrons 
51: and the localized $f$-electrons gives rise a non-vanishing excitonic
52: $\langle f^{+}d\rangle$-expectation value even in the limit of vanishing
53: hybridization $V \rightarrow 0$. As an applied (optical) electrical field
54: provides for excitations between d- and f-states and thus for a polarization
55: expectation value $P_{fd}=\langle f_i^{+}d_i\rangle$, the finding of a spontaneous
56: $P_{fd}$ (without hybridization or electric field) has been interpreted as
57: evidence for electronic ferroelectricity. This result stimulated further
58: theoretical studies of the model. 
59: Analytical calculations within well controlled approximation (for
60: $U$ small) performed by Czycholl~\cite{Cz} in infinite dimensions did
61: not confirm the existence of electronic ferroelectricity. In contrast to 
62: results obtained by Portengen et al.~\cite{P1,P2} he found that the FKM
63: in the symmetric case ($n_f=n_d=0.5$) does not allow for a 
64: ferroelectric ground state with a spontaneous polarization,
65: i.e., there is no nonvanishing $\langle f^{+}d\rangle$-expectation value in the
66: limit of vanishing hybridization.  The same conclusion has been also 
67: obtained independently by extrapolation of small-cluster 
68: exact-diagonalization and density matrix renormalization group (DMRG) 
69: calculations in the one dimension for both 
70: intermediate and strong interactions~\cite{F1}. In these regions 
71: the finite-size effects are negligible and thus the results can be 
72: satisfactory extrapolated to the thermodynamic limit. 
73: 
74: Hybridization between the itinerant $d$ and localized $f$ states, however,
75: is not the only way to develop $d$-$f$ coherence. Recent theoretical works
76: of Batista et al.~\cite{B1,B2} showed that the ground state with 
77: a spontaneous electric polarization can also be induced by $f$-$f$ hopping
78: for dimensions $D>1$.
79: In the strong coupling limit this result has been proven by mapping the
80: extended FKM into the $xxz$ spin 1/2 model with a magnetic field along 
81: the $z$-direction, while in the intermediate coupling regime the
82: ferroelectric state has been identified numerically by constrained path 
83: Monte Carlo (CPMC) technique. 
84: On the base of these results the authors postulated 
85: the following conditions that favor the formation of the electronically 
86: driven ferroelectric state: (a) The system must be in a mixed-valence regime 
87: and the two bands involved must have different parity. (b) It is best, 
88: though not necessary, if both bands have similar bandwidths. (c) A local 
89: Coulomb repulsion ($U$) between the different orbitals is required. 
90: 
91: In the present paper we study the extended FKM (the spinless FKM with
92: $f$-$f$ hopping) in the HF approximation with the charge-density-wave (CDW) 
93: instability. For reasons mentioned above we restrict our studies
94: on dimensions $D>1$. First we show that the HF solutions
95: with the CDW instability reproduce perfectly the ground-state phase 
96: diagram obtained by the CPMC method ($D=2$) for 
97: intermediate Coulomb interactions~\cite{B2}. This "calibration" allows us to 
98: extend calculations to the case of small values of the $f$-electron hopping
99: integral $|t_f| < 0.1$, that has been omitted in the CPMC phase diagram 
100: for numerical problems. Just in this region 
101: we have found a new phase that corresponds to the inhomogeneous solution    
102: for $\langle f^+d\rangle$-expectation value. This result completes the ground-state 
103: phase diagram of the two-dimensional FKM extended by $f$-$f$ hopping for 
104: intermediate couplings. The same calculations we have performed also in 
105: the weak coupling limit (for $D=2$) as well as in three dimensions.  
106: We have found that the ferroelectricity remains robust with respect to the
107: reducing strength of the coupling as well as with respect to the increasing
108: dimension of the system. 
109: 
110: \section{The model}
111: 
112: The extended FKM for the spinless fermions on a $D$-dimensional hypercubic
113: lattice is
114: \begin{equation}
115: H=-t_d\sum_{<ij>}d^+_id_j-t_f\sum_{<ij>}f^+_if_j+U\sum_if^+_if_id^+_id_i
116: +E_f\sum_if^+_if_i,
117: \end{equation}
118: where $f^+_i$ $(f_i)$ and $d^+_i$ $(d_i)$
119: is the creation (annihilation) operator  of heavy ($f$)  and light ($d$) 
120: electron at lattice site $i$.
121: 
122: The first two terms of (1) are the kinetic energies corresponding to
123: quantum-mechanical hopping of $d$ and $f$ electrons 
124: between the nearest neighbor sites $i$ and $j$
125: with hopping probabilities $t_d$ and $t_f$, respectively.
126: The third term represents the on-site
127: Coulomb interaction between the $d$ electrons with density
128: $n_d=\frac{1}{L}\sum_id^+_id_i$ and the $f$ electrons with density 
129: $n_f=\frac{1}{L}\sum_if^+_if_i$, where $L$ is the number of lattice sites. 
130: Usually, the hopping integral of the $d$ electrons is taken to be the unit 
131: of energy $(t_d=1)$ and the $f$-electron hopping integral is considered in the 
132: limit $|t_f| < 1$. This is a reason why the $d$ electrons are called
133: light and the $f$ electrons heavy.  
134: 
135: 
136: In our HF study of the extended FKM we go beyond the usual HF
137: approach~\cite{Leder} in which
138: only homogeneous solutions are postulated. In accordance with~\cite{Brydon} 
139: we consider here also inhomogeneous solutions modeled by  a periodic 
140: modulation of the order parameters:
141: \begin{eqnarray}
142: \langle n_i^f\rangle=n^f+\delta_f\cos({\bf Q}\cdot {\bf r}_i)\ ,\\
143: \langle n_i^d\rangle=n^d+\delta_d\cos({\bf Q}\cdot {\bf r}_i)\ ,\\
144: \langle f_i^+d_i\rangle=\Delta+\Delta_P\cos({\bf Q}\cdot {\bf r}_i)\ .
145: \end{eqnarray}
146: where $\delta_{d}$ and $\delta_{f}$ is the order parameter of the CDW state 
147: for the $d$- and $f$-electrons and $\Delta$ is the excitonic average.
148: The nesting vector ${\bf{Q}}=(\pi,\pi)$ for $D=2$ and 
149: ${\bf{Q}}=(\pi,\pi,\pi)$ for $D=3$.
150: 
151: Using expressions for $\langle n_i^f\rangle$,$\langle n_i^d\rangle$ 
152: and $\langle f_i^+d_i\rangle$ the HF Hamiltonian of the extended
153: FKM can be written as
154: \begin{eqnarray}
155: {\cal H}=&-&t_d\sum_{\langle i,j\rangle} d^+_id_j - t_f\sum_{\langle i,j\rangle}f^+_if_j
156: + E_f\sum_i n_i^f
157:       + U\sum_i (n^f+\delta_f\cos({\bf Q}\cdot {\bf r}_i))n_i^d
158: \nonumber\\
159:       &+& U\sum_i (n^d+\delta_d\cos({\bf Q}\cdot {\bf r}_i))n_i^f
160:       -U\sum_{i} ( \Delta+\Delta_P\cos({\bf Q}\cdot {\bf r}_i))
161:       d_i^+f_i+h.c. 
162: \end{eqnarray}
163: Following the work of Brydon et al.~\cite{Brydon} the effective HF Hamiltonian
164: is diagonalized by canonical transformation 
165: \begin{eqnarray}
166: \begin{array}{ccc}
167: \gamma_k^m=u_k^md_k + v_k^md_{k+{\bf Q}} + a_k^mf_k + b_k^mf_{k+{\bf Q}}\ , &   &
168: m=1,2,3,4\ ,
169: \end{array}
170: \end{eqnarray}
171: where $a_k^m,b_k^m,u_k^m,v_k^m$ are solutions of the associated Bogoliubov-de Gennes (BdG)
172: eigenequations:
173: 
174: \begin{equation}
175: H_k\Psi_k^m=E_k^m\Psi_k^m\ , 
176: \end{equation}
177: with
178: \begin{eqnarray}
179: H_k=\left( \begin{array}{cccc}
180:     \epsilon_k^d+Un^f &  U\delta_f &   -U\Delta &     -U\Delta_P \\
181:      U\delta_f  & \epsilon_{k+Q}^d+Un^f & -U\Delta_P &    -U\Delta\\
182:      -U\Delta   &  -U\Delta_P   &  \epsilon_{k}^f+Un^d+E_f & U\delta_d \\
183:      -U\Delta_P &  -U\Delta    &  U\delta_d  & \epsilon_{k+Q}^f+Un^d+E_f\\
184:             \end{array}          
185: \right)
186: \end{eqnarray}
187: and
188: \begin{eqnarray}
189: \Psi_k^m=\left(\begin{array}{c}
190:                u_k^m\\
191:                v_k^m\\
192:                a_k^m\\
193:                b_k^m
194:                 \end{array}
195: \right) 
196: \end{eqnarray}
197: 
198: The corresponding energy dispersions $\epsilon_k^d$ and $\epsilon_k^f$ 
199: can be obtained directly by the Fourier transform of the $d$- and
200: $f$-electron hopping amplitudes and for the case of hypercubic lattice
201: they are given by ($\alpha=d,f$):
202: 
203: \begin{eqnarray}
204: \epsilon_k^{\alpha}&=&-2t_{\alpha}(\cos(k_x)+\cos(k_y)), \mbox{ for D=2},\\
205: \epsilon_k^{\alpha}&=&-2t_{\alpha}(\cos(k_x)+\cos(k_y)+\cos(k_z)), \mbox{
206: for D=3}.
207: \end{eqnarray}
208: 
209: 
210: The HF parameters $n_d,\delta_d,n_f,\delta_f,\Delta,\Delta_P$ can be written
211: directly in terms of the Bogoliubov-de Gennes eigenvectors:
212: \begin{equation}
213: n^d=\frac{1}{N}\sum_k{}'\sum_m \{ u_k^mu_k^m + v_k^mv_k^m\}f(E_k^m)\ .
214: \end{equation}
215: \begin{equation}
216: \delta_d=\frac{1}{N}\sum_k{}'\sum_m \{ v_k^mu_k^m + u_k^mv_k^m\}f(E_k^m)\ .
217: \end{equation}
218: \begin{equation}
219: n^f=\frac{1}{N}\sum_k{}'\sum_m \{ a_k^ma_k^m + b_k^mb_k^m\}f(E_k^m)\ .
220: \end{equation}
221: \begin{equation}
222: \delta_f=\frac{1}{N}\sum_k{}'\sum_m \{ b_k^ma_k^m + a_k^mb_k^m\}f(E_k^m)\ .
223: \end{equation}
224: \begin{equation}
225: \Delta=\frac{1}{N}\sum_k{}'\sum_m \{ a_k^mu_k^m + b_k^mv_k^m\}f(E_k^m)\ .
226: \end{equation}
227: \begin{equation}
228: \Delta_P=\frac{1}{N}\sum_k{}'\sum_m \{ b_k^mu_k^m + a_k^mv_k^m\}f(E_k^m)\ .
229: \end{equation}
230: where the  prime denotes summation over half the Brillouin zone and 
231: $f(E)=1/\{1+\exp[\beta(E-\mu)]\}$ is the Fermi distribution function.
232: 
233: The same approach has been used recently by Brydon et al.~\cite{Brydon} 
234: to study the interplay between excitonic effects and the CDW instability in the FKM with  on-site as well as 
235: non-local hybridization. Here we use the zero temperature variant of this
236: procedure to describe ground-state phase diagram of the spinless
237: FKM with $f$-$f$ hopping.
238: 
239: \section{Results and Discussion}
240: To determine the ground-state phase diagram of the extended FKM in the 
241: $E_f$-$t_f$ plane (corresponding to selected $U$) the HF equations are solved 
242: self-consistently for each pair of $(E_f,t_f)$ values.
243: We use an exact diagonalization method to solve the Bogoliubov-de Gennes  
244: equation. We start with an initial set of order parameters. 
245: By solving Eq.~(7), the new
246: order parameters are computed via Eqs.~(12) to
247: (17) and are substituted back into
248: Eq.~(7). The iteration is repeated until a desired
249: accuracy is achieved.
250: 
251: First we have examined the two-dimensional extended FKM model in the 
252: intermediate coupling regime and $t_f$ negative. For this case there 
253: exists the 
254: comprehensive phase diagram of the model obtained by a CPMC 
255: technique~\cite{B2}  for $f$-electron hopping integrals $|t_f| \geq 0.1$.
256: According these Monte-Carlo studies the phase diagram of the extended FKM
257: consists of only three main phases, and namely, (i) the integer-valent 
258: state ($n_f=0,1, n_d=1,0$), (ii) the mixed-valent CDW state ($n_f=n_d=0.5$), 
259: and (iii) the mixed-valent ferroelectric state that is stable for remaining
260: values of $n_f$ ($n_d$).     
261: 
262: In Fig.~1 we have displayed typical examples of our HF solutions 
263: obtained for $n_d,\delta_d,n_f,\delta_f,\Delta,\Delta_P$ in the intermediate 
264: coupling regime $U=2$. 
265: It is seen that the extended FKM in the HF approximation with the
266: CDW instability exhibits non-vanishing excitonic  $\langle
267: f^+d\rangle$-expectation value for all $f$-electron densities except the
268: case when $n_f=0, 1/2$ and 1. Thus in accordance with the Quantum
269: Monte-Carlo studies~\cite{B2} we have found that the ferroelectric
270: ground state with the spontaneous polarization is stabilized when the system
271: is in the mixed valence regime and the sign of the $f$-electron hopping
272: integral is opposite to the sign of the $d$-electron one. The fact that
273: HF solutions can describe the existence of ferroelectric
274: ground-state with spontaneous polarization is not surprising, since this
275: state has been found already in the homogeneous HF solution of the
276: conventional FKM ($t_f=0$) in the limit of vanishing hybridization 
277: $V\rightarrow 0$~\cite{P1,P2}, even for all
278: $f$-electron concentrations (for all values of $E_f$ from the $d$-electron
279: band). What is however surprising is that the HF solutions with the CDW
280: instability reproduce perfectly the ground-state phase diagram obtained by
281: CPMC method for all examined values of $f$-electron hopping ($|t_f|\geq
282: 0.1$). This is clearly demonstrated in Fig.~2, where both phase diagrams are
283: compared.
284: 
285: The fact that the HF approximation with the CDW instability can describe
286: qualitatively as well as quantitatively ground-state properties of the FKM
287: with $f$-electron hopping motivated us to extend our HF study on cases that 
288: have been not described by Quantum Monte-Carlo simulations. At first 
289: this is the case of
290: small $f$-electron hopping integrals ($|t_f|<0.1$) that has been not
291: considered in the original work of Batista et al.~\cite{B2} because 
292: numerical difficulties which appear in the Quantum Monte-Carlo simulations
293: for small $t_f$ (the limitations in the numerical accuracy). The second
294: interesting case that we would like to study here within the HF theory is
295: the three-dimensional case for which the numerical results are very rare due
296: to numerical limitations on the size of clusters. 
297: 
298: Let us first discuss our two-dimensional results obtained in the limit of
299: small values of $f$-electron hopping integrals. In Fig.~3 we present
300: results of detailed HF analysis  performed in this limit for 
301: $\Delta,\Delta_P$ and $\delta_d$. It is seen that the non-vanishing 
302: excitonic $\langle f^+d\rangle$
303: expectation value persists also for small values of $|t_f|$ but now the
304: inhomogeneous solution $\Delta_P\neq 0$ (with AB-sublattice oscillations in
305: the excitonic and charge order parameters) is stabilized
306: against the homogeneous one ($\Delta_P=0$). The effect is especially strong
307: when we approach the $t_f=0$ limit. This is clearly demonstrated in Fig.~4,
308: where the complete intermediate-coupling phase diagram of the FKM with
309: $f$-$f$-hopping is displayed. Five different phases depicted in Fig.~4 as 
310: $\alpha$ (the full $f$ band), $\beta, \beta'$ (the excitonic phases),
311: $\gamma$ (the CDW phase) and $\epsilon$ (the full $d$ band) correspond
312: to following HF solutions: 
313: \begin{eqnarray}
314:  \begin{array}{cccccll}
315:   \alpha \mbox{ phase:} & \Delta=0, & \Delta_P=0,   
316:   & \delta_f=0, & \delta_d=0,  & n_f=1
317:   \nonumber\\
318:   \beta \mbox{ phase:} & \Delta>0, &  \Delta_P=0,   
319:   & \delta_f=0, & \delta_d=0,  & 0<n_f<n_f^c & \mbox{for $E_f>0$} 
320:   \nonumber\\
321:     & &  
322:   & &  & 1-n_f^c<n_f<1 & \mbox{for $E_f<0$} 
323:   \nonumber\\
324:   \beta' \mbox{ phase:} & \Delta > 0, & \Delta_P< 0,   
325:   & \delta_f<0, & \delta_d>0, & n_f^c<n_f<1/2 & \mbox{for $E_f>0$}  
326:   \nonumber\\
327:     & &  
328:   & &  & 1/2<n_f<1-n_f^c & \mbox{for $E_f<0$} 
329:   \nonumber\\
330:   \gamma \mbox{ phase:} & \Delta=0,  &\Delta_P=0,   
331:   & \delta_f<0, & \delta_d>0,  & n_f=1/2
332:   \nonumber\\
333:   \varepsilon \mbox{ phase:} & \Delta=0, & \Delta_P=0,   
334:   & \delta_f=0, & \delta_d=0,  & n_f=0
335:   \nonumber\\
336:  \end{array}
337: \end{eqnarray}
338: The stability of different HF solutions was also checked numerically by
339: calculating the total energy and it was found that all phases presented in
340: the ground-state phase diagram represent the most stable HF solutions. 
341: To determine the type of transitions between different phases we have
342: performed an exhaustive numerical study of the $E_f$ dependence of the HF
343: order parameters (the typical examples are shown in Fig.~1 and Fig.~3). 
344: At a first glance it seems that there are both first-order ($t_f$ large) and 
345: second-order ($t_f$ small) phase transitions in the extended FKM with $f$-$f$
346: hopping. However, a more detailed analysis of numerical data (with much 
347: higher resolution than used in Fig.~1 and Fig.~3) showed that the $\beta'$ 
348: phase persists also for large $t_f$, although its stability region is now 
349: considerably reduced (see insets in Fig.~1). Thus there is no difference 
350: between the case of small and large values of $t_f$. In both cases the HF
351: order parameters change continuously indicating that the phase transitions
352: between different phases presented in the ($E_f$-$t_f$) ground-state phase 
353: diagram are of the second order.     
354: 
355: 
356: The same calculations we have performed also in the weak coupling limit
357: ($U \leq 1$). We have found that the phase diagrams obtained in the weak and
358: intermediate coupling regime have qualitatively the same form and only one
359: difference between them is that the ferroelectric domain ($\beta$) is 
360: stabilized against remaining phases with decreasing Coulomb 
361: interaction (see Fig.~4). Of course, this fact does not imply automatically 
362: that the excitonic $\langle f^+d\rangle$ expectation value persists also 
363: for vanishing $U$ and that the Coulomb interaction $U$ is not necessary 
364: for a stabilization of the ferroelectric state, what should be in 
365: a contradiction with conclusions  based on the CPMC simulations. 
366: Indeed, calculations that we have performed for different values of $t_f$ 
367: at the selected $f$-electron density $n_f=1/4$ showed (see Fig.~5) that the excitonic 
368: $\langle f^+d\rangle $ expectation  value is zero for $U=0$,
369: rapidly increases with increasing $U$ and tends to the saturated state for
370: $U$ sufficiently large. This confirms independently the third postulate of
371: Batista et al.~\cite{B2} and namely, that the local Coulomb interaction 
372: between the different orbitals is required in order to stabilize the 
373: ferroelectric state with the spontaneous polarization.
374: 
375: Before discussing the case of positive $t_f$ let us show explicitly the HF
376: solution for the limit of the conventional FKM ($t_f=0$). For this case we 
377: have found that $\Delta=\Delta_P=0$ in the $\alpha, \gamma$ and $\varepsilon$ 
378: phase, while $\Delta=-\Delta_P$
379: in the $\beta'$ phase. The last solution implies that the extitonic $\langle
380: f^+_id_i\rangle$-expectation value is equal to $2\Delta$ on the $A$
381: sublattice of the hypercubic lattice, while $\langle f^+_id_i\rangle =0$ on
382: the $B$ sublattice. For the symmetric case $E_f=0$ our solutions are fully
383: consistent with the Czycholl's ones obtained in the limit of infinite
384: dimensions~\cite{Cz}. On the other hand both these inhomogeneous solutions
385: fully differ from the homogeneous one~\cite{P1,P2} that predicts a non-zero
386: excitonic $\langle f^+_id_i\rangle$-expectation value for all $E_f$ from the
387: mixed valence regime with maximum of $\langle f^+_id_i\rangle$ at $E_f=0$.
388: 
389: Similar calculations as for $t_f<0$ we have performed also for $t_f>0$. We
390: have found that the ground-state phase diagram for $t_f>0$  has exactly the
391: same form as for $t_f<0$, however five different phases $\alpha, \beta, 
392: \beta', \gamma$ and $\varepsilon$
393: are now characterized by:
394: \begin{eqnarray}
395:  \begin{array}{cccccll}
396:   \alpha \mbox{ phase:} & \Delta=0, & \Delta_P=0,   
397:   & \delta_f=0, & \delta_d=0, & n_f=1
398:   \nonumber\\
399:   \beta \mbox{ phase:} & \Delta = 0, & \Delta_P < 0,   
400:   & \delta_f=0,  & \delta_d=0, & 0<n_f<n_f^c  & \mbox{for $E_f>0$}
401:   \nonumber\\
402:    & & 
403:   &  & & 1-n_f^c<n_f<1  & \mbox{for $E_f<0$}
404:   \nonumber\\
405:   \beta' \mbox{ phase:} & \Delta > 0, & \Delta_P< 0,   
406:   & \delta_f<0, & \delta_d>0, & n_f^c<n_f<1/2 &  \mbox{for $E_f>0$}
407:   \nonumber\\
408:    & & 
409:   &  & & 1/2<n_f<1-n_f^c  & \mbox{for $E_f<0$}
410:   \nonumber\\
411:   \gamma \mbox{ phase:} & \Delta=0,  &\Delta_P=0,   
412:   & \delta_f<0, & \delta_d>0, & n_f=1/2
413:   \nonumber\\
414:   \varepsilon \mbox{ phase:} & \Delta=0, & \Delta_P=0,   
415:   & \delta_f=0, & \delta_d=0 & n_f=0
416:   \nonumber\\
417:  \end{array}
418: \end{eqnarray}
419: Thus the main difference between the phase diagrams obtained for 
420: negative and positive $t_f$ is that the ferroelectric domain $\beta$ at $t_f<0$ 
421: is replaced by the antiferroelectric one at $t_f>0$. These two large domains
422: are separated by a relatively narrow $\beta'$ domain within which the
423: sublattice excitonic averages ($P^{A}_{fd},P^{B}_{fd}$) change 
424: continuously (see Fig.~6) from the ferroelectric case 
425: ($P^{A}_{fd}=P^{B}_{fd}$) to the antiferroelectric case 
426: ($P^{A}_{fd}=-P^{B}_{fd}$). 
427: 
428: Qualitatively the same picture we have observed also for the
429: three-dimensional case. This is illustrated  in Fig.~7, where the
430: ground-state phase diagrams of the extended FKM are plotted for two different
431: values of Coulomb interaction ($U=2$ and $U=4$). These results indicate that
432: ferroelectricity remains robust with respect to the increasing dimension of
433: the system, what should be important for an application of HF
434: solutions on a description of real three dimensional systems.
435: 
436: In conclusion, we have calculated the ground-state phase diagram 
437: of the spinless FKM with $f$-$f$ hopping in the HF approximation 
438: with the CDW instability. We have found that the HF solutions with the CDW 
439: instability reproduce perfectly the two-dimensional intermediate coupling
440: phase diagram of the extended FKM calculated by CPMC method. Using this 
441: fact we have extended our HF study on cases that have been not described by 
442: CPMC and namely, the case of small values of $f$-electron hopping integrals, 
443: the case of weak Coulomb interactions and the three-dimensional case. 
444: We have found that the ferroelectric ground state with the spontaneous 
445: polarization remains stable in all examined cases. 
446: 
447: 
448: \vspace{0.5cm}
449: This work was supported by the Slovak Grant Agency for Science
450: under grant No. 2/7057/27 and the Slovak APVV Grant Agency under Grant
451: LPP-0047-06. I would also like to acknowledge H. \v Cen\v carikov\'a for a
452: technical help during the preparation of manuscript.
453: \\
454: {\it Note added.-} After submitting this work we came to know about the work
455: of Schneider and Czycholl~\cite{Schneider} who studied the extended FKM in
456: the limit of infinite dimensions and obtained results similar to ours.
457: 
458: \newpage
459: \begin{thebibliography}{99}
460: 
461: \bibitem{Falicov} L.M. Falicov and J.C. Kimball, Phys. Rev. Lett.
462: {\bf 22}, 997 (1969).
463: 
464: \bibitem{Freericks} J. K. Freericks and L. M. Falicov, Phys. Rev. B {\bf
465: 41}, 2163 (1990).
466: 
467: \bibitem{Gruber} Ch. Gruber, J. L. Lebowitz and N. Macris, Phys. Rev. B {\bf
468: 48}, 4312 (1993); R. Lemanski, J.K. Freericks, G. Banach, J. Stat. Phys.
469: {\bf  116}, 699 (2004). 
470: 
471: \bibitem{Farkas} P. Farka\v sovsk\'y, Eur. Phys. J. B {\bf 20}, 209 (2001); 
472: P. Farka\v sovsk\'y, H. \v Cen\v carikov\'a, Eur. Phys. J. B {\bf  45}, 479 
473: (2005).  
474: 
475: \bibitem{Cz} G. Czycholl, Phys. Rev. B {\bf 59}, 2642 (1999).
476: 
477: \bibitem{F1} P. Farka\v{s}ovsk\'y, Phys. Rev. B {\bf 59}, 9707 (1999); Phys.
478: Rev. B {\bf 65}, 081102(R) (2002).
479: 
480: \bibitem{Zl} V. Zlati\'c, J.K. Freericks, R. Lemanski, G. Czycholl, 
481: Philosophical magazine B {\bf 81}, 1443 (2001); J. K. Freericks, V.
482: Zlati\'c, Rev. Mod. Phys. {\bf 75}, 1333 (2003).
483: 
484: \bibitem{B1} C. D. Batista, Phys. Rev. Lett. {\bf 89}, 166403 (2002).
485: 
486: \bibitem{B2} C. D. Batista, J. E. Gubernatis, J. Bonca, and H. Q. Lin, 
487: Phys. Rev. Lett. {\bf 92}, 187601 (2004). 
488: 
489: \bibitem{Yin} W. G. Yin, W. N. Mei, Ch. G. Duan, H. Q. Lin, and J. R.
490: Hardy, Phys. Rev. B {\bf 68}, 075111 (2003). 
491: 
492: \bibitem{P1} T. Portengen, T. \"Ostreich, L.J. Sham, Phys. Rev. Lett.
493: {\bf 76}, 3384 (1996).
494: 
495: \bibitem{P2} T. Portengen, T. \"Ostreich, L.J. Sham, Phys. Rev. B
496: {\bf 54}, 17452 (1996).
497: 
498: \bibitem{Leder} H.J. Leder, Solid State Commun. {\bf 27}, 579 (1978).
499: 
500: \bibitem{Brydon} P. M. R. Brydon, J. X. Zhu, M. Gulacsi, A. R. Bishop,
501: Phys. Rev. B {\bf 72}, 125122 (2005).
502: 
503: \bibitem{Schneider} C. Schneider and G. Czycholl, cond-mat/0802.1828.
504: \end{thebibliography}
505: 
506: \newpage
507: 
508: \centerline{\bf Figure Caption}
509: 
510: \vspace{0.5cm}
511: Fig.~1. Dependence of the HF parameters $n_f,\delta_f,n_d,\delta_d,\Delta$
512: and $\Delta_P$ on the $f$-level energy $E_f$ calculated (with step $\Delta
513: E_f=0.005$) for three different values of $t_f$ ($t_f=-0.2, -0.5,-0.8$) 
514: and $U=2$. Insets show the $t_f=-0.5$ case at much higher resolution (the
515: numerical data have been obtained with step $\Delta E_f=0.00005$). The case
516: of $t_f=-0.8$ is analogous to $t_f=-0.5$.
517: 
518: \vspace{0.5cm}
519: Fig.~2. The HF ($\bullet$) and CPMC~\cite{B2} ($\Box$) phase diagram of the two
520: dimensional FKM with $f$-$f$ hopping obtained for $U=2$.
521: 
522: \vspace{0.5cm}
523: Fig.~3. Dependence of the HF parameters $\Delta$, $\Delta_P$ and $\delta_d$ 
524: on the $f$-level energy $E_f$ calculated for different values of $t_f$ 
525: ($t_f=0, -0.01, -0.02, -0.05$) and $U=2$.
526: 
527: \vspace{0.5cm}
528: Fig.~4. The complete HF phase diagram of the two-dimensional extended FKM in
529: the intermediate ($U=2$) and weak coupling ($U=1$) regime. 
530: 
531: \vspace{0.5cm}
532: Fig.~5. Dependence of the HF parameter $\Delta$ on the Coulomb interaction
533: $U$ calculated for different values of $t_f$ and $n_f=1/4$.
534: 
535: \vspace{0.5cm}
536: Fig.~6. Dependence of the excitonic expectation value 
537: $P_{fd}=\langle f^{+}_id_i\rangle$ on $t_f$ calculated 
538: for $E_f=0.7$ and $U=2$.
539: 
540: \vspace{0.5cm}
541: Fig.~7. The complete HF phase diagram of the three-dimensional extended FKM
542: calculated for $U=2$ and $U=4$.
543: 
544: \end{document}
545: