1: \input{epsf}
2: \documentclass [aps,pre,twocolumn,showpacs]{revtex4}
3: %\documentclass[preprint,aps,showpacs]{revtex4}
4: %\topmargin 0.1in
5: \usepackage[final]{graphics}
6: \usepackage{amssymb}
7: \usepackage{amsfonts}
8: \usepackage{epsfig}
9:
10:
11: \begin{document}
12:
13: \title{Probability distribution of the order parameter}
14:
15: \author{P. H. L. Martins$^{1,2}$\footnote{Email: pmartins@uft.edu.br} and
16: J. A. Plascak$^{2}$\footnote{Email: pla@fisica.ufmg.br} }
17:
18: \affiliation{$^1$Universidade Federal do Tocantins,
19: Caixa Postal 111, 77001-970, Palmas, TO - Brazil \\
20: $^2$Departamento de F\'{\i}sica, Instituto de Ci\^encias Exatas,
21: Universidade Federal de Minas Gerais, Caixa Postal 702, 30123-970,
22: Belo Horizonte, MG - Brazil }
23:
24: \date{\today}
25:
26:
27: \begin{abstract}
28:
29: The probability distribution of the order parameter is exploited
30: in order to obtain the criticality of magnetic systems.
31: Monte Carlo simulations have been employed by using single
32: spin flip Metropolis algorithm aided by finite-size scaling and
33: histogram reweighting techniques. A method is proposed
34: to obtain this probability distribution even when the transition temperature
35: of the model is unknown. A test is performed on the two-dimensional
36: spin-1/2 and spin-1 Ising model
37: and the results show that the present procedure can be quite efficient
38: and accurate to describe the criticality of the system.
39:
40:
41: \end{abstract}
42:
43: \pacs{05.10.Ln, 02.50.Ng, 64.60.-i}
44:
45: \maketitle
46:
47: \section{Introduction}
48:
49: The order parameter distribution function has been proved to be a
50: powerful tool for studying not only magnetic systems
51: \cite{Binder,Bruce,Nicolaides,Plascak,Tsypin},
52: but also the liquid-gas critical point
53: \cite{Wilding}, the critical point in the unified theory of weak and
54: electromagnetic interactions\cite{Rummuk}, and the critical point in quantum
55: chromodynamics\cite{Alex}. For the specific case of magnetic
56: systems the order parameter
57: can be chosen as the magnetization per spin, namely
58: $m = \frac{1}{N} \sum_{i=1}^{N} S_i$, where $N$ is the total
59: number of spins and $S_i$ is the spin at site $i$. In finite-size
60: systems, the magnetization $m$ is a fluctuating quantity,
61: characterized by the probability distribution $P(m)$
62: \cite{Binder,Bruce}. In Ising-like models undergoing a second-order
63: phase transition it is known that
64: at temperatures lower than the critical
65: temperature $T_c$, the distribution $P(m)$ has a double peak,
66: centered at the spontaneous magnetizations $+m$ and $-m$. At temperatures
67: greater than $T_c$, $P(m)$ has a single peak at zero magnetization, and
68: exactly at $T_c$ a double-peak shape is observed\cite{Binder}.
69: Analogously to the usual finite-size scaling assumptions \cite{Fisher71},
70: one then expects that, for a large finite system of linear
71: dimension $L$ at the critical point, $P(m)$ takes the form
72: %
73: \begin{equation}
74: %
75: P(m) = b P^* (\tilde{m}),
76: %
77: \label{pm}
78: \end{equation}
79: %
80: where $b = b_0 L^{\beta/\nu}$, $\beta$ and $\nu$ are critical exponents,
81: $\tilde{m}= bm$, $b_0$ is a constant, and $P^*(\tilde{m})$ is a universal
82: scaling function, normalized to unit norm and unit variance.
83: Scaling functions,
84: such as $P^*$, are characteristic of the corresponding universality class.
85: Systems belonging to the same universality class share the same
86: scaling functions. Thus, from the precise knowledge of $P^*(\tilde{m})$
87: one can characterize critical points and also identify universality
88: classes. This is what has been done so far in the literature, with
89: the distribution for the spin-1/2 Ising model being the standard $P^*$
90: function\cite{Plascak,Wilding} for this universality class.
91: For instance, it is shown in Fig. \ref{spins1-3} the normalized
92: distribution $P^*(\tilde{m})$ for the two-dimensional spin-1/2, spin-1,
93: and spin-3/2 Ising model at criticality. Simulations have been done
94: on square lattices with $L=32$
95: at the exact $T_c$ for spin-1/2, at $T_c = 1.6935$ for spin-1,
96: according series expansions \cite{Adler} and Monte Carlo simulations
97: \cite{Pla02}, and at $T_c=3.28794$ for the spin-3/2 model \cite{Pla02}.
98: The universal aspect of these systems can be easily noted.
99: %
100: %%%%%%%%%%%%%%%%%%%%%%%% FIGURE 1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
101: %
102: \begin{figure}[htb]
103: \includegraphics[clip,angle=0,width=6.5cm]{fig1.eps}
104: \caption{\label{spins1-3}Scaling function $P^*(\tilde{m})$
105: for the two-dimensional
106: spin-1/2, spin-1, and spin-3/2 Ising model on square lattices with $L=32$.
107: Simulations were performed at the exact $T_c$ for spin-1/2, at $T_c = 1.6935$
108: for spin-1, according series expansions \cite{Adler} and
109: Monte Carlo simulations
110: \cite{Pla02}, and at $T_c=3.28794$ for the spin-3/2 model \cite{Pla02}.
111: The error bars are smaller than the symbol sizes. After Ref. \cite{Pla02}}
112: \end{figure}
113: %
114: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
115: %
116:
117:
118: Monte Carlo simulations seem to be the most effective method to
119: obtain results as those shown in Fig.\ref{spins1-3}, where the probability
120: distribution $P(m)$
121: corresponds to the fraction of the total number of realizations in
122: which the system magnetization is $m$, i.e.,
123: %
124: \begin{equation}
125: %
126: P(m) = \frac {N_m} {{\cal N}_{MCS}},
127: %
128: \label{pm2}
129: \end{equation}
130: %
131: where $N_m$ is the number of times that magnetization $m$ appears and
132: ${\cal N}_{MCS}$ is the total number of Monte Carlo steps. To compute the
133: normalized distribution $P^*(\tilde{m})$ via Eq. (\ref{pm}) one has to
134: evaluate the pre-factor $b$. This can be easily done by noting that
135: $b = 1/\sigma$, where $\sigma$ is the square root of the magnetization
136: variance ($\sigma^2 = \langle m^2 \rangle - \langle m \rangle^2$).
137: Thus, one obtains the universal function $P^*(\tilde{m})$ by simply rescaling
138: the magnetization and by using Eq. (\ref{pm}).
139:
140: In general, the probability distribution is used for studying models in
141: which the critical temperature or even the distribution function is exactly
142: (or high-precisely) known. That is in fact what has been done in the study
143: of several systems. When this distribution, as well as the critical
144: temperature and critical exponents, are not known, one can of course
145: do first a
146: canonical simulation in order to get the critical values (universal and
147: non universal) and compute, afterwards, the desired distribution.
148: The present approach is different from this conventional one in the sense
149: that it does use the order parameter distribution itself in order to
150: obtain the criticality of the system.
151: The procedure, as well as the results obtained for the spin-1/2 and spin-1
152: Ising model, are discussed in section II and
153: the conclusions are presented in the final section.
154:
155:
156: \section{Approach and results}
157:
158: We have performed extensive Monte Carlo simulations (up to $10^7 - 10^8$
159: Monte Carlo steps per spin after $2.0 - 5.0 \times 10^4$ steps for
160: thermalization) on square $L \times L$ lattices with periodic boundary
161: conditions for systems of length $12 \le L \le 64$. For a given $L$, the
162: simulation ran at a fixed temperature, evolving according the standard
163: Metropolis algorithm. A histogram reweighting technique
164: \cite{ron,Ferrenberg88}
165: was used to obtain thermodynamic information in the vicinity of the
166: simulated temperature.
167:
168: Let us first discuss the spin-1/2 Ising model. Figure \ref{Fig2} shows the
169: distribution $P^*$ as a function of the normalized magnetization
170: $\tilde{m}$ for temperatures different from the critical value $T_c$.
171: As expected, one can
172: see that for a temperature lower than $T_c$ (Fig. \ref{Fig2}a),
173: the maximum value of $P^*$ increases when the lattice size $L$ increases,
174: while for a temperature greater than $T_c$, an increase of $L$ leads to a
175: decrease of the corresponding peaks of $P^*$ (see Fig. \ref{Fig2}b).
176: %
177: %%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE 2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
178: \begin{figure}[htb]
179: \includegraphics[clip,angle=0,width=11.0cm,height=9.5cm]{fig2.eps}
180: \caption{\label{Fig2}Scaling function $P^*(\tilde{m})$ for the spin-1/2
181: Ising model
182: with $L$ = 16, 32, and 64 at a fixed temperature $T$:
183: (a) lower than $T_c$ ($T=2.2472$) and
184: (b) greater than $T_c$ ($T=2.2831$). The error bars have been ommited
185: for clarity.}
186: \end{figure}
187: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
188: %
189: In other words,
190: suppose we have a distribution function $P^*(\tilde{m})$ for a given $L$
191: (say for example, $L=16$) at a fixed temperature $T_{L=16}$. If
192: $T_{L=16} < T_c$, a similar distribution will be obtained for a bigger
193: lattice (e.g., $L=64$) at a different temperature $T_{L=64}$ such that
194: $T_{L=16} < T_{L=64} < T_c$. Analogously, if $T_{L=16} > T_c$, we will
195: have $T_{L=16} > T_{L=64} > T_c$.
196: This fact suggests a mechanism to obtain the critical temperature,
197: as well as the exponent $\nu$ and the universal distribution, for the system
198: under study. Table \ref{Tab1} shows the temperatures of several lattice
199: %
200: %%%%%%%%%%%%%%%%%%%%%%%%% TABLE 1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
201: %
202: \begin{table}[htb]
203: \caption{Temperature for different lattice sizes at which the distribution
204: $P^*(\tilde{m})$ for $L=16-48$ is the same as that obtained for $L=64$
205: at the shown temperatures (spin-1/2). Error in
206: parentheses affects the last digits. The second and third columns correspond
207: to temperatures greater than the critical one, and the two following columns
208: correspond to temperatures lower than the critical one.
209: The last column represents the data when $P^*(\tilde{m})$ for
210: $L=64$ is obatined at $T_c$.}
211: \begin{tabular}{cccccc}
212: \colrule
213: \colrule
214: \empty{Size} & \multicolumn{5}{c}{Temperature (in units of $J/k_B$)} \\
215: \colrule
216: 16 & 2.3923(11) & 2.3272(8) & 2.2477(8) & 2.1901(10) & 2.27221(52) \\
217: 20 & 2.3666(8) & 2.3154(8) & 2.2502(8) & 2.2036(10) & 2.27092(52) \\
218: 24 & 2.3502(8) & 2.3073(5) & 2.2528(5) & 2.2134(7) & 2.27015(52) \\
219: 32 & 2.3288(5) & 2.2973(5) & 2.2563(5) & 2.2262(5) & 2.26963(26) \\
220: 48 & 2.3089(5) & 2.2878(5) & 2.2604(5) & 2.2399(5) & 2.26937(26) \\
221: 64 & 2.2989 & 2.2831 & 2.2624 & 2.2472 & 2.269184 \\
222: \colrule
223: \colrule
224: \end{tabular}
225: \label{Tab1}
226: \end{table}
227: %
228: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
229: %
230: sizes we have used to obtain the distributions displayed in Fig. \ref{Fig3}.
231: These temperatures were evaluated as follows. For $L=64$ and a given
232: temperature, for instance $T=2.2989$ in Table \ref{Tab1}, we compute the
233: corresponding probability distribution of the order parameter, which will be
234: the ``reference'' distribution. For other values of $L$, we search for the
235: temperature $T_L$
236: which gives a distribution equivalent to the reference one. In this way,
237: we obtain the data shown in the second column of Table \ref{Tab1}.
238: Taking a different reference distribution,
239: obtained at a different temperature for $L=64$, we have another
240: set of $T_L$, and so on. All the distributions so obtained are depicted in Fig.
241: \ref{Fig3}. It means that each curve in Fig. \ref{Fig3} is in fact
242: a superposition of six different distributions taking at the
243: temperatures given in Table \ref{Tab1}.
244: %
245: %%%%%%%%%%%%%%%%%%%%%%%% FIGURE 3 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
246: %
247: \begin{figure}[htb]
248: \includegraphics[clip,angle=0,width=6.5cm]{fig3.eps}
249: \caption{\label{Fig3}Normalized distribution $P^*(\tilde{m})$ for systems
250: with lattice sizes and temperatures shown in Table \ref{Tab1}. Each curve
251: is a supperposition of six different distributions taking with the
252: data from this table.}
253: \end{figure}
254: %
255: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
256: %
257:
258: %
259: %%%%%%%%%%%%%%%%%%%%%%%% FIGURE 4 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
260: %
261: \begin{figure}[htb]
262: \includegraphics[clip,angle=0,width=12.0cm,height=9.0cm]{fig4.eps}
263: \caption{\label{Fig4}(a) Temperature $T_L$ as a function of $L^{-1/\nu}$
264: with $\nu=1$. Different symbols correspond to different choices of the
265: reference distribution (from top to bottom we have the results from
266: the second to fifth columns of Table \ref{Tab1}). Error bars are smaller
267: than the symbol sizes.
268: (b) Temperature $T_L$ as a function of $L^{-(1+\theta)/\nu}$
269: with $\nu=1$ and $\theta=2$ taking the data of the last column of
270: Table \ref{Tab1}.
271: }
272: \end{figure}
273: %
274: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
275: %
276:
277: Since one expects that the difference
278: $|T_L - T_c|$ scales as $L^{-1/\nu}$, where $\nu$ is the correlation
279: length critical exponent, a finite-size scaling analysis can be done to
280: estimate the critical values of the infinite system.
281: In Fig. \ref{Fig4}a, we have a plot of
282: $T_L$ vs. $L^{-1/\nu}$, with $\nu=1$, using the values of the first
283: five columns of Table \ref{Tab1}, which confirms the exact exponent
284: $\nu=1$ and gives
285: $T_c = 2.267(2)$. Another choice for this model is, of course, the
286: corresponding critical temperature. In this case
287: (very close to $T_c$, last column of Table \ref{Tab1}), however,
288: it is known that
289: $|T_L - T_c|$ scales as $L^{-(1+\theta)/\nu}$, where $\theta$ is the
290: correction to scaling exponent \cite{Binder}.
291: In Fig. \ref{Fig4}b, we plot the estimates $T_L$ as a
292: function of $L^{-(1+\theta)/\nu}$ with $\nu=1$ and
293: $\theta=2$ \cite{Calabrese00}. Linear regression gives $T_c=2.2693(1)$ for
294: the infinite system, which is in fact quite close to the exact one.
295:
296: %The procedure of using distributions very close to $T_c$
297: %have been largely employed in literature. Nevertheless, we show in this paper
298: %that distributions at temperatures different from $T_c$ are also useful. We can
299: %also observe two distinct forms of scaling, as $L^{-(1+\theta)/\nu}$ or
300: %$L^{-1/\nu}$, if the system is or not close to the transition temperature.
301:
302:
303: In order to measure the applicability of the present mechanism for
304: obtaining the transition temperature from
305: non-universal distributions, we also study the spin-1 Ising model.
306: To have an idea of
307: the value of $T_c$, one just performs short simulations in a range of
308: temperatures to
309: check whether the probability distribution $P(m)$ has single or double peak.
310: Then, one
311: proceeds according to the same manner already discussed for the spin-1/2
312: case. We fix
313: the temperature and verify how the peaks of the distribution change if
314: the lattice size $L$
315: increases. Figure \ref{Fig-sp1a} shows the distributions obtained for
316: lattice sizes
317: $L = 12$, 16, 24, and 32 at two different temperatures: $T = 1.660$ and
318: $T=1.720$.
319: In the former case an increasing lattice size leads to increasing peaks,
320: while in the latter case
321: the height of the peaks decreases when the lattice becomes larger. Thus,
322: we conclude
323: that the transition temperature is between 1.660 and 1.720, and hence we
324: perform longer simulations in this temperature range.
325:
326: %
327: %%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE 5 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
328: %
329: \begin{figure}[htb]
330: \includegraphics[clip,angle=0,width=6.5cm]{fig5.eps}
331: \caption{\label{Fig-sp1a}Scaling function $P^*(\tilde{m})$ for the spin-1 model on
332: square lattices with $L = 12, 16, 24$, and 32 at a fixed temperature $T$:
333: (a) lower than $T_c$ ($T=1.660$) and
334: (b) greater than $T_c$ ($T=1.720$).}
335: \end{figure}
336: %
337: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
338: %
339:
340: The procedure now is the same as that we have done for the spin-1/2 model.
341: We fix the
342: temperature and compute $P^*(\tilde{m})$ for the lattice with $L=32$ (reference
343: distribution). For a different $L$, we search for the temperature that gives a
344: distribution equal to the reference one. Table \ref{Tab-sp1a} shows the
345: temperatures so obtained. For each set of temperatures (which corresponds
346: to each
347: column of Table \ref{Tab-sp1a}), we plot $T_L$ vs. $L^{-1/\nu}$ and vary the
348: exponent
349: $\nu$ until we get a straight line. Thus, each column of Table \ref{Tab-sp1a}
350: gives an
351: independent estimate of $\nu$ and also of the critical temperature $T_c$.
352: Figure
353: \ref{Fig-sp1b} illustrates this procedure. By taking the mean value of these
354: quantities, one obtains $\nu = 1.0(1)$ and $T_c=1.6933(16)$, where the
355: latter agrees well with the value $T_c=1.6935(10)$ \cite{Adler}.
356: After we have evaluated $T_c$, we ran
357: a longer simulation on a larger lattice to determine, by this way, the
358: universal distribution $P^*(\tilde{m})$. Figure \ref{Fig-sp1c} shows the
359: distribution
360: $P^*(\tilde{m})$ on a $L=64$ lattice for spin-1/2 and spin-1 models, and
361: confirms
362: the fact that both systems belong to the same universality class, as
363: already expected.
364:
365: %
366: %%%%%%%%%%%%%%%%%%%%%%%%% TABLE 2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
367: %
368: \begin{table}[htb]
369: \caption{Temperature for different lattice sizes at which the distribution
370: $P^*(\tilde{m})$ for $L=12-24$ is the same as that obtained for $L=32$
371: at the shown temperatures (spin-1). Error in
372: parentheses affects the last digits. The second and third columns correspond
373: to temperatures greater than the critical one, and the two following columns
374: correspond to temperatures lower than the critical one.}
375: \begin{tabular}{ccccc}
376: \colrule
377: \colrule
378: \empty{Size} & \multicolumn{4}{c}{Temperature (in units of $J/k_B$)} \\
379: \colrule
380: 12 & 1.598(4) & 1.652(2) & 1.708(2) & 1.765(1) \\
381: 16 & 1.624(2) & 1.664(2) & 1.705(2) & 1.747(1) \\
382: 24 & 1.649(2) & 1.675(1) & 1.702(1) & 1.730(1) \\
383: 32 & 1.660(2) & 1.680(1) & 1.700(1) & 1.720(1) \\
384: \colrule
385: \colrule
386: \end{tabular}
387: \label{Tab-sp1a}
388: \end{table}
389: %
390: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
391: %
392: %
393: %%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE 6 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
394: %
395: \begin{figure}[tb]
396: \includegraphics[clip,angle=0,width=6.5cm]{fig6.eps}
397: \caption{\label{Fig-sp1b}Temperature $T_L$ as a function of $L^{-1/\nu}$ (spin-1).
398: Different symbols correspond to different choices of the reference distribution
399: (see Table \ref{Tab-sp1a}). The values of $\nu$ that give the best linear fit
400: were (from top
401: to bottom): $\nu = 0.9(1)$, $\nu = 0.9(1)$, $\nu = 1.1(1)$, and $\nu = 1.1(1)$.
402: Error bars are smaller than the symbol sizes.
403: }
404: \end{figure}
405: %
406: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
407: %
408: %
409: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE 7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
410: %
411: \begin{figure}[htb]
412: \includegraphics[clip,angle=0,width=6.5cm]{fig7.eps}
413: \caption{\label{Fig-sp1c}Universal function $P^*(\tilde{m})$ for lattice size
414: $L = 64$ at temperatures obtained in this work: $T=2.2693$ for spin-1/2 and
415: $T=1.6933$ for spin-1.
416: }
417: \end{figure}
418: %
419: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
420: %
421:
422: %Concerning the correction to scaling exponent $\theta$, we did not evaluate it for
423: %the spin-1 model. In the critical region, the statistical errors are large.
424: %This makes difficult to choose different temperatures for different lattice sizes.
425: %For example, at $T=1.693$, the distributions for $L=24$ and $L=32$ are equivalent,
426: %within the errors. Thus, the evaluation of $\theta$ for this model needs longer
427: %simulations.
428:
429: \section{Conclusions}
430:
431: The present approach, using just the order parameter distribution, seems to
432: be a robust way to obtain the criticality of magnetic systems, regarding
433: its universal and non-universal aspects. There is a clear distinction between
434: the finite-size behavior of $P^*$ close to the
435: critical temperature (scaling with $L^{-(1+\theta)/\nu}$) or away from it
436: (scaling with $L^{-1/\nu}$), as depicted in Figs. \ref{Fig4} and \ref{Fig-sp1b}.
437: It also seems, at first sight, that there is a limitation regarding the size
438: of the lattices which could be considered.
439: For instance, we have used here lattice sizes which are
440: smaller than that from the reference distribution.
441: Nevertheless, this limitation is not so drastic since in the spin-1 model we
442: considered the reference distribution for $L=32$ and with lattices smaller
443: than this value the results prove quite accurate. One can, of course,
444: consider
445: lattices larger than that of the reference distribution. We feel, however,
446: that reweighting the distribution is more easily done for smaller systems.
447: Application of the present procedure to other models (pure and random),
448: as well as to multicritical behavior, will be very welcome; some are
449: now in progress.
450:
451:
452: \acknowledgments
453:
454: We would like to thank J. G. Moreira and R. Dickman
455: for fruitful discussions. Financial support from the Brazilian agencies
456: CNPq, CAPES, FAPEMIG and
457: CIAM-02 49.0101/03-8 (CNPq) are gratefully acknowledged.
458:
459:
460:
461:
462: \begin{thebibliography}{99}
463:
464: \bibitem{Binder}
465: K. Binder,
466: Z. Phys. B {\bf 43}, 119 (1981).
467:
468: \bibitem{Bruce}
469: A.D. Bruce,
470: J. Phys. C {\bf 14}, 3667 (1981).
471:
472: \bibitem{Nicolaides}
473: D. Nicolaides and A.D. Bruce,
474: J. Phys. A {\bf 21}, 233 (1988).
475:
476: \bibitem{Plascak}
477: J.A. Plascak and D.P. Landau,
478: Phys. Rev. E {\bf 67} 015103 (R) (2003).
479:
480: \bibitem{Tsypin}
481: M.M. Tsypin and H.W.J. Bl\"ote,
482: Phys. Rev. E {\bf 62}, 73 (2000).
483:
484: \bibitem{Wilding} A. D. Bruce and N. B. Wilding, Phys. Rev. Lett.
485: {\bf 68}, 193 (1992); N. B. Wilding and A. D. Bruce, J. Phys.: Condens.
486: Matter {\bf 4}, 3087 (1992); N. B. Wilding, Phys. Rev. E {\bf 52},
487: 602 (1995).
488:
489: \bibitem{Rummuk} K. Rummukainen, M. Tsypin, K. Kajantie, M. Laine, and
490: M. Shaposhnikov, Nucl. Phys. B {\bf 532}, 283 (1998).
491:
492: \bibitem{Alex} C. Alexandrou, A. Borici, A. Feo, P. de Forcrand,
493: A. Galli, F. Jegerlehner, and T. Takaishi, Phys. Rev. D {\bf 60},
494: 034 504 (1999).
495:
496:
497: \bibitem{Fisher71}
498: M.E. Fisher in
499: {\it Critical Phenomena},
500: edited by M.S. Green (Academic, New York, 1971).
501:
502: \bibitem{Adler}
503: J. Adler and I.G. Enting,
504: J. Phys. A {\bf 17}, L275 (1984).
505:
506: \bibitem{Pla02}
507: J.A. Plascak, A.M. Ferrenberg, and D. P. Landau,
508: Phys. Rev. E {\bf 65}, 066702 (2002).
509:
510: \bibitem{ron}
511: R. Dickman and W.C. Schieve, J. Physique {\bf 45}, 1727 (1984).
512:
513: \bibitem{Ferrenberg88}
514: A.M. Ferrenberg and R.H. Swendsen,
515: Phys. Rev. Lett. {\bf 61}, 2635 (1988).
516:
517: \bibitem{Calabrese00}
518: P. Calabrese, M. Caselle, A. Celi, A. Pelisseto, and E. Vicari,
519: J. Phys. A {\bf 33}, 8155 (2000);
520: J. Salas and A. Sokal, J. Stat. Phys. A {\bf 98}, 551 (2000);
521: S.L.A. de Queiroz, J. Phys. A {\bf 33}, 721 (2000).
522:
523: \end{thebibliography}
524:
525:
526: \end{document}
527: