cond-mat0001136/art.tex
1: \documentstyle[preprint,aps,epsfig]{revtex}
2: \tightenlines
3: \begin{document}
4: \draft
5: \title{
6: Phase transitions in generalized chiral or Stiefel's models
7:        }
8: \author{D. Loison}
9: \address{
10: Institut f\"ur Theoretische Physik, Freie Universit\"at Berlin,
11: Arnimallee 14, 14195 Berlin, Germany\\
12: Damien.Loison@physik.fu-berlin.de
13: }
14: \maketitle
15: \begin{abstract}
16: We study the phase transition in generalized
17: chiral or Stiefel's models using Monte Carlo simulations. 
18: These models are characterized by a breakdown of symmetry $O(N)/O(N-P)$.
19: We show that the phase transition is clearly 
20: first order for $N \ge 3$ when $P=N$ and $P=N-1$, 
21: contrary to predictions based 
22: on the Renormalization Group in $4-\epsilon$ expansion but in agreement
23: with a recent non perturbative  Renormalization Group approach.
24: \end{abstract}
25: \vspace{1.cm}
26: P.A.C.S. numbers: 05.50.+q, 75.10.Hk,05.70.Fh, 64.60.Cn, 75.10.-b 
27: %\narrowtext
28: 
29: \section{INTRODUCTION}
30: 
31: The critical properties of frustrated spin systems are still under
32: discussion \cite{Diep2}. In particular no consensus exists about the
33: nature of the phase transition in canted magnetic systems.
34: One  example is 
35: the  stacked triangular lattice with the nearest neighbor antiferromagnetic 
36: interactions (STA) with vector spins $O(N)$ where $N$ is the number of 
37: spin components  which is always controversial 
38: \cite{LoisonSchotteXY,LoisonSchotteHei}. 
39: The non-collinear ground state due to the frustration leads to a breakdown of 
40: symmetry (BS) from $O(N)$ in the high temperature to $O(N-2)$ in the low 
41: temperature. This is different from ferromagnets in which the ground state
42: is collinear and the BS is $O(N)/O(N-1)$. 
43: Based on the concept of universality, the class of the 
44: transition would be different in the two models. 
45: We generalize this chiral model for a BS of the 
46: type $O(N)/O(N-P)$. We obtain the STA model for $P=2$  while 
47: we obtain new BS for $N\ge P \ge 3$. For example, the case $N=P=3$ should 
48: correspond to real experimental systems, 
49: this is also applicable in spin glasses where some disorder is present. 
50: Several authors have already studied
51: these generalized chiral models applying the Renormalization 
52: Group  technic \cite{Saul,Zumbach93,Kawamura90}.
53: In mean field \cite{Zumbach93} for $N> P$ the model shows a usual second 
54: order type, but for $N=P$ the transition shows a special behavior.
55: This  last  result could be interpreted with the BS in this case 
56: being $Z_2 \otimes SO(N)$  and the coupling between the two symmetries 
57: leading to some special behavior (for example, the case $N=P=2$ in two
58: dimensions (d) is always very debated \cite{XYfrustre2d}). 
59: The $d=4-\epsilon$ expansion gives more information. 
60: The picture is very similar for all $N \ge P \ge 2$ (for details see 
61: \cite{LoisonSchotteHei} and references therein).
62: At the lowest order in $\epsilon$, there are up to four fixed points, 
63: depending on the values of $N$ and $P$. Amongst them are the trivial 
64: Gaussian fixed point and the standard isotropic $O(NP)$ Heisenberg fixed point.
65: These two fixed points are unstable. In addition, a pair of new fixed points,
66: one stable and the other unstable, appear 
67: if the case is $N\ge N_c(d)$ with 
68: \cite{Kawamura90} 
69: \begin{eqnarray}
70: N_c(d)=5P+2+2\sqrt{6(P+2)(P-1)}-\bigg{[}5P+2
71: +{25P^2+22P-32 \over 2\sqrt{6(P+2)(P-1)}}\bigg{]}\,\epsilon \,.
72: \end{eqnarray}
73: For $P=2$ we find the standard result $N_c=21.8-23.4\,\epsilon$. 
74: On the other hand, 
75: for $P=3$ we obtain $N_c=32.5-33.7\,\epsilon$ and 
76: for $P=4$ we obtain $N_c=42.8-43.9\,\epsilon$.
77: A "tricritical" line exists which divides a second order region for 
78: low $d$ and large $N$ from a first order region for large $d$ and small $N$. 
79: From these results Kawamura, using $\epsilon=1$ ($d=3$), obtained that 
80: $N_c(d=3)<0$ for all $P$. Thus he concluded that the experimental or 
81: numerical accessible systems ($N>0$) were in the second order region. 
82: Unfortunately it has been proved that the results of $4-\epsilon$ are, at best,
83: asymptotic \cite{LeGuillou}.
84: They have to be resummed to obtain reliable results.
85: Indeed for $P=2$ the calculation of the next order in 
86: $\epsilon$, combined with a resummation technic, leads the experimental
87: accessible systems for $N=2$ or $N=3$ in the first order region 
88: \cite{Antonenko 94,Antonenko2}. We  believe that the same applies for 
89: $P\ge 3$. 
90: In order to verify our assumption, we have done some simulations 
91: for $P=3$ and $P=4$
92: with $N=P$ and $N=P+1$. The most interesting case is $P=N=3$ 
93: with some possible experimental realizations and connection with 
94: the spin glasses. Moreover it is meaningful to study the generalized model 
95: in order to have a better overview. 
96: The system we analyze is the Stiefel model \cite{Kunz}. This model is
97: constructed to have the needed BS. It is closely
98: connected to real systems with complicating interactions, 
99: which are characterized by the same 
100: BS (for the case $P=2$ see \cite{LoisonSchotteHei} and
101: reference therein). From the principle of universality, models with the same 
102: BS should belong to the same universality class. Moreover we have shown 
103: that the use of the Stiefel model allows us to avoid problems 
104: which are seen 
105: in standard models, such as the presence of a complex fixed point 
106: (or minimum in the flow) \cite{LoisonSchotteXY,LoisonSchotteHei}. 
107: 
108: 
109: 
110: 
111: In the following section II the studied models are presented, we 
112: describe 
113: the details of the simulations and the finite size scaling analysis.
114: Results will be given in section III and the last section is devoted 
115: to the conclusion.
116: 
117: \section {Stiefel's models, Monte Carlo simulations and first order transitions}
118: In this section we introduce  different models studied in this work.  
119: 
120: First the $V_{3,3}$ model which is represented in Fig. \ref{figure1}.
121: The energy of the model is 
122: \begin{equation}
123: \label{tata2}
124: H = J \sum_{ij}  \sum_{k=1}^P 
125: \Big{[} \ {\bf e}_{k}(i)\cdot{\bf e}_{k}(j) \ \Big{]}
126: \end{equation}
127: where the $P$ mutual orthogonal $N$ component unit vectors ${\bf e}_k(i)$
128: at lattice site $i$ interact with the next $P$ vectors 
129: at the neighboring sites $j$. The interaction constant is here
130: negative to favor alignment of the vectors at different sites. 
131: Taking a strict orthogonality between the vectors is similar to removing
132: "irrelevant" modes corresponding to the variation between the spins inside
133: the cell. For example, in the case of a triangular lattice with 
134: antiferromagnetic interactions (STA) we force the three spins of each cell
135: to have a rigidity constraint with the sum of all the spins being always zero.
136: The obtained model is equivalent to the STA at the critical temperature
137: and can easily be transformed into the Stiefel's $V_{3,2}$ model (for more 
138: detail see \cite{LoisonSchotteXY,LoisonSchotteHei}).
139: 
140: We did not use the clusters algorithm \cite{LoisonSchotteHei}  
141: because it gives worse results than the standard Metropolis algorithm
142: for first order transitions.
143: 
144: The method for choosing the random vector depends on the number of components $N$.
145: For $N=3$ we follow the method explained in \cite{LoisonSchotteHei} for the 
146: direct-trihedral model. We construct two orthogonal vectors ${\bf e_1}$ and 
147: ${\bf e_2}$, and the third 
148: vector is constructed by the vector product of the first two:
149: \begin{eqnarray}
150: {\bf e_3} = \sigma \ {\bf e_1} \times {\bf e_2} \ 
151: \end{eqnarray}
152: where $\sigma$ is a random Ising variable, corresponding to the Ising symmetry
153: present in the $V_{3,3}$ model. This is the difference with the 
154: direct-trihedral model defined in \cite{LoisonSchotteHei}, where no Ising 
155: symmetry is present. 
156: 
157: To simulate the $V_{4,3}$ model we follow a similar procedure. 
158: We construct now three orthogonal unit vectors,
159: ${\bf e}_k=(e_k^1,e_k^2,e_k^3,e_k^4)$ with $k=1,\,2,\,3$, 
160: randomly in four 
161: dimensions using six Euler angles. 
162: The first $\theta_0$ must be chosen with probability 
163: $\sin(\theta_0)^2\, d \theta_0$,
164: two other with probability $\sin(\theta_{1,2})\, d \theta_{1,2}$
165: and the rest three $\theta_{3,4,5}$ with probability $\theta_{3,4,5}$.
166: We obtain for ${\bf e_1}$:
167: \begin{eqnarray}
168: e_1^1=&-\cos(\theta_3)*\cos(\theta_1)*\sin(\theta_5)
169: *\cos(\theta_2)*\cos(\theta_4)&\nonumber \\ \nonumber
170: &- \cos(\theta_3)*\cos(\theta_1)*\cos(\theta_5)*\sin(\theta_4)\\ \nonumber
171: & + \cos(\theta_3)*\sin(\theta_1)*\sin(\theta_5)*\sin(\theta_2)*\cos(\theta_0)\\
172: \nonumber
173: & + \sin(\theta_3)*\sin(\theta_5)*\cos(\theta_2)*\sin(\theta_4)\\ \nonumber
174: & - \sin(\theta_3)*\cos(\theta_5)*\cos(\theta_4) \\ \nonumber
175: e_1^2=& - \sin(\theta_3)*\cos(\theta_1)*\sin(\theta_5)
176: *\cos(\theta_2)*\cos(\theta_4)\\ \nonumber
177: & - \sin(\theta_3)*\cos(\theta_1)*\cos(\theta_5)*\sin(\theta_4)\\ \nonumber
178: & + \sin(\theta_3)*\sin(\theta_1)*\sin(\theta_5)*\sin(\theta_2)*\cos(\theta_0)\\
179: \nonumber
180: & - \cos(\theta_3)*\sin(\theta_5)*\cos(\theta_2)*\sin(\theta_4)\\ \nonumber
181: & + \cos(\theta_3)*\cos(\theta_5)*\cos(\theta_4) \\ \nonumber
182: e_1^3=& - \sin(\theta_5)*\sin(\theta_2)*\sin(\theta_0) \\ \nonumber
183: e_1^4=&   \sin(\theta_1)*\sin(\theta_5)*\cos(\theta_2)*\cos(\theta_4)\\
184: \nonumber
185: & + \sin(\theta_1)*\cos(\theta_5)*\sin(\theta_4)\\ \nonumber
186: & + \cos(\theta_1)*\sin(\theta_5)*\sin(\theta_2)*\cos(\theta_0)  \nonumber 
187: \end{eqnarray}
188: for ${\bf e_2}$:
189: \begin{eqnarray}
190: e_2^1=&\sin(\theta_0)*\sin(\theta_1)*\cos(\theta_3) \nonumber \\ \nonumber
191: e_2^2=&\sin(\theta_0)*\sin(\theta_1)*\sin(\theta_3)  \\ \nonumber
192: e_2^3=&\cos(\theta_0)  \\ \nonumber
193: e_2^4=&\sin(\theta_0)*\cos(\theta_1)  
194: \end{eqnarray}
195: and for ${\bf e_3}$:
196: \begin{eqnarray}
197: e_3^1=&  \cos(\theta_3)*\sin(\theta_2)*\cos(\theta_4)*\cos(\theta_1)\nonumber 
198: \\ \nonumber
199: &+ \cos(\theta_3)*\cos(\theta_2)*\cos(\theta_0)*\sin(\theta_1)\\ \nonumber
200: &- \sin(\theta_2)*\sin(\theta_4)*\sin(\theta_3) \\ \nonumber
201: e_3^2=&  \sin(\theta_3)*\sin(\theta_2)*\cos(\theta_4)*\cos(\theta_1)\\ \nonumber
202: &+ \sin(\theta_3)*\cos(\theta_2)*\cos(\theta_0)*\sin(\theta_1)\\ \nonumber
203: &+ \sin(\theta_2)*\sin(\theta_4)*\cos(\theta_3) \\ \nonumber
204: e_3^3=&- \cos(\theta_2)*\sin(\theta_0) \\ \nonumber
205: e_3^4=&- \sin(\theta_2)*\cos(\theta_4)*\sin(\theta_1)\\ \nonumber
206: &+ \cos(\theta_2)*\cos(\theta_0)*\cos(\theta_1) \ .
207: \end{eqnarray}
208: 
209: From this model we can easily create the direct-quadrihedral model
210: and the $V_{4,4}$, these two models being composed of four orthogonal vectors 
211: with four components. 
212: The differences between the two models are the presence 
213: of an Ising variables in the $V_{4,4}$ where right handed and left handed
214: are allowed while only one possibility exists in the direct-quadrihedral. 
215: The direct-quadrihedral and the $V_{4,3}$ models are topologically equivalent,
216: they should have the same low energy physics and therefore belong to
217: the same universality class.
218: The connection between the direct-trihedral 
219: and the $V_{3,2}$ models \cite{LoisonSchotteHei} is very similar to the above.
220: We form a fourth
221: vector from the vector product of ${\bf e_1}$, ${\bf e_2}$ and ${\bf e_3}$:
222: \begin{eqnarray}
223: {\bf e_4} = {\bf e_1} \times {\bf e_2} \times {\bf e_3} \
224: \end{eqnarray}
225: for the direct-quadrihedral model, and we add a  random Ising 
226: variable $\sigma$ to the $V_{4,4}$ model:
227: \begin{eqnarray}
228: {\bf e_4} = \sigma \ {\bf e_1} \times {\bf e_2} \times {\bf e_3} \ .
229: \end{eqnarray}
230: 
231: We follow the standard Metropolis algorithm to update one $P$-hedral
232: after the other.
233: In each simulation between 20\thinspace 000 to 100 \thinspace 000 
234: Monte Carlo steps are made for
235: equilibration and averages. Cubic systems  of linear dimensions
236: from $L=10$ to $L = 25$ are simulated. 
237: 
238: The order parameter $M$ for this model is
239: \begin{eqnarray}
240: M = {1 \over P\,L^3 }\, \sum_{i=1}^P \,\big|\,M_{i}\big|
241: \end{eqnarray}
242: where $M_i$ is the total magnetization
243: given by the sum of the vectors ${\bf e}_i$ over all sites and $L^3$ 
244: is the total number of sites.
245: 
246: For $N=P$ we define a chirality order parameter:
247: \begin{eqnarray}
248: \kappa = {1 \over L^3 }\, {\bf e_P}.(\, \prod_{i=1}^{P-1}\,{\bf e_i}\,)
249: \end{eqnarray}
250: where ${\prod}$ means the vector product $\times$.
251: 
252: We use the histogram MC technique
253: developed by Ferrenberg and Swendsen \cite{Ferren88}
254: which  
255: is very useful for identifying a first order transition.
256: 
257: The finite size scaling (FSS) for a first order
258: transition has been  extensively studied
259: \cite{Privman,Binder2,Billoire2}.
260: A first order transition can be identified by some properties and in 
261: particular by the following:
262: \begin{itemize}
263: \item[a.] The histogram $P(E)$ has a double peak.
264: \item[b.] The magnetization, the  chirality and the energy have hysteresis.
265: \end{itemize}
266: The double peak in P(E)
267: means that at least two states
268: with different energies coexist in the system at one temperature. 
269: 
270: \section{Results}     
271: 
272: We now present our results for the different models. The $V_{3,3}$
273: and the $V_{4,4}$
274: show a strong first order transition. The hysteresis in $E$ 
275: and $<M>$ are shown in Fig.~\ref{figure7}  and \ref{figure8} for the
276: $V_{3,3}$ model, and in $E$ and $\kappa$ for the $V_{4,4}$ model in
277: Fig.~\ref{figure9}  and \ref{figure10}. This is in accordance with the
278: negative $\eta$ exponent for the $V_{3,3}$ model found in \cite{Kunz} 
279: which describes a first order transition because $\eta$ must be positive
280: \cite{LoisonSchotteHei,Patashinskii,Zinn89}.
281: This result is understandable because there is a coupling between the Ising 
282: symmetry and the $SO(N)$ symmetry. We notice that the $V_{2,2}$ model 
283: is also of first order \cite{LoisonSchotteXY} and that the models
284: have a stronger first order transition if $N$ is greater (we obtain the
285: same hysteresis for $L=20$ for the $V_{3,3}$ as for $L=10$ for the
286: $V_{4,4}$). 
287: Thus we can generalize our
288: result that the transition is always of first order for $N=P$.
289: 
290: The $V_{4,3}$ model shows no hysteresis. However a double peak structure 
291: appears in the energy histogram and becomes more apparent when the size
292: increases (Fig.~\ref{figure11}). 
293: For greater sizes the two peaks are well
294: separated by a region of zero probability, the transition time
295: from one state to the other grows exponentially with the size of
296: the lattice.
297: We should obtains hysteresis in the thermodynamic quantities
298: when the simulation is not too long. 
299: The $V_{4,3}$ model
300: has a first order transition but weaker than the direct-quadrihedral
301: model which, for similar sizes, shows hysteresis.
302: As explained above the two models belong to the same  
303: universality class, i.e. a first order transition, 
304: similar to the  dihedral model $V_{3,2}$ and the
305: direct-trihedral model \cite{LoisonSchotteHei}.
306: The addition of the fourth leg to the $V_{4,3}$ model allows 
307: the first order behavior to be more clearly visible. 
308: In Fig.~\ref{figure17} we have plotted our hypothesis for 
309: the RG diagram flow.
310: Following the initial point, the flow could be under 
311: the influence of a "complex" fixed point (or minimum of the flow
312: \cite{Zumbach93}) and the system  mimics a second order transition. 
313: Well outside the influence of this fixed point the transition is strongly
314: of first order and in the crossover between these two regions the transition
315: is weakly first order. For a more developed discussion see 
316: \cite{LoisonSchotteHei}. 
317: 
318: \section {Conclusion}
319: 
320: We have tried to give a general picture of the transition with a 
321: $O(N)/O(N-P)$ breakdown of symmetry. We have shown by numerical 
322: simulations that for 
323: $N=P=3$ and $N=P=4$, the transition is clearly of first order.   
324: We have generalized our result for all $N=P$.
325: A similar conclusion is obtained for $N=4$ and $P=3$. Using the fact 
326: that for $N=3$ and $P=2$ the transition is also of first order,
327: we can generalize our result for all $P=N-1$. 
328: This is in contradiction with 
329: to the conclusion of Kawamura \cite{Kawamura90} which is based on two loops 
330: of a $4-\epsilon$ expansion. As we have noted the $\epsilon$
331: expansion has to be resummed to obtain reliable results. We can try
332: to achieve this by forming simple Pad\'e approximants. For a function 
333: $f=a+b\,\epsilon$ we obtain the approximation $f=1/(1-\epsilon \, b/a)$
334: which we apply to $\epsilon=1$ ($d=3$) and $P=$2, 3 and 4. We obtain 
335: $N_c(P=2)\sim 10$, $N_c(P=3)\sim 16$ and $N_c(P=2)\sim 21$. 
336: Unfortunately the results can not be perfect and in 
337: particular the result for $P=2$ is not close enough to the result 
338: including the next order expansion $N_c=3.39$ \cite{Antonenko2}
339: which demonstrates that the lower-order $\epsilon$ expansions are useless 
340: in this case. However we remark that the $N_c$ "resummed" increases
341: with $P$ which is in agreement with our result, i.e. that the initial
342: point in the renormalization flow is farther away from the mimic of the
343: second order region \cite{LoisonSchotteHei}. Thus the systems will show a 
344: stronger first order transition. 
345: This result matches with a recent study on the case $P=N=3$ 
346: which is based on a non perturbative Renormalization Group procedure
347: \cite{Delamotte99}.
348: We conclude that transitions for $N=P$ and $N=P+1$ are of first order 
349: for all $N$. 
350: 
351: 
352: 
353: \section {Acknowledgments}
354: This work is supported by the Alexander von Humboldt Foundation.
355: The authors are grateful to Professors B. Delamotte, G. Zumbach,  and 
356: K.D. Schotte for discussions.
357: 
358: 
359: 
360: \begin{references}
361: \bibitem{Diep2} {\it  Magnetic Systems with Competing Interactions 
362: (Frustrated Spin Systems)}, edited by H.T. Diep (World Scientific, Singapore,
363: 1994).
364: \bibitem{LoisonSchotteXY} D. Loison and K.D. Schotte, 
365: Eur. Phys. J. B {\bf 5}, 735 (1998)
366: \bibitem{LoisonSchotteHei} D. Loison and K.D. Schotte, 
367: to be published in Eur. Phys. J. B, accessible at
368: http://www.physik.fu-berlin.de/$^\sim$loison/articles/reference10.html
369: \bibitem{Saul} L. Saul, Phys. Rev. B {\bf 46}, 13847 (1992)
370: \bibitem {Zumbach93} G. Zumbach Nucl. Phys. B {\bf 413},771 (1994). 
371: \bibitem{Kawamura90} H. Kawamura, J. Phys. Soc. Jpn {\bf 59}, 2305 (1990)
372: \bibitem{XYfrustre2d} E. Granato and M.P. Nightingale 
373: Physica B {\bf 222}, 266 (1996), P. Olson Phys. Rev. B {\bf 55} 3585 (1997).
374: \bibitem{LeGuillou} J.C. Le Guillou and J. Zinn-Justin, J. Phys. (Paris)
375: Lett. {\bf 46}, L137 (1985)
376: \bibitem{Antonenko 94} S.A. Antonenko and A.I. Sokolov, Phys. Rev. B
377: {\bf 49}, 15901 (1994).
378: \bibitem {Antonenko2} S.A. Antonenko, A.I. Sokolov and V.B. Varnashev,
379: Phys. Lett. A {\bf 208}, 161 (1995).
380: \bibitem{Kunz} H. Kunz and G. Zumbach,J. Phys. A {\bf 26}, 3121 (1993). 
381: \bibitem{Ferren88}  A. M. Ferrenberg and R. H. Swendsen, Phys. Rev. Let.
382: {\bf 61}, 2635 (1988), Phys. Rev. Let. {\bf 63}, 1195 (1989).
383: \bibitem {Privman} V. Privman and M.E. Fisher, J. Stat. Phys. {\bf 33},
384: 385 (1983).
385: \bibitem {Binder2} K. Binder, Rep. Prog. Phys. {\bf 50}, 783 (1987)
386: \bibitem {Billoire2} A. Billoire, R Lacaze and A. Morel, Nucl. Phys. B
387: {\bf 370} 773 (1992).
388: \bibitem {Patashinskii} A.Z. Patashinskii and V.I. Pokrovskii,
389: {\it Fluctuation Theory of Phase Transitions}, (Pergamon press 1979), \S VII,
390: {\bf 6} , {\it The S-matrix method and unitary relations}.
391: \bibitem {Zinn89} J. Zinn-Justin, 
392: {\it Quantum Field Theory and Critical Phenomena}, 
393: (Oxford University Press, Oxford, 1996), \S 7.4 {\it Real-time quantum field
394: theory and S-matrix}, \S 11.8 {\it Dimensional regularization, minimal 
395: subtraction: calculation of RG functions}. 
396: \bibitem {Delamotte99} M. Tissier, D. Mouhanna, B. Delamotte, cond-mat/9908352
397: \end{references}
398: 
399: \newpage
400: \twocolumn
401: 
402: \begin{figure}
403: 
404: \vskip 1cm
405: \centerline{
406: \psfig{figure=fig1.eps,width=6.5cm,height=3.2cm} }
407: \vskip 1cm
408: \caption{\label{figure1}
409: Stiefel's model  $V_{3,3}$
410: and their interactions. ${\bf e}_1(i)$ interacts only with
411: ${\bf e}_1(j)$, neither with ${\bf e}_2(j)$ which interacts
412: with ${\bf e}_2(i)$,
413: nor with ${\bf e}_3(j)$ which interacts with ${\bf e}_3(i)$.
414: }
415: 
416: 
417: \vskip 1cm
418: \centerline{
419: \psfig{figure=fig7.eps,width=6.5cm} }
420: \vskip 1cm
421: \caption{\label{figure7}
422: Internal energy per spin $E$ versus T for the $V_{3,3}$ model.
423: Lines are guides to the eye. The arrows indicate if the MC simulation
424: is cooling (circle) or heating (square) the system. The system size
425: is $L=20$.
426: A hysteresis is clearly visible.
427: }
428: 
429: \vskip 1cm
430: \centerline{
431: \psfig{figure=fig8.eps,width=6.5cm} }
432: \vskip 1cm
433: \caption{\label{figure8}
434: Magnetization versus T for the $V_{3,3}$ model.
435: The size of the system is $L=20$.
436: See comments in Fig.~\ref{figure7}
437: }
438: 
439: 
440: \vskip 1cm
441: \centerline{
442: \psfig{figure=fig9.eps,width=6.5cm} }
443: \vskip 1cm
444: \caption{\label{figure9}
445: Internal energy per spin $E$ versus T for the $V_{4,4}$ model.
446: The system size is $L=10$.
447: See comments in Fig.~\ref{figure7}
448: }
449: \newpage
450: 
451: %\rule{5cm}{0.2mm}\hfill\rule{5cm}{0.2mm}
452: \vskip 1cm
453: %\rule{5cm}{0.2mm}\hfill\rule{5cm}{0.2mm}
454: \centerline{
455: \psfig{figure=fig10.eps,width=6.5cm} }
456: \vskip 1cm
457: \caption{\label{figure10}
458: Chirality versus T for the $V_{4,4}$ model.
459: The system size is $L=10$.
460: See comments in Fig.~\ref{figure7}
461: }
462: 
463: 
464: \vskip 1cm
465: \centerline{
466: \psfig{figure=fig11.eps,width=6.5cm} }
467: \vskip 1cm
468: \caption{\label{figure11}
469: Energy histogram $P(E)$ as a function of the energy per site $E$ for the
470: $V_{4,3}$
471: model for various sizes $L$ at different temperatures of simulation $T_L$:
472: $T_{15}=1.1802$, $T_{20}=1.1771$,
473: $T_{25}=1.1758$.
474: }
475: 
476: %\vspace{2cm}
477: 
478: \vskip 1cm
479: \centerline{
480: \psfig{figure=fig12.eps,width=6.5cm} }
481: \vskip 1cm
482: \caption{\label{figure12}
483: Internal energy per spin $E$ versus T for the direct-quadrihedral model.
484: The system size is $L=20$.
485: See comments in Fig.~\ref{figure7}
486: }
487: 
488: \vskip 1cm
489: \centerline{
490: \psfig{figure=fig17.eps,width=6.5cm} }
491: \vskip 1cm
492: \caption{\label{figure17}
493: Hypotheses of Hamiltonian flows induced by
494: renormalization-group transformations.
495: The arrows (dashed line) indicate the direction of flow under iteration.
496: }
497: 
498: 
499: \end{figure}
500: 
501: \end{document}
502: 
503: