1: %%%\documentstyle[seceq]{jpsj}
2: \documentstyle[twocolumn,epsfig]{jpsj}
3: %%%\documentstyle[seceq,short]{jpsj}
4: %%%\documentstyle[seceq,preprint]{jpsj}
5:
6: %%%\def\sf{\rm}
7: %%%\renewcommand\figureheight[1]{\vspace{24pt}\mbox{\rule{0cm}{#1}}}
8:
9: \def\runtitle{}
10: \def\runauthor{Kouji {\sc Ueda}, Ryota {\sc Otani}, Yukinobu {\sc Nishio},
11: Andrej {\sc Gendiar}, and Tomotoshi {\sc Nishino}}
12:
13:
14: \title{Critical Point of a Symmetric Vertex Model}
15:
16: \author
17: {Kouji {\sc Ueda}, Ryota {\sc Otani}, Yukinobu {\sc Nishio},
18: Andrej {\sc Gendiar}$^{2)}$, and Tomotoshi {\sc Nishino}}
19:
20: \inst
21: {Department of Physics, Faculty of Science, Kobe University, Kobe 657-8501\\
22: $^{2)}$ Institute of Electrical Engineering, Slovak Academy of Sciences, D\'ubravsk\'a cesta 9,
23: SK-841 04 Bratislava, Slovakia}
24:
25: \begin{document}
26: \sloppy
27: \maketitle
28:
29: The vertex models have been intensively studied and their thermodynamic
30: properties are well known inside the solvable parameter area.~\cite{Lieb,Baxter}
31: On the other hand, outside the solvable area the model is not
32: fully analyzed. Takasaki et al. numerically investigated a vertex model, which
33: allows 7 vertex configurations and contains 2 parameters.~\cite{Takasaki} They observed
34: a phase transition that belongs to the Ising universality class.
35: In this article we study a symmetric vertex model, that allows 10 vertex
36: configurations, by use of the corner transfer matrix renormalization
37: group (CTMRG),~\cite{CTMRG}
38: a variant of the density matrix renormalization group.~\cite{DMRG,DMRG2,DMRG3}
39: As we report in the following the model has a critical point.
40:
41: \begin{figure}
42: \epsfxsize=40mm
43: \centerline{\epsffile{Fig1.eps}}
44: \caption{Allowed vertex configurations and their Boltzmann weights.
45: The open and black circles, respectively, denote spins that take 0 and 1.
46: The shaded area will be used when we plot the spin configuration
47: snapshots in Fig.2 and 5.}
48: \label{fig:1}
49: \end{figure}
50:
51: Let us introduce the vertex model that we study in the following.
52: Consider a square lattice, where there is a 2-state spin variable
53: ($\sigma = 0$ or $1$) on each bond. Thus a lattice point is surrounded by
54: 4 spins. We impose a local constraint for these 4 spins as shown in Fig.1.
55: We assign the following Boltzmann weights
56: %
57: \begin{equation}
58: W_0^{~} = 1 \, , \,\,\,\,
59: W_1^{~} = x \, , \,\,\,\,
60: W_2^{~} = a x^2_{~} \, , \,\,\,\,
61: W_4^{~} = x^4_{~}
62: \end{equation}
63: %
64: for these 10 local configurations,
65: where the subscripts denote the sum of 4 spin variables. The parameter $x$
66: is positive. The expectation value of a spin
67: %
68: \begin{equation}
69: \rho( x, a ) \equiv \langle \sigma \rangle = ( P_1^{~} + 2 P_2^{~} + 4 P_4^{~} ) / 4 \, ,
70: \end{equation}
71: %
72: where $P_{\ell}^{~}$ is the probability to observe vertices with ${\ell}$ numbers of 1,
73: is the increasing function of $x$.
74:
75: \begin{figure}
76: \epsfxsize=40mm
77: \centerline{\epsffile{Fig2.eps}}
78: \caption{Spin configuration snapshot at $a = 0$ and $x = 1.23$. For each
79: vertex, the shaded area in Fig.1 is drawn. This configuration snapshot is
80: obtained by recently developed extension of CTMRG.~\cite{Snap,Snap2}}
81: \label{fig:2}
82: \end{figure}
83:
84: When the parameter $a$ is $0$, only 6 vertex configurations are allowed.
85: Under this strong constraint, rectangular areas inside which all spins
86: are $\sigma = 1$, appear separately in the `sea' of $\sigma = 0$, as
87: shown in Fig.2. The expectation value $\rho( x, a = 0 )$
88: gradually increases with $x$ till $x = 1.23$. At this point
89: the system shows first order phase transition, and $\rho$
90: jumps to unity. (See Fig.3.)
91:
92: \begin{figure}
93: \epsfxsize=65mm
94: \centerline{\epsffile{Fig3.eps}}
95: \caption{Expectation value $\rho( x, a )$ for $a = 0$, $0.3599$,
96: and $0.6$.}
97: \label{fig:3}
98: \end{figure}
99:
100: In contrast, when $a = 0.6$ there is no discontinuity nor singularity in
101: $\rho( x, a = 0.6 )$. Thus there should be a critical point at certain
102: values of $a$ and $x$. We trace the line of the first order transition in the
103: parameter space as shown in Fig.4. The critical value of $a$ is roughly
104: determined as 0.36. We further perform scaling analysis for $\rho( x, a )$,
105: and find one of the parameter line
106: %
107: \begin{equation}
108: a = - 0.2736 ( x - x_{\rm C}^{~} ) + a_{\rm C}^{~}
109: \end{equation}
110: %
111: shown in the inset, that passes the critical point at
112: $( x_{\rm C}^{~}, a_{\rm C}^{~} ) = ( 1.31438, 0.3599 )$.
113: Figure 5 shows the configuration snapshot at the criticality.
114:
115: \begin{figure}
116: \epsfxsize=65mm
117: \centerline{\epsffile{Fig4.eps}}
118: \caption{Phase Diagram. The line of the first order phase transition ends
119: at the critical point $( x_{\rm C}^{~}, a_{\rm C}^{~} ) = ( 1.31438, 0.3599 )$.
120: Cross marks represent data points used for the snapshot observation in Fig.2
121: and Fig.5. Inset: the parameter line used for the scaling analysis in Fig.6.}
122: \label{fig:4}
123: \end{figure}
124:
125: \begin{figure}
126: \epsfxsize=40mm
127: \centerline{\epsffile{Fig5.eps}}
128: \caption{Snapshot at the criticality.}
129: \label{fig:5}
130: \end{figure}
131:
132: In order to determine the spin expectation value
133: $\rho( x_{\rm C}^{~}, a_{\rm C}^{~} )$ at the criticality, we
134: assume a scaling form
135: %
136: \begin{equation}
137: \rho( x, a ) - \rho( x_{\rm C}^{~}, a_{\rm C}^{~} ) \propto
138: \left| x_{\rm C}^{~} - x \right|^\mu_{~}
139: \end{equation}
140: %
141: on the parameter line that passes through the critical point,~\cite{tangent}
142: and plot $\rho( x, a )$ on the parameter line given by Eq.(3) with respect to a
143: various power of $| x_{\rm C}^{~} - x |$. As a result we confirm that $\rho( x, a )$
144: shows linear dependence with $| x_{\rm C}^{~} - x |^{1/15}_{~}$ as shown in Fig.6,
145: where the two fitted lines are represented by the following equations
146: %
147: \begin{eqnarray}
148: 0.369 ( x - x_{\rm C}^{~} )^{1/15}_{~} &+& 0.716
149: \,\,\,\,\,\,\,\,\,\, ( x > x_{\rm C}^{~} ) \nonumber\\
150: -0.422 ( x_{\rm C}^{~} - x )^{1/15}_{~} &+& 0.716
151: \,\,\,\,\,\,\,\,\,\, ( x < x_{\rm C}^{~} ) \, .
152: \end{eqnarray}
153: In conclusion the exponent $\mu$ in Eq.(4) is $1/15$ and the value
154: $\rho( x_{\rm C}^{~}, a_{\rm C}^{~} )$ is $0.716$. Thus the
155: critical point belongs to the same universality class as the 2D Ising model,
156: whose magnetization at critical temperature obeys the scaling form
157: %
158: \begin{equation}
159: M( T_{\rm C}^{~}, h ) \propto | h |^{1/\delta}_{~}
160: \end{equation}
161: %
162: where $\delta = 15$.
163:
164: \begin{figure}
165: \epsfxsize=65mm
166: \centerline{\epsffile{Fig6.eps}}
167: \caption{Spin expectation value $\rho$ on the parameter line given by Eq.(3).
168: The exponent $\mu$ is fixed to $1/15$.}
169: \label{fig:6}
170: \end{figure}
171:
172: To summarize, we investigated the phase transition of a vertex model
173: with 10 local configurations, and found a critical point that belongs
174: to the Ising universality class. A possible extension of the current
175: study is toward the direction of the lattice polymer.\cite{Polymer,Polymer2} For example,
176: introducing two new vertices to those vertices shown in Fig.1,
177: one obtains a unified model with the 7-vertex case studied by Takasaki
178: et al. that corresponds to a straight line polymer
179: in two dimension.~\cite{Takasaki}
180:
181: The authors thank to E.~Kaneshita for valuable discussions.
182: A.G. is supported by the VEGA grant No. 2/3118/23, and also by
183: Japan Society for the Promotion of Science (P01192).
184:
185:
186: \begin{thebibliography}{99}
187: \bibitem{Lieb} E.M.~Lieb and F.Y.~Wu, {\it Phase Transitions and Critical Phenomena,}
188: eds. C.~Domb and M.S.~Green (Academic Press, London, 1972), Vol.1, 331.
189: \bibitem{Baxter} R.J.~Baxter:
190: {\it Exactly Solved Models in Statistical Mechanics} (Academic Press, London, 1982).
191: \bibitem{Takasaki} H.~Takasaki, T.~Nishino, and Y.~Hieida:
192: J. Phys. Soc. Jpn {\bf 70} (2001) 1429.
193: \bibitem{CTMRG} T.~Nishino and K.~Okunishi:
194: J. Phys. Soc. Jpn. {\bf 65} (1996) 891;
195: J. Phys. Soc. Jpn. {\bf 66} (1997) 3040.
196: \bibitem{DMRG} S.~R.~White:
197: Phys. Rev. Lett. {\bf 69} (1992) 2863;
198: Phys. Rev. B {\bf 48} (1993) 10345.
199: \bibitem{DMRG2} {\it Density-Matrix Renormalization
200: - A new numerical method in physics -,} eds.
201: I.~Peschel, X.~Wang, M.~Kaulke and K.~Hallberg, (Springer Berlin, 1999),
202: and references there in.
203: \bibitem{DMRG3} T.~Nishino:
204: J. Phys. Soc. Jpn. {\bf 64} (1995) 3598.
205: \bibitem{Snap} K.~Ueda, R.~Otani, Y.~Nishio, A.~Gendiar, and T.~Nishino:
206: to appear in a supplement issue of J. Phys. Soc. Jpn.
207: \bibitem{Snap2} For the calculation to obtain Fig.2 and 5 we keep $m = 20$ block spin
208: states, the number which is sufficient for drawing snapshot of the size $50$ by $50$.
209: \bibitem{tangent} To the tangential direction of the first order phase transition line
210: in Fig.4 one observes another critical exponent $\beta$.
211: \bibitem{Polymer} H.~Saleur: J. Phys. A Math. Gen. {\bf 19} (1986) 2409.
212: \bibitem{Polymer2} G.I.~Menon, and R.~Pandit: Phys. Rev. E {\bf 59} (1999) 787.
213: \end{thebibliography}
214: \end{document}
215: