1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %FEL.tex
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \documentclass[12pt]{article}
5: \usepackage[dvips]{graphicx}
6: \usepackage{cite}
7: %%Latex definitions%%%
8: %]
9: \setlength{\topmargin}{-1.4cm}
10: %{c
11: \setlength{\textheight}{24cm}
12: %{
13: \setlength{\textwidth}{15.5cm}
14: %y[WR> \setlength{\oddsidemargin}{0.2cm}
15: %y[WER> \setlength{\evensidemargin}{0.2cm}
16: %\renewcommand{\baselinestretch}{2.0}
17: %sR> \renewcommand{\baselinestretch}{1.5}
18: %%%end of latex definitions%%%
19: \begin{document}
20: \title{Construction of the free energy landscape by the density functional
21: theory}
22: \author
23: {Takashi Yoshidome, Akira Yoshimori and Takashi Odagaki\\
24: Department of Physics, Kyushu University, Fukuoka 812-8581, Japan}
25: \date{ }
26: \maketitle
27: \begin{abstract}
28: On the basis of the density functional theory, we give a clear definition of
29: the free energy landscape. To show the usefulness of the definition, we
30: construct the free energy landscape for rearrangement of atoms in an FCC
31: crystal of hard spheres. In this description, the cooperatively rearranging
32: region (CRR) is clealy related to the hard spheres involved in the saddle
33: between two adjacent basins.
34: A new concept of the simultaneously rearranging region (SRR) emerges naturally
35: as spheres defined by the difference between two adjacent basins. We show that
36: the SRR and the CRR can be determined explicitly from the free energy
37: landscape.
38: \end{abstract}
39: %\keywords{Suggested keywords}
40: %\pacs{64.70.Pf, 61.42.+h, 61.12.Bt}
41:
42: \newpage
43: \section{Introduction}
44: The thermodynamics and the dynamics of glass forming substances exhibit
45: anomalous behaviors near the glass transition temperature, $T_g$.
46: As the sample which was prepared by rapid cooling below $T_g$ is heated,
47: the specific heat exhibits an abrupt increment at
48: $T_g$\cite{hinetuexp_yamamuro}.
49: It is also known that supercooled liquids near $T_g$ show several
50: relaxation processes and that the relaxation time of the slowest process
51: tends to diverge at a temperature below $T_g$\cite{rev_glass}.
52: It is, therefore, desirable to construct a theory for the glass transition
53: which gives a unified explanation for these dynamic and thermodynamic
54: anomalies.
55:
56: The mode coupling theory (MCT)\cite{MCT} for the glass transition
57: describes the liquid dynamics based on the generalized Langevin equation
58: and predicts an ergodic-to-nonergodic transition at a critical temperature,
59: $T_c$ due to the nonlinear coupling between various modes. However,
60: it is now believed that $T_c$ is much higher than $T_g$.
61: In addition, MCT cannot treat the thermodynamic singularities.
62: Mezard and Parisi estimated thermodynamic properties of glassy state
63: in quasi-equilibrium with the replica method \cite{Mezard}.
64: In this approach, the Kauzmann temperature, $T_K$, becomes an ideal glass
65: transition temperature. However, $T_K$ is much lower than $T_g$ and
66: the replica method cannot treat the dynamics either.
67: Both theories thus can not give a unified explanation for the thermodynamic
68: and dynamic singularities of supercooled liquids near $T_g$.
69:
70: We have proposed the free energy landscape picture by which
71: a unified understanding is given for the singularities
72: near $T_g$\cite{odagaki04}.
73: The free energy landscape is defined in the configurational space
74: as a function of fictitious localized atomic positions and
75: each point in the free energy landscape is determined by coarse-graining of
76: the microscopic motion in the configurational space.
77: The free energy landscape is expected to have many basins and the saddle
78: points at low temperatures. The separation of the dynamics can be
79: understood by the separation of the dynamics among basins and the dynamics
80: within a basin.
81: The thermodynamic and dynamic singularities near $T_g$ can also be explained
82: by the free energy landscape picture. The trapping diffusion model\cite{TDM}
83: and its generalization to dynamics in a landscape\cite{odagaki04} provide a unified
84: understanding of the dynamical singularities near the glass transition.
85: This model showed that the dynamical transition occurs at the temperature
86: where the mean waiting time for escaping from a basin diverges.
87: Tao et al.\cite{hinetu_tao} proposed a model based on the free energy
88: landscape picture to explain the specific heat anomaly at the glass transition.
89: They showed that the anomaly of the specific heat occurs when the mean waiting
90: time exceeds the observation time at low temperatures.
91:
92: It is now believed that the glass transition can be phenomenologically well
93: explained by the free energy landscape picture. It is thus an important
94: question how one can explicitly define and construct the landscape
95: on the basis of the microscopic Hamiltonian. Because of this lack of
96: an explicit theory, the analysis of the dynamics in the landscape has
97: been detered so far.
98:
99: The aim of the present paper is to define explicitly the free energy landscape
100: using the microscopic Hamiltonian. To this end, we exploit the density
101: functional theory (DFT)\cite{HM,Oxtoby,Singh,Yoshimori}.
102:
103: One of the key concepts to understand the glass transition singularity
104: is the cooperatively rearranging region (CRR) proposed by Adam and
105: Gibbs\cite{AG}.
106: Adam and Gibbs succeeded in explaining the non-Arrenius behavior in the
107: temperature dependence of the viscosity for fragile glass forming liquids.
108: However there is no clear common understanding of the physical meaning for
109: the CRR. From the present definition of the free energy landscape,
110: the clear meaning of the CRR emerges. In addition to the CRR,
111: the present construction of the landscape leads to a new concept of
112: rearranging region, namely simultaneously rearranging region (SRR)
113: which is defined by the difference between the atomic
114: configuration in two adjacent basins.
115:
116: This paper is organized as follows. We give a definition of the free energy
117: landscape by the DFT in Sec 2. To confirm the usefulness of this definition,
118: we construct the free energy landscape for rearrangement of atoms in an FCC
119: crystal of hard spheres in Sec 3. A relation between the free energy landscape
120: and the CRR is given in Sec 4. We also discuss the SRR in Sec. 4.
121: Results are summarized in Sec 5.
122: %
123: \section{Definition of the free energy landscape}
124: We define the free energy by the partition function which is obtained by
125: the partial sum of the fast motion in the configurational space.
126: The free energy becomes a function of the fictitious atomic positions
127: and we call this free energy as the landscape.
128: We exploit the density functional theory (DFT)\cite{HM,Oxtoby,Singh,Yoshimori}
129: to calculate the free energy. In the DFT the grand potential,
130: $\Omega[\rho(\textbf{r})]$, is expressed as a functional of the density
131: field, $\rho(\textbf{r})$, and the equilibrium density field,
132: $\rho_{eq}(\textbf{r})$, is determined as a minimum of the grand potential.
133: If the density field is given as a function of a particle configuration,
134: $\{ \textbf{R}_i \} \equiv \{\textbf{R}_1,\textbf{R}_2,\cdots,\textbf{R}_N \}$
135: where $\textbf{R}_i$ is the position of the \textit{i}-th particle,
136: the density functional $\Omega[\rho(\textbf{r})]$ becomes a function of
137: $\{ \textbf{R}_i \}$ which can be considerd as the free energy landscape
138: defined above.
139: We employ a sum of Gaussians as the density field:
140: \begin{eqnarray}
141: \rho(\textbf{r})=\Bigl(\frac{\alpha}{\pi} \Bigr)^{\frac{3}{2}} \sum_i
142: \exp[-\alpha(\textbf{r}-\textbf{R}_i)^2].
143: \label{eq:gauss_hennbunn}
144: \end{eqnarray}
145: Here $\alpha$ and $\textbf{R}_i$ are the degree of the spread of the density
146: distribution and the position of the \textit{i}-th particle, respectively.
147: This density field has been used for the investigations such as the
148: liquid-solid transition\cite{Oxtoby,Singh,Barrat, Mohanty} and the glass
149: transition\cite{Wolynes1,Baus,Lowen,Das,Kim}.
150: The density field (\ref{eq:gauss_hennbunn}) means that the distribution of
151: the vibrational motion around $\{ \textbf{R}_i \}$ is approximated by Gaussian
152: functions. The vibrational motion within
153: $| \textbf{r}-\textbf{R}_i| \le 1/\sqrt{\alpha}$ is coarse-grained by $\alpha$.
154: By using eq. (\ref{eq:gauss_hennbunn}), $\Omega[\rho(\textbf{r})]$ becomes a
155: function of $\alpha$ and $\{\textbf{R}_i \}$:
156: \begin{eqnarray}
157: \Omega[\rho(\textbf{r})] \equiv \Omega(\alpha, \{ \textbf{R}_i \}).
158: \label{eq:def_fel}
159: \end{eqnarray}
160:
161: We use the grand potential (\ref{eq:def_fel}) as a definition of the free
162: energy landscape. Our definition of the free energy landscape is different
163: from the definition by Dasgupta and Valls \cite{Dasgupta_FEL}. They defined
164: the free energy landscape as a functional of the density field,
165: $\rho(\textbf{r})$.
166: %
167: \section{Calculation of the free energy landscape}
168: \subsection{Model}
169: We use the approximation proposed by Ramakrishnan and Yussoff to calculate
170: $\Omega[\rho(\textbf{r})]$\cite{RY}. The result of the approximation is as
171: follows:
172: \begin{eqnarray}
173: \beta \Delta \Omega[\rho(\textbf{r})]&\equiv&\beta \Omega[\rho(\textbf{r})]-
174: \beta \Omega[\rho_l] \nonumber \\
175: &=& \int_{V} d \textbf{r} \rho(\textbf{r}) \log\Bigl[\frac{\rho(\textbf{r})}
176: {\rho_l} \Bigr]-\int_V d \textbf{r} (\rho (\textbf{r})-\rho_l) \nonumber \\
177: & &-\frac{1}{2} \int_V d \textbf{r}_1 \int_V d \textbf{r}_2 c(|\textbf{r}_1-
178: \textbf{r}_2|)(\rho (\textbf{r}_1)-\rho_l)(\rho (\textbf{r}_2)-\rho_l) .
179: \label{eq:RY_basic}
180: \end{eqnarray}
181: Here $\beta=(k_BT)^{-1}$ where $k_B$ is the Boltzmann constant and $T$ is the
182: temperature, and $\rho_l$ denotes the density of the uniform liquid.
183: The coefficient $c(|\textbf{r}|)$ is the direct correlation function of the
184: uniform liquid at $\rho_l$.
185:
186: We substitute eq. (\ref{eq:gauss_hennbunn}) into eq. (\ref{eq:RY_basic}):
187: \begin{eqnarray}
188: \beta \Delta \Omega(\alpha,\{\textbf{R}_i\}) &=& F_{id}(\alpha,
189: \{\textbf{R}_i\})+F_0-\frac{N_s}{2} I_0(\alpha)-\frac{1}{2} \sum_{i}
190: \sum_{j \neq i} I(\alpha,\textbf{R}_{ij})
191: \label{eq:RY_koeki_alpha_R}
192: \end{eqnarray}
193: where
194: \begin{eqnarray}
195: & & F_{id}(\alpha,\{\textbf{R}_i\})= \Bigl(\frac{\alpha}{\pi}
196: \Bigr)^{\frac{3}{2}} \sum_i \int_V d \textbf{r} \exp[-\alpha(\textbf{r}-
197: \textbf{R}_i)^2] \log \left \{\frac{1}{\rho_l} \Bigl(\frac{\alpha}{\pi}
198: \Bigr)^{\frac{3}{2}} \sum_j \exp[-\alpha(\textbf{r}-\textbf{R}_j)^2]
199: \right \} \nonumber \\ \\
200: & & F_0=N_l \biggl( \rho_s-\frac{\rho_l}{2} \biggr) \int_V d \textbf{r} c(r) \\
201: & & I_0(\alpha)=\Bigl(\frac{\alpha}{\pi} \Bigr)^3 \int_V d \textbf{r}_1
202: \int_V d \textbf{r}_2 c(|\textbf{r}_1-\textbf{r}_2|)
203: \exp[-\alpha \textbf{r}_1^2] \exp[-\alpha \textbf{r}_2^2] \\
204: \label{eq:I0}
205: & & I(\alpha,\textbf{R})=\Bigl(\frac{\alpha}{\pi} \Bigr)^3 \int_V d
206: \textbf{r}_1 \int_V d \textbf{r}_2 c(|\textbf{r}_1-\textbf{r}_2|)
207: \exp[-\alpha \textbf{r}_1^2] \exp[-\alpha(\textbf{r}_2+\textbf{R})^2] .
208: \label{eq:IR}
209: \end{eqnarray}
210: The average density $\rho_s$ is defined by
211: \begin{eqnarray}
212: \rho_s=\frac{1}{V} \int_V d \textbf{r} \rho(\textbf{r}) ,
213: \end{eqnarray}
214: and $N_s \equiv \rho_s V$, $N_l \equiv \rho_l V$.
215:
216: In the present paper, we treat a system of the hard spheres.
217: We use Percus-Yevick approximation for the direct correlation function of the
218: hard sphere \cite{HM}:
219: \begin{eqnarray}
220: c(r)&=&
221: \left \{
222: \begin{array}{lr}
223: B_0+B_1r+B_2r^3 &(r < \sigma) \\
224: 0 & (r > \sigma). \\
225: \end{array}
226: \right.
227: \label{eq:PY}
228: \end{eqnarray}
229: where $\sigma$ is the diameter of the hard sphere.
230: The coefficients $B_0$, $B_1$, and $B_2$ in eq. (\ref{eq:PY}) are given by
231: \begin{eqnarray}
232: B_0&=&-\frac{(1+2 \eta)^2}{(1-\eta)^4} \\
233: B_1&=&6 \eta \frac{(1+\eta/2)^2}{(1-\eta)^4} \\
234: B_2&=&-\frac{\eta}{2} \frac{(1+2 \eta)^2}{(1-\eta)^4},
235: \end{eqnarray}
236: and $\eta=\pi \rho_l/6$. By using eq. (\ref{eq:PY}), $F_0$, $I_0(\alpha)$,
237: and $I(\alpha,\textbf{R})$ can be calculated
238: analytically\cite{Mohanty,keisannhouhou}.
239:
240: We use $\alpha \sigma^2=478$ correponding to the liquid-solid transition
241: point \cite{Barrat}. For large $\alpha$, $F_{id}(\alpha,\{\textbf{R}_i\})$
242: reduces to
243: \begin{eqnarray}
244: F_{id}(\alpha,\{\textbf{R}_i\}) & \simeq & N_s
245: \Bigl(\frac{\alpha}{\pi})^{\frac{3}{2}} \int_V d\textbf{r}
246: \exp[-\alpha \textbf{r}^2] \log
247: \Bigl[\frac{\exp[-\alpha \textbf{r}^2]}{\rho_l}
248: \Bigl(\frac{\alpha}{\pi}\Bigr)^{\frac{3}{2}} \Bigr] \nonumber \\
249: &=&N_s \Bigl[\frac{3}{2}\log
250: \Bigl(\frac{\alpha}{\pi}\Bigr)-\log(\rho_l)-\frac{5}{2} \Bigr]+\rho_l.
251: \label{eq:L_Fid}
252: \end{eqnarray}
253: Though Jones and Mohanty proposed that $I_0(\alpha)$ and $I(\alpha,\textbf{R})$
254: also reduce to more simple form \cite{Mohanty} for large $\alpha$, we do not
255: use their expression, since it can be used only for large $\alpha$, and is
256: not applicable for the entire region of $\alpha$. We use the analytical form
257: which can be used for all $\alpha$ \cite{keisannhouhou}.
258: %
259: \subsection{Forced relaxation in an FCC lattice}
260: In order to show the usefulness and the validity of the present definition
261: of the free energy landscape, we calculate the free energy landscape for a
262: forced structural change in an FCC crystal of hard spheres.
263: Since the grand potential (\ref{eq:def_fel}) has
264: too many parameters, $\alpha$ and $\{ \textbf{R}_i \}$, we calculate the free
265: energy landscape for a process in which we force to move a specific group of
266: particles and to relax other particles to minimize
267: $\Omega(\alpha,\{\textbf{R}_i\}) $ in the $\{ \textbf{R}_i \}$ space.
268: We calculated the free energy landscape when we force to rotate two adjacent
269: particles with their distance fixed (Figure 1).
270: The axis of rotation is the center of the two particles and two particles are
271: rotated around an axis in $<100>$ direction.
272: The arrangement of other particles is changed to minimize
273: $\Omega(\alpha,\{\textbf{R}_i\}) $ in the $\{ \textbf{R}_i \}$ space with
274: the positions of the two particles fixed.
275: We exploit the steepest decent method to find the minimum.
276:
277: We imposed periodic boundary conditions in all directions.
278: The number of the particles in the system is 108.
279: We set $\rho_s \sigma^3=1.0$ and $\rho_l \sigma^3=0.967$ which correspond to
280: the metastable state to the liquid state\cite{Barrat}.
281: Figure 2 shows the free energy landscape as a function of rotation angle
282: $\theta$. The free energy landscape has a saddle point at $\theta=\pi/2$ and
283: two minima at $\theta=0$ and $\theta=\pi$. Two minima at $\theta=0$ and
284: $\theta=\pi$ correspond to the FCC configuration (Figures 3 (a) $\sim$ (e)).
285: The height of the saddle point is about $57k_BT$.
286: %
287: \section{CRR and SRR in the free energy landscape}
288: \subsection{Definition of CRR and SRR}
289: We define the simultaneously rearranging region (SRR) as the difference in the
290: configuration at the minimum between the two adjacent basins which are connected directly via a saddle point. Apparently, the number of the particles in SRR is
291: 2 in the system studied in Sec 3.
292:
293: The CRR corresponds to the particles which are moved at the saddle point
294: from the configuration at the minimum of the original basin.
295: Our definition of the CRR is thought to be closely related to the idea of the
296: CRR by Adam and Gibbs\cite{AG}, since the particles which are moved at the
297: saddle point are thought to give the configurational entropy for the
298: rearrangement proposed by Adam and Gibbs\cite{AG}.
299: %
300: \subsection{Calculation of the CRR}
301: Since we use $\alpha \sigma^2=478$, the degree of the coarse-graining is about
302: $0.05 \sigma$. Hence we assumed that the particle in the CRR for the system
303: calculated in Sec 3 is the particle displacing more than $0.05 \sigma$ at
304: $\theta=0.5 \pi$ from the position at $\theta=0$.
305:
306: We found that the number of the particles in the CRR is $46$.
307: The particles which are shown by the bold circles at $\theta=\pi/2$ in
308: Figure 3 are those which satisfy the criterion.
309: %
310: \section{Summary}
311: In the present paper, we have given a clear definition of the free energy
312: landscape using the density functional theory.
313: As a test of the definition of the free energy landscape, we have calculated
314: the free energy landscape for rearrangement of atoms in an FCC crystal of
315: hard spheres.
316:
317: We have clarified the concept of the cooperatively rearranging region (CRR) by
318: giving a relation between the free energy landscape and the CRR. We also have
319: introduced a new region, simultaneously rearranging region (SRR) which is
320: different from the CRR. We have calculated the number of the particles in the
321: SRR and the CRR from the calculated free energy landscape.
322:
323: The present frame work can be easily generalized for glass-forming substances.
324: Hence we can calculate the number of the particles in the CRR and the height
325: of the saddle point near the glass transition from the free energy landscape.
326: Details will be studied in a forthcoming paper.
327:
328: \section*{Acknowlegdement}
329: This work was supported in part by the Grant-in-Aid for Scientific Research
330: from the Ministry of Education, Culture, Sports, Science and Technology.
331: %
332: \begin{thebibliography}{99}
333: %
334: \bibitem{hinetuexp_yamamuro}
335: O. Yamamuro, I. Tsukushi, A. Lindqvist, S. Takahara, M. Ishikawa, and
336: T. Matsuo, J. Phys. Chem. B \textbf{102} 1605 (1998).
337: \label{hinetuexp_yamamuro}
338: %
339: \bibitem{rev_glass}
340: M. D. Ediger, C. A. Angell, and S. R. Nagel, J. Phys. Chem. \textbf{100}
341: 13200 (1996).
342: \label{rev_glass}
343: %
344: \bibitem{MCT}
345: R. Schilling, Theories of the Structural Glass transition, in
346: \textit{Collective Dynamics of Nonlinear and Disorderd Systems}, edited
347: by G. Radons, W. Just, and P. H\"aussler (Springer Verlag, 2004).
348: \label{MCT}
349: %
350: \bibitem{Mezard}
351: M. Mezard and G. Parisi, J. Chem. Phys. \textbf{111} 1076 (1999).
352: \label{Mezard}
353: %
354: \bibitem{odagaki04}
355: T. Odagaki, Phys. Rev. Lett. \textbf{75} 3701 (1995), T. Odagaki, Prog. Theo.
356: Phys. Suppl. \textbf{126} 9 (1997), T. Odagaki, presented at "UNIFYING CONCEPTS IN GLASS PHYSICS III"
357: (Bangalore, India 2004).
358: \label{odagaki04}
359: %
360: \bibitem{TDM}
361: T. Odagaki, and Y.Hiwatari, Phys. Rev. A. \textbf{41} 929 (1990).
362: \label{TDM}
363: %
364: \bibitem{hinetu_tao}
365: T. Tao, A. Yoshiomori, and T. Odagaki, Phys. Rev. E. \textbf{64} 046112 (2001),
366: \textit{ibid}, \textbf{66} 041103 (2002), T. Odagaki, T. Tao, and
367: A. Yoshiomori, J. Non-Cryst. Solids, \textbf{307} 407 (2002),
368: T. Odagaki, T. Yoshidome, T. Tao, and A. Yoshimori, J. Chem. Phys. \textbf{117}
369: 10151 (2002), T.Tao, T. Odagaki, and A. Yoshimori, J. Chem. Phys. \textbf{122}
370: 044505 (2005).
371: \label{hinetu_tao}
372: %
373: \bibitem{HM}
374: J. P. Hansen, and L. R. Mcdonald, \textit{Theory of Simple Liquids} ,
375: Academic Press (1986).
376: \label{HM}
377: %
378: \bibitem{Oxtoby}
379: D. Oxtoby, Crystallization of liquids: a density functional approach, in
380: \textit{liquids, freezing and glass transition}, edited by J. P. Hansen,
381: D. Levesque, and J. Zinn-Justin (North-Holland, Amsterdam, 1989).
382: \label{Oxtoby}
383: %
384: \bibitem{Singh}
385: Y. Singh, Physics Reports. \textbf{207} 351 (1992).
386: \label{Singh}
387: %
388: \bibitem{Yoshimori}
389: A. Yoshimori, J. Theor. Comput. Chem. \textbf{3} 117 (2004).
390: \label{Yoshimori}
391: %
392: \bibitem{AG}
393: G. Adam, and J. H. Gibbs, J. Chem. Phys \textbf{43} 139 (1965).
394: \label{AG}
395: %
396: \bibitem{Barrat}
397: J. L. Barrat, J. P. Hansen, G. Pastore, and E. W. Waisman, J. Chem. Phys.
398: \textbf{86} 6360 (1987).
399: \label{Barrat}
400: %
401: \bibitem{Mohanty}
402: G. L. Jones, and U. Mohanty, Mol. Phys. \textbf{54} 1241 (1985).
403: \label{Mohanty}
404: %
405: \bibitem{Wolynes1}
406: Y. Sigh, J. P. Stoessel, and P. G. Wolynes, Phys. Rev. Lett. \textbf{54}
407: 1059 (1985).
408: \label{Wolynes1}
409: %
410: \bibitem{Baus}
411: M. Baus, and J. L. Colot, J. Phys. C \textbf{19} L135 (1986).
412: \label{Baus}
413: %
414: \bibitem{Lowen}
415: H. L\"owen, J. Phys.: Condens. Matter. \textbf{2} 8477 (1990).
416: \label{Lowen}
417: %
418: \bibitem{Das}
419: C. Kaur and S. P. Das, Phys. Rev. Lett. \textbf{86} 2062 (2001), C. Kaur and
420: S. P. Das, Phys. Rev. E. \textbf{65} 026123 (2002).
421: \label{Das}
422: %
423: \bibitem{Kim}
424: K. Kim and T. Munakata, Phys. Rev. E. \textbf{68} 021502 (2003) .
425: \label{Kim}
426: %
427: \bibitem{Dasgupta_FEL}
428: C. Dasgupta, and O. T. Valls, Phys. Rev. E. \textbf{58} 801 (1998),
429: \textit{ibid}, \textbf{59} 3123 (1999), C. Dasgupta, and O. T. Valls,
430: J. Phys. :Condens. Matter \textbf{12} 6553 (2000).
431: \label{Dasgupta_FEL}
432: %
433: \bibitem{RY}
434: T. V. Ramakrishnan and M. Yussouff, Phys. Rev. B. \textbf{19} 2775 (1979),
435: A. D. J. Haymet and D. W. Oxtoby, J. Chem. Phys. \textbf{74} 2559 (1981).
436: \label{RY}
437: %
438: \bibitem{keisannhouhou}
439: T. Yoshidome, A Yoshimori, and T. Odagaki (In preparation)
440: \label{keisannhouhou}
441: \end{thebibliography}
442:
443: \newpage
444: \begin{center}
445: {\Large Figure Captions}
446: \end{center}
447: \begin{description}
448: \item[Fig 1]
449: A schematic representation of the model under consideration.
450: Two particles are rotated with their distance fixed. The arrangement of other
451: particles are changed to minimize $\Omega(\alpha,\{\textbf{R}_i\}) $ in
452: the $\{ \textbf{R}_i \}$ space.
453: \item[Fig 2]
454: The free energy landscape when we force to rotate the two particles.
455: The horizontal axis is rotation angle $\theta$. The vertical axis is the
456: difference between $\Omega(\alpha, \{ \textbf{R}_i \})$ of a certain rotation
457: angle and $\Omega(\alpha, \{ \textbf{R}_i \})$ of FCC. The number of the
458: particles in the system is 108. We set $\rho_s \sigma^3=1.0$,
459: $\rho_l \sigma^3=0.967$, and $\alpha \sigma^2=478$. Although we write the
460: horizontal axis by one dimension, the actual free energy landscape has
461: $108 \times 3$ dimensions.
462: \item[Fig 3]
463: The configurations at the plane which two particles rotate. Each figure
464: corresponds to (a) $\theta=0$, (b) $\theta=0.3\pi$, (c)
465: $\theta=0.5\pi$, (d) $\theta=0.7\pi$, and (e) $\theta=\pi$.
466: Two particles with the bold circle are the particles rotating.
467: The particles with the bold circle at $\theta=0.5\pi$ are the particles in CRR.
468: \end{description}
469: %
470: %Fig 1
471: \begin{figure}[h]
472: \vspace{1cm}
473: \begin{center}
474: \fbox{Fig 1}
475: \end{center}
476: \vspace{1cm}
477: \includegraphics[width=15cm,clip]{model_2rot.eps}
478: \end{figure}
479: %
480: %Fig 2
481: \begin{figure}[h]
482: \vspace{1cm}
483: \begin{center}
484: \fbox{Fig 2}
485: \end{center}
486: \vspace{1cm}
487: \includegraphics[width=15cm,clip]{omega.eps}
488: \end{figure}
489: %
490: %Fig 3
491: \begin{figure}[htbp]
492: \begin{center}
493: \fbox{Fig 3}
494: \end{center}
495: \begin{center}
496: \vspace{1cm}
497: \begin{minipage}{6.0cm}
498: \includegraphics[width=5cm,clip]{haiti0_2.eps}
499: \vspace{1cm}
500: \end{minipage}
501: \hspace{1cm}
502: \begin{minipage}{6.0cm}
503: \includegraphics[width=5cm,clip]{haiti30_2.eps}
504: \vspace{1cm}
505: \end{minipage}
506: %
507: \begin{minipage}{6.0cm}
508: \includegraphics[width=5cm,clip]{haiti50_2.eps}
509: \vspace{1cm}
510: \end{minipage}
511: \hspace{1cm}
512: \begin{minipage}{6.0cm}
513: \includegraphics[width=5cm,clip]{haiti70_2.eps}
514: \vspace{1cm}
515: \end{minipage}
516: %
517: \includegraphics[width=5cm,clip]{haiti100_2.eps}
518: \end{center}
519: \end{figure}
520: \end{document}
521: \end
522:
523: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
524: